An Integrated Method to Assess Consumer
Motivation in Difficult Market Niches:
A Case of the Premium Car Segment in Russia
vorgelegt von Master of Science in Economics
Marina Shcheglova aus Nowosibirsk, Russland
Von der Fakultät VII ─ Wirtschaft und Management der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Wirtschaftswissenschaften Dr. rer. oec.
genehmigte Dissertation Promotionsausschuss: Vorsitzender: Prof. Dr. Hans G. Gemünden Berichter: Prof. Dr. Volker Trommsdorff Berichter: Prof. Dr. Frank Jacob Tag der wissenschaftlichen Aussprache: 21. September 2009
Berlin 2009 D 83
II
Content
List of figures...........................................................................................................V
List of tables...........................................................................................................VI
List of abbreviations ..............................................................................................VII
INTRODUCTION ........................................................................................................ 1
1. Research problem............................................................................................... 1
2. Research objectives............................................................................................ 2
3. Thesis structure................................................................................................... 4
I. THEORY.................................................................................................................. 7
1. Motives and consumer motivation...................................................................... 7
1.1. Definition of the concepts of motivation and motive ......................................... 7
1.2. Motivation research in consumer behavior..................................................... 11 1.2.1. Origin of consumer motivation research ............................................................11
1.2.2. Contemporary directions and applied methods in motivation research ..............12
1.2.3. Common limitations of contemporary motivation theories..................................14
1.3. Motives in consumer behavior........................................................................ 15 1.3.1. Goal-oriented nature of consumer motives........................................................15
1.3.2. Hierarchical network of consumer motives ........................................................16
1.4. Summary........................................................................................................ 17
2. Consumer motivational sphere ......................................................................... 19
2.1. Theoretical backgrounds for modeling consumer motivational sphere........... 20 2.1.1. Cognitive structures...........................................................................................20
2.1.2. Means-end chain theory....................................................................................23
2.2. Conceptualization of the consumer motivational sphere ................................ 28
2.3. Summary........................................................................................................ 33
3. Premium car market ........................................................................................... 35
3.1. The notion of ‘premium’ in the car market ...................................................... 35 3.1.1. The meaning of the term ‘premium’ ...................................................................35
3.1.2. Premium, luxury and volume car brands ...........................................................36
3.2. Consumer behavior in the premium car market ............................................. 39 3.2.1. Types of consumer buying behavior..................................................................40
3.2.2. Perceived product value....................................................................................44
3.2.3. Reasons for purchasing premium cars ..............................................................46
3.3. The role of society in consumer behavior....................................................... 47 3.3.1. Consumer society .............................................................................................47
3.3.2. A person within the social environment .............................................................48
3.3.3. Social status......................................................................................................49
3.3.4. Symbols generated by premium brands ............................................................51
3.3.5. Aspirational brands ...........................................................................................52
3.4. Summary........................................................................................................ 54
III
II. METHODOLOGY ................................................................................................. 56
4. Repertory grid technique within the personal construct psychology ........... 56
4.1. Introduction to methodology........................................................................... 56
4.2. Personal construct theory .............................................................................. 58
4.3. Repertory grid technique: basic description ................................................... 62 4.3.1. Topic .................................................................................................................63
4.3.2. Elements ...........................................................................................................64
4.3.3. Constructs.........................................................................................................68
4.3.4. Scaling methods................................................................................................70
4.4. Application of the repertory grid technique in marketing and other fields ....... 72 4.4.1. Application experience ......................................................................................72
4.4.2. Criticism of the applicability of the repertory grid ...............................................76
4.4.3. Criticism solution ...............................................................................................77
4.5. Repertory grid technique: elicitation procedure.............................................. 80 4.5.1. Knowledge elicitation phase: element selection.................................................81
4.5.2. Knowledge elicitation phase: construct elicitation ..............................................83
4.5.2.1. Construct elicitation approaches.................................................................84
4.5.2.2. Construct specification ...............................................................................89
4.5.2.3. Elicited and supplied constructs..................................................................91
4.5.3. Rating grid or element comparison phase .........................................................93
4.5.4. Eliciting personal values....................................................................................96
4.5.4.1. Capturing personal values: laddering up technique ....................................96
4.5.4.2. Prioritizing personal values: resistance-to-change technique......................97
4.6. Analysis of a repertory grid........................................................................... 100 4.6.1. Qualitative methods for repertory grid analysis................................................102
4.6.1.1. Frequency counts .....................................................................................102
4.6.1.2. Content analysis.......................................................................................103
4.6.1.3. Honey’s content analysis ..........................................................................107
4.6.2. Quantitative methods for repertory grid analysis..............................................115
4.6.2.1. Hierarchical cluster analysis .....................................................................117
4.6.2.2. Principal component analysis ...................................................................121
4.6.2.3. Synergy of principal component and hierarchical cluster analysis.............128
4.7. Reliability and validity of the repertory grid................................................... 130
4.8. Summary...................................................................................................... 134
5. Modeling the consumer motivational sphere................................................. 141
5.1. Product characteristics................................................................................. 143
5.2. Product attributes ......................................................................................... 145
5.3. Specific expected consequences ................................................................. 147
5.4. Abstract expected consequences ................................................................ 150
5.5. Values .......................................................................................................... 153
5.6. Situational and environmental incentives ..................................................... 154
5.7. Summary and conclusion............................................................................. 156
IV
III. FIELD RESEARCH ........................................................................................... 161
6. Research design............................................................................................... 161
6.1. Research objectives..................................................................................... 161
6.2. Research design .......................................................................................... 162 6.2.1. Structure of the sample ...................................................................................162
6.2.2. Survey instrument: repertory grid design .........................................................166
7. Analysis of repertory grids .............................................................................. 177
7.1. Preparation of data....................................................................................... 177
7.2. Content analysis........................................................................................... 177 7.2.1. The generic content analysis procedure..........................................................177
7.2.2. Honey’s content analysis.................................................................................182
7.3. Analysis of relationships within a single grid ................................................ 187 7.3.1. Preparation and input of data into the software package .................................188
7.3.2. Systematization of elicited meanings...............................................................189
7.3.3. Systematization of elements perceptions ........................................................202
7.4. Segmentation based on the comparison of the findings from the single grids analyses.............................................................................................................. 209
7.5. Multi-grid analysis ........................................................................................ 218
7.6. Additional analysis and findings for managerial implications........................ 228 7.6.1. Segmentation according to aims for purchasing an auto .................................228
7.6.2. Stereotypes of car attribute perceptions ..........................................................233
7.7. Summary...................................................................................................... 235
CONCLUSION ....................................................................................................... 240
1. Theoretical contribution................................................................................... 240
2. Methodological contribution ............................................................................ 242
3. Practical contribution and implication .............................................................. 245
4. Limitations....................................................................................................... 250
5. Potential for future research............................................................................ 252
References ............................................................................................................ 254
Appendix ............................................................................................................... 271
Appendix 1: Car classification ............................................................................. 271
Appendix 2: Minkowski metric ............................................................................. 272
Appendix 3: Repertory grid form and questionnaire developed for interviews .... 273
Appendix 4: Categorized elicited constructs ....................................................... 281
Appendix 5: Customers’ values and most preferred car attributes...................... 289
Appendix 6: Brands of cars named for element categories................................ 290
V
List of figures
Figure 1: Means-end chain original structure............................................................ 24
Figure 2: Consumer cognitive structure by Olson and Reynolds .............................. 24
Figure 3: Consumer motivational sphere.................................................................. 29
Figure 4: Public coverage by premium and luxury brands........................................ 38
Figure 5: Components of product utility by Vershofen .............................................. 42
Figure 6: Components of total product value ............................................................ 44
Figure 7: Reasons for purchasing premium brand products ..................................... 46
Figure 8: Aspirational brands audience .................................................................... 53
Figure 9: The assumptive structure of the personal construct theory ....................... 60
Figure 10: Display of elicited repertory grid ............................................................ 118
Figure 11: Graphic plot (dendrogram) of HCA of the grid ....................................... 119
Figure 12: Graphic plot of PCA of the grid .............................................................. 123
Figure 13: Element perception on different dimensions.......................................... 127
Figure 14: Synergy effect of PCA and HCA............................................................ 129
Figure 15: Consumer motivational sphere .............................................................. 141
Figure 16: The CMS schema extended with sources ............................................. 142
Figure 17: Display of the 14th repertory grid............................................................ 188
Figure 18: Graphic plot of a PCA of the 14th grid .................................................... 190
Figure 19: HCA: Component 1 including constructs 5, 9, 10 .................................. 193
Figure 20: HCA: Component 2 including constructs 5, 9, 10 .................................. 194
Figure 21: Relationships between elements ........................................................... 203
Figure 22: HCA producing element and construct dendrograms ............................ 204
Figure 23: HCA conducted within the first PCA dimension ..................................... 205
Figure 24: HCA conducted within the second PCA dimension ............................... 205
Figure 25: HCA of the collective grid ...................................................................... 225
Figure 26: PCA of the collective grid....................................................................... 226
Figure 27: Relationships between elements ........................................................... 226
Figure 28: Aims of purchasing the current or next auto .......................................... 229
VI
List of tables
Table 1: The repertory grid elicitation form ............................................................... 63
Table 2: Scaling methods used in repertory grid based research............................. 71
Table 3: An example of an elicited repertory grid ..................................................... 81
Table 4: An example of a ladder derived from one construct.................................... 97
Table 5: Subsequent comparison of personal values ............................................... 99
Table 6: Frequency of value preferences over other ones and a value hierarchy .. 100
Table 7: Relationship between two constructs about six elements......................... 109
Table 8: Relationship between two constructs showing a reversal ......................... 110
Table 9: Repertory grid using Honey’s technique ................................................... 114
Table 10: Links between constructs ....................................................................... 119
Table 11: Links between elements ......................................................................... 120
Table 12: Element sort after cluster rearrangement ............................................... 120
Table 13: Construct sort after cluster rearrangement ............................................. 120
Table 14: Percentage of variance accounted for by each component .................... 123
Table 15: Ratings given on constructs of the same product attribute. .................... 146
Table 16: Personal estimation of welfare................................................................ 166
Table 17: Category based questions for eliciting a set of elements........................ 169
Table 18: Repertory grid record of the 14th respondent .......................................... 176
Table 19: Content analysis procedure .................................................................... 179
Table 20: Homogeneity and the importance of the derived meanings.................... 181
Table 21: Honey’s technique calculations............................................................... 183
Table 22: Intervals for H-I-L indices........................................................................ 184
Table 23: Honey’s content analysis of all grids....................................................... 185
Table 24: Correspondence between category poles and construct poles............... 189
Table 25: Percentage of variance accounted for by each component. ................... 189
Table 26: Construct loadings on each component.................................................. 191
Table 27: Links in the dendrogram of Figure 19 ..................................................... 193
Table 28: Links in the dendrogram of Figure 20 ..................................................... 194
Table 29: Individual structure of construct meanings.............................................. 196
Table 30: Individual schema of a consumer motivational sphere ........................... 201
Table 31: Links between adjacent elements in the dendrogram of Figure 22......... 204
Table 32: Interrelationships between element perceptions..................................... 207
Table 33: Customer segmentation.......................................................................... 214
Table 34: Collective construct ‘comfortable – uncomfortable’................................. 220
Table 35: Constructs of the collective grid .............................................................. 220
Table 36: Collective RG with average ratings......................................................... 221
Table 37: Fundamental core perceptual dimensions based on construct loadings. 222
Table 38: Correlation matrix of collective constructs .............................................. 223
Table 39: Correlation matrix of element categories ................................................ 225
Table 40: Common customer abstract consequences............................................ 230
Table 41: Stereotypes of car attribute perceptions among the respondents........... 235
VII
List of abbreviations
CMS Consumer Motivational Sphere
dim. dimension
e.g. exempli gratia (for example)
et al. et alii (and others)
etc. et cetera (and so forth)
HCA Hierarchical Cluster Analysis
i.e. id est (in other words)
M match
MEC Means-End Chain
p. page
PCA Principal Component Analysis
PCP Personal Construct Psychology
PCS Product Construct System
PCT Personal Construct Theory
PSS Percentage Similarity Score
RG Repertory Grid
RGT Repertory Grid Technique
SDI Subjective Discretionary Income
SUV Sport Utility Vehicle
var. variance
Introduction
1
“I don’t know anything more interesting in life, than motives of human actions.”
Maxim Gorky
Introduction
1. Research problem
Understanding consumer purchasing behavior is one of the major challenges for
companies to succeed in highly competitive and overloaded markets. To understand
the reasons for actual purchase, retention, failure, recommendation and other
essentials of consumer behavior, marketers need to explore psychological dynamics
that determine consumer behavior. With increased competition, diversity of
assortment and importance of social factors, consumer behavior becomes more
complex and differentiated being influenced by very different aspects of consumers’
knowledge, cognitive representations and external environmental incentives. People
choose products according to the lifestyle they aspire to, often acting irrationally, and
value things that are rather distant from a single product but have a strong impact on
the preference and final choice for a certain product or brand. Therefore, the
challenge is to figure out which processes, factors and incentives, among everything
that a consumer experiences and is surrounded with, define particular consumer
behavior within a given situation and circumstance. Marketers need to understand
the principal motivation behind product consumption in order to precisely target
current and potential customers.
Motivation refers to the processes that cause people to behave as they do. From a
marketing research perspective the study of motivation represents a search for
principles that can help in understanding why people initiate, choose or persist in
specific actions and in specific circumstances. Motivation research targets motives
for specific behavior and tries to analyze them.
The concept of consumer motivation represents a long lasting debate between
researchers who are trying to create commonly agreed upon and adopted definition,
however, to this date there is still no consensus regarding the definition and
measurement of consumer motivation. There are certain limitations and lack of
consensus in contemporary consumer motivation research which provide an
inducement for this dissertation.
Most concepts of consumer motivation suggest universal sets or categories of
motives, needs or goals which can be applied to any type of product, consumer and
Introduction
2
pattern of behavior. The major disadvantage of these sets is that they appear unable
to account for specific actions and to recognize the situational variability of consumer
behavior.
Most theory based concepts of consumer motivation do not provide certain
supportive methodology which would enable samples to be drawn in investigations of
consumers and their attitudes towards particular products, brands or product
categories.
The concept of consumers' motives based on the hierarchical nature of their
interrelations distinguishes between too-specific motives, middle-range motives and
too-general motives. It is assumed that too-specific motives explain a very specific
purchasing behavior and often cannot be approximated either on other products or
within a group of people. One the other hand, too-general motives are rather distal
determinants which do not explain by themselves much about consumers’ particular
buying motivation. Therefore, it is important to obtain the full sequence of motives
where motives in the middle-range can be defined and measured so that they allow
us to approximate and to find commonalities within a group of people while, however,
not being too general to be interpreted in terms of product features and perceptions.
In consumer research a long-standing dispute between qualitative and quantitative
approaches to motivation research resulted in attempts to develop a compromise
approach, however, the true synthesis of motivation research with quantitative
practice has remained obscure.
These limitations and research dilemmas cause the major disadvantage of most
analysis of motives which are conducted as purely atheoretical predictive exercises
resulting in an ad hoc inductive explanation. Therefore, a theoretical model of
motivation and causality underpinning an active guidance for assessment algorithms
is required.
2. Research objectives
The goal of this dissertation is to contribute to the understanding of the consumer
motivation which explains the psychological dynamics determining consumer
behavior. Considering limitations and the potential for contribution described above,
the general objective of this thesis is to propose a logical construction
conceptualization of consumer’s cognitive structure of knowledge and perceptions
Introduction
3
which defines his motivated consuming behavior and to provide a methodological
guidance for its practical implementation. As recognized by researchers, it is a key
factor in attaining a useful understanding of customer behavior because such
cognitive structures bear a major influence on behavior as they are essential in
explaining information-processing behavior and, in turn, explicit behaviors like
purchasing choices that are functions of those cognitive processes.
The theoretical objective of this thesis is to provide a concept of a consumer
motivational sphere built on existing theoretical statements, assumptions and
propositions. It is based on the notion that motivated behavior is goal-directed and
that the reasons for choosing a particular goal are represented in a hierarchical
network of motives.
In order to make the concept applicable in practical investigations of consumers, two
methodological objectives of the thesis are defined. The first objective is to suggest
and describe an appropriate interpretive research framework for exploring
relationships in the content and structure of consumers’ subjective meaning systems.
The second objective is to develop a methodology for estimating how customers‘
preferences are influenced by motives and external factors which would be
supportive and congruent with the theoretical conceptualization of consumer
motivational sphere and would integrate both qualitative and quantitative
approaches. A significant peculiarity of the proposed methodology is that it aims at
the exploration and aggregation of personal meanings expressed in a person’s own
language and comparative estimates. Furthermore, the focus is put on the definition
and assessment of ‘motives of the middle range’ in which a special interest of
marketing theory is shown.
The practical objective of the dissertation is to demonstrate the applicability of the
suggested methodology and to provide some recommendations for its practical
implication. To achieve this goal, an exploratory research study of customers’
motivation mechanism and preferences was conducted on the Russian automotive
market with the focus on consumers of premium cars. Thus, a special focus of the
study is put on motivation research of difficult groups of consumers characterized by
often irrational and, therefore, behavior that is difficult both to access as well as to
predict.
Introduction
4
3. Thesis structure
The dissertation consists of three major parts.
The theoretical part is devoted to the development of the concept of a consumer
motivational sphere that further serves as a theoretical framework for which an
assessment methodology is proposed. The theoretical part consists of three
chapters.
The first chapter contains a review about theoretical conceptions of consumer
motivation and motives as described in literature. This chapter presents definitions,
consumer motivation research experience and its contemporary state and the nature
of motives in consumer behavior.
The second chapter is devoted to the conceptualization of the consumer motivational
sphere based on the existing theoretical statements and propositions. At first, this
chapter provides an overview of the theoretical background for the concept and then
presents a schematic model of the consumer motivational sphere.
The third chapter describes the peculiarities of consumer behavior in the markets of
premium products in general and on the premium car market in particular. It focuses
on the consumer behaviors influenced by the social environment and describes
concepts which consider such factors as prestige, social status, social perception,
etc. and their role in a consumer’s decision making process and perceived product
value.
The methodological part provides a thorough description of a qualitative technique
of structured interviewing named ‘repertory grid’ based on the personal construct
psychology and describes the proposed methodology for the assessment of the
consumer motivational sphere. This part consists of two chapters.
The first chapter presents the repertory grid technique, which provides an interpretive
research framework. The format of the data gathered by means of the repertory grid
technique and its further analysis enables the disclosure of each component of the
schematic model suggested in the theoretical part of the work. This chapter
thoroughly describes a procedure of repertory grid elicitation, considering its
numerous variations, and presents different methods for analyzing the data.
Introduction
5
The second chapter describes a methodology developed to derive each level of the
consumer motivational sphere from the data elicited by means of the repertory grid
technique. The suggested methodology represents an integrated method based on
qualitative and quantitative approaches which is aimed at the investigation of the
individual construing system.
The empirical part of the thesis is devoted to the field research that is conducted
and presented in order to demonstrate the applicability of the suggested
methodology. This part consists of two chapters.
The first chapter describes the objectives, peculiarities, circumstances and a sample
of the conducted survey and, based on that, the repertory grid form appropriate for
this research is designed.
The second chapter presents analysis of the data. A single-grid analysis
demonstrates disclosure and construction of an individual consumer motivational
sphere from a single grid. Customer segmentation is performed based on all of the
grid analysis conducted by analogy. Afterwards, multi-grid analysis based on the
collective grid is conducted and inferences are presented.
The empirical study conducted with an application of suggested interviewing
techniques and analysis methods was requested by a major German car
manufacturer. All findings that have been made during the study and presented to
the company could not be presented in this dissertation due to either secondary
relevance to the scientific aspect of the dissertation and proposed methodology or
the confidentiality restrictions imposed by the company. Therefore, the empirical part
of the work is mainly concentrated on the demonstration of the proposed
methodology, however, several additional inferences made for managerial implication
are included.
All chapters (except the sixth chapter which describes objectives and the sample of
the empirical study) are provided with a summary listing the main issues discussed in
the context. The dissertation ends with a discussion of its theoretical, methodological
and practical contributions, limitations and potential for future research.
Throughout the context of this dissertation pronouns referred to a person, an
individual, a consumer, a respondent, etc. or referred to an indefinite antecedent,
Introduction
6
such as ‘one’ or ‘anybody’, are written in male sex form. This form stands for both
male and female where, for example, ‘he’ implies ‘he or she’. According to English
grammar such writing is applied when a phrase refers to a person of unknown or
unspecified sex. The issue of political correctness should not be addressed to this
work considering the female sex of its author.
1. Motives and consumer motivation
7
I. Theory
1. Motives and consumer motivation
Since many decades marketers try to understand why customers act or react as they
do. This question is one of the major questions the theory of consumer behavior
pursues to explore and disclose. There are a lot of questions researchers have to ask
themselves when trying to understand certain purchasing behaviors. Why customers
prefer certain brands or products, why do they often act irrationally, why some
consumers aspire to a certain lifestyle which others reject and why they value not
only an assortment and quality of products but also a distinctive purchasing
experience, etc. (Kroeber-Riel et al., 2009, p. 167). Marketers need to understand
the principal motivation behind each product consumption to accurately target current
and potential customers.
This chapter provides an overview of theoretical conceptions of consumer motives
and motivation. In section 1.1 definitions of motivation and motives are presented
followed by a description of the nature of motives. Section 1.2 is aimed at providing
an overview of the motivation research, its history and contemporary agreements and
limitations found in the literature. The last section of the chapter describes motives
within the theory of consumer behavior. It is focussed on the notion and the nature of
consumer motives rather than on general motives of human behavior. Considering
agreements and limitations described in the literature, the chapter ends with a
discussion of the direction in which consumer motives will be further considered and
used in this work.
1.1. Definition of the concepts of motivation and motive
Motivation refers to the processes that cause people to behave as they do. The
concept of motivation is aimed at explaining reasons for a certain behavior. From
marketing research perspective to understand motivation is to understand why
consumers do what they do. Study of motivation represents search for principles that
can help in understanding why people initiate, choose or persist in specific actions
and in specific circumstances. Motivation research targets motives for specific
behavior and tries to analyze it. Especially important next to the basic, physical
motives are social motives (Mook, 1987, p. 4).
1. Motives and consumer motivation
8
Motivation is a goal-oriented driving process which consists of two major components
namely emotional component and cognitive component (Kroeber-Riel et al., 2009, p.
177). The emotional component represents a trigger for the action, whereas the
cognitive component defines the direction and the way the action is to be realized.
Often instincts are considered separately as part of motivation construct which can
trigger and influence the behavior. Instincts are inborn and, therefore, not learned
behavior patterns which are activated by basic, inner stimuli and automatically result
in certain actions. Thus, instincts are biologically preprogrammed and can be also
understood as physiological deficiency, e.g., hunger, which activates a certain
behavior to liquidate this deficiency (Puca and Langens, 2008, p. 194).
Therefore, motivation appears from an interaction between affective and cognitive
processes which results in the goal definition and a certain way of acting (Kroeber-
Riel et al., 2009, p. 169).
Notion of motivation and motive are closely interrelated. The term ‘motive’ is often
used as identification of an outlasting, latent disposition, whereas the notion of
motivation refers to the process of actualization of motives (Kroeber-Riel et al., 2009,
p. 170).
Motives identify qualities of people which they can be more or less aware of, to
respond emotionally to the goals related to a certain topic such as, for example,
affiliation, power or achievement. Motivation describes the state of purposeful
behavior which can be characterized by such parameters as direction, intensity and
duration (Puca and Langens, 2008, p. 224).
Trommsdorff (2009, p. 32) also defines motives as latent states which, in case of
actualization, drive the behavior with a certain intensity and in a certain direction. The
actualization can be resulted by deficiency or by external incentives. Thus, motives
can be seen on a wide range between ‘extremely uncontrolled, emotional’ to
‘extremely controlled, rational’.
In the literature motives are classified in two directions, first, according to their origin
and necessity and, second, according to their extent of consciousness.
Regarding their origin and necessity, motives are subdivided into primary and
secondary motives (Kroeber-Riel et al., 2009, p. 170):
1. Motives and consumer motivation
9
• Primary motives are coursed by inborn needs such as hunger or thirst which
each person has to satisfy in order to exist. Thus, instincts refer to the primary
motives.
• Secondary motives are acquired during socialization processes. These
motives arise as a result of learning through interacting with other people and
society and adapting to the different situations and circumstances. Secondary
motives are aimed at satisfying needs that are not vital and indispensable to
life but essential for individual esteem and importance to self and to the
society.
Maslow (1954) proposed the most popular but, at the same time, controversial and
debatable pyramid of needs in attempt to classify motives. He emphasized five levels
of needs where each level refers either to the primary or “deficiency” needs or to the
secondary or “growth” needs. It is argued that as soon as the needs of a lower level
are satisfied the next level of needs is activated (Kroeber-Riel et al., 2009, p. 170).
Second classification of motives distinguishes between conscious and unconscious
motives.
• Conscious motives refer to the individual’s conscious functioning.
Consciousness is associated with an individual experience gained from both
outside and inner worlds and, thus, associated with own identity (Kiefer, 2008,
p. 156). Conscious processes bear on the individual attention or the capability
to notice and reflect events in the individual’s mind (Kroeber-Riel et al., 2009,
p. 171). Motive that precedes the action is conscious if the person is aware of
the cause of this action.
The informational environments that the brain has to handle are too complex and
appear beyond the capacity of consciousness. Therefore, especially frequently
performed actions and behaviors become automatic being guided to a large degree
by unconscious motives.
• Unconscious motives refer to the individual’s unconscious processing.
Unconsciousness implies the individual’s capability to very quickly derive
meanings from different signs in the environment. However, the main
difference with conscious formulation of meanings is that the individual has no
1. Motives and consumer motivation
10
or very incomplete insight of how this happens. Although, once formed, these
meanings become rather authoritative and guide individual’s behavior without
him being entirely aware of it or choosing that this should happen (Sohlberg,
2006, p. 151). The motive is unconscious if it activates behavior without
individual’s realization of its cause.
People cannot be conscious about everything what they do and how they do.
However, much of automatic, unconscious processing can become conscious by an
effort of recall or attention. For example, if to ask a person who goes every day to the
same store and automatically buys the same milk why he chooses this milk, he might
immediately answer that he likes its taste, however, if he takes time to recall the
whole process and to think about actual reason for the purchase, he might
understand that he also chooses this milk because in its advertising the person who
got the milk looked very healthy, good and happy. That is why the person
unconsciously stored the image of the person from the advertising and tried to
associate himself with this image by purchasing the promoted milk. This example
also illustrates how advertising can influence consumer behavior even without
consumer’s awareness of it. The behavior remains automatic and unconscious until
the person does not need to pay attention to what is guiding his behavior (Sohlberg,
2006, p. 151).
Motives, as results of inner processes, can be unconscious and influence the
behavior outside of awareness, however, they can often be made conscious by
individual own reflection or by means of certain questioning techniques, especially
those, which apply indirect questioning approaches (Kroeber-Riel et al., 2009, p. 175;
Trommsdorff, 2009, p. 108; Jacob, 2009, p. 64).
The notion of motives should be distinguished from notions of needs and wants
which are often used as synonyms for each other.
Concept of needs exists on a different theoretical level than concept of motives.
Needs are, indeed, the elicitors for motives, characterizing a level of deprivation.
Motives are already explained, i.e. they are goal-oriented, while needs are not. Thus,
with a real sensation of deprivation a need attains a grade of motive possessed of its
defined direction (Trommsdorff, 2009, p. 108).
1. Motives and consumer motivation
11
Wants are also often used in the context without any difference to motives. Wants
are on the same theoretical level as needs. While needs explaining what is really
needed, wants are equally to wishes or desires. They explain what people wish to
achieve or to obtain. Wants (wishes or desires) represent specific implementation of
a goal that motivates consumer to buy a particular product (Lass, 2002, p. 593).
1.2. Motivation research in consumer behavior
1.2.1. Origin of consumer motivation research
Ernest Dichter, considered by many a founding father of motivation research
(Haugtvedt et al., 2008, p. 20), was a successful market researcher who began to
develop the theory of a “dynamic psychology”, which later became the conceptual
basis of motivation research. Thereby, Dichter and his colleagues, psychologists,
brought a discovery from two perspectives. “On the one hand, Dichter began to
conceptualize the human soul as a hidden “realm of desires”, full of taboos and
secrets. On the other hand, Dichter discovered the “soul of the products”, which was
also structured as a space of complexes and taboos” (Gries, 2005).
Dichter was the first who employed one-on-one format of in-depth interviews based
on the indirect questioning, as opposed to the direct approach typically used in early
quantitative research. He also applied projective techniques borrowed from the
clinical psychology. The questioning was stressed on revealing the conditions
surrounding different purchase and consumption occasions and on finding the
evidence of a symbolic expression of hidden motivations (Pincus, 2004).
Dichter sought clues into consumer motivation by questioning selected individuals
who would be proactive in providing insights into product usage. This information he
used to provide a “psychological inventory” of basic motives for specific product
purchases (Dichter, 1964 cited in Haugtvedt et al., 2008, p. 19). The purchasing
motives he proclaimed were derived from application of psychoanalytic theories and
methods to study consumer behavior (Trommsdorff, 2009, p. 109; Pincus, 2004).
Thus, for many different products he identified the most dominant purchasing
motives. He meant that this information about certain products would help marketers
and advertisers develop messages that would directly address the consumer
motives.
1. Motives and consumer motivation
12
Dichter’s work caused the initial split in consumer research between qualitative
motivation research and more typical quantitative research which, since then, have
competed against each other, resulting in attempts to develop a compromise
approach, however, true synthesis of motivation research with quantitative practice
has remained obscure (Pincus, 2004).
1.2.2. Contemporary directions and applied methods in motivation research
In the literature two basic directions of motivation research can be emphasized which
compete against each other to be a better concept of motivation explanation.
The first direction is psychological motivation research on cognition. This type of
research is stressed on the cognitive component of the motive construct and, thus,
on the conscious reasonable goal-oriented consumer behavior (Kroeber-Riel et al.,
2009, p. 176). Consumers’ attention and mechanisms of information selection as well
as perceptive processes are studied. The motivation is considered as being based on
a subjective perception of means-goal (end)-coherence and subjective expected and
desired quality of the goal achievement.
One of the most well-known theories consistent with cognitive motivation is called
“means-end chain theory” which studies subjective conscious processes of goal
definition and desirable or expected ways of its reaching within the given situation.
Therefore, the derived motivation is restricted within the conscious drivers and
incentives aimed at an achievement of goals resulted in gained experience and
desires.
Researchers who work in the cognitive-oriented direction mostly use so called
laddering technique to reveal and explain causes of action (e.g., Botschen et al.,
1999; Lin, 2002; Reynolds and Gutman, 1988; Reynolds and Whitlark, 1995; Huber
et al., 2004; Grunert and Grunert, 1995; Wansink, 2003; Valette-Florence and
Rapacchi, 1991). This technique is based on the questioning procedure where a
respondent is urged to express perceived linkages between his individual goals and
means to reach these goals (or ‘ends’). Laddering technique and the underpinning
means-end chain theory are discussed further in this work as a part of the suggested
methodology related to the cognitive and conscious oriented exploration of the
consumer behavior.
1. Motives and consumer motivation
13
The second research direction is psychological motivation research on emotions.
Advocates of this direction argue that the behavior is not always caused by a
cognitively realized goal orientation but in the foreground can be influenced and
defined by inner driving forces that activate the behavior (DeCharms and Shea, 1976
cited in Kroeber-Riel et al., 2009, p. 177). This kind of research is focused on the
investigation of emotions and instincts that activate and direct actions often without a
cognitive awareness of their causes.
Methods widely used in emotions oriented research are adopted from the
psychoanalytical theory, clinical social psychology and cultural anthropology. Most
widely used methods are Thematic Apperception Test (TAT) and one of its
modifications Rorschach Test (Kroeber-Riel et al., 2009, p. 179). These methods as
well as their further modifications represent projective techniques which use pictures,
stories, cartoons, photos, collages, etc. to reach respondent’s unconscious and
reveal reserved aspects of personality such as motives, needs, desires, attitudes,
etc. (see Gröppel-Klein and Königstorfer, 2007 for more about projective methods).
The respondent is assessed according to his projects onto the indefinite and
uncertain images or stories, the way he interprets them and associates with self or
ideal self. Thus, respondent creates his stories, i.e., interpretations which must be
further carefully analyzed to uncover underlying meanings and patterns of reaction.
Projective methods attracted much criticism regarding findings validity and reliability
and it is suggested that a combination of different types of methods should be
applied in order to overcome this criticism (see Kroeber-Riel et al., 2009, pp. 179-183
for more detail).
Motivation research on cognition refers to a considerable extent to the quantitative
type of research often aimed at empirical testing of hypotheses, whereas motivation
research on emotions is a qualitative research which explores the behavior on a
deep psychological level (Trommsdorff, 2009, p. 110). Thus, Dichter’s work gave the
first basis for psychological motivation research on emotions and caused the first
rivalry between two research directions.
Overall, the definition of motivation based on consistence of both emotional and
cognitive components where motives differ on the driving intensity of each
1. Motives and consumer motivation
14
component requires research methods and underpinning theories which would
comprise exploration of both cognitive and emotional driving forces.
1.2.3. Common limitations of contemporary motivation theories
Research activity in the area of motivation as an explicit focus for marketing research
has got cyclic intensity having declined significantly after the fisrt half of the twentieth
century (Trommsdorff, 2009, p. 107; Paulssen, 1999; Fennell, 1975). Although it has
gained interest in the last decades and a lot of researchers and authors proposed
their concepts of understanding and explaining the construct of consumer motivation
(e.g., Berkman and Gilson, 1986; Rossiter and Percy, 1991; Fennell, 1978), there is
still no common agreement on this construct reached. Moreover, the theory of
consumer motivation lacks methodological approaches based on its underpinning
theory explaining the impact of motivational mechanism on the purchasing behavior.
Bagozzi, in his article (2003), overviews motivation theories which propose within
their own framework different sets of motives that drive human behavior. The list
comprises theories of such authors as Murray (1938), Maslow (1954), Alderfer (1972)
and McClelland (1987) who suggested categories of motives (Murray) and needs
(Maslow, Alderfer, McClelland) as well as more recent approaches to categorization
of general goals and evaluating them on different dimensions provided by Ford and
Nichols (1987) and Novacek and Lazarus (1990) (see Bagozzi, 2003 for more detail).
In the field of marketing research authors also suggest categories or types of motives
aimed at explaining consumer behavior (e.g., Rossiter and Percy, 1991; Fennell,
1978; Trommsdorff, 2009, p. 114; Vigneron and Johnson, 1999). Thus, the author
points out that to date most concepts of consumer motivation suggest universal sets
or categories of motives, needs or goals which can be applied to any type of
products, consumers and patterns of behavior. These sets are not so general as
goals of highest levels which stand for values and, therefore, can be applied to a
particular purchasing behavior.
However, despite differences among their content these sets are all similar in sense
of being given a priori and specifying rather general motives that are presumed to
apply universally across contexts. This defines their major disadvantage which is
their inability to account for specific actions and to recognize the situational variability
of consumer behavior. Therefore, they appear helpless in deciding on particular
strategies aimed at influencing a specific behavior.
1. Motives and consumer motivation
15
Finally, even if a researcher agrees with a particular concept and tends to apply its
definition and categorization of consumer motives, as already mentioned, most
concepts do not provide certain supportive methodology which would enable to
sample it in a certain investigation of consumers and their attitudes towards particular
products, brands or product categories (Bagozzi, 2003).
1.3. Motives in consumer behavior
1.3.1. Goal-oriented nature of consumer motives
Since many decades marketers try to explore and answer the question why
consumers often behave differently in the same situations. However, the question is
too general to be explored and clearly answered so that it will cover all aspects of
motivation study aroused from its different perspectives. Up to date no universal
answer is found and it is reasonable, in attempts to understand consumer motivation,
to specify some of its aspects and define the direction in which and purpose for which
this question is to be answered (Trommsdorff, 2009, p. 107).
The direction of motivation study conducted in this work is based on the often cited
statement provided by James (1980): “The pursuance of future ends and the choice
of means for their attainment are the mark and criterion of the presence of mentality
in a phenomenon” (cited in Austin and Vancouver, 1996). This statement declares a
purposive and goal-oriented view of human behavior (Paulssen, 1999).
According to Bagozzi (2003) “one of the first ways to specify motives is to identify
which goals are more important to people as ends for their own sake”. General goals
provide a basis for action, however, being rather distal determinants. At the same
time the reasons for choosing a particular goal are represented in a hierarchical
network of motives. Therefore, to be able to more accurately predict particular action
or behavior a consideration of more precise, context-specific motives is required
(Mowday and Sutton, 1993 cited in Bagozzi, 2003).
The concept of direction in this context defines the notion of motivated behavior as a
goal-directed and further in this work goal-directed action or behavior is understood
identically to the motivated action or behavior (Heckhausen et al., 1985, p. 6).
According to Heckhausen and colleagues (1985, p. 7): “motivation consists of the
anticipation of possible actions expected to lead to an outcome that will have certain
consequences, which will possibly bring the person closer to a superordinate goal”.
1. Motives and consumer motivation
16
1.3.2. Hierarchical network of consumer motives
Trommsdorff (2009, p. 108) proposes hierarchical relationship between consumer
motives according to their level of abstraction. He argues that for a practical use
further differentiation of the motive concept is reasonable. It will allow a formation of
at least rough motive classes identified already with regard to specified products,
product categories or consumers.
On top of his hierarchy, Trommsdorff positions ‘too-general’ motives such as, for
example, self-actualization. These motives are too general to explain a consumer
preference towards specific product or brand, for example, how the consumer
discriminates between Mercedes and Porsche and why he prefers one to another.
Therefore, too-general motives on their own are quite useless for marketing interests
and purposes.
On the other hand, on the bottom of the motive hierarchy ‘too-specific’ motives are
disposed. These motives explain very specific purchasing behavior which usually can
be related only to a particular product or brand, for example, “Mercedes-buying-
motive” which explains why Mercedes customer purchases Mercedes and this motive
cannot be used for other products.
To overcome this dilemma, market researchers need to solve two tasks
(Trommsdorff, 2009, p. 109).
First of all, “motives of the middle range” which will balance between too-specific and
too-general motives need to be defined and investigated. Motives have to be sought
that can explain consumer behavior in a broader sense, i.e., enable revelation of
behavior patterns. These motives should be goal-oriented, emotional and cognitive
controlled, relative-universally and theoretically reasonable drivers of consumer
behavior. It is also important that motives of the middle range are meaningful for
different target groups and can explain consumer behavior regarding different
products. That is, even if motives are identified during a research conducted for a
particular product or product category, the level of motive abstraction should, on the
one hand, be sufficient to apply them towards other product categories and, on the
other hand, still be associated with the consumer behavior regarding a certain
class(es) of products. It is hard to define motives that will not be too general and, at
the same time, will be applicable to the consumption of all products a person
1. Motives and consumer motivation
17
purchases because different types of products are considered or purchased due to
different drivers depending on the extent of involvement, product intended
application, extent of conspicuity, etc.
As a second task, a measurement instrument needs to be developed to support
researchers in solving actual problems. Motives cannot simply be educed by
inferences made about consumers’ behavior. In this way hundreds of motives would
be produced without any specific validity. Different criteria or systemized approaches
are needed to methodize motives.
1.4. Summary
♦ The study of motivation represents a search for principles that can help in
understanding why people initiate, choose or persist in specific actions and in
specific circumstances.
♦ Motivation appears from an interaction between affective and cognitive processes
which results in a person’s goal definition and a certain way of acting.
♦ Motives represent latent states which, in case of actualization, drive the behavior
with a certain intensity and in a certain direction.
♦ Motives identify an outlasting, latent disposition, whereas motivation reflects the
process of actualization of motives.
♦ Motives, as results of inner processes, can be unconscious and influence the
behavior outside of awareness, however, they can often be made conscious by
individual own reflection or by means of indirect questioning approaches.
♦ In consumer research a long-standing dispute between qualitative and
quantitative approaches to motivation research resulted in attempts to develop a
compromise approach, however, the true synthesis of motivation research
with quantitative practice has remained obscure.
♦ Most developed concepts of consumer motivation do not provide certain
supportive methodology which would enable to sample it in a particular
consumer investigation.
1. Motives and consumer motivation
18
♦ Most concepts of consumer motivation suggest universal sets or categories of
motives, needs or goals which appear unable to account for specific actions
and to recognize the situational variability of consumer behavior.
♦ The major disadvantage of most ‘motives’ analysis is that they are conducted as
purely atheoretical predictive exercises resulting in an ad hoc inductive
explanation. A theoretical model of motivation and causality underpinning an
active guidance for statistical algorithms is required.
♦ Motivated behavior is goal-directed where reasons for choosing a particular
goal are represented in a hierarchical network of motives.
♦ Goals represent motives at their highest level of abstraction. They are too-
general to explain consumer preference towards a specific product or brand.
However, too-specific motives which explain very specific purchasing behavior
cannot be approximated on other brands or products even within the same
product category.
♦ To overcome the dilemma between too-specific and too-general motives two
tasks need to be solved:
� “motives of the middle range” which will balance between too-specific and
too-general motives need to be defined and investigated;
� a measurement instrument which would provide some criteria or systemized
approach needs to be developed to support researchers in actual problems.
♦ In order to investigate consumer motives for implementation within the practical
marketing objectives a conceptualization of consumer motivational sphere
providing a certain representation of consumer knowledge structure is required
which would be:
� a theory based conception of content and structure of consumer’s mental
representations and their interrelations;
� construed with respect to a methodology underpinning theoretical
assumptions and, thereby, allowing for further practical application by
marketers and researchers.
2. Consumer motivational sphere
19
2. Consumer motivational sphere
The goal of this chapter is to propose a conceptualization of consumer motivational
sphere based on the existing theoretical statements and propositions which will
further be evolved with the methodological approach. Following a brief introduction,
section 2.1 provides an overview of the theoretical background underpinning the
proposed conceptualization. In section 2.2 a model of consumer motivational sphere
is conceptualized.
Exploration of consumer motivation is a key to understanding the psychological
dynamics that determine actual purchase, retention, failure, recommendation and
other essentials of consumer behavior. The major disadvantage of most ‘motives’
analysis is that they are conducted as purely atheoretical predictive exercises
resulting in an ad hoc inductive explanation. In order to explain causal linkages that
motivate consumer behavior a deductive approach in which the stated causal theory
would be tested is necessary. It means that a researcher must begin with a
theoretical model of motivation and causality which can provide an active guidance
for statistical algorithms (Pincus, 2004).
The purpose of this study is to propose a logic construction conceptualization for
consumer’s cognitive structure of knowledge and perceptions and to provide a
methodological guidance for its practical implementation. As recognized by
researchers it is a key factor in attaining a useful understanding of customer
behavior. Such cognitive structures bring to bear a major influence on behavior as
they are essential in explaining the information-processing behavior and, in turn, the
explicit behavior like purchase choice, that is a function of those cognitive processes
(Olson and Reynolds, 1983, p. 77).
In the marketing literature discussions about knowledge structures or cognitive
structures are being held by researchers for a long time. There are numerous studies
devoted, for instance, to product perceptions, brand attitudes, brand images, etc. but
all of them, at the end, refer to kinds of knowledge about products and brands. When
this knowledge is to be interrelated, associated and organized it can be considered
as a structure of knowledge or a cognitive structure (Olson and Reynolds, 1983, p.
78).
2. Consumer motivational sphere
20
Person’s knowledge structure is captured in person’s schemas. Schemas are
“learned, internalized patterns of thought-feeling that mediate both the interpretation
of on-going experience and the reconstruction of memories” (Strauss, 1992).
Bagozzi (2003), in his article, discusses schemas as an appropriate framework for
thinking about motives within which they can be hierarchically organized. He argues
that: “a person’s focal goal in any situation is explained by his superordinate motives
and is achieved through implementation of subordinate goals”. Although his study
was conducted in the context of human motivation in working settings, some of its
notions and assumptions can be taken over in the marketing research.
2.1. Theoretical backgrounds for modeling consumer motivational sphere
2.1.1. Cognitive structures
In economic theory most models are built on assumptions such as the consumer
rationality which implies utility maximizing behavior and the perfect information. To
explain consumer behavior the models use relative prices and disposable
income/budget as explanatory variables and consider other influences such as
social, cultural or quality perception as latent or unobservable variables (Zanoli and
Naspetti, 2002).
In marketing, however, most consumer behavior analysis deal with consumer
perceptions, preferences and the way preferences are represented and structured in
the consumer’s mind. Marketing approaches to consumer behavior can be
distinguished as cognitive versus behavioral (Zanoli and Naspetti, 2002).
The behavioral approaches tend to explain consumer behavior through observation
assuming that our environment is what causes us to behave differently. Cognitive
approaches look at thinking processes and other unobservable activities with the
belief that such processes affect the way in which consumers behave. Thus,
behavioral approaches focus on the consumer visible or observable behavior which
depends on the external stimuli without trying to understand the internal processes
that create it, whereas cognitive approaches are focused on how consumers think
and construe their mental representations, studying consumer product knowledge,
memory, perception, involvement, motivation and other mental and emotional
processes which define consumer behavior (Zanoli and Naspetti, 2002; Foxall,
2. Consumer motivational sphere
21
1998). In its turn, analysis of thinking processes does not imply an ignorance of
external influences such as environmental or social characteristics. It is rather
assumed that a person perceives external influences and information and processes
it mentally acquiring his own construction of world perception. Both approaches are
widely accepted and used in consumer behavior analysis being complemented and
adjusted (Peter and Olson, 1990, p. 27).
This study is concentrated on the cognitive approach and focus on knowledge,
meanings and beliefs that consumers develop from their experience and store in their
memories.
The general concept of cognition implies that the human cognitive system has two
major functions. The first function is to make sense of, interpret or understand
significant aspects of people’s experience in their environment. To do so, the
cognitive system creates symbolic or cognitive representations which represent the
subjective meanings reflecting each person’s individual interpretation of the stimuli in
question. Thus, the second function of the cognitive system is to use these cognitive
representations or meanings to carry out cognitive tasks such as identifying
objectives and goals, developing and evaluating alternative courses of action to meet
those goals, choosing a particular course of action and carrying out the behaviors.
Cognition processes comprise mental processes of understanding, learning,
remembering, evaluating, interpreting, deciding, and planning (Peter and Olson,
1990, p. 45).
Not all cognitive processes are conscious. Consumers often are unaware of their
cognitive processes. However, consumers are often aware of the cognitive meanings
produced by those processes, although still being unable to reason the actions or
perceived meaning (Peter and Olson, 1990, p. 48). For example, a person has a
cognitive meaning or representation of his behavior: “I eat a lot of apples” or of a
symbolic meaning: “this style of watches is appropriate for older people” but he
actually cannot explain or does not realize why he eats so much apples or why he
thinks that way about that watches. That is why consumer behavior represents, in
essence, a consequence of consumption relevant cognitive structure (Grunert and
Grunert, 1995).
2. Consumer motivational sphere
22
Cognitive structures described by authors in different disciplines including marketing
and psychology differ fundamentally depending on the general basis for taxonomy of
cognitive representations. That means it is necessary to define the foundation for
distinguishing between cognitive representations, i.e., on which basis they differ and
what brings them to a certain level of the taxonomy. In this study the cognitive
representations are discriminated and put to a schema according to their level of
abstraction.
Gutman and Reynolds (1979) used as a basis to decide on the level of abstraction of
a certain mental representation its relationship to self. That is, the stronger and more
direct the association between representation and one’s self-concept the higher is
the level of abstraction (Olson and Reynolds, 1983, p. 80).
Another presentation of concept of abstraction was suggested by Geistfeld, Sproles,
and Badenhop (1977) and Cohen (1979). It states that representations that are
relatively direct reflections of physical features of the product (for example, color)
may be considered as concrete or low in abstraction. Representations that are
reflections of several concrete features (for example, style) involve higher-order
meanings and are more abstract (Olson and Reynolds, 1983, p. 80). Product
features at some higher levels of abstraction are usually represented by functional
and psychosocial consequences of product use. In turn, a positive outcome brought
by functional or psychosocial consequences is called benefits or desirable
consequences, otherwise they bring undesirable consequences.
At this point of the concept closer linkages referred to self as also proposed by
Gutman and Reynolds (1979) can be recognized. At the highest level of abstraction
the product can be represented in terms of values that may be reached by its
purchase and use. Such values are very abstract and have much to do with
individual ideas and image of self. Although values stand in the hierarchy associated
with a product, they are several levels away from any physical referent (Olson and
Reynolds, 1983, p. 80).
The taxonomy proposed in this study orders representations hierarchically such as
product characteristics, attributes, consequences and values based on levels of
abstraction connected through means-end linkages. It is built on the assumptions of
means-end chain (MEC) theory and some of its modifications and on the concept of
2. Consumer motivational sphere
23
consumers’ motives proposed by Trommsdorff (2009) described earlier in this
chapter.
The proposed taxonomy is, therefore, based on the concepts which are already
known and have proved to be solid by applications in many marketing as well as
other context studies. For example, Bagozzi and his colleagues (2003) in their study
of hierarchical representation of motives in human goal setting used means-end
relationships to disclose superordinate motives for striving for a personal goal. Similar
approach has been used to generate hierarchies of causal attributions (e.g., Antaki,
1989), goals (e.g., Antaki, 1988; Bagozzi and Dabholkar, 1994; Pieters et al., 1995),
and values (e.g. Bagozzi and Dabholkar, 2000), however, “no studies have
investigated hierarchical schemas for motives” (cited in Bagozzi et al., 2003).
Within the marketing context MEC was typically applied to analyze and interpret
consumers’ product and brand perceptions. Thus, MEC is being more broadly viewed
as representing the relationship between self and products (Walker and Olson, 1991;
Walker et al., 1987). Next section is devoted to the application of MEC in marketing
research.
2.1.2. Means-end chain theory
The means-end chain theory was originally developed in order to relate consumers’
product knowledge to their self knowledge (Gutman, 1982; Olson and Reynolds,
1983, p. 80). In a broader sense it is assumed that knowledge structure represents a
hierarchy where concrete thoughts are linked to more abstract thoughts in a
sequence progressing from means to ends.
Gutman (1982) defines MEC as: “Means are objects (products) or activities in which
people engage (running, reading). Ends are valued states of being such as
happiness, security, and accomplishment. A means-end chain is a model that seeks
to explain how a product or service selection facilitates the achievement of desired
end states.”
In sense of product consumption, MEC represents a simple knowledge structure that
links product attributes to the consequences produced by these attributes (Gutman,
1982). This approach is based on the assumption that consumers see products as
means to important ends. That means that product characteristics can produce
personal consequences which are self-relevant and important for a person. And thus,
2. Consumer motivational sphere
24
the person perceives a product with its certain characteristics as instrumental in
achieving important for him consequences or values (Mulvey et al., 1994). Personal
values are the end consequences produced by the product attributes. They are
mental representations of important life goals that consumers are trying to achieve.
Values are more abstract than functional or psychological consequences and its
achievement tends to be quite subjective and intangible personal experience (Peter
and Olson, 1990, p. 75). Figure 1 illustrates MEC originally consisting of three levels
of abstraction.
Figure 1: Means-end chain original structure
In general, MEC represents consumer’s cognitive structure interrelating consumer’s
knowledge about a product which is in turn defined as a bundle of salient attributes
and benefits and consumer’s self knowledge which consist of personal psychological
and social consequences and values (Mulvey et al., 1994). Thus, the means-end
approach assumes that products are purchased and used not for themselves or their
characteristics but for meanings they produce in minds of consumers (Reynolds and
Gutman, 1988). Although products being selected for fairly concrete features such as
their characteristics and attributes and for the benefits which they are capable of
providing, they are perceived subconsciously as aimed at and associated with the
achievement of personal values (e.g., Solomon, 2004, p. 133).
A number of extensions of the original conception of MEC have received a solid
attention of researchers and being used in conduction of rather fine analysis of the
types of mental representations. One of such extensions was suggested by Olson
and Reynolds (1983, p. 81). They have made finer distinctions in terms of abstraction
by dichotomizing each of three levels of Figure 1 (see Figure 2).
Figure 2: Consumer cognitive structure by Olson and Reynolds
The extended model suggests that each basic level of abstraction can be subdivided
leading to distinct categories of abstraction: attributes into concrete and abstract,
Concrete attributes
Abstract attributes
Functional con-
sequences
Psychosocial con-
sequences
Instrumental values
Terminal values
Attributes Consequences Values
2. Consumer motivational sphere
25
consequences into functional and psychosocial, and personal values into
instrumental and terminal.
Attributes are at the lowest level in the chain. Both concrete and abstract attributes
are relatively concrete meanings that represent physical or perceptible characteristics
in a product (Gengler et al., 1999). Concrete attributes are defined as the directly
perceptible physical characteristics of a product, for example color, size, and weight
(Vriens and Hofstede, 2000), while abstract characteristics refer to relatively
intangible characteristics (e.g., style and image) or perceived value (Botschen et al.,
1999). Botschen, Thelen and Pieters (1999) define attributes as characteristics of
products which may be preferred or sought by consumers, however, they do not
explain per se for what reasons the product is or might be bought. They also
subdivide attributes into concrete and abstract emphasizing that they both give
additional information about the product itself but do not provide discovering of any
underlying reasons why the product is selected or purchased.
Consequences have more abstract meanings which reflect perceived benefits or
perceived risks associated with attributes (Gengler et al., 1999). Peter and Olson
(1990, p. 73) define consequences as specific events that happen to a consumer
when the product is bought and used. This definition seems to reflect a common
understanding of consequences more accurately unlike definition provided by Lin
(2002) which states that consequences are what the customer feels after consuming
the product. Vriens and Hofstede (2000) define consequences as characteristics that
are less directly perceptible in a product or brand and represent the result of various
attributes combinations and the product use by the customer.
In this study it is agreed that consequences may represent emotions, feelings,
evaluations, events (Peter and Olson, 1990, p. 74), and in some situations behaviors
(Valette-Florence and Rapacchi, 1991) experienced by consumers. It is assumed
that differentiation between attributes and consequences lies not in “what” is
expressed by a customer but rather “what kind of information” is underlying in
expression. Adapted definitions for attributes and consequences and distinction
between them are those proposed by Botschen and his colleagues (1999). According
to them, while attributes give additional information about the product itself,
consequences explain why people are looking for certain attributes: “for customers it
2. Consumer motivational sphere
26
is not the product’s attributes which count, but the problem solution - the benefit
sought – which they derive from a certain combination of attributes”.
It is important to distinguish between positive and negative consequences. Positive
consequences or benefits represent the desirable consequences which are the
advantages consumers enjoy from the usage or consumption of products (Gutman,
1982) and when selecting a certain product or brand they are looking for and
expecting to receive these benefits. Negative consequences or perceived risks
represent the undesirable consequences which consumer seeks to avoid when
purchasing and consuming a product (Peter and Olson, 1990, p. 74).
Some marketers consider consumers as thinking about products as bundles of
benefits. That is, in their researches the intermediary MEC consists of benefits
subdivided into two levels of abstraction not taking into account perceived risks (e.g.,
Young and Feigin, 1975). In general, consideration of negative consequences in
marketing research depends on the purpose of the research. For example, for an
efficient product positioning, advertising and distribution it is preferred to develop
benefit segmentation, that is, to divide consumers into homogeneous segments
based on their interest in particular product benefits (Peter and Olson, 1990, p. 74;
Botschen et al., 1999). To further appeal to segments by their benefits sought, this
approach is based on the assumption of MEC theory stating that consumer behavior
is driven by the ‘true’ benefits sought which cause the desire of preference for certain
attributes (Botschen et al., 1999). However, for example, for a successful new
product development and launch it is also important to elicit and analyze consumers’
perceived risks when they select the product. It can help product developers to focus
on the characteristics which cause negative consequences and to introduce possible
improvements (Herm and Gall, 2008).
Olson and Reynolds (1983, p. 81) differentiate consequences on functional and
psychosocial levels. Functional consequences are less abstract outcomes of product
use that act directly on the consumer from the time the product is consumed (Peter
and Olson, 1990, p. 73; Valette-Florence and Rapacchi, 1991). They include the
immediate physiological, physical, tangible, performance outcomes of using or
consuming the product, for example, a car gets a certain number of miles per gallon
or an ink pen writes smoothly without skipping (Peter and Olson, 1990, p. 74).
2. Consumer motivational sphere
27
The term ‘psychosocial consequences’ comprises two types of an outcome
psychological consequences and social consequences (Peter and Olson, 1990, p.
74). Psychological consequences are less tangible and more personal outcomes of a
product use which reflect how the usage of product makes a person feel; for
example, wearing clothes of a certain brand might make a person feel more attractive
or stylish. Psychosocial consequences also include intangible social consequences
which a customer might be aware of, such as “my partners will respect me if I buy
this auto” or “people from my surrounding will consider me as smart and successful if
I buy this house”.
Personal values are at the most abstract level in the chain representing beliefs and
relatively stable cognitions that have a strong emotional impact (Vriens and Hofstede,
2000). Rokeach (1973) suggested subdividing of values into instrumental and
terminal values. Instrumental values are defined as cognitive representations of
preferred modes of behavior which lead to an achievement of terminal values.
Terminal values, in turn, represent final preferred states of existence, i.e., they are
mental representations of the most fundamental goals and end states that
consumers seek in life. As an example for terminal values can be self-achievement,
prosperity, happiness, whereas for instrumental values some authors give examples
such as having a good time, being independent, being resourceful and ambitious
(Peter and Olson, 1990, p. 75; Veludo-de-Oliveira et al., 2006).
Although the authors propose 6-level chain, they emphasize that not all the links
have to be presented in every case. For example, a consumer may be aware of a
product, but yet not knowing its performance consequences of certain attributes or
the valued end state the attribute can help to reach and, thus, the person can
possess little abstract meaning for this product attribute (Olson and Reynolds, 1983,
p. 81). In such case certain levels of abstraction in the scheme might be missing or
the person may express untruly meanings just to satisfy the researcher or pretending
being aware of what he is talking about. Both cases are damaging for the findings,
especially when a group of individuals is analyzed and the data is to be aggregated
to find similarities and differences in representations.
Another problem with such a complex MEC is that it might be difficult to differentiate,
for example, psychosocial consequences and some instrumental values. Peter and
Olson (1990, p. 79) argue that making such fine distinctions is actually seldom
2. Consumer motivational sphere
28
necessary and original three-level means-end structure is sufficient for many
marketing purposes.
It is also pointed that not every MEC leads to an instrumental or terminal value. The
end of the chain can be a consequence at any level of abstraction (Peter and Olson,
1990, p. 80). When product attributes have no associations with consequences or
values, for example, if a person doesn’t know what the attribute is good for or if he is
aware of the consequence but it is not important for him this might have a small
impact on his purchasing decision.
Therefore, at first, it is important to identify the importance of each product feature for
a consumer to produce veritable, not artificial chains linking product characteristics to
values.
In marketing research methods such as repertory grid technique and laddering are
commonly applied to perform MEC analysis. However, the repertory grid technique is
mostly used only to elicit relevant product characteristics as means from which ends
are further revealed through the laddering technique (Jacob, 2009, p. 64). This
approach does not use the full potential of the repertory grid technique thus
diminishing the richness of analysis (Marsden and Littler, 2000a). As a result, the
produced outcome does not show, for example, the difference between
characteristics’ importance and the role of each particular product characteristic or
attribute in consumer’s purchase decision. This analysis identifies product perception
structures, however, it is rather limited in prediction of consumer purchase decision
(Jacob, 2009, p. 68).
In this dissertation the potential of the repertory grid technique is discussed in the
methodological part in order to show how analysis and findings can be enriched by
this versatile tool.
2.2. Conceptualization of the consumer motivational sphere
In this study the general conception of MEC and its original structure consisting of
three major levels of abstraction is adhered. Further extensions are proposed
building on the existing theoretical assumptions and propositions found in the
literature. At the end, the suggested taxonomy is aimed at providing a certain
representation of consumer knowledge structure which, on the one hand, is a theory
based conception of content and structure of consumers mental representations and
2. Consumer motivational sphere
29
their interrelations and, on the other hand, is construed with respect to the
methodology underpinning theoretical assumptions and, thereby, allowing for further
practical application by marketers and researchers. The extensions and modifications
are made with an attempt to overcome some limitations of traditional approaches in
construing and measuring consumers’ cognitive structures and to suggest a
congruent methodology which gives an insight to other aspects of consumers’ mind
than being traditionally explored.
Incorporating the results of the preceding discussion of structuring consumers’
mental representations the schema proposed in this work brings some extensions to
the basic conceptualization. The following Figure 3 illustrates the proposed extended
model.
Product characteristics Product attributes Specific consequences Abstract consequences
"too specific motives" "motives of
middle range"
"too general motives"
MOTIVES
Expected consequences
(desirable/undesirable)Product
Values
Situational and Environmental incentives
Figure 3: Consumer motivational sphere
The most apparent modification lies not in the consistence of components of the
main structure but in the underlying meanings of each component. Three main
substructures of the model can be emphasized.
The first substructure comprises first two levels of abstraction that refer to a product.
Product characteristics which represent product physical characteristics are linked to
product attributes representing product perceptible characteristics (Gengler et al.,
1999). These two levels reflect features offered by a product which are in complex
may offer a person benefits that he seeks for.
The second substructure of the chain consists of three main levels of abstraction
which represent the subjective motivational cognitive structure related to a product
purchasing decision (Trommsdorff, 2009, p. 104). These three levels reflect product
related consequences and values. Consequences are subdivided according to level
2. Consumer motivational sphere
30
of abstraction. At the bottom lay specific expected consequences which are very
personal and hard to approximate in order to reach a common set of expected
consequences. The third, highest level of abstraction reflects values which are also
personal but already very general ends which themselves cannot show what is it
expected from a product consumption that would bring a person or a group of people
to the wished state of being. And in between, the meanings lay that express
consequences on the level which, on the one hand, allows to approximate and to find
commonalities within a group of people and, on the other hand, they are not too
general and can be interpreted in terms of product features and perceptions.
Concerning the issue about positive and negative outcomes, both desirable and
undesirable consequences are implied in the term ‘expected consequences’.
Interviewing technique and assessment methodology further suggested in this work
allow revealing both types of consequences followed by also disclosing consumer
preferences which already indicate desirable outcomes. For this reason, in the
general model the term expected consequences is used and it is, then, depends on
the research objectives whether researcher considers undesirable states or he needs
to consider just consumer preferences, i.e., desirable consequences.
The third substructure consists of only one component which is situational and
environmental incentives which have a certain impact on the choice made by
consumer. People consume products in a variety of situations and environments
(Gutman, 1982; Yang et al., 2002) which certainly effect their attributes preferences
and benefits they seek for. According to Clancey (1989), knowledge is a capacity to
behave adaptively within an environment; it cannot be reduced to representations of
behavior or the environment (cited in Shaw and Gaines, 2005). Numerous
discussions in literature are devoted to the issue of situational and environmental
incentives, their impact on the consumer behavior and how they should be
considered and linked within the main cognitive structure of mental representations.
Gutman (1982) defines product-use situation as “any situation that involves the use
of a commercially available product or service”. He states that consumers learn over
time which choices in given situations produce their desirable consequences and
which do not. He also points that within the time spent by a consumer to adopt with
regard to a given situation the importance of the consequences may adjust. That
means that over time consumers learn in which types of situations which products
2. Consumer motivational sphere
31
they would use. As an example, Gutman considers purchasing of hot dog which can
be consumed during regular lunch, occasional lunch or at baseball game. Each
usage situation would modify the importance of consequences the person seeks for.
For example, if person eats hot dogs every single lunch he may look for the ones that
contain less fat, salt, etc., or he may consider the preparation process and choose
the one that he likes. But if a person buys it at a baseball game, what happens rarely,
he probably would like to have it big and full of ketchup, mustard and mayonnaise to
enjoy its taste. Thus usage situation produces a set of consequences for particular
consumption situation.
Fennell (1978) also discusses the effect of situations in which consumers find
themselves. She gives an example when three different consumers, motorists, ask
for the power however each of them is willing to satisfy different wants by that and so
the consequences of obtaining power differ. One may want power because his
habitual driving pattern makes it necessary to be able to enter fast-moving traffic,
while second one wants to have a powerful car because it reflects his masculine self-
image. And third consumer might enjoy a sensory experience of driving a powerful
machine. This example demonstrates the importance to understand the conditions
that lead the consumer to ask for specific product attributes and consequently to
expect particular benefits. These conditions can be found in consumers’ personality
and their life situation. According to Fennell: “the person is viewed as a unique
composite of genetic inheritance and past experience, now in a current set of life
circumstances (family, job, circle of friends, avocational activities, geographic
location) that are perceived through the filter of his or her individuality”. As long as
person moves from one activity or state to another, the environmental stimuli
interconnect with person systems forming an actual situational perception.
Walker and Olson (1991) discuss the importance of decision situation. They assume
that personal self-schema contains an enormous amount of information about
oneself which is stored in personal long-term memory and only a small portion of
consumer’s self-related knowledge is activated at any given time in his or her working
memory. Only the activated self-meanings can influence consumer’s cognitions and
behaviors. Therefore, it is essential to identify factors that influence which self-related
ends are activated. One of important determinants of self in working memory is the
decision situation. Thus, features and characteristics of situations largely determine
what self-meanings are accessed and, in turn, the abstractness and importance of
2. Consumer motivational sphere
32
these meanings affect how relevant and important that situation is to the consumer
and what product knowledge is activated (Walker and Olson, 1991; Kihlstrom and
Cantor, 1984).
Fennell (1978), proposing her brand choice model, includes internal, relatively
permanent personal aspects like genetic inheritance and past experience and
external, relatively temporary life circumstances in the notion of product-use situation.
Therefore, she views consumer perception of product-use situation as an activator
for product purchase which also directs the choice among available brands. Later, by
modeling variation in brand preference, Fennell and her colleagues distinguish
between personal systems and environmental systems, arguing that intersection of
these components “allocates an individual’s resources to a domain of action and
direction of desired adjustment within that domain”. In turn, the result of such
intersection produces motivating conditions that specify the kinds of attribute an
individual finds valuable (Yang et al., 2002).
In this study, considering described concepts of situational aspect in consumers’
cognition and behavior, it is agreed that consumer’s “core” self which includes
genetic inheritance and past experience as well as personal systems is relatively
stable (Walker and Olson, 1991; Fennell, 1978) and generates basic desirable
consequences. However, there are environmental systems and current life
circumstances, as described by Fennell (1978), which person faces at a given time
that may significantly affect the perception of desirable ends and choice criteria
related to the product attributes. Environmental systems and current life
circumstances are relatively temporary and unstable, i.e., they can be changed when
person achieves another social level, moves to another place, changes family status
or work, etc.
Building on ideas of Fennell (1978) and Walker and Olson (1991) it is assumed that
the expected consequences having origin from “core” self are influenced by rather
temporary situational and environmental incentives. Consumer expresses his
meanings and perceptions already considering these incentives. Therefore, it is
needed to identify where he points these incentives in order to understand better why
and in which circumstances consumer prefers certain attributes and expects certain
benefits.
2. Consumer motivational sphere
33
Overall, the proposed structure of consumer motivational sphere which links product
related knowledge to person related knowledge serves as a basic framework within
which an assessment methodology is to be proposed in this study. An integrated
method for assessing each level of abstraction in consumer motivational sphere
based on data gathered by means of repertory grid interviewing technique is
suggested. Therefore, further in this work each level of abstraction will be discussed
in terms of its disclosing. The suggested methodology underpinning this concept is
proposed as an expansion of the traditional analysis of consumer knowledge
structure in marketing research.
2.3. Summary
♦ The purpose of this study is to propose a logic conceptualization for consumer’s
cognitive structure of knowledge and perceptions and to provide a
methodological guidance for its practical implementation.
♦ A person’s knowledge structure is captured in person’s schemas which represent
learned, internalized patterns of thought-feeling that mediate both the
interpretation of an on-going experience and the reconstruction of memories.
Schemas represent an appropriate framework for thinking about motives within
which they can be hierarchically organized.
♦ Cognitive representations are discriminated and put to a schema according to
their levels of abstraction.
♦ Representations such as product characteristics, attributes, consequences and
values are ordered hierarchically based on levels of abstraction connected
through means-end linkages. The taxonomy is based on the assumptions of the
means-end chain theory, some of its modifications and on the concept of
consumers’ motives suggesting three levels of motives: too-specific motives,
motives of the middle range and too-general motives.
♦ Schema of consumer motivational sphere consists of three main substructures:
� The first substructure comprises first two levels of abstraction which refer to
the person’s product knowledge and links product characteristics with
product attributes representing physical and perceptible characteristics of the
product;
2. Consumer motivational sphere
34
� The second substructure consists of three levels which refer to the
person’s self knowledge:
• Specific expected consequences are very personal and hard to
approximate in order to reach a common set of expected consequences.
• Abstract expected consequences allow to approximate and to find
commonalities within a group of people, however, being not too
general to be interpreted in terms of product features and perceptions.
• Values represent very general ends which themselves cannot show what
is it expected from a product consumption that would bring a person or a
group of people to the desired state of being.
� The third substructure consists of only one component reflecting situational
and environmental incentives which currently influence a consumer’s choice.
♦ Expected consequences have origin from the “core” self which is relatively stable
including the genetic inheritance, past experience and personal systems but they
are also influenced by rather temporary and unstable situational and
environmental incentives.
♦ Product characteristics and attributes give additional information about the
product itself, whereas consequences and values explain why people look for
certain attributes, that is the problem solution or the benefit which they derive
from a certain combination of attributes.
♦ The schema of consumer motivational sphere which links the product related
knowledge with the person related knowledge serves as a basic framework
within which the assessment methodology is to be proposed in this study.
3. Premium car market
35
3. Premium car market
This chapter is devoted to the peculiarities of consumer behavior in the markets of
premium goods in general and in the premium car market. Although theoretical
concepts of this chapter do not refer to the theoretical background of the elaboration
of methodological approach suggested in this work, they still underpin the
methodology in a sense that the method is best appropriate for exploration research
of a relatively small sample. In turn, the sample represents customers whose
consumer behavior is not obvious, often irrational in terms of commonly considered
factors influencing purchasing decision and with its peculiarities which need to be
disclosed.
This chapter presents concepts which consider such factors as social status, social
perception and its role in the consumer decision process and perceived product
value. Section 3.1 is devoted to the notion of ‘premium’ in general and in the car
market. It also specifies differences between premium, luxury and volume car brands.
In section 3.2 different types of consumer behavior in the premium car market are
described. Also in this section product components which form a product utility and a
product value in the premium markets are presented. The last section, 3.3, is
devoted to the role of society in consumer behavior; it presents a man within the
society and his social environment and discusses such aspects as social status, its
symbolic expression and how products may reflect personal attitudes and
aspirations, distinguish the person from others and emphasize personal status and
belonging to certain social groups. In this chapter peculiarities of the consumer
behavior in the Russian premium car market regarding some aspects are presented
in order to give a clue on which base some research interpretations of the empirical
part of the work are made.
3.1. The notion of ‘premium’ in the car market
3.1.1. The meaning of the term ‘premium’
At first, the general understanding of the term ‘premium’ needs to be discussed to
further specify this notion towards automotive market.
The definition of the word ‘premium’ as an adjective in English language was first
indicated in 1926, originally in reference to a grade of motor fuel (Harper, 2001). It
qualifies an object “of exceptional quality or of greater value than others of its kind”,
3. Premium car market
36
i.e., something superior. As an example, it can be spoken of a wine made of premium
grapes and whose quality would be particularly high. From a purely mercantile point
of view, this adjective would characterize an object of higher price or more expensive
(Rundell and Fox, 2007, p. 1167).
The word ‘premium’ as a noun was first introduced in 1601 and referred to a “reward
given for a specific act”. It came from the Latin praemium “reward, profit derived from
booty” (Harper, 2001). Today the word has retained its meaning as a reward, a
bonus but the criminal origin of its sense has evolved and definition of premium is “a
sum of money or bonus paid in addition to a regular price, salary, or other amount”
(Soukhanov et al., 2000). The increase of its price by means of a bonus places the
premium product in the category of products with prices above the average of the
market.
In the context of the relationship between a retailer and a customer it is said that
customers who purchase a premium product are ready to spend more money on it
than they would for a regular product in order to benefit from premium superiority.
3.1.2. Premium, luxury and volume car brands
This section is aimed at providing a clear difference between the notions of premium,
luxury and volume car brands. At first, the distinction between premium and luxury
car brands is discussed.
Premium and luxury car brands
According to the meaning of the term ‘premium’ described in the previous section,
the first characteristic of a premium car brand is the price-premium, defined as a
positive price difference as compared with the market average. The price- premium
can also be defined as a positive difference in prices to the average price on the
global market or on a specific market segment (Diez, 2001, p. 3).
The first characteristic does not help to make a clear distinction of the nature
between premium and luxury brands as a luxury brand also works with a price-
premium and its products are sold at a price higher than the average price on the
market segment. The difference between premium and luxury brands would then
only consist in the amount of the price-premium, i.e., the luxury brands’ products
being more expensive than the premium brands’ ones.
3. Premium car market
37
An essential characteristic of a premium brand on the car market which distinguishes
it from a luxury one is that premium brand has a focus on innovation through which it
is also able to set trends and standards in the industry in terms of technology. This
focal point is communicated through advertising and company performance and
brings the brand a premium image (Rosengarten and Stürmer, 2005, p. 26).
Premium brands aim at being leaders on the car market as far as new technologies
are concerned and care to have top quality products near technical perfection. This
perfection is searched in all tangible functions and properties of a car. A premium car
is a standard car which tends to represent the highest quality, especially oriented
towards the functions which it is intended to fulfill (Diez, 2001, p. 11). These
innovations can contribute, for instance, to a better security of passengers or to an
improvement of the road holding of a car in difficult driving conditions. This, however,
does not mean premium brands do not care about the aesthetics of a car and its
features but the aesthetics is subordinated to the functionality.
On the contrary, luxury brands bet on sensational features and obtain their status
either from their heritage related to the respect of brand tradition or from a very
strong marketed image. Thus, the idiosyncrasy of luxury brands is the focus on the
increase of sophistication-related features which are outside of rationality rather than
on the permanent improvement of features related to the primary function of a
product (Rosengarten and Stürmer, 2005, p. 26). As an example, Rolls Royce does
not exactly represent a model of technological and ergonomic perfection but its
luxury image relies more on its aesthetics and the legendary history.
While luxury car brands rely mostly on the strength of image, premium brands can
rely on two focal strengths: an image of a high-class product and a high innovation
potential (Figure 4). This enables premium brands to target more potential customers
than luxury brands as they can resort to more sales arguments and of different
natures (Rosengarten and Stürmer, 2005, p. 27). Today, though, when customers
are relatively experienced and aware of the existing technologies the distinction
between luxury and premium becomes vaguer as luxury brands cannot keep the
competitiveness without advancing in the technological and functional perfection.
However, the essential difference between the notions can be further supported as
luxury brands may often adopt most recent but already developed technologies,
3. Premium car market
38
whereas, premium brands concentrate on innovations in order to appeal to
customers as providing leading and new technologies.
Figure 4: Public coverage by premium and luxury brands
When talking about luxury brands it is sometimes important to bear in mind that there
is a certain difference between considerations of luxury on the European and
American car markets. On the European market luxury is based on the refinement in
car details and on the brand history, whereas on the American market perception of
luxury is based on comfort and size of a vehicle (Rosengarten and Stürmer, 2005, p.
27).
This is explained by the peculiarity of the country which is the large territory and,
therefore, long distances which influence auto exploitation and general car culture
including perceptions and attitudes towards notions of premium and luxury.
In America large territory of the country and, therefore, long distances influence auto
exploitation and general car culture. Consumers tend to buy big cars in terms of size
to make long rides. Therefore, people put a lot of value on comfort and space in their
cars.
As a result, car classifications used in Europe and North America differ especially
regarding entry-level luxury segment (see Appendix 1 for most commonly used car
classifications). However, today, car classification is rather subjective because many
vehicles either fall into multiple categories or do not fit well to any.
Peculiarities of Russian car market when compared with American and European
are closer to the American in sense of territory and distances, variety of climate
conditions as well as consumer attitudes towards size, space, comfort and consumer
3. Premium car market
39
exploitation art implying often long rides. The Russian or, better to say, Soviet
classification of cars applied to the Russian autos does not correspond to the
contemporary foreign classifications. At the same time, there are no autos produced
in Russia which are considered as premium or luxury. Some companies attempt to
position their brands as of an upper class, however, technologies and design are still
far from the world progress and, therefore, it will probably take long time until
customers of premium and luxury autos will want to buy domestic cars.
Premium and volume car brands
The notion of premium referred to car brands also enables to differentiate premium
car brands from volume car brands. As their name indicates, volume brands aim at
selling as much products as possible in order to increase their revenue. Thus, they
position themselves on the market with relatively low prices. To enable this they
pursue low-cost production process reducing in the first ranks research and
development costs. That is, volume brands do not innovate as much as premium
brands and benefit from innovations of the latter, for instance, by adopting them with
a delay. However, today, the delay tends to reduce itself as evolution in the car
market is becoming faster resulting in shortening of products’ life cycle. Doing that,
volume brands cannot display a status of technological leadership. They also do not
have the same care concerning quality of materials and precision of the finish. This
constitutes the main distinctions between premium and volume brands. A common
point, though, is their focus on the improvement of a car in its functional use
(Rosengarten and Stürmer, 2005, p. 30).
3.2. Consumer behavior in the premium car market
In this section reasons for consumption of premium products with the focus on the
consumer behavior in the premium car market is discussed. Further, in this study, the
notion of premium will be used implying both premium and luxury cars in a sense that
they both relate to the price-premium car segment, possessing both quality and
image components on a higher level comparing to the volume brands. Thus,
purchasing behavior towards price-premium cars is opposed to the purchasing
behavior towards volume brand cars rather than purchasing behavior towards
premium is opposed to the purchasing behavior towards luxury.
3. Premium car market
40
The aim is to disclose what actually attracts consumers in the premium segment in
order to define the strengths of premium brands and how they can communicate on
them. Specifically, it is aimed at showing that even if objective reasons play a
significant role in the process of customer decision making, there is also a great role
of subjective reasons, and notably a search for prestige, which can explain the
preference of consumers for a premium brand rather than for a volume brand.
3.2.1. Types of consumer buying behavior
One of the principal categorization of consumer buying behavior, displayed in the
literature is a categorization according to the degree of cognitive buyer involvement.
It implies that decisions with a comparatively high degree of cognitive involvement
can be distinguished from forms with comparatively low cognitive involvement
(Assael 1995, p. 152 cited in Esch et al., 2003). This results in four types of
consumer buying behavior (Esch et al., 2003).
1. Routine decision making implies buying decision with low cognitive and
emotional involvement which is made neither with effort nor considering. It is a
widely automatic process without assimilating much of external information.
This decision has been solved before as a result of learning processes from
repeated behavior. This type of decision is made towards frequently
purchased, low-involvement and relatively low cost items, for instance,
newspapers, milk, snack foods, etc.
2. Limited decision making implies buying decision characterized by simplified
cognitive processes. This decision is taken when consumer is familiar with a
product category but not familiar with a brand or doesn’t show any clear
preference for a certain brand. The customers are assumed to be emotionally
low involved and hence as soon as an alternative satisfies the needs, the
consumer terminates the decision making process and purchases this
alternative, for example, clothes in case of unfamiliar brand.
3. Impulse buying behavior implies a decision making process in which
customers are only little cognitively involved but usually show a high degree of
emotional involvement. It occurs when consumers feel an unexpected, sudden
and urgent need to buy. Thus, impulse purchases are unplanned and without
an evaluation of need. They often occur in favorable situations with strong
3. Premium car market
41
stimuli like POS (place of sale) displays. This type of decision making usually
relates to such categories of products as clothes and delicacies, however, it
also depends on the level of income consumers dispose of.
4. Extensive decision making implies customer’s both cognitive and emotional
high involvement. This decision making requires a comparatively long period
of time for seeking wide information and deciding between alternatives which
are thoroughly examined and evaluated. This type of behavior is followed
during infrequent purchases of expensive products often with high risk degree
such as housing, car, computer or education.
The type of occurring decision making process in concrete purchase depends on
several factors which are: price of product, frequency of product purchases,
consumer involvement and the perceived risk related to the purchase. However, in
practice these four types of decision making do not usually occur in their pure form.
Depending on situation and circumstances some combinations and overlaps are
likely, and other factors may appear to be important. Even the purchase of the same
product in different circumstances may elicit different types of buying behavior.
Basically, product consumption is aimed at satisfying certain consumer needs. In
order to answer the question why consumers choose a particular product it is useful
to turn to the concept of utility.
Concept of utility developed by Vershofen (1940) expresses an individualistic
measure of need satisfaction (Vershofen, 1959 cited in Orth and Krska, 2002). Its
main assumption states that total (product) utility is made of partial, i.e., attribute level
utilities. In order to offer a proper product marketing managers need to know what
partial utilities product attributes (e.g., design, price) and levels on which these
attributes are performed contribute to an overall product utility (Orth and Krska,
2002).
Two types of partial utilities are recognized ‘basic’ and ‘additional’ (Figure 5). Basic
(functional) utility of a product is derived from its physical-chemical-technical
attributes, either concrete or abstract (Bauer, et al., 1998) and refers to the technical-
rational intended purpose. Additional utilities aimed at satisfaction of any need which
exceeds the actual use purpose, they relate to a capacity to impart psychological
added value on a person in oneself (Haury, 2007, p. 44). Essentially, additional
3. Premium car market
42
utilities of a product are designed to satisfy the craving for social recognition of the
owner. Cars relate to the type of products which are characterized as socially
remarkable products for which sufficient performance of additional utilities is
particularly important.
Basic utilities Fulfillment of technical-functionally
intended use
Additional utilities Fulfillment of socio-psychological
intended use
Edification utilities Satisfaction of needs originated
from personal images
Esteem utilities Satisfaction of needs originated
from social interaction
� Social affiliation integration
� Social distinction
� Social recognition, prestige
� Aesthetic needs
� Self-realization
� Well-being
Product utility
Figure 5: Components of product utility by Vershofen
(own interpretation from Vershofen, 1940 cited in Haury, 2007, p. 92)
In its turn, additional utilities are sub-divided into ‘edification’ and ‘esteem’ utilities.
Edification utilities serve to provide personal emotional use images such as aesthetic,
self-realization and well-being. Esteem utilities serve to satisfy more socially oriented
emotional use images such as social affiliation and integration, social distinction,
recognition and prestige. Product benefit perceived by consumers arises from the
sum of all product utility components.
When considering premium products, the fulfillment of needs provided by additional
utilities comes to the forefront and withdraws the competitive role of basic utilities.
The progressive saturation of material basic needs causes an increase of the relative
importance of social consumption motives such as striving for prestige and status.
Therefore, needs for self-esteem, recognition and respect represent important factors
by the purchasing decision.
3. Premium car market
43
Following Vershofen’s distinction between components of product utility, Belz (1994)
emphasizes two need categories which refer to the consumption of premium and
luxury products: the need for prestige and the need for self-esteem.
The need for prestige is based on the consumer crave to obtain own position within
the society through owned products and visible consumer behavior. Here, the
product fulfills a staging function demonstrating symbols of the position a consumer
possesses or wants to possess. The need for prestige represents, therefore, a
socially-oriented need and is satisfied by a high esteem utility of the product.
The need for self-esteem is based on personal aspiration to consume products which
correspond to and underline own personality. The consumer prefers products by
which means he can identify himself and which embody his values, becoming
symbols of a life position or of an attitude to life which consumer the considers as his
own. This need represents a personally-oriented need of the consumer and is
satisfied by a high edification utility of the product.
According to the given definitions premium and luxury cars are similar in terms of
their higher price than on the volume brand cars which show comparable tangible
functions. The difference between premium and luxury brands and products implies
focus of premium brands on technological excellence, innovation and ability to set
the trends and standards (Rosengarten and Stürmer, 2005, p. 26). However, this
differentiation does not deny importance of emotional and image perception of
premium brands. Moreover, premium in American classification refers to the entry
level of luxury cars (see Appendix 1) and, therefore, more similarities can be
expected in consumer behavior towards premium and luxury cars rather than
differences. When considering luxury products the concept of utility focuses on the
additional utilities which correspond to the fulfillment of socio-psychological needs.
Thus, the concept underestimates the possibility to appeal to a customer by
technological achievements and innovations referred to the basic or functional
utilities according to their definition applied in the concept. For this reason another
approach to explore the perceived product value from the consumer standpoint is
also considered in this work.
3. Premium car market
44
3.2.2. Perceived product value
According to the extent and the way a product can satisfy individual needs a
consumer perceives the product more or less valuable. From consumer’s standpoint
there are three elements that compose consumer’s overall perception of product
value: the prime value, the labor value and the symbolic value (Karmasin, 1998, p.
249).
Technology and material
Production process and country of origin
Labor Value
Symbolic Value
Psychographic value of the brand
Prime Value
Figure 6: Components of total product value
• Prime value stands for the value of a product determined by its performed
level of technology and materials used.
• Labor value reflects the value of a product resulting from its production
process. In a narrow sense, an applied method of product production is
considered. In a broader sense, the place where a product is made, especially
a country of origin, also belongs to the labor value. Very often consumers
have particular judgements about a product according to its place of
production.
• Symbolic value consists in the emblematic signification of a product.
Predominantly brands embody some signification and the value of a particular
brand results from its anchoring in consumers’ minds.
On automotive market these three types of values, when being added or
strengthened, allow to explain why consumers are ready to pay a higher price for the
premium brand car rather than to buy a volume brand car if both are assumed to be
comparable products in the sense that they fulfill the same tangible functions being
both means of conveyance.
3. Premium car market
45
Considering prime value of a car, if the product affords innovative technologies that
improve its attributes from the consumer’s standpoint then it increases the perceived
value of the product. These technologies can, for instance, improve comfort, driving
performance or reliability of the car. This is also true for the quality of the materials
used in an auto design, i.e., the use of high-class materials which can be perceived
by the consumer can lead to an increase of the product value from the consumer’s
point of view and, thus, increase his willingness to pay.
Labor value in automotive industry presents a significant component in overall
perceived product value. The origin of a product is an important influencing factor for
the value appeal, especially for products which production is perceived as complex
by consumers. On the automotive market the so called “country-of-origin-effect”
represents an important issue that one must not underestimate (Hausruckinger and
Helm, 1996). For example, often people considering cars produced by German or
Japanese companies already have prejudicial view concerning its high level of
quality. The same occurs regarding different product categories such as, for
example, Swiss watches or Swiss banks, however, in some cases place of origin
may impart negative value to a product. Another source that can also bring additional
labor value is a method of production. In automotive industry it is a particular issue to
explicitly denominate products which have benefit from a special care (e.g.,
handmade finish) in the production process which gives a more noble impression to
the consumer and, thereby, produces a greater value.
The symbolic value of premium brands is stronger than the average, i.e., the one of
volume brands and, moreover, the lower the potential for an increase of the prime
and the labor value, the stronger the symbolic value is. Thus, in relatively simple
product categories such as T-shirts or sunglasses the premium price is essentially
based on the symbolic value. On the contrary, on the automotive market there are
certainly possibilities to increase product value due to technology, quality and
production process. However, the strength of the symbolic value is also undisputed.
An illustration of this is the consumers’ willingness to pay for almost identical cars in a
sense of functionality, equipment, body structure, etc. sold under different brand
names. This reflects the importance of the symbolic value for the general value
appeal of autos (Diez, 2001, p. 8). The strength of the symbolic value is based on six
factors: creative innovation, a clear brand code which enables the brand to
3. Premium car market
46
differentiate, attractiveness, credibility through authenticity, mastery of the elite codes
and continuity.
Overall, premium nature of premium brands and/or products is based on the above
the average perception of prime, labor and symbolic values by consumers. Thus, the
premium attribute is not an objective feature which comes with the product in itself
but rather a subjective attitude perceived by consumers.
3.2.3. Reasons for purchasing premium cars
Another approach to explain consumer purchasing behavior in the premium car
market described by Diez (2001, p. 14) distinguishes between two types of purchase
reasons, on the one hand, reasons related to the product in itself and to the function
it fulfills and, on the other hand, reasons based on the personal benefit in terms of
image and self-fulfillment which consumer pursues to obtain by owning the product.
Figure 7: Reasons for purchasing premium brand products
Reasons related to the product in itself and to the function it fulfills can be
schematized as follows. Consumers who opt for a premium brand car want to have a
qualitatively high-class product and this because they want to make a profitable
investment, reduce purchase risks or benefit from a technologically and aesthetically
well executed product (or all of these reasons at the same time).
3. Premium car market
47
• Make a profitable investment:
Premium brand products are often considered as more reliable and more
durable than volume products. And that is particularly true in the car market
where premium brand cars definitely have, for instance, a higher residual
value. Therefore, buying a premium brand car with a higher price is not seen
as a waste but rather as a rational decision based on economic realities.
• Reduce purchase risks:
Purchase of a premium brand car can also be the sign of a search for security.
In the opinion of lots of consumers buying a premium brand car prevents from
technical-qualitative risks as well as social risks as premium brands have
originally a good reputation.
• Benefit from a technologically and aesthetically well executed product:
Premium brand products represent products which received a lot of care from
the car makers during the development phase and, therefore, are sometimes
seen by consumers as works of art and trigger enthusiasm.
Reasons based on the personal profit in terms of image and self-fulfillment, comprise
two main outcomes which consumer obtains with the product ownership. These
outcomes include the pursuit of social prestige and the pursuit of a meaning of life.
These social functions are seen as major purchase reasons in the premium car
market. It corresponds to a consumer’s wish to assert himself in the society and not
to mingle with the crowd.
3.3. The role of society in consumer behavior
3.3.1. Consumer society
Certain tendencies inherent in our modern society make it reasonable and
meaningful to define our society as a consumer society. Firstly, with the world
globalization the market has become saturated with a variety of products where
people being a part of the economic process cannot avoid consumption activities
(Feemers, 1992, p. 15).
3. Premium car market
48
Secondly, during the period of a dynamic movement between social classes an
earlier accepted traditional system of status is no longer presented in most of the
countries. Thus, the traditional symbols of status like titles, medals, uniforms, etc. are
either lost or maintained rather formally without giving actual prestige value to the
owner within the modern perception of the social status and prestige. With this
tendency people miss the feeling of being certain and secure in their social status,
whereas the human ambition, nevertheless, retains the desire for social classification
where a person can fulfill his ambition to social appreciation and differentiation
(Rinsche, 1961, p. 127; Kluth, 1957, p. 10). Thus, in the period of growing disposable
income and reduction of traditional social boundaries, economic goods attained the
quality of status symbols reflecting individual performance and success as well as
being an instrument for satisfaction of a need for recognition (Rinsche, 1961, p. 129).
Also, the population agglomeration resulted in the relative anonymity of an individual
within the society. For example, the professional performance of an individual is
rarely visible outside the firm. Therefore, individual success and material prosperity of
a person who is anonymous to the majority of his fellow men can be socially effective
only if it becomes outwardly visible (Rinsche, 1961, p. 130).
For these reasons the act of purchase and consumption is no more neutral but a
particular form of social behavior, especially regarding products which have public
resonance, i.e., which have an added benefit of recognition next to their original
intended function. This explains difficulties and the complexion of determination of
consumers’ purchase reasons and motives which are originated from very different
natures and not obviously rational.
A man as a social entity is influenced by his environment what defines the concept of
socialization. Socialization is a process of a mutual influencing between a person and
his fellow men, a process that leads to an acceptance and an adaptation of the social
behavior patterns (Fichter, 1968, p. 23).
3.3.2. A person within the social environment
People are influenced, consciously or not, in their acts and decisions by their
environment, that is, they live in relationship with other fellow men or belong to circles
of people. For example, a person works in a team, can be a part of an association or
a part of a social class. All these circles of people can be defined as groups which
3. Premium car market
49
have an impact and influence on the life and behavior of a person who belongs to it.
The groups differ from one another according to their size, stability, structure, level of
organization, goals, function and the way they integrate new members (Feemers,
1992, p. 21). Groups bring together people who interact with one another and who
share similarities on essential points. Inside a group there are norms and values
which can be formal or informal but which influence and direct the behavior of the
members. These norms are not motionless, they evolve with the time.
The groups to which a person refers are called reference groups. The norms
established in the group very often include consumption norms. These norms enable
to differentiate members of the group from people who do not belong to this group.
Therefore, when a person wants to be associated with a group, he needs to present
the same attributes as other members. This behavior represents the search for
conformity which can bring the person the recognition of the group and assert his
belonging to the group. The conformity has to be strong but not total so that the
person keeps his individuality. Thus, consumption norms should be considered as a
frame which indicates what kind of products corresponds to the group but the
consumer remains a master of his decision and can demonstrate his individuality
while respecting the norms of his group.
Another type of the influencing circles of groups is called aspiration groups.
Aspiration group is a group to which a person wants to belong. As a result, the
person orients his consumption towards the standards of that group, especially
regarding visible goods. This represents a conspicuous consumption aimed at
gaining the recognition of the society (Wiswede, 1995, p. 324; Beckmeier-Feuerhahn,
1998, p. 133).
3.3.3. Social status
Social status, according to the theory of a social stratum, describes classified
positions in the social hierarchy. The higher the associated appreciation and
recognition of the person’s position by others, the higher his social status (Adlwarth,
1983, p. 36). Since the ‘internal’ characteristics of a person such as intellect, values,
etc. are difficult to evaluate at first sight, the association of a person with a social
status is based on the evaluation of exterior characteristics that are immediately
visible and can be appreciated. Thus, criteria for a social status can be a family
lineage, education, profession, income, standard of living, etc. In most of the
3. Premium car market
50
developed countries the consistency of status demonstrates a high correlation
between these characteristics of a person. In order to make these characteristics
visible for everyone in the society, the person turns to symbols which stand for
denotation of his social status or the status he strives for.
However, sometimes inconsistency of status may occur when people, for example,
have a high income but have a lower standard of education or are regarded as
having an inferior job. According to the theory of cognitive dissonance, these people
try to compensate their incongruity of status by demonstrating strong and often
overreaching consumer conspicuous behavior (Wiswede, 1973, p. 150). This kind of
behavior explained by status inconsistency is very common in Russia, where the
economic and political situation of the last decade of 20th century gave a lot of
opportunities for rapid enrichment of people who were rather ambitious and very risky
than educated and professionally experienced. A lot of young people of that
generation decided in favor of opened business opportunities rather than education
what caused later a feeling of deficiency of intellectual capital.
Achievement of a high social status is accompanied by obtaining a prestige which is
a person’s respect, reputation and recognition provided from his social environment
(Schuster, 1994, p. 108). Prestige represents a strong social motive which stimulates
the behavior aimed at gaining appreciation by other people (Trommsdorff, 2009, p.
117). Therefore, the motive of prestige influences consumer behavior regarding
noticeable purchases which can be perceived by others.
The prestige value pertains to the products of a high quality or exclusive on account
of their ‘prestige pricing’. Therefore, a person can be recognized and identified as a
member of a certain (high) social class through the prestige value of products or
property he possesses. Prestige value, especially provided by product exclusiveness,
is to a large extent based on and associated with the product positioning. Products
with prestige value stand for symbols of a social status and, therefore, their decisive
characteristics are visibility and comparative and absolute higher price (Veblen,
1981, p. 24). By purchasing premium and luxury goods the consumer aspires to a
higher status and prestige and tries, on the one hand, to distinguish himself from the
lower class and, on the other hand, to belong to the upper class (Mason, 1981, p.
146).
3. Premium car market
51
After 1900 autos have become a typical example of a status symbol along with
places for vacation, watches, wine, clothes, etc. Since autos are used in the public
area, they can easily be perceived by others in the society, i.e., they represent a
great heraldic product with a high social visibility (Diez, 2001, p. 17). Moreover, the
auto’s brand name is an integrated component which can immediately be seen and
perceived by observers.
3.3.4. Symbols generated by premium brands
As described above, together with the instrumental function premium brand autos
have a social function which enables to provide the consumer, firstly, a social
prestige and, secondly, a sense of life he aspires to.
Brands and, in particular, premium brands can become some kind of objects of
devotion which then represent landmarks in the consumer’s life towards which the
consumer orients his consumption behavior (Lucerna, 1999, p. 359). As soon as the
person can afford this ‘object’ he becomes gratified and gets a sense of
accomplishment in reaching or becoming closer to the way of life he wants to live.
Premium brands generate symbols through which a consumer obtains social prestige
and acknowledgement of his social status. The symbolic meaning can reflect
different aspects of an individual.
• Premium brands as a symbol of wealth and power:
According to their definition, premium brand products are products with prices above
the average and, moreover, relatively high prices. Thus, especially regarding
premium brand autos not everybody can afford them. For this reason premium
brands work as a symbol of wealth and power. The disposal of resources is the
classical basis of social prestige in the modern society which is typically hierarchically
structured (Diez, 2001, p. 18).
• Premium brands as a symbol of taste and connoisseurship:
Other sources of social prestige are taste and connoisseurship. According to
Bourdieu, it implies instead of standing out by the financial well-being a
demonstration to the society that one has a “cultural capital” (Bourdieu, 1982 cited in
Diez, 2001, p. 20). The cultural capital is the ability of a person to stylize his life, to
3. Premium car market
52
search for aesthetics in every part of his life even in the objects of the everyday life.
Premium brand autos fully refer to the category of products that can provide
aesthetics as their designers try to pay attention to every detail and to make of each
auto a high-class product in terms of quality and design. Connoisseurship results
from taste and can be defined as an ability to differentiate high quality products from
average products. This is a knowledge which only experienced consumers can
usually have what gives them the ‘entrance ticket’ to a selected circle of people and,
thereby, brings respect, recognition, i.e., social prestige.
• Premium brands as a symbol of a lifestyle opened to an experience:
Another source of prestige in the society is the quality of life. By seeking for this
symbol people want to show others that they have a happy life full of success and,
more important, full of experience and adventure. This is demonstrated by the
ostentatious consumption of products with an ‘experience value’ (Schulze, 1992, p.
186). That is, premium brand products, especially autos, represent high-class
products with an orientation towards innovation and performance of features that
have never been seen. Thereby, premium brand products stand for challenge and
creativity and bring the consumer this ‘experience’ image resulting in the acquiring of
social prestige.
3.3.5. Aspirational brands
In consumer marketing a brand or a product is premium, i.e., can command the price
premium in the marketplace over a commodity (volume) brand until it refers to the
category of aspirational brands (or products).
Aspirational brand (or product) means that a large part of its exposure audience
wishes to own it but at present cannot afford it for economical reasons. The exposure
audience includes the audience which wishes to own the brand but can never afford
it and the target audience of the aspirational brand (Figure 8). In its turn, target
audience comprises brand’s aspirational audience and consumption audience.
Aspirational audience represents the part that cannot purchase the brand or the
product today but thinks of itself as having a fair probability of at a certain point in the
future being able to do so. Consumption audience represents the part that already
can afford the brand (or the product) (o. V. 2008).
3. Premium car market
53
Figure 8: Aspirational brands audience
As a general rule, the smaller the size of the product's consumption audience
compared to the exposure audience, the more the product satisfies the consumer’s
need for conspicuous consumption, and the higher the premium that such a
consumer is prepared to pay. Besides, the larger the ratio of aspirational to
consumption consumers in the target audience, the higher the brand's price premium
is. To keep the premium level of a brand high, the portion of consumption audience
should not exceed 30% of the aspirational audience (o. V. 2008).
Becoming an aspirational brand is important to make it perceived differently from
competitive offerings and, moreover, to acquire consumers’ loyalty which provides
stronger resistance to change purchasing decisions towards competitive brands.
Overall, nowadays, instead of vertical differentiation between social classes, the
society is prone to the horizontal differentiation between various lifestyles which are
in a large extent reflected in the consumer behavior (Schuster, 1994, p.111). People
use the perceived features of consumed products to demonstrate their personality
and lifestyle and expect that the valued product characteristics are projected onto
themselves (Adlwarth, 1983, p. 49). By purchasing premium and luxury goods people
want to benefit from their functional and quality superiority as well as from social
perception of goods which results in social prestige and self-esteem.
Therefore, in order to be successful, premium brands need to provide their
customers, on the one hand, with the high-quality and innovative products and, on
the other hand, with the social prestige. That means that the products must have a
positive resonance among the society and people who own such products should be
associated with the socially perceived attributes of the brand.
3. Premium car market
54
3.4. Summary
♦ Customer who purchases a premium product is ready to spend more money on it
than he would for a regular product in order to benefit from its premium
superiority.
♦ The progressive saturation of material basic needs causes an increase of the
relative importance of social consumption motives such as striving for prestige
and status.
♦ The total product utility is made of partial, i.e., attribute level utilities.
♦ There are two need categories which refer to the consumption of premium and
luxury products: the need for prestige and need for self-esteem.
♦ From consumer’s standpoint there are three elements that compose consumer’s
overall perception of the product value:
� Prime value stands for the value of a product determined by its performed
level of technology and materials used.
� Labor value reflects the value of a product resulting from its production
process, applied method and place of production.
� Symbolic value consists in the emblematic signification of a product
emphasizing personal status.
♦ Premium attribute is not an objective feature which comes with the product in
itself but rather a subjective attitude towards prime, labor and symbolic values
perceived by consumers.
♦ Personal profit in terms of image and self-fulfillment results in the pursuit of
social prestige and pursuit of a meaning of life. These social functions are
evaluated as major purchase reasons in the premium car market.
♦ The act of purchase and consumption is no more neutral but a particular form of
the social behavior, especially regarding products which have public
resonance, i.e., which have an added benefit of recognition next to their original
intended function.
3. Premium car market
55
♦ Reference group is a group to which a person belongs. It establishes norms
including consumption norms which enable to differentiate members of the
group from others.
♦ Aspiration group is a group to which a person wants to belong. As a result, the
person orients his consumption towards standards of this group, especially
regarding visible goods. It represents a conspicuous consumption aimed at
gaining recognition of the society.
♦ Together with the instrumental function premium brand autos have the social
function which enables to provide a consumer, firstly, with the social prestige
and, secondly, with the sense of life he aspires to.
♦ Premium brands generate symbols through which a consumer obtains the social
prestige and acknowledgement of his social status.
♦ Premium brand or product refers to the category of aspirational brands or
products which implies that a large part of its exposure audience wishes to own it
but at present cannot afford it for economical reasons.
♦ To be successful, premium brands need to provide their customers with high-
quality and innovative products and with the social prestige.
4. Repertory grid technique within the personal construct psychology
56
II. Methodology
4. Repertory grid technique within the personal construct psychology
This chapter presents a technique which provides an interpretive research framework
which format of gathered data and its further analysis enables to disclose each
component of the model suggested in the first part of the work. An introduction to
methodology is followed by the personal construct theory underpinning the repertory
grid technique. Afterwards, parts 4.3 and 4.4 provide a basic description of the
technique and its origin and also discuss a solution for some criticism as a result of
various experience of the technique application. Part 4.5 thoroughly describes a
procedure of a grid elicitation considering its numerous variations. Next part provides
different methods for analyzing the data elicited by the technique. It shows how both
qualitative and quantitative approaches can be used. This part ends with the
proposed way of application of existing methods so that a certain synergy effect can
be reached in structuring and presenting the information. The last part 4.7 is devoted
to the questions of reliability and validity of the repertory grid technique and to the
results revealed by its application.
4.1. Introduction to methodology
Individual perception partiality, life stands and experience cause a diversity of
perceptions of an investigated object. Therefore, a problem of describing this specific
psychological reality – subjective picture of the world arises.
For uncovering subjective mental representations with respect to the theoretical
conceptualization, first of all, a proper data collection technique is needed that
enables to extract these underlying, conscious, subconscious and often unconscious
meanings with a minimum of interviewer bias. Besides, a research of cognitions and
underlying meanings requires the use of a methodological technique that would be
able to capture respondents’ mental representations without a bias of predetermined
questions on a questionnaire that may also influence and frame their responses. To
really bring to the surface how people see, interpret and make sense of their worlds
using their own “theories in use”, in their own language and within their own
environment the research needs to go further beyond the supplied and
predetermined questions of a survey instrument (Wright and Cheung, 2007). The
4. Repertory grid technique within the personal construct psychology
57
whole point, as expressed by Fennell (1985), is that: “at base, the marketer’s task is
not to assess perceived value in objects that already exist but to identify the
antecedents of perceived value so that valuable offerings can be fashioned,
produced and made available for sale” (cited in Saegert and Fennell, 1991).
There are certain limitations in the application of a questionnaire survey. The problem
of impropriety of using this traditional method for solving a whole number of problems
facing researchers is not new. Yadov (1991), famous Russian scientist in sociology
and psychology, emphasizes a list of reasons why a researcher has to refuse
questionnaire survey in favor of methods related to the qualitative paradigm. One of
his statements is that quantitative methods are good to prove already beforehand
formulated hypotheses. But when a research concerns peoples’ conceptions about
complex things and phenomena, their attitudes and motives and etc., the
interpretation of responses based on offered beforehand formulated multiple-choice
questions won’t adequately reflect the whole variety of meanings projected in them.
The researcher induces a respondent to use system of meanings, which can be
partly or even fully not concurrent with the respondent’s one. It should be also taken
into consideration that in different systems of conceptions and relations the same
attitude or the same motive can have opposite meanings. This often causes
misleading interpretations, “noisiness” of investigated components and future
development forecast errors.
The repertory grid (RG) is a cognitive mapping instrument designed to capture
dimensions and a structure of personal meanings. It is aimed at eliciting and
describing the ways in which people give meanings to their experience in their own
terms, in their own language. Its methodology allows a researcher to go deep into
respondents’ theories in use to provide a new insight on how they think and perceive
life experience and things around. The technique implies the qualitative approach for
data collection designed so that gathered data is appropriate for both qualitative and
quantitative analysis methods (Marsden and Littler, 2000a). Therefore, RG
represents a versatile tool where an inquiry design and further applied analysis
methods are very much determined by research objectives.
Studies conducted by means of RG are considered qualitative in a sense that they
are aimed at eliciting and exploring personal meanings rather than to testify
predetermined hypothesis about personal representations. Although theoretically the
4. Repertory grid technique within the personal construct psychology
58
qualitative approach represents an alternative to the quantitative one, in practical
perspective more and more researchers recognize a significant potential of methods’
complementarity which synthesis can bring new insights into the exploration and
explanation of consumer behavior and consumer motivation (Pincus, 2004). By using
the RG technique in consumer research interpretations and findings can be made by
an application of both qualitative and quantitative analysis methods, however, the
qualitative paradigm on which the technique and the underlying theory are based
needs to be borne in mind.
In this chapter, at first, personal construct theory on which basis the RG technique is
grounded is presented. Afterwards, the RG technique and its consistent qualitative
paradigm are thoroughly discussed.
4.2. Personal construct theory
George A. Kelly developed the repertory grid technique (RGT) (originally ‘The Rep
Test’) as an instrument for elicitation of personal constructs. This technique is derived
directly from Kelly's (1955) personal construct theory (PCT) (see, e.g., Fransella et
al., 2004, p. 1). The PCT is considered to be a predecessor of the cognitive approach
currently dominating the field of clinical and social psychology. Although difficult to
classify, Mahoney's (1991; Mahoney and Gabriel, 1987) idea of considering PCT as
a cognitive constructivist approach seems suitable (Neimeyer and Mahoney, 1995)
because it would respect the phenomenological inclination (Rychlak, 1981) of Kelly's
theory. This definition not only has the advantage of linking PCT to other cognitive
approaches, but it also differentiates it from more rationalistic approaches, which may
not be epistemologically compatible.
Kelly (1955) describes humans as “scientists” whose existence consists essentially of
trying to predict and control events in their environment. A person, as well as a
scientist, develops his individual set of concepts by continuously evaluating and
revaluating own experience. Kelly calls this ‘construction system’ and says that
individuals use it to classify the objects and events in their environment, to interpret
them and to use their own interpretations for understanding, controlling and
predicting the world around them.
People do not go through life without goal or direction, but are constantly and actively
engaged in improving the cognitive system they use to make their predictions
4. Repertory grid technique within the personal construct psychology
59
(Meyer, 1997). People observe the results of their experiments (they live with the
outcomes of their behavior), they modify their theory (i.e., they change their minds,
and they change themselves), and so the cycle continues (after Fransella et al.,
2004, p. 6). Thus people used to make sense of their material and social
environments as Mick and Buhl (1992) point out: “Philosophers such as Merleau-
Ponty and Sartre and psychologists such as Allport and Kelly have observed that
people structure their goals and means in an effort to create coherence in their lives”.
In contrast to the dominant deterministic theories of consumer behavior, PCT
maintains that a person’s understanding of the world is the result of an active,
constructive process rather than a passive reaction to some external reality
(Mahoney, 1988; Ross and Nisbett, 1991 cited in Marsden and Littler, 2000a).
PCT suggests that this understanding is achieved through the process of contrast
and similarity, what Kelly (1955/1991, Vol.1, p. 51/1991) termed ‘construing’. As
Eden and Jones (1984) explain: “We construe situations by seeking to differentiate
them from others and see them as similar to others; it is only through such a process
that we give meanings to events, that they have significance”.
One of the central assumptions of the theory is that reality and what people make of
it is built up of contrasts rather than absolutes (Jankowicz, 2004, p. 11). Thus, Kelly
emphasizes an important characteristic of process of individual meaning making:
when we separate, name, argue, etc., we are always also aware of an opposite
meaning to what is being actualized at the moment (also in Kelly, 1969).
Similarity and contrast in events or objects that are being observed by an individual
stimulate a creation of his personal classification standards – constructs. By Kelly,
construct represents a basic unit similar to the concept introduced before by
Vygotsky (Vygotsky, 1989; Wertsch, 1985 cited in Shaw and Gaines, 1992).
However, unlike Vygotsky’s notion of concept which reveals from objects some
common characteristic, leaving the difference aside, in the notion of Kelly’s construct
both generalization and difference are presented concurrently. By introducing a
notion of construct, Kelly combines two mental functions – the function of
generalization (i.e., determination of similarity, abstraction) and the function of
contrasting or contraposition.
4. Repertory grid technique within the personal construct psychology
60
In order to fully understand and properly evaluate the RGT in consumer research it is
important to examine the assumptions of the theory that underpins it.
Figure 9: The assumptive structure of the personal construct theory
Kelly expresses his PCT in the form of a fundamental postulate which is elaborated
by means of eleven corollaries (Kelly, 1955/1991, Vol. 2, p. 4/1991). The corollaries
or clarifying statements extended the theory and added more elaboration to how the
theory impacts and is used (Fisher and Savage, 1999). Figure 9 presents a summary
of the assumptive structure of the PCT.
Kelly's Fundamental Postulate is that: A person's processes are psychologically
channelized by the way in which he anticipates events (Kelly, 1955, p. 46).
A postulate represents an accepted basic assumption on which the rest of the theory
is built. This language is very confusing on first encounter but Kelly selected every
word accurately to fully reflect an underpinning nature of the concept (Atherton,
2007):
‘Person’ indicates that Kelly is referring to the individual person as a whole and
not to any part or processes of the person.
With ‘processes’ Kelly indicates that a person is not an object that is temporary
in a moving state, but is himself a form of motion always making efforts to
understand and always acting on and in the world.
‘Psychologically’ indicates that it’s not about the brain, or culture, but the mind:
it is quite clear about its level of analysis and its “range of convenience”.
4. Repertory grid technique within the personal construct psychology
61
By ‘channelized’ Kelly points that a person’s behavior takes place in a flexible
and frequently altering network of rather than in a vacuum. This network is
structured what, in turn, facilitates and restricts a person’s range of action.
‘Ways’ indicate that the channels are outcomes of an individual’s own
deliberate choices and, thereby, are set up as means to ends.
With ‘he’ Kelly emphasizes that it is not an ideal or a perfect but rather an
individually chosen way in which the person operates.
‘Anticipates’ denote that PCT is future-oriented. The person is only a product
of memory and learning to the extent he makes use of these in his construct
system. He is a scientist in formulating hypotheses (his constructs) about the
world he perceives, and testing them by acting on them.
By ‘events’ Kelly shows that a person seeks to anticipate real events. That is,
person wants the future reality to be best or better performed.
• Construction Corollary: A person anticipates events by construing their
replications
• Individuality Corollary: Persons differ from each other in their constructions of
events
• Organization Corollary: Each person characteristically evolves for his
convenience in anticipating events, a construction system embracing ordinal
relationships between constructs
• Dichotomy Corollary: A person's construct system is composed of a finite number
of dichotomous constructs
• Choice Corollary: A person chooses for himself that alternative in a dichotomized
construct through which he anticipates the great possibility for the elaboration of
his system.
• Range Corollary: A construct is convenient for the anticipation of a finite range of
events only.
4. Repertory grid technique within the personal construct psychology
62
• Experience Corollary: A person's construction system varies and he successively
construes the replications of events.
• Modulation Corollary: The variation in a person's construction system is limited by
the permeability of the constructs within whose ranges of convenience the
variants lie.
• Fragmentation Corollary: A person may successively employ a variety of
construction subsystems which are inferentially incompatible with each other.
• Commonality Corollary: To the extent that one person employs a construction of
experience which is similar to that employed by another, his processes are
psychologically similar to those of the other person.
• Sociality Corollary: To the extent that one person construes the construction
processes of another he may play a role in the social process involving the other
person (Kelly, 1955, pp. 50-95).
A detailed explanation of each corollary is not included in this context (see Boeree, o.
J. for detailed explanation of each statement). However, certain corollaries will be
discussed further in chapter within the descriptions and statements made with
respect or built on the corollaries.
4.3. Repertory grid technique: basic description
Grounded in the PCT, the RGT is a widely used method for studying personal and
interpersonal systems of meanings and their organization (Neimeyer, 2002). ‘Grid’ is
a generic term for a number of simple rating-scale procedures. They are all used for
arriving at straightforward descriptions of how a person views the world around him,
or some smaller part of it, in his own terms (Jankowicz, 2004, p. 8). The RG is a
qualitative technique that is often compared with the structured interviewing. Such
type of an interview is often appropriate when the nature and extent of participant
opinions on the researched topic are not known beforehand and cannot be quantified
easily. The technique is no longer inextricably tied to Kelly’s theory of personal
constructs and its use as an analytical tool does not require an acceptance of the
model of man which Kelly proposed (Slater, 1976) However, the terms Kelly used
have become standard (Dillon and McKnight, 1990).
4. Repertory grid technique within the personal construct psychology
63
The resulting form of the elicitation procedure in whole represents a set of rating
scales printed one above the other, with the ratings arranged in rows and columns
into a table or grid.
E1 E2 E3 E4 Q Em
1 E1, E3, E4 beautiful 2 4 5 1 @ 1 unattractive design
2 E2, E4, Em for business purposes 3 1 4 1 Q 5 for leisure time
Q Q Q
n
ElementsConstruct emergent pole (1) Construct implicit pole (5)Construct № Triad
Table 1: The repertory grid elicitation form
Every grid consists of four main components (Table 1):
• Topic;
• Elements;
• Constructs;
• Ratings.
The basic unit of description and analysis is called a construct. People construe
things by means of constructs. To construe means to make sense of something; to
have a personal understanding of it; to find meaning in it (Jankowicz, 2004, p. 10).
Earlier used expressions ‘viewing the world’ and ‘ways of seeing’ mean construing.
Describing what is meant by a construct is more difficult than describing other
components of RG because construct has properties and Kelly has offered several
definitions of it as well as certain assumptions (Fransella et al., 2004, p. 15). For this
reason, at first, other three components are described followed by a comprehensive
description of the notion of construct.
4.3.1. Topic
A grid is always conducted about a particular topic with an intention of eliciting just
those constructs which the person uses in making sense of that particular field of
discourse, i.e., that particular piece of his experience.
By revealing the constructs a researcher discovers how person thinks and what
meanings he usually distinguishes about a particular topic. The person can possess
many more other constructs about his experience concerning other topics that
4. Repertory grid technique within the personal construct psychology
64
researcher is not dealing with. Therefore, by filling out a grid on particular topic, only
the related part of a person’s repertoire can be discovered.
RGT allows to disclose and to understand participant’s private language including his
jargons and some terms which he uses when talks on particular topic (Jankowicz,
2004, p. 12). When a number of grids of different people who are clustered in one
group or segment based on a certain criteria is being analyzed, it is important to
reveal common terms, words and phrases that have got entrenched into their
language related to the investigated topic. It allows further to appeal to the group
applying its own jargons and meanings and so making it easier for them to
understand and perceive a message. This can be especially useful when conducting
a market research with a purpose of revealing customers’ needs, motives and
perceptions and then to use it in positioning and communication planning.
“A grid, then, is a highly focused technique, in which the topic must always be clearly
specified in advance.” (Jankowicz, 2004, p. 13)
4.3.2. Elements
Kelly defined elements as ‘the things or events which are abstracted by a construct’
and are seen as one of the ‘formal aspects of a construct’ (Kelly, 1955/1991, Vol.1, p.
95/1991).
As formulated by Jankowicz (2004, p. 13), “an element is an example of, exemplar
of, instance of, sampling of, or occurrence within, a particular topic”. By providing a
set of elements related to a certain topic, researcher stimulates participant’s display
of constructs which he uses to perceive the realm of discourse. Thus, a small sample
of that person’s construing of the world can be studied. A researcher can investigate
the system of constructs of an individual consciousness and get a notion of individual
underlying (unconscious, nonverbalized) constructs. Besides, the researcher can
investigate individual peculiarities in perception of elements, study the relationships
between the elements within individual insight and build up a subjective semantic
space of individual meanings associated with the elements (Fransella et al., 2004, p.
21).
Since elements represent objects of a particular topic, it becomes crucial to choose
the right set of elements in doing a grid because elements that are chosen to present
a given topic will influence the types of elicited constructs and consequently the
4. Repertory grid technique within the personal construct psychology
65
gained insight into the ‘language’ of the participant in that particular subject matter
(Jankowicz, 2004, p. 13; Fransella et al., 2004, p. 21).
Depending on the topic to be explored by means of RGT, elements that are used can
be almost anything. For instance, if marketers want to find out how people construe
meanings about snack food the elements might be different types of snack food
which fully represent the range of common snacks available or if marketers are
interested in particular list of types, only these types can be represented as elements.
Originally, Kelly’s RGT was developed for use in clinical psychology as a technique
for exploring individuals’ personal constructs about interpersonal relationships in the
context of psychotherapy. He used interpersonal relationships as his context and
developed role-title list providing people as elements. However, the technique has
been further applied extensively to the elicitation of personal constructs for many
other purposes and a wide variety of elements have been used (Steed and
McDonnell, 2003; Fransella et al., 2004, p. 19). Some other examples provided by
Fransella and her colleagues (2004, p. 21) include photographs of people first used
by Bannister (1962), feelings by Fransella and Adams (1966), situations by Fransella
(1972), diseases by Orley and Leff (1972), rooms by Honikman (1976), shops by
Hudson (1974), foreign countries by Lemon (1975), classes of Spirit in Ganda
mythology by Orley (1976), brightly colored stand-up models by Salmon (1976),
architectural maps by Stringer (1974), an artist’s paintings, drawings of different
motion phenomena in physics by Winer and Vázquez-Abad (1997), and perceptions
of different scents by Williams, Whittlestone and Martin (1992). R. Neimeyer (1985)
and further R. Neimeyer and Stewart (1996) used different stages in life which the
person construed as important elements in his ‘biographical grid’. Fransella (1978)
used specially designed elements – a standard body shape altered by an artist to
range from the extreme thinness of the person with anorexia nervosa to extreme
obesity (Fransella et al., 2004, p. 21). There are many more other examples and they
all prove the ability of repertory grid to analyze individual construing system about
very different aspects of personal life and world around.
Three guiding principles that can help and should be considered when selecting the
elements are (Stewart et al., 1981, p. 28):
4. Repertory grid technique within the personal construct psychology
66
• Elements should be as precise and specific as possible. Imprecise elements
when being compared to produce a construct are likely to induce vagueness
of contrast and, therefore, not enough clear construct.
• A nonuniform scatter over the elements area is acceptable, i.e., elements do
not have to be evenly distributed over the available area.
• When one is interested in the boundary between the elements, both sides of
the bordered area need to be covered by elements. For example, if researcher
is interested in what consumer likes and dislikes in different types of snack
food, elements which person does and does not like should be included.
Moreover, there are several criteria for choosing the right set of elements in a grid
(Stewart et al., 1981, p. 29):
• Elements should be discrete. Most often elements are people, objects, events
or activities – in other words, nouns or verbs. Nouns should be quite specific
and indicate tangible things or subjects, e.g., specific objects or particular
people. Abstract nouns such as “good person” or “leadership” should be
avoided. Verb elements such as events and activities should be defined as
close as possible in space and time. The verbs should be, if possible,
converted into a verbal noun trying to express them as activities, each ending
with ‘-ing’ since it is more user-friendly and easier to handle while presenting
the elements (Jankowicz, 2004, p. 29). For example, instead of “to read” it is
better to offer “reading” or “to offer” should be replaced by “offering”. Loose
descriptions that cannot be specified clearly, such as “thinking” or “negotiating”
- should also be avoided.
• It is also recommended to avoid adjectival and adverbial phrases as elements.
It is better to stick to nouns and verbs or even verbal nouns. Another fault is to
use features of elements as elements, but not elements themselves. It can be
recognized if an ‘opposite’ can be produced to the element. It would show that
things that are used as elements should really emerge during the construct
elicitation phase (Stewart et al., 1981, p. 30).
• Elements should be within the range of convenience of the constructs to be
used. It does not mean that the constructs are already known beforehand
4. Repertory grid technique within the personal construct psychology
67
when elements are being chosen, but that the context determines the range of
elements which can be reasonably used. When participant discriminates
people, events or objects by means of constructs, he applies these
discriminations only to a limited number of people, events or objects. It is not
needed to have old people in among the elements while conducting a grid
about youth of today because they may appear to be outside the range of
convenience of the youth-of-today type of construct (Fransella et al., 2004, p.
18).
• Elements should be homogeneous. Classes of elements should not be mixed,
i.e., people should not be mixed with things or things with activities and etc.
Yorke (1985) showed that heterogeneous elements are likely to result in range
of convenience problems as well as in decreasing grid validity (Fransella et al.,
2004, p. 19).
• Elements should not be subsets of other elements. That means one element
must not comprise another and the whole set should consist of mutually
exclusive items. It would prove impossible to work with an element set where
one element describes the rest or at least some other elements, for example
set that would include elements: ‘cats’, ‘dogs’, ‘budgerigars’, ‘Siamese’, and
‘German shepherd’ (Jankowicz, 2004, p. 30).
• Elements should not be evaluative. Some elements, especially events or
activities, may contain an implicit evaluative nuance what can make it difficult
to handle them during an interview (Stewart et al., 1981, p. 30). For instance,
elements like ‘leading a team’, or ‘drinking alcoholic beverages’ may be
attached with rather different meanings depending on the understanding what
‘leading’ implies or on the extent of alcohol consumption, i.e., whether it is a
glass of wine during the meal or it is an alcohol addiction.
Described principles and criteria for selecting the elements do not sometimes have to
be all followed. However, only an experienced interviewer, who has gained a good
feeling and intuition for suitable and easily-handled elements, should experiment with
element set kind.
4. Repertory grid technique within the personal construct psychology
68
4.3.3. Constructs
A construct always represents a contrast and the contrast is needed to be spelled out
before an interviewer can be sure of the meaning intended to the whole construct
(Jankowicz, 2003). Among different formulations Kelly (1955/1991, Vol.1, p. 5/1991)
defines construct as “a way in which two or more things are alike and thereby
different from a third or more things”.
Constructs form the building blocks of our 'personality' and as such come in various
shapes and sizes. According to the Organization corollary some constructs are more
important than others. The most important constructs are those which are “core” to
our sense of being. These are very resistant to change and include things like moral
code, religious beliefs, etc. and cause significant psychological impact if they are
threatened in any way (Fisher and Savage, 1999). The other constructs are called
“peripheral” constructs and a change to them does not have such a strong impact.
Construct theory offers descriptive ways of categorizing constructs, firstly, according
to the type of thinking a person uses to produce a construct and, secondly, according
to the way constructs are interrelated within the construing system (Fransella et al.,
2004, p. 12).
According to the type of thinking a person applies to produce a construct, three types
of constructs can be defined:
• Pre-emptive constructs – constructs that are applied in an all or nothing way.
They provide very black and white type of thinking – if this is a table, then it is
nothing else but just a table (Fransella et al., 2004, p. 12; Fisher and Savage,
1999). Thus, when a person provides pre-emptive constructs he obliges
himself to choose between one of its two poles (alternatives).
• Constellatory constructs – stereotyping constructs – if this is a table, it must be
square, made of wood and located in the kitchen. In other words, the
subordinate constructs are strongly associated with superordinate construct.
Such constructs bring a lot of supplementary baggage with them, no matter
whether it is right or wrong as individual perceptions of elements and
construing system are considered.
4. Repertory grid technique within the personal construct psychology
69
• Propositional constructs – the most open form of construct which carries no
additional labels or implications. Use of propositional constructs allows people
to construe the elements circumscribed by their constructions in various ways.
Propositional construing represents the opposite of pre-emptive and
constellatory construing.
In terms of the extent to which constructs imply each other within a construing
system, Kelly differentiates between ‘tight’ and ‘loose’ constructs. That is, given
where a person rates an element on one construct, to what extent does that predict
where he will place the same element on other constructs (Smith, 2000).
A loose construct is one which may or may not lead to the same behavior every time
(Fisher and Savage, 1999). This makes a researcher unable to predict the
construer’s actions consistently. On another hand, a tight construct leads to the same
behavior.
According to Kelly (1955), a loose construct system leads to varying predictions,
whereas, a tight construct system leads to unvarying predictions. People who provide
tight construct system obtain firmly held views concerning considered elements.
People mostly start off with loose constructs, trying things out for different
characteristics, seeing what works and what doesn’t. As we move towards the new
we tighten up our construing, narrowing down our experimentation and so we begin
making clearer associations and developing more clearly the 'new' (Fisher and
Savage, 1999). Thus, the essence of Kelly’s argument is that we loosen and then
tighten and then loosen our thinking in a cyclic manner. Our aim is, first of all, to gain
perspective and then become concrete enough to define our themed operationally
and so regain a new perspective. As an example of loose and tight construct
systems, in a very tight system an element perceived as ‘honest’ might also be
perceived as ‘generous’, ‘kind’, ‘intelligent’ and so on. A looser system might allow
the possibility that ‘honest’ people are considered as ‘unkind’ or ‘mean’ (Smith,
2000).
Bieri and colleagues (1966, p. 185) invented a notion of ‘cognitive complexity’ which
is defined as “Qthe capacity to construe social behavior in a multidimensional way. A
more cognitively complex person has available a more differentiated system of
dimensions for perceiving others’ behavior than does a less cognitively complex
4. Repertory grid technique within the personal construct psychology
70
individual.” Thus, in terms of cognitive complexity the more loosely knitted constructs,
the more cognitively complex the person’s construing system.
4.3.4. Scaling methods
George Kelly, physicist, mathematician and would-be engineer, loved mathematics.
He regarded mathematics as the “purest form of construing” (Hinkle, 1970 cited in
Fransella et al., 2004, p. 1).
After constructs are elicited a question arises which contrast pole of each construct is
being applied by the participant to which element and in what extent (Fransella et al.,
2004, p. 54). “But we can look beyond words. We can study contextsQ The answers
to questions such as these may give us an understanding of the interviewing of the
client’s terminology and provide us with an understanding of his outlook which no
dictionary can offer” (Kelly, 1955/1991, Vol.1, p. 189/1991).
Kelly offered a methodology for deriving a mathematical description of part of a
person’s psychological space. His basic assumption was that a mathematical
relationship between a person’s judgements reflects psychological assumptions
underlying those judgements (Fransella et al., 2004, p. 54).
Mathematical relationships in this context actually represent the structural
relationships between the constructs and the set of elements (Marsden and Littler,
2000a). To examine these relationships various scaling methods are suggested.
Most common, accepted and discussed methods are (Feixas and Alvarez, 1996):
1. Dichotomous method;
2. Ordinal method;
3. Rating scale method (Table 2).
1. The dichotomous method implies placing all the elements on either one pole or
the other of the construct.
In Kelly’s original grid form after having elicited the constructs, to complete the grid
form the person is asked to place a tick under the name of each element to which the
construct applies. The final form of grid then appears in the form of matrix which
consists of a number of ticks and blanks. However, Kelly faced a problem with this
4. Repertory grid technique within the personal construct psychology
71
kind of grid. A person often would perceive hardly any other element as being
characterized by one pole of the construct. At first, to overcome this difficulty Kelly
suggested to simply eliminate those rows from the grid calculations. This conclusion
didn’t appear to be satisfactory because of losing part of elicited context information
and, so, other approaches to complete repertory grids were developed (Fransella et
al., 2004, p. 55).
E1 E2 E3 E4 E5 E6
beautiful + - + + - + unattractive
E1 E2 E3 E4 E5 E6
beautiful 3 6 5 2 4 1 unattractive
E1 E2 E3 E4 E5 E6
beautiful 5 5 3 4 2 4 unattractive
Rating scale method (rating)
Construct emergent pole (1)Elements
Construct implicit pole (5)
Dichotomous method (dichotomizing)
Ordinal method (ranking)
Construct emergent pole (1)Elements
Construct implicit pole (6)
Construct emergent pole (+) Construct implicit pole (-)Elements
Table 2: Scaling methods used in repertory grid based research
2. The ordinal method implies ordering the elements from one pole of the construct
to the other. This method emerged as an attempt to overcome difficulties presented
in the dichotomous approach (Feixas and Alvarez, 1996).
It was suggested by Phillida Salmon and first described by Bannister (1963) to ask
the person to rank the elements in terms of each personal construct. The task is to
rank all the elements according to how each element is subsumed under the
construct poles. Ranking grids proved to be very popular allowing different forms of
analysis to be applicable such as simple arithmetic or Spearman’s rank order
correlation. However, again they have been found too restrictive because they force
the elements to be uniformly distributed across the construct, not allowing any form
of lopsidedness, even if it is appropriate (Fransella et al., 2004, p. 59).
3. In the rating scale method “each element is assigned a value in a Likert-type
scale delimited by both poles of the constructs” (Feixas and Alvarez, 1996).
Finally, Bannister returned to Kelly’s original method but now using a longer scale
then 2-point scale which Kelly used. In this grid each element is rated on a scale
4. Repertory grid technique within the personal construct psychology
72
defined by the two construct poles. It is this “rating” method which is much in use
today (Fransella et al., 2004, p. 59).
Now grid allows to identify exactly what the person means when he uses certain
terms in constructs. Thus by rating each element on each construct the person
provides an exact picture of what he has got in mind about the elements and topic in
general (Jankowicz, 2004, p. 14).
The length of the scale can be anything from Kelly’s 2-point scale (tick or blank) to 10
or even 16. Longer scales are seen as providing more scope for people to express
their views and judgements therefore producing more comprehensive data. A
commonly used length is the 5-point or 7-point scales as they also give the mid-point.
According to the conducted researches aimed to find out whether the length of scale
might well affect grid measures, it doesn’t proved to matter greatly what length of
scale is used (Fransella et al., 2004, p. 63).
The rating scale is most widely used and recommended by authors (Jankowicz,
2004; Marsden and Littler, 2000a; Björklund, 2005; Dillon and McKnight, 1990).
More detailed consideration of RG components will be described in 4.5. By now, it
can be resumed that RG is “simply a set of rating scales which uses the individual’s
own constructs as the subject matter on which ratings are carried out” (Jankowicz,
2004, p. 14). One could argue about this definition. There are more different ways of
thinking and defining RG and it also confirms one of the fundamental assumptions
made by Kelly which he called ‘constructive alternativism’ when he developed the
Role Construct Repertory Test (Kelly’s original name of RG interviewing technique).
By this he meant that people construe the same thing differently and also that a
single person always has the option of construing the same thing in different ways on
separate occasions. Therefore, it is also possible to accept a variety of different
definitions of a repertory grid (Jankowicz, 2004, p. 15).
4.4. Application of the repertory grid technique in marketing and other fields
4.4.1. Application experience
Developed within the PCT, the RGT with its various modifications represents a widely
used technique for studying personal and interpersonal systems of meanings.
Because of its flexibility the technique was used for a broad variety of matters,
4. Repertory grid technique within the personal construct psychology
73
however, its most consistent area of application was probably in the clinical domain
where grids were used to assess the properties of meaning systems of different
groups of persons (e.g., those diagnosed as “thought disordered” or “agoraphobic”)
and how these change over the course of treatment (Neimeyer, 2002).
Among other clinical and non-clinical applications can be mentioned: studying
neurotics by Ryle (1976); magistrates’ decision making by McKnight (1981);
categorization analysis by Coltheart and Evans (1982); job analysis by Hassard
(1988) and application to the domain of human computer interaction, particularly with
respect to elicitation of knowledge in the development of expert systems (Shaw and
Gaines, 1987). This list can be widely extended by names and topics.
Jankowicz (2004, p. 9) provides a comprehensive list of applications in which RG can
be used, mentioning also that the list still doesn’t cover all the applications. His list is
divided into application groups according to the fields where the repertory grid was
used. Applications, in which its developmental and didactic implications are most
profound (Ravenette, 1997 cited in Jankowicz, 2001), include educational,
occupational and clinical applications. He also emphasizes a group of general
applications which contains examples of using RGT mostly concerning interpersonal
relationships and/or exploration of personal meanings and attitudes about different
matters. One example of use which demonstrates the value of technique is in
heightened risk field of bank commercial lending and venture capital investment
decision making which are thoroughly discussed by Jankowicz (see his 2001 for
more detail).
The RGT was also widely employed as a qualitative method in market research,
consumer research and new product development in identifying those attributes of
products which are not self-evident (Frost and Braine, 1967; Hallsworth, 1988 listed
in Marsden and Littler, 2000b; Jankowicz, 1990; Stewart et al., 1981; Jankowicz,
2001; Marsden and Littler, 2000a). RGT procedure enables researchers to
investigate their products, markets and customers allowing to pursue different
research purposes. Most widely pursued objectives in marketing research conducted
by means of RGT are related to capturing market and product perception, i.e.,
learning what customers think about certain products, product categories or group of
products in their own language. Some common objectives of market studies
conducted by means of RGT include investigation of:
4. Repertory grid technique within the personal construct psychology
74
� How company’s product compares with others.
Customers draw their cognitive map by describing and rating the differences and
similarities between company’s product and its competitors, what allows the
researchers to:
• Define customer’s ideal product and ideal attributes;
• Assign priorities between attributes as well as between different brands;
• Experiment with different scenarios or product packages;
• Reveal whether customers’ view of the product and things they
consider to be important are the same as with those of company’s staff
responsible for product development, promotion and sale.
� How customers experience a certain product.
Customers provide constructs which reflect their experience with a company’s
product and its rivals allowing revelation of critical incidents and defining company’s
competitive edge, especially in the markets with only a few highly competitive
players.
� Specification or design of a new product.
Obtaining customers’ cognitive maps of existing products or services and then
experimenting with ‘what if’ scenarios allows to build a picture of the key features of a
new product and to examine priorities and trade-offs.
� Customer perception tracking.
It can be useful to obtain a database of key customer perceptions in a particular
sector or to track customer perceptions of products over time (o. V., 2007).
� Positioning opportunities of new and existing products.
By identifying dimensions on which consumers evaluate products, marketers can
elaborate the benefits and most important consumers’ attributes.
� Development of accurate distribution strategies
4. Repertory grid technique within the personal construct psychology
75
Understanding the different ways in which consumers search for products permits to
develop more precise distribution strategies.
� Development of communication strategies
Marketers can design communication programs based on the RGT findings which
incorporate the consumers’ language and vocabulary and therefore increase the
probability of mutual understanding. Besides the complexity levels of consumers’
construct systems can serve as an index to decide on the amount and type of
information that should be presented in communication programs. It is also possible
to link by means of communication strategies product benefits to the important beliefs
and values held by consumers.
� Segmentation
Commonalities in subjective meanings can provide a useful basis for identifying
concrete segments of consumers (Marsden and Littler, 2000b).
Marsden and Littler (2000b) emphasize five basic components of the network of
subjective meanings that consumers attach to their consumption experience which
can be explored by means of RGT. Authors have termed these components as
consumers’ product construct systems (PCSs) which include:
1. Consumption domains: how do customers categorize different products and
services?
2. Decision rules: what search strategies and evaluative criteria are employed for
each category?
3. Values: what core beliefs underpin different decision rules?
4. Construct complexity: how discriminating are consumer’s decision rules and
values?
5. Construct commonalities: what are the similarities and differences in
consumers’ PCSs and how are they mediated by their demographic
backgrounds?
4. Repertory grid technique within the personal construct psychology
76
Thus, RGT is not new to market research, however, despite a solid amount of market
studies conducted within the application of RGT many studies failed to account for
the diversity and richness of consumer experience because they did not employ the
technique in its entirety and full capacity (Marsden and Littler, 2000b).
4.4.2. Criticism of the applicability of the repertory grid
Despite its relatively widespread use the RGT was constantly attracting much
criticism. Among other issues, it was criticized for generating “utterly valueless” and
“irrelevant” information from consumers (Sampson, 1972; Gordon and Langmaid,
1988 listed in Marsden and Littler, 2000b), for its “inherent complexity” and “slavish
adherence” to the notion of bipolar constructs (Reynolds and Gutman, 1984; Frost,
1982), and for not being based on an “adequate theory” of human behavior (Grunert
et al., 1995).
Thereby, at the beginning of 90s it was agreed that the RGT has “limited use” as a
qualitative method in consumer research (Rice, 1993 cited in Marsden and Littler,
2000a). At that time, in response to this criticism, Chisnall (1992, p. 178) said:
“Further consideration of construct theory in the marketing sphere is called for; the
intellectual attractions of Kellian techniques need to be tested in some length before
they can be expected to win support of practical researches” (cited in Marsden and
Littler, 2000a).
Marsden and Littler (2000a) argue that the reason for a failure of the RGT to reach its
full potential in the field of consumer research appeared from a “limited and
somewhat confused way in which qualitative methods in general have traditionally
been used and evaluated in the field” (confer Murray and Evers, 1989). The authors
(Marsden and Littler, 2000b) emphasize four reasons why the RGT was not used to
its full potential as a qualitative method in market research in general and in
consumer research in particular:
1. In market research RGT was mostly confined to the exploratory stages
because qualitative methods played a secondary role to quantitative methods
(Hunt, 1994). Hence, the way in which the RGT can be applied throughout the
market research process and how its findings can be reported on in their own
right was not adequately addressed (Stewart, 1990).
4. Repertory grid technique within the personal construct psychology
77
2. Kelly’s PCT which underpins and provides general guidelines for using and
evaluating the RG was largely ignored (Chisnall, 1992). Instead, the RGT was
incorporated into the theoretical framework of what is called the natural
science, or onto the dominant “positivist” paradigm” what resulted in various
operational and conceptual misunderstandings (Marsden and Littler, 2000b;
Fransella et al., 1988).
3. Traditionally the RGT combines an integrated set of elicitation methods which
include triading, laddering and pyramiding. However, these methods were
rarely employed together in any systematic manner (Earl, 1986; Fransella et
al., 1988). Therefore, the RGT was often criticized for incompleteness and
inadequacy of its findings for developing marketing strategy (e.g., Frost, 1982;
Gordon and Langmaid, 1988; Rice, 1993; Sampson, 1972 listed in Marsden
and Littler, 2000b).
4. The RGT was used to explore various structural components of consumers’
subjective meaning systems where each component was often considered in
isolation from one another. The nature of such meaning systems was
assumed to be a simply quantitative sum of their component parts (i.e.,
1+1=2) while later it was argued that the totality of consumers’ meaning
systems is qualitatively different to any of their individual component parts (i.e.,
1+1=3). In this sense, the findings constituted a rather fragmented body of
knowledge in consumer research (Stewart, 1990).
In response to these limitations Marsden and Littler (2000a, 2000b) showed how the
RGT and theory can be employed in market research as a qualitative method in its
own right with its potential usefulness.
4.4.3. Criticism solution
According to Shrivastava (1985), evaluations of the research methods should be
grounded in the assumptions of their underlying paradigm. Personal construct
psychology (PCP) and, respectively, theory derives from the
“interpretive/constructivist” paradigm which assumptions differ from positivist
paradigm regarding the nature of consumer experience and its representation
(Marsden, 1997 cited in Marsden and Littler, 2000a). The traditional positivist
assumption such as the existence of an objective reality which is verifiable and
4. Repertory grid technique within the personal construct psychology
78
quantifiable is largely irrelevant to the underlying assumptions of the RGT (Marsden
and Littler, 2000a; Highlen and Finley, 1996).
By resuming a solid number of statements and findings related to the PCT and RGT
offered in literature, Marsden and Littler (2000a) provided detailed evidence that
interpretive paradigm is the one the underlying assumptions of PCT and RGT
methodology are consistent with. Hence, further development of the assumptions of
interpretive paradigm is complemented and facilitated by assumptions of PCP by its
emphasis on the process of meaning construction (construing), the structure of
individual and shared meanings (hierarchical bipolar construct systems) and
advocacy of pluralistic methods of investigation (RGT).
Originally, the interpretive paradigm emerged in consumer research in the early
1980s. It was an alternative to deterministic theories such as cognitive, behavioral
and trait which obtained a growing criticism for depicting consumers as passive
organisms simply responding to marketing incentives. Besides, it was also a
response to the criticism of reductionistic methods of inquiry, i.e., quantitative
associated with the positivist paradigm which was criticized for ignoring much of the
complexity and richness of experience (Kassarjian, 1994; Dholakia and Arndt, 1985).
By examining the relationship between the assumptions of PCT and those of the
interpretive paradigm researchers pointed that the PCT is consistent with the
interpretive paradigm’s focus on exploring the psychological processes that people
use to make sense of their material and social environments. Moreover, based on the
process of construing which implies an achievement of understanding through a
process of contrast and similarity, the PCT offers a more integral understanding of
the process of meaning making (Neimeyer, 1993; Marsden and Littler, 2000a).
Another assumption of the PCT that contributes to the further development of
interpretive paradigm is that subjective meanings take the form of bipolar constructs
which are organized into a system of superordinate and subordinate relationships
(Earl, 1986; Droge and Calantone, 1984). The elicitation methods including triading,
laddering up and laddering down or pyramiding which in complex produce
hierarchically organized constructs, when used together are capable of eliciting a
wide range of meanings from participants at different levels of psychological
abstraction, from which fairly holistic representation of consumer experience can be
4. Repertory grid technique within the personal construct psychology
79
produced (Marsden and Littler, 2000b). Therefore, not only content but also structure
of meanings is considered within PCT which offers a concrete analytical basis for
exploring nature and form of subjective meaning systems (Zinkin and Biswas, 1988).
Interpretive paradigm was criticized for being “solipsistic” because of its focus on the
individual’s subjective consciousness. While some authors argue that interpretations
are highly individualistic and cannot be generally agreed and that any given situation
is going to be interpreted differently by each individual because all people have got
different experiences, the PCT avoids the blame of solipsism by clarifying the
conceptual basis for examining the commonalities between individuals’ subjective
meaning systems (Droge and Calantone, 1984). It assumes the possible similarity
between individuals not in their gained experience but in the way people construe
their experiences (Neimeyer, 1993). Duck (1994, p. 125) maintains that people can
similarly construe their interpretive maps of the world around though having different
experiences. He points out: “psychological similarity is founded on the similar
construction of experience where experience is considered as subjective
interpretation, but not on similarity of experience itself where experience is equated
with events”.
PCT and RGT as interpretive theory and method correspondently are limited in
application for understanding the social context of construing at a macro level of
abstraction and, thus, inadequate for understanding the ideological dimensions of
subjectivity. However, their usefulness for exploring the micro-processes of meaning
construction, interconnection and organization is undoubted (Burkitt, 1991 cited in
Marsden and Littler, 2000a).
Another advantage of PCT which contributes to the development of interpretive
paradigm is a capability of RGT to incorporate both qualitative and quantitative
approaches. It is demonstrated that RGT allows to interleave concept formation, data
collection and qualitative dimensions of analysis throughout the entire research
process. As mentioned before, in consumer research the RGT was often employed
within the dominant positivist paradigm what caused much confusion. The RGT was
applied in the preliminary stages of research prior to the forming of standardized
questionnaires (e.g., Hallsworth, 1988) and it resulted in suspicion of either deceiving
of data contributors or in producing an alienating distance instead of maintaining the
close interpersonal relationship.
4. Repertory grid technique within the personal construct psychology
80
Since the RGT is diametrically opposed to such objective research strategies it is
suggested to consider the RGT as a ‘participatory exercise’ where researchers and
participants talk ‘with’ but not ‘to’ one another (Fewtrell and O’Connor, 1995).
Interpretive paradigm suggests that researchers should consider the meaning of
phenomena from the consumer’s perspective and this is maintained in the RGT by
focusing on the language as the medium of psychological inquiry (Ozanne and
Anderson, 1989).
The RGT avoids the criticism of an adherence to “incommensurable thesis” which
means mutual exclusiveness of different theories and methods such as, for instance,
qualitative and quantitative (Heath, 1992; Kavanagh, 1994). The potential of
supplement of different methods has been increasingly recognized and RGT has
become more attractive for incorporating the virtues of both approaches.
Overall, it is proved that RGT and PCT offer a potentially useful approach to
conducting qualitative market research. However, it is also concluded that “their full
implications for marketing strategy will only be fully realized along with their further
elaboration and refinement at both the theoretical and methodological methods”
(Marsden and Littler, 2000b).
4.5. Repertory grid technique: elicitation procedure
Kelly developed the RGT as a way by which individual’s construction systems can be
uncovered and enables a researcher to obtain a mental map of how a participant
views the world.
The technique represents a content-free procedure for exploring and forming
personal constructs. The assumption underlying the technique is that people can
represent their experiences, i.e., the situations with which they are faced, by placing
alternative constructions upon them. A grid conversation encourages a person to
make it clear what for him distinguished one experience from another (Steed and
McDonnell, 2003).
The grid is not so much a test but rather a methodology involving highly flexible
techniques and variable applications. What all applications have in common is their
seeking to understand how people impose meaning in their worlds (Solas, 1991).
The grid can be used for various purposes but only one purpose at a time can
prevail. A researcher must first decide on the topic according to his purpose for doing
4. Repertory grid technique within the personal construct psychology
81
a grid. An example of a completed RG is shown in Table 3. Subsequent sections
describe elicitation of a grid and its components by various procedures.
Topic: Computers I might buy for my work
Elements: 4 different computers and 1 generalized image of 'my ideal' computer
Constructs: 7 constructs elicited from the buyer and 1 construct supplied by the interviewer
Ratings: on a 5-point scale
emergent pole (1) PC Mac G3 iMac G4 eMac Ideal implicit pole (5)
Looks boxy and 'standard' 1 2 5 4 5 The looks are to die for
Large range of software 1 2 4 2 1 Smaller range of software
Slow performer 1 3 5 2 5 Fast
Easy to set up 5 1 1 2 1 Difficult to set up
Good build quality 5 2 1 3 1 Flimsy build
Easy to upgrade 2 3 1 1 1 Upgrade is a dealer job
Difficult to move 1 1 4 5 5 Transportable
Appropriate for my work 5 4 1 2 1 Doesn't fit properly
Table 3: An example of an elicited repertory grid
The procedure of RG elicitation basically consists of two main phases: knowledge
elicitation phase and rating grid or element comparison phase. In turn, knowledge
elicitation phase comprises two stages: element selection and then construct
elicitation (Marsden and Littler, 2000a).
4.5.1. Knowledge elicitation phase: element selection
Representing the topic which is an inquiry focus a research, the first stage in applying
the RGT is to choose a set of elements which are consistent with the objectives of
the research and the targeted (sub)system of constructs to be elicited from
participants (Stewart et al., 1981, p. 32).
There are 3 strategies for element selection (Stewart et al., 1981, p. 32):
1. Interviewer provides all or some part of the elements.
Interviewer chooses the elements based on his own knowledge, experience and
reason for doing the grid. This strategy can be used when interviewer needs
certain elements to be presented in a grid to perceive individual meanings and
judgements specifically about them. However, if it happens that the participant is
not familiar with provided elements, the value of the interview will be reduced.
Another problem may appear when part of the elements is provided by the
interviewer. In this case, the respondent may consider them as more important or
4. Repertory grid technique within the personal construct psychology
82
interesting for the interviewer what may result in bias of his judgements (Stewart
et al., 1981, p. 32). Supplying elements may also run the risk of omitting elements
which are important to the interviewee (Jankowicz, 2004, p. 30).
However, very often it can be valuable to work with supplied elements when only
one or two are provided as additional to elicited ones. This permits a comparison
of how the interviewee construes elements he has provided with these key
elements. The most common key elements of this kind are ‘self’ and ‘ideal self’.
By making the interviewee rating himself on the constructs, it can be revealed
where his personal actual and desirable images are among the elements being
construed.
The ‘self’ or the ‘ideal self’ elements do not need to be related to the individual
person’s actual self, current or ideal. If a person is considering different jobs
elements could usefully include “my current job” and “my ideal job”. A customer
choosing between different autos as elements could have ‘my current auto’ and
‘my ideal auto’ as elements (Jankowicz, 2001). Therefore, ‘self’ elements
represent individual perceptions and images of what respondent actually has or
wants when talking about things, events or people enclosed in the elements set.
Although being provided, ‘self’ elements avoid the problem of unfamiliarity of
elements to the respondent. Usage of such key elements permits to significantly
extend the findings from further analyses. For instance, ratings of the ‘ideal’
element allow to discover which element is closest to the ideal and which
construct poles are preferable (Jankowicz, 2004, p. 58).
2. Respondent generates the list of elements himself.
Once the interviewer explained to the respondent what class of elements is
wanted and on which matter, the respondent names the elements himself what
ensures that the topic is represented from respondent’s point of view. The
disadvantage of this strategy is that participant’s responses will have a definite
bias with regard to those elements he personally values little, Thus he may omit
elements and, hence, issues that the interviewer is interested in (Jankowicz,
2004, p. 11; Stewart et al., 1981, p. 33).
3. Interviewer elicits elements.
The interviewer provides general categories which cover the range of the topic
and formulates questions based on the categories, answers to which appear as
4. Repertory grid technique within the personal construct psychology
83
elements. On the one hand, the respondent names the elements himself and,
hence, is familiar with them. On the other hand, the interviewer by providing the
categories and directing the choice by his questions reveals elements that are
interesting for him from the light of the respondent’s prism.
This strategy has a significant advantage of enabling to offer the same categories
of elements to a sample of respondents. Although each of them will be
considering different named instances, the existence of common categories helps
in analyzing the grids as a set (Jankowicz, 2004, p. 32) and, therefore, allows
comparison of perceptions and meanings within the group of people.
In Kelly’s generalised variant of his original Role Construct Repertory Test the third
strategy with elicited-element form was applied. He used twenty four standard
categories that he found useful in his clinical work (Stewart et al., 1981, p. 35).
Once the suitable set of elements is defined it is time to specify the constructs.
4.5.2. Knowledge elicitation phase: construct elicitation
Constructs represent individual scales that a person uses when thinking about the
elements, where poles of a construct represent scale limits. There are many ways for
eliciting constructs depending on purpose of the research, the nature of elements,
the amount of respondents and type of analysis to be applied further. A
comprehensive review of different approaches of construct elicitation is presented by
Epting, Probert and Pittman (1993).
A construct is a way in which some things are construed as being alike and yet
different from others. Therefore, it is essential in the nature of the construct that it is
bipolar (Steed and McDonnell, 2003). The bipolar form implies a construct to be
composed of two poles, emergent or similarity pole and implicit or differentiation pole.
The poles present a pair of contrast descriptors in their own right, rather than of a
phrase and its dictionary opposite. When an assertion and its bare, grammatical
negative are stated it cannot be considered as a construct because it does not
actually provide an indication of what is meant by the similarity pole (Jankowicz,
2001). Therefore, it is essential for investigator to reveal both ends of the construct
equally clear, what usually means avoiding attaching a simple negation to the
similarity pole (Stewart et al., 1981, p. 16).
The researcher has to decide first whether to provide a list of previously determined
constructs or to elicit them directly from respondent. Although both options can be
4. Repertory grid technique within the personal construct psychology
84
performed in RGT, the fundamental definition of the RG assume that researcher aims
at eliciting the respondent’s own constructs which reflect own meanings (Jankowicz,
2004, p. 56). This represents probably the greatest value of the technique and should
be pursued when applying RGT to benefit from its full capacity.
Rarely appears to be enough to use only suggested or supplied constructs. For
example, if the researcher provides the constructs for the study, then it is likely that
his interest lies in discovering commonalities between the ways people evaluate
elements on certain dimensions. Even in such case of supplying constructs, authors
still recommend carrying out a series of pilot studies involving free elicitation of
constructs in order to facilitate the selection of the most representative ones (see
Neimeyer, 1993 for review). Further, in 4.5.2.3, the issue about reasonability for
partly supplying constructs for particular objective is discussed. Here the context is
stressed on the basic procedure of construct elicitation.
4.5.2.1. Construct elicitation approaches
Two approaches to construct elicitation are common in the personal construct
literature: dyadic elicitation and triadic elicitation (Caputi and Reddy, 1999).
The dyadic elicitation implies consideration of two elements at a time. The
interviewee is asked to specify how the two elements are alike or different in some
particular way. If the two elements are perceived to be similar, the contrasting pole is
determined by considering remaining elements from the elements set, identifying an
element which is different from the considered pair and describing the difference
(Caputi and Reddy, 1999).
In triadic elicitation the interviewee is presented with three elements from the
elements set at a time. He is then asked to specify some important way in which two
of the three elements are alike and thereby different from the third. A word or phrase
may be used to represent first the similarity which forms an emergent pole. The
respondent then generates a contrast, i.e., implicit pole by stating the difference of
two elements from the third one (Caputi and Reddy, 1999). Therefore, the construct
elicited appears to be strongly determined by the particular triad.
Grids employing dyadic elicitation tend to provide lower ordination scores, more
functionally independent constructs, and lower level of cognitive complexity than
grids using triadic elicitation (Caputi and Keynes, 2001).
4. Repertory grid technique within the personal construct psychology
85
By using dyadic approach the elicited grids do not reflect the full complexity and
variety of constructs potentially available. This approach works well where the basic
triadic procedure is too difficult and confusing for the interviewee, as in working with
children under 10 to 12 years of age (Salmon, 1976), individuals with learning
difficulties (Barton, Walton and Rowe, 1976) and the deaf (Baillie-Grohman, 1975).
However, if the triadic procedure can be used adequately, it is suggested to do so
(Jankowicz, 2004, p. 53; Fransella et al., 2004, p. 28).
The implicit pole of the construct can also be elicited in two different ways: with an
opposite method or with a difference method. Each method can be used in both,
dyadic and triadic elicitation approaches.
In difference method the contrast pole appears from the comparison of two similar
elements with the third element of the triad or, in case of dyadic elicitation, from
comparison of two elements if they appear to be different for respondent or from
comparison of similar elements with one of the remaining elements in a set. Anyway,
in difference method contrast pole emerges by discovering the difference between
the elements.
In opposite method contrast pole is elicited by asking the opposite of the word or
phrase expressed in the initially elicited emergent pole (Fransella et al., 2004, p. 28).
• Triadic difference: implies presenting three elements at a time and asking
“How are two alike in some way, but different from the third?”
• Triadic opposite: implies presenting three elements at a time and asking “How
are two alike in some way?” followed by “What is the opposite of that?”
• Dyadic difference: presenting two elements and asking “How are these two
alike or different?” If a difference is given, this is taken to be a contrast pole of
the construct. If a similarity is given, the person is asked to find among
remaining elements one that would represent a difference.
• Dyadic opposite: presenting two elements and asking “How are these two
alike or different?” If a difference is reported, this specifies the two poles of the
construct. If a similarity is reported, the person is asked for the opposite of that
similarity (Fransella et al., 2004, p. 30).
Kelly’s original method utilized a triadic difference elicitation procedure. This method
produces constructs that are less functionally independent and more meaningful in
4. Repertory grid technique within the personal construct psychology
86
that they are better able to discriminate among discussed elements. The construct
sets revealed by this method are more cognitively complex (Caputi and Reddy,
1999).
Hagans, Neimeyer and Goodholm (2000) pointed out that when using the triadic
opposite method the contrast pole may not directly apply to any of the elements in
the grid but the similarity pole was elicited under condition of presence of the third
element. Therefore, “the third element may well play a part in the verbal label
elicited”. However, an important consideration, further provided by Hagans and his
colleagues (Hagans et al., 2000) is that asking for the ‘opposite’ of the elicited
similarity pole produces more extreme ratings, i.e. more highly polarized constructs
than if a person states how the third element in the triad is different from other two.
For example, if the opposite to beautiful is not so attractive the person can give an
extreme rating, e.g., ‘5’ regarding the car he is not really attracted to. But if the
opposite is ugly the rating could only be ‘3’. Providing opposites may produce
construct poles that are too negative to be applied to the elements. The consequence
is that a greater number of elements are assigned to a single pole of the construct,
thereby decreasing differentiation of elements along constructs and provoking an
artificial deflation of differentiation measures (Hagans et al., 2000).
Another important consideration with regard to both the triadic and dyadic opposite
methods is that their instructions to produce an ‘opposite’ allow the possibility that the
third element or at least one element in the full element set is not applicable to the
construct and may fall outside the range of convenience of the construct. In contrast,
both triadic and dyadic difference methods require that the contrast pole is developed
from, and therefore applicable to, at least one element in the grid (Hagans et al.,
2000).
Based on obtained evidence (see Hagans et al., 2000; Caputi and Reddy, 1999) it is
not yet possible to advocate on behalf of any one preferred form of repertory grid
procedure. However, researchers need to be aware that application of opposite
methods of personal construct elicitation may cause development of a greater
number of positive emergent poles, more extreme, negative contrast poles, and
significantly lower levels of construct system differentiation.
According to Hagans, Neimeyer and Goodholm (2000), the triadic difference method
represents traditional, most frequently used method for construct elicitation applied
within the RGT.
4. Repertory grid technique within the personal construct psychology
87
Kelly offered six assumptions that underlie his original Role Construct Repertory Test
but which are equally applicable to subsequent grid modifications and to construct
elicitation in general (cited in Fransella et al., 2004, p. 23):
• The elicited constructs should be permeable. This means that each elicited
construct can be applied by the respondent to all other elements from the set
besides the three ones from which the construct was elicited.
• Pre-existing constructs should be elicited. It is assumed that there is a certain
remaining degree of permanence in the constructs elicited and it rarely
happens when respondent produces a new (for himself) meaning and,
therefore, construct during the process of elicitation.
• The verbal labels attached to the constructs should be communicable. The
interviewer should have some reasonably accurate idea of what the
respondent intends.
Other three assumptions were made regarding using people as elements. However,
essential ideas should be suitable for other grids.
• The construct should represent the respondent’s understanding regardless of
the judgement whether it is right or wrong from other people’s point of view.
• The constructs elicited should be explicitly bipolar. By stating what a thing or a
person is, respondent is also saying that which it or he is not.
• People should not dissociate themselves entirely from the elements or from
the constructs elicited. They must be able to see themselves somewhere
along the construct dimensions.
Although the last assumption relates only to the case when people are used as
elements, it can be resumed that respondents should not provide objectivism but
rather subjectivism by seeing themselves and their personal perceptions and
attitudes related to the discussed subject matter.
The following description of RG elicitation procedure is focused on the approaches
and methods that are most appropriate with respect to the current research
objectives and, therefore, applied in this work. Description of further numerous
modifications of grid usage depending on its objectives and field of research can be
found in literature (e.g., Fransella et al., 2004; Jankowicz, 2004).
4. Repertory grid technique within the personal construct psychology
88
In the context of this study Kelly’s original triadic difference method of elicitation is
applied. To make sense from similarity and contrast bearing in mind that constructs
are dichotomous an interviewee needs to consider three elements at a time and so
provide two contrasting or dichotomous poles of a construct.
Practically, it is proved to be easier and friendlier to think and compare elements
when each element is written on a separate card. Respondent is presented with
appropriate three cards at a time, laid out in a triangle to provide one construct. Many
people find it helpful to think about the constructs if they have something tangible to
move around on a table (Jankowicz, 2004, p. 33; Eden and Jones, 1984).
Respondents are expected to produce one construct per triad. However, sometimes
a person may produce several constructs from one triad by reorganizing the three
elements in different contrasts (Dillon and McKnight, 1990). This need should not be
discouraged. Although the process normally continues by offering further selected
triads, this process is, indeed, designed to help articulation and provide the analyst
with a sense of order rather than being a rigid method (Eden and Jones, 1984).
The amount and sequence of selected triads is an important issue when talking
about grid reliability. Triads should be selected either randomly or according to a
reasonably determined consequence. It is also important that a person considers all
of the elements at one time or another.
The number of constructs to be elicited depends on the research.
The procedure depends on the purpose, given time for interview and whether further
analysis to be conducted within a single grid or within several grids. When time for
interviewing is not limited and the purpose of investigation is to obtain an in-depth
detailed view of a single person, the researcher is aimed at more information and
would provide all possible combinations of elements revealing maximum constructs
and living all of them even if there is a potential overlap and some constructs seem
very similar (Jankowicz, 2004, p. 45). All this is done in order to reach as full picture
of personal construing capturing his tacit knowledge as possible. In these conditions
the RG elicitation procedure is relative free and elicited information and details
depend a lot on the negotiation between interviewer and interviewee.
Thereby, sometimes constructs are elicited by considering different triads until no
more constructs seem to be emerging, for instance, when constructs start to repeat
or if the respondent cannot supply any more constructs (e.g., Steed and McDonnell,
4. Repertory grid technique within the personal construct psychology
89
2003; Eden and Jones, 1984). However, the researcher has to be careful with
recognizing among simple repeats those meanings that bare a certain importance to
the respondent. It was found from an experiment that people repeat some constructs
because they are important to them. In the test people were allowed to repeat
constructs and then to rate the constructs in terms of importance. As a result, the
repeated constructs were rated more highly signifying their higher importance
(Fransella et al., 2004, p. 46).
However, when several grids are to be analyzed together or in order to compare
single grids assessment methods often require same amount of constructs and,
moreover, constructs should be elicited from the same set of triadic combinations.
This not only increases the validity and reliability of findings but also allows for
additional inferences regarding, for example, ways people differentiate same
elements.
4.5.2.2. Construct specification
According to the “Organization Corollary” of Kelly’s PCT personal constructs do not
operate in isolation from each other but form an integrated construction system
containing ordinal relationships between constructs (Jankowicz, 2004, p. 186) Thus,
elicitation of knowledge implies both eliciting the content as well as the hierarchical
structure of the subjective meanings that people attach to the set of elements. It is
essential to expose as much of this construction as it is necessary for the purpose of
an interview.
By using the RGT individual's constructs can be elicited at different levels of
psychological abstraction. Four interviewing procedures designed for this purpose
are triading, laddering down, pyramiding and laddering up (Banister et al., 1994).
Bipolar constructs elicited through triading by asking a person how two elements of a
triad are alike and different from the third can be stated in more or less detail – in
rather abstract terms or in very precise terms (Jankowicz, 2004, p. 64). However, “a
‘good’ construct is also one which is appropriately detailed” (Jankowicz, 2004, p. 33).
The interviewer has to reveal the real respondent’s meanings attached to expressed
labels.
Two procedures which take the person either ‘up’ a ladder or ‘down’ a ladder allow to
elicit the more abstract values or, in contrast, more concrete meanings associated
4. Repertory grid technique within the personal construct psychology
90
with constructs obtained from triading (Marsden and Littler, 2000a; Fransella et al.,
2004, p. 44).
Laddering down provides an opportunity to make a person express the construct in
sufficient additional detail so that interviewer understands the intended meaning as
precisely as possible. After the respondent stated a construct he is asked “how do
you mean?” or “what do you mean?” to encourage him to express each pole of the
original construct in greater detail being more explicit about the content involved
(Jankowicz, 2004, p. 64).
The interviewer always has a choice: whether to accept the originally expressed
construct or whether to ladder the person ‘down’ to look for more detail. Sometimes
people may provide constructs which are actually clichés. This can happen when a
person has never thought about the discussed issue before. Therefore, the clichés
may reflect insufficient consideration and don’t represent interviewee’s actual view.
However, they might also represent his views precisely. Also what may seem
superficial or vague to the interviewer may appear to be neither superficial nor vague
to the respondent. Free-flow discussion, interviewer’s own empathy and sensitivity as
well as awareness of respondent’s background, experience, vocabulary, etc. can be
helpful during elicitation and interpretation of personal meanings (Jankowicz, 2004, p.
34; Fransella and Bannister, 1997).
The procedure of revealing clear, operationally defined, non-clichéd constructs can
be conducted differently depending on the topic of the grid or on the circumstances in
which the topic is investigated. If the interviewer needs to identify constructs at a
detailed level he would use simple laddering down procedure. Practically this is done
by asking the question “how” or “what” first about the emergent pole and then
providing the same question about the implicit pole of the original construct.
Laddering down is used to express the construct more precisely.
Sometimes it is needed to investigate the variety of a person’s construing i.e., the
range of person’s points of view. In this case an alternative to laddering down
procedure called ‘pyramiding’ is used.
Pyramiding was suggested by Landfield (1971) and implies elicitation of increasing
subordinate constructs from the original one. The person is first asked the question
“how” or “what” about the emergent pole of the original construct. Then he needs to
state opposite or contrast to the just expressed pole. Afterwards, the same procedure
4. Repertory grid technique within the personal construct psychology
91
is done regarding the implicit pole of the original construct. The shape of the
gathered data structure at the end reminds a pyramid, where each pole of
superordinate construct is expanded by the subordinate bipolar construct (Jankowicz,
2004, p. 67; Fransella et al., 2004, p. 43).
To summarize, the main difference between laddering down and pyramiding
procedures lies in the purpose of constructs specification. In laddering down the
superordinate-subordinate relationship between constructs is described by seeking
ever-more specific ways of expressing constructs. Whereas in pyramiding this
relationship is elaborated in ever-expanding detail by examining and expanding on
each pole separately (Jankowicz, 2004, p. 66)
In the current study laddering down technique was appropriate to be used in the
interviews. Very often it was needed to understand the meanings attached to some
expressions or jargons used by respondents. Therefore, the concern was to get a
clear, operationally defined and non-clichéd construct.
Pyramiding was less appropriate because of time consuming and the fact that
respondents represent intellectual, self conscious and determined people who are
able to clearly express their meanings which are enough determined and precise.
Laddering up is in a sense an opposite of laddering down or pyramiding. The
procedure is designed to elicit superordinate constructs. Hence, when laddered
down, constructs are more precise and detailed, whereas when laddered up, the
interviewer arrives at more abstract meanings associated with original constructs
(Jankowicz, 2004, p. 186; Marsden and Littler, 2000a).
Laddering up technique is developed for eliciting means and ends, i.e., mental
representations of higher levels of abstraction than constructs. In order to keep up
with the sequence of the elicitation procedure, laddering up procedure is described
further (4.5.4.1) as it enables to reach personal values (Jankowicz, 2004, p.188) from
some of the elicited constructs selected after elements are rated on all of the
constructs.
4.5.2.3. Elicited and supplied constructs
The repertory grid procedure fundamentally assumes elicitation of interviewee’s own
constructs. However, the purpose of the procedure is to understand how the person
construes the topic in question. One issue that emerged in grid usage is whether or
not supplied constructs give the same disclosure of the matter as elicited constructs.
4. Repertory grid technique within the personal construct psychology
92
For some purposes it can be useful to supply interviewee with constructs, at least in
part in order to enhance an extent of complexity and richness of interview (Fransella
et al., 2004, p. 46). For example, the interviewer needs to find out how interviewee’s
own constructs relate to some other construct which may not be in the interviewee’s
repertoire but which is still important to the interviewer. Interviewer may provide a
construct to see which of the interviewee’s constructs are used in a similar way, i.e.,
are rated the same way (Jankowicz, 2004, p. 56).
However, the interviewer may face a problem when the supplied construct appear
meaningless to the interviewee. By providing a construct an interviewer is supplying
a verbal label to which the person will attach his personal construct. Provided
construct labels may range from those that are identical with constructs used by the
interviewee to constructs that are nonsense to him. As example for the later may be
constructs with verbal labels in a foreign language, or from specialized field
containing jargons. Therefore, it is important to bear in mind when supplying a
construct that the labels should be meaningful to a person.
According to Kelly’s Individuality Corollary “persons differ from each other in their
construction of events” what implies that constructs are personal in the sense that the
person is able to attribute them to events and make something of them. Elicited own
constructs may appear not as useful for an interviewer as constructs he could have
provided himself, but he can usually make some sense of them (Fransella et al.,
2004, p. 46).
The matter of elicited versus supplied constructs attracted a substantial research
attention. Some researches provided evidence that although supplied constructs,
selected in a proper way, can be meaningful to other people, own personal
constructs are more meaningful and useful. Assuming that the more meaningful and
important the construct is the more extremely people rate it, Bonarius (1977) showed
that people provide more extreme ratings on constructs elicited from them than on
supplied constructs. It was also showed that elicited constructs produce more
differentiation or cognitive complexity (for review of some studies see Neimeyer,
1992).
The research conducted within this study is aimed, among others, at discovering
personal meanings and revealing expressions which customers use when talking
about the discussed matter. Therefore, most of the constructs were elicited from
respondents.
4. Repertory grid technique within the personal construct psychology
93
However, two so called “overall” constructs were supplied in order to apply further a
certain kind of content analysis developed by Honey (1979). Overall construct implies
obtaining respondent’s ‘overall summary’ on the topic in question by asking him to
provide ratings on a construct which summarises his overall view on the discussed
topic (Jankowicz, 2004, p. 56). For example, if a grid is designed to explore
consumers’ perception of a variety of refreshing beverages in order to make
improvements in taste, the grid might involve a supplied construct which assessed
the extent to which they slaked their thirst: ‘overall, I fully slaked my thirst – overall, I
did not satisfy my thirst’.
Honey’s content analysis justifies an enclosure of supplied constructs but the
purpose of its providing implies possibility to make additional inferences about
elicited constructs, their interrelations and relation to the topic rather than elicitation
of personal meanings. Thus, Honey did not suggest to use supplied construct instead
of elicited ones but to add one or two supplied constructs in addition to the elicited to
expand the variety of findings available from analysis of RG data. Honey’s content
analysis is described further in the chapter devoted to the analysis methods applied
for RG data (see 4.6.1.3).
4.5.3. Rating grid or element comparison phase
Once a representative lists of elements and constructs are obtained the next phase is
the construction of a grid, i.e., a matrix of elements and constructs. The elements are
placed vertically on the top half of the grid and constitute the columns. The constructs
are placed horizontally so that on the left side of the grid the emergent poles and on
the right side of the grid the implicit poles of constructs are placed. Thereby,
constructs constitute the rows of the grid. The arrangement corresponds to the RGT
implication where each construct can be related to every element via a rating scale.
Therefore, it is necessary that all the elements fall within the range of convenience of
the constructs (Feixas and Alvarez, 1996).
As soon as columns and rows of the grid are attributed with meanings, participants
are asked to rate all elements in terms of each construct (Cassell et al., 2000).
Each participant is introduced to the idea that the bipolar constructs should be
considered as the two ends of a rating scale (Fransella et al., 2004, p. 59). The end
of the construct which describes the similarity, i.e., emergent pole should be written
down on the left and the rating assigned to this pole on a 1-5 point scale is ‘1’. The
4. Repertory grid technique within the personal construct psychology
94
contrasting end, i.e., implicit pole is written on the right with a ‘5’ rating assigned
(Jankowicz, 2004, p. 48).
It was discussed in 4.3.4 that it does not appear to matter greatly what length of scale
should be used. However, it is suggested to use scale containing a mid-point as
neutral. This is supported by numerous applications of repertory grid using ‘5’ or ‘7’
point scales (e.g., Jankowicz, 2004, p. 37; Fransella et al., 2004, p. 63; Marsden and
Littler, 2000a; Björklund, 2005). In the current study ‘5’ point scale is used, therefore,
further descriptions regard to this length of scale.
Several issues may arise when asking for ratings.
The first issue namely ‘direction’ of rating is whether each construct should be rated
on all elements in turn, as Kelly first suggested, or whether each element should be
rated on all constructs in turn. However, at the present time there is no consistent
evidence that the direction of rating affects grid measures. Hence, it is suggested to
follow Kelly’s chosen way of rating each construct on all elements before moving to
the next construct (Fransella et al., 2004, p. 64). This way also seems more
meaningful from the point that respondent rather compares elements on different
dimensions than simply evaluates each element on those dimensions. This is
followed from the initial procedure of the triadic or dyadic elicitation where constructs
are elicited by comparison of three or two elements respectively.
Second consideration is whether it would make sense and would be more friendly for
a respondent if the left end of the construct is the one that positively evaluated and
the right end is with a negative contrast. It sounds right until a construct has a clear
positively and negatively evaluated end, for example ‘nasty versus nice’ or ‘evil
versus good’. However, not all constructs do have such clear ends and it is hard to
judge which end is better, e.g., ‘light versus dark’ or ‘speedy versus calm’. Also,
sometimes, one end can be slightly better, but only a little, and vice versa. This can
lead to enormous confusion during constructs analysis. Thereby, the form of
recording should always follow the rule ‘emergent on the left, and implicit on the
right’, regardless of ‘which way round is positively evaluated’ (Jankowicz, 2004, p.
48).
Third consideration arises when a construct does not apply to all of the elements
from the element set. How this should be handled in order to conduct further analysis
properly, especially, when a certain software is used which is not able to cope with
4. Repertory grid technique within the personal construct psychology
95
missing ratings? Researcher’s task is to let the interviewee express his meanings as
precisely as possible and when he cannot provide a rating on a particular element,
the researcher should not insist on a rating. Otherwise, it would deform the intended
meaning. It is suggested to leave the cell blank without any rating. However, if the
software is used which demands the ratings in each cell of the grid the least damage
would be to put a neutral mid-point rating (Jankowicz, 2004, p. 49).
Fourth issue concerns the turn in which the elements should be rated on a construct.
That is, whether interviewer should always ask for the ratings of the elements in the
triad first, before getting the ratings of the remaining elements or whether to simply
run along the row each time. Jankowicz (2004, p. 47) suggests that, at first, elements
from the triad should be rated and then remaining elements. He finds it more logical
to the interviewee and simplifies the recording for unpracticed interviewer. Still, he
admits that both ways are acceptable.
During conduction of the current study it was assumed that when first asking the
respondent to rate the elements from the triad it is likely that respondent would
assign extreme ratings to the elements, especially to the element which differs from
other two. However, during the rating of remaining elements there could be elements
which he would rate more extremely than the one from the triad but, by that moment,
he would already used the most extreme rating. To avoid this confusion, the way of
rating along the row was applied. By obtained results it is seen that very often
elements from a triad were rated less extremely than remaining ones, although still
without reversing the ratings unintentionally.
Sometimes, rating value assigned to element from presented triad was neutral, i.e.,
‘3’ despite that the contrast pole was defined by consideration of this element. This
happens because, at the end of the day, as was pointed by Stewart and colleagues
(1981): “the rating figures carry no inherent meaning in themselves, but simply
provide a way in which consumers can position the elements in relative terms on
each of their construct dimensions, thus, providing the researcher with a richer
picture of the overall structure of their construct system” (cited in Marsden and Littler,
2000a).
In general, grid form enables to get closer to the functional meanings of elements
and constructs. The interview form and the procedure allow to see how each element
is rated on each construct and how each construct is being used, thereby, allowing
comparison of both the elements and the constructs respectively.
4. Repertory grid technique within the personal construct psychology
96
4.5.4. Eliciting personal values
Elicitation of personal values within the PCT and RGT is supported by Kelly’s
‘Organization Corollary’ which assumes that personal constructs do not operate in
isolation from each other, but form an integrated system comprising ordinal
relationships between the constructs (Jankowicz, 2004, p. 186). Therefore, eliciting of
personal values implies at first, arriving to the most abstract mental representations
which represent personal values and secondly, revealing the relationships between
derived personal values.
4.5.4.1. Capturing personal values: laddering up technique
Laddering refers to an in-depth, one-on-one interviewing technique used to develop
an understanding of how consumers translate product related knowledge about
characteristics and attributes into person or self related knowledge which is
meaningful associations with respect to self, following the MEC theory (Reynolds and
Gutman, 1988). Laddering up procedure, first described by Hinkle (1965), is
designed to elicit increasingly superordinate constructs – constructs of a higher order
of abstraction then those elicited from the original triads or dyads of elements.
Hence, by giving an interviewee the laddering up task the interviewer arrives at more
abstract meanings associated with original constructs (Jankowicz, 2004, p. 186;
Marsden and Littler, 2000a).
The procedure represents a discussion between interviewer and respondent using
repetitive and interactive questions to dig deeply into the product associative
meanings. Based on the elicited respondent’s construct which refers to product
characteristics and attributes the value hierarchy begins to be built, where the
interviewer derives the reasons for preferences pointed out by the respondent, and
thus, moves the answers to the abstraction levels corresponding to consequences
and values. The researcher leads the respondent to abstraction by repeatedly asking
him “why is this important to you” concerning each arisen level. By sequentially
answering this question, the respondent expresses consequences derived from
attributes, and personal values that arise from consequences (Veludo-de-Oliveira et
al., 2006). Questions based on the question “why is it important to you” are
continually asked until a personal value is revealed (Wansink, 2000).
For example, the interviewer can ask an auto buyer: “Why presence of comfort
electronics and technologies in an auto is important to you?” The respondent can
4. Repertory grid technique within the personal construct psychology
97
answer: “Because I can feel myself comfortable and relaxed during the ride”. The
interviewer continues: “Why is it important for you to feel relaxed during the ride?”
The respondent’s answer can be: “Because it helps me to refresh my strength and,
therefore, stay always effective at work”. Interviewer: “Why is it important to stay
always effective at work?” Respondent: “Because I want to become successful in my
career and realize my professional goals and ambitions”. Thereby, the interviewer
arrived to the laddering shown in Table 4.
Personal value professional success and self realization
stay always effective at work
feel comfortable and relaxed during the ride
Product characteristic / attribute comfort electronics and techniques / comfort
Consequences
Table 4: An example of a ladder derived from one construct
The procedure implies that the product characteristic from which ladder is derived is
important to the respondent so that he can associate it with more abstract
consequences and, even, with personal values. Therefore, it is another question of
the interviewing procedure how to identify important characteristics and attributes
among all elicited. This is one of the critical issues which are often discussed by
researchers with respect to the laddering up technique. In this work the process of
recognizing most preferred and important constructs and construct poles towards
which laddering task is to be applied is discussed in 5.3 as it relates to the proposed
methodology. In the current part the description of laddering up basic procedure from
any specific meaning is discussed in order to illustrate how a personal value can be
revealed.
4.5.4.2. Prioritizing personal values: resistance-to-change technique
One of the most important concepts in Rokeach’s (1973) theory of human values
implies that once a value is learned by a person, it becomes part of personal value
system in which each value is ordered in priority relative to each other (cited in
Kamakura and Novak, 1992). It is assumed that most situations in life activate more
than one value and often involve a conflict between values, for example, striving for
hedonic pleasure and professional achievement. In this case the person uses his
4. Repertory grid technique within the personal construct psychology
98
value system as a tool for conflict resolution and decision making so that self-esteem
can be maintained or enhanced. Therefore, value system consisting of prioritized
personal values, rather than a single value, can provide a more complete
understanding of motivational forces driving consumer attitudes and behavior
(Kamakura and Novak, 1992).
The resistance-to-change procedure is a technique designed for identifying which
personal values are more important and central and which are less so and, thereby,
describing an individual’s hierarchy of personal values.
It is argued that core meanings representing personal values are relatively resistant
to change. That is, the more important and central an individual belief, the more
important it is for a person to see the world in that way, and the less likely an
individual is to change his mind about it. Therefore, in order to recognize which
personal values are more important for the individual and which less, he is asked to
make a choice between values presented in a way which focuses attention on their
relative desirability. The choice is made by a person according to the dominance of
desirability of one option over another. Hence, the respondent appears to be ready to
compromise on some choices indicating less important values which are relatively
more open to change, and he is not ready at all to compromise on the others
indicating the bedrock, most important values which are highly resistance to change
(Jankowicz, 2004, p. 199).
At first, the first two personal values identified by the laddering technique are
presented to the respondent as a choice between ‘experiencing one at the cost of
another’ and visa versa. For example, if the first value is ‘A’ versus its contrast ‘-A’
and second value is ‘B’ versus its contrast ‘-B’, such as follows:
A -A B -B
achievement and success
aspirations failure
enjoyment and pleasure of life
no feel of joy and gratification
_ _
The respondent is presented with a choice between ‘A at the cost of –B’, or ‘B at the
cost of –A”. The choice actually makes the respondent to think about how strongly he
feels about each personal value and which option has the greatest reward with the
least pain.
4. Repertory grid technique within the personal construct psychology
99
In this case the respondent is asked to imagine the choice as an existential one and
make a choice between ‘imagine in the future, even at the end of your life, you
understand that you have achieved everything you have strived for and gained
success, however, you also realize that you do not enjoy your life and there is no
gratification and joy in it’, or ‘imagine in the future, even at the end of your life, you
understand that your life is full of pleasure, joy and excitement but you have not
managed to achieve everything you have been striving for to realize your ambitions
and aspirations’. By making the choice between these options the respondent points
the value which he prefers over the other. For instance, he chooses the second
option being more prepared to compromise on success rather than on the enjoyment
of life. On the record sheet the option ‘B at cost of –A’ is marked with ‘yes’ indicating
preference of value B over value A (Table 5).
Procedure of resistance-to-change technique implies presenting the respondent in
turn with all pairs of values revealed through laddering up technique. Thus, if there
are three revealed personal values, the respondent is asked then to choose between
‘A at the cost of –C’, or ‘C at the cost of –A’ and, afterwards, he is asked to choose
between ‘B at the cost of –C’, or ‘C at the cost of –B’. And if there are more values
then each value is compared to all the remaining ones as shown in Table 5:
A -B B -A Yes
A -C Yes C -AA -D Yes D -AB -C Yes C -BB -D Yes D -BC -D D -C Yes
Either Or
Table 5: Subsequent comparison of personal values
When all the values are compared with each other, the number of times that each
value is preferred over another is counted. In Table 6 the number of ‘yes’ responses
against each personal value is counted and the value hierarchy based on the
revealed order is built up. In case of two or more personal values chosen an equal
number of times, they are recorded side by side, rather than one above the other.
The hierarchy revealed through resistance-to-change technique may also identify
values that are not sincere and central for a person, but were expressed rather
fictitious to seem more meaningful. Such values, if any, are likely to appear at the
bottom of the hierarchy.
4. Repertory grid technique within the personal construct psychology
100
Personal Value Times chosen Hierarchy of personal valuesA 2 BB 3 AC 0 DD 1 C
Table 6: Frequency of value preferences over other ones and a value hierarchy
Another important inference that can be made from the revealed hierarchical value
system relates to the subordinate constructs that draw on values and are particularly
representatives of these values. The higher the personal value in the hierarchy, the
higher its resistant to change and, thus, the representative subordinate constructs,
including product knowledge related levels of abstraction, i.e., product attributes, are
tend to be similarly resistant to change (Jankowicz, 2004, p. 203).
4.6. Analysis of a repertory grid
Once the RG is elicited, the resulting data matrix is potentially as complex as the
rating system allows, and its size depends on the ultimate number of provided or
elicited elements and constructs used (Feixas and Alvarez, 1996). There can be a lot
of these data, for example, a 4 x 4 grid contains 28 pieces of data and an 8 x 8 grid
contains 88 pieces of data which consists of the element figures, construct poles
labels and numbers which indicate the relationships between elements and
constructs (Fransella et al., 2004, p. 82). The matrix format (Table 3) is a compact
representation of a deceptively large amount of information contained in the RG. A
researcher is, therefore, faced with a great deal of information which must be
processed so that the initial structure can be retained without too much loss of
information (Feixas and Alvarez, 1996).
One of the most interesting and attractive features of the RGT is the wide variety of
different types of analysis that can be applied to the elicited information (Hessenzahl
and Wessler, 2000). Many efforts made to mathematically synthesize the basic
structure of RG data so that it can become useful to a researcher, starting from non-
parametric factor analysis proposed by Kelly himself to the more modern multivariate
analyses (Feixas and Alvarez, 1996). This chapter briefly covers the most significant
methods proposed in the literature.
Analyzing grid data can be conducted by means of one of or a combination of a
number of ways that can broadly be categorized as quantitative or qualitative.
4. Repertory grid technique within the personal construct psychology
101
Furthermore, in order to better understand what is being said in a grid a blend of
both, qualitative and quantitative approaches are very often appear useful to be
applied.
The RGT permits to capture the meanings where each meaning is expressed by both
words and numbers (or symbols). People need words to express and communicate a
construct – a dimension through which meaning can be expressed; and they need
numbers to describe the elements with respect to that dimension, i.e., “to ascribe the
meaning attached to the elements by their positions on the various constructs in the
grid” (Jankowicz, 2004, p. 72).
In general, in grid analysis the main task is dual: first is to identify a participant’s
meanings, and second, is to draw whatever implications seem to be appropriate to
the investigator depending on his research objectives (Jankowicz, 2004, p. 72).
Originally, Kelly proposed the first system for summarizing grid data using
dichotomous scores (described in 4.3.4) with an implication of a form of non-
parametric factor analysis that could be done by hand. Although Kelly’s creative
mathematical contribution was admitted and appreciated, it is only applicable to the
dichotomically scored grids what certainly limits overall RGT’s potential.
Nowadays, the amount of approaches and techniques to be used for grid analysis is
numerous however most of them present original as well as forms or modifications of
four principle methods of analysis that are traditionally being used (Stewart et al.,
1981):
� Frequency counts;
� Content analysis;
� Cluster analysis;
� Principal component analysis.
The first two methods can be described as qualitative approaches which are
concerned with analyzing the contents of the grid. Two remaining are quantitative
approaches which also analyze the inter-relationships contained in the grid. Choice
of the method will depend on the purpose of the grid and what aspects of the
4. Repertory grid technique within the personal construct psychology
102
person’s construct system a researcher wants to understand (Gutierrez, 1989 cited in
Marsden and Littler, 2000b; Stewart et al., 1981).
The first two methods are usually used for analyzing more than one grid. They are
aimed at finding commonalities and coincidences between grids either from different
people or from one person but elicited on several occasions or in different time.
The remaining twp methods are mainly intended for analyzing relationships between
constructs and/or elements within a single grid. They include more mathematical and
statistical procedures and mainly deal with the structure of the information captured
in repertory grid rather than its content (Jankowicz, 2004, p. 93; Stewart et al., 1981).
4.6.1. Qualitative methods for repertory grid analysis
4.6.1.1. Frequency counts
Frequency counts allow to count the number of times a particular element or a
construct is mentioned. When elements and constructs are not supplied during the
elicitation procedure, the number of times a certain element or construct has been
stated in response to eliciting question gives a useful insight in itself and can be
valuable when comparing different people or groups of people and revealing
common trends. It can be especially useful when elements are elicited by using
identical elicitation categories (described in 4.5.1) revealing elements which can be
different as instances, i.e., different brands, products, people or events but present
the same set of categories. The number of mentioned elements which can be, for
example, different brands of a certain product category may already contribute to the
understanding the topic in question (Stewart et al., 1981).
Defining the index of frequency of constructs is often more complicated and requires
complementary methods because different interviewees may rarely produce identical
constructs in terms of used wording. As one of the solutions, one can select a fairly
common construct or group of constructs and make a pattern to see how they are
used but he needs to be careful to bring misleading when similarly sounded
constructs imply absolutely different meanings revealed during the laddering
procedure (Stewart et al., 1981). Therefore, frequency counts often appear more
appropriate regarding elements, whereas, classification of different constructs as well
as of different types of constructs provided by interviewees is usually conducted by
means of content analysis (Jankowicz, 2004, p. 146).
4. Repertory grid technique within the personal construct psychology
103
Some authors, for example Stewart, V., Stewart, A., and Fonda, N. (1981)
differentiate frequency analysis and content analysis putting them in one list of
variety of methods used for repertory grid analysis (Stewart et al., 1981). Green
(2004), however, assigns frequency count analysis as one of three basic approaches
to content analysis. The procedure of frequency analysis basically presents the
simplified content analysis where the content itself doesn’t need to be interpreted and
coded before counting for coincidences. Still, as frequently counts method has got its
own procedure based on certain terms, it does have a right to be considered as an
independent method. In this study it is adhered to the Stewart’s and her colleagues’
definition but this question is only a matter of definitions and it is up to the reader
which definition to accept.
4.6.1.2. Content analysis
Content analysis is a set of techniques that have a potential to assist in examining
different kinds of verbalized or other meaningful data including grid content. It is a
common type of category generation that involves revealing of patterns in the data
and then assigning each pattern into a category (Highlen and Finley, 1996).
According to Krippendorf (2004), “content analysis is a research technique for
making replicable and valid inferences from texts (or other meaningful matter) to the
contexts of their use” (cited in Green, 2004). This definition contains three
fundamental features of content analysis, namely (Green, 2004):
a. The findings from a content analysis should be able to be replicated by others;
b. The analysis should measure what it claims to measure;
c. Content analysis is not limited to textual data.
A variety of different approaches to content analysis is available to researchers,
however, three basic ones can be emphasized. The first is to examine the co-
occurrence of words. That means to count how many times a combination of certain
words is used together. The second is to code the text units (e.g., words, sentences
or paragraphs) using some form of coding scheme (Green, 2004).
Green (2004) describes three basic approaches including frequency counts and two
described in the previous paragraph. The first two approaches, i.e., frequency counts
and words co-occurrence are more likely to be suited to relatively larger texts
4. Repertory grid technique within the personal construct psychology
104
because the likelihood of the same word repeating a number of times or two words
appearing together will always be less than the likelihood of either word appearing
individually because of the relatively small number of words provided in the grid in
general.
Coding of text units is an approach which is mostly used in PCP research (Green,
2004). The analysis of data is a non-linear simultaneous process of brining order,
structure and meaning to the data in search for general statements about
relationships among categories of this data (Highlen and Finley, 1996). The central
here is ‘meaning’ and the objective is to try to understand the content and complexity
of those meanings derived from participants rather than take some measures of
frequency (Smith, 1995).
Some concerns regarding interpretation of individual meaning made researchers
either to refuse application of content analysis or, as later, to find explanation and
solution to overcome confusion.
The first concern suggested by Kelly (1995) is that constructs are not necessarily
equivalent to word labels. Four possible relationships that may exist between
construct meanings and word labels are emphasized (Shaw, 1994):
a. Agreement between construct meanings and word labels;
b. Different words being used for the same construct meanings;
c. The same word being used for different construct meanings;
d. Different words being used for different construct meanings.
Second concern is that individual meaning cannot be readily categorized by another
person, especially in the absence of elaboration or understanding of context and
application (Yorke, 1989). Kelly (1995) meant that in order to avoid serious
misinterpretation of respondent’s meaning, a researcher should realize that he is
setting out to learn a new language rather than relying on presumption that the
respondent agrees with the dictionary (cited in Green, 2004).
Thus, researchers need to find the proper ways or improve the existing ones to
interpret personal meanings more accurately. One of the defining aspects in
considering the trustfulness and precision of interpretation is related to the nature of
4. Repertory grid technique within the personal construct psychology
105
coding process. That is, an important difference in application of content analysis
provides whether the categories utilized are theory driven or data driven (Simon and
Xenos, 2004; Stewart et al., 1981).
Theory driven content analysis categorize data in terms of categories developed
beforehand on the basis of theory or on empirical grounds, whereas data driven
approach develops the categories from the raw data.
“Theory driven approaches are more transparent, more readily applied by others and
make explicit assumptions, which may not be as apparent in data driven approaches”
(Green, 2004).
Theory driven approaches have acquired a certain criticism regarding its limitations.
The main problem is that it is hard to develop a category coding scheme which will
cover all the constructs elicited afterwards. In most studies utilizing theory driven
code schemas a major issue is how constructs that come from different domains
should be treated. Categories, especially when derived from not RGT based data,
are often formulated in one or two words which can bear rather universal meanings
or originally imply different meanings than those current researcher wants to attach
and ascribe constructs to them (Green, 2004). In the RGT constructs are usually
coded as bipolar entity rather than both poles are coded separately, and the problem
arises when construct poles appear to relate to different aspects and, therefore,
coding such a construct into a single predefined category becomes problematic.
Another disadvantage of theory driven categorization is that it limits the range of
purposes for which the RGT can be used. Stewart with her colleagues (1981)
provides an example of a study which purpose is to test whether the predetermined
categories in the questionnaire developed to examine the meanings of co-workers
really reflect the aspects used by these people when talking about the matter. The
findings showed that four aspects in the questionnaire were not used by people at all,
whereas four other aspects not presented in the questionnaire were revealed.
Data driven categories employed in the content analyses are much more flexible and
appealing because categories are derived from and, thus, closer to the raw data. A
researcher does not need to force data into predetermined categories that may
sometimes not be applicable to the data at all (Green, 2004).
4. Repertory grid technique within the personal construct psychology
106
The major advantage of data driven approach is that categories reflect the constructs
they were developed from. This enables to recognize new patterns, domains and
categories which are specific for current topic and, maybe, were overlooked during
previous research studies.
There are, however, concerns which make researchers look for solutions.
One of the disadvantages of data driven approach is the potential for low replicability
by others, especially, when there is no accurate and apparent technique for coding
development being adopted.
Another concern is about the fact that data driven approaches involve construing the
constructs of others by a certain person (or several people). This fact gives rise to
following concern about the reliability of construing. When one is authorized to
interpret and develop categories based on the raw data, it is obvious that very often
the results will meet some disagreements from others because, according to Kelly’s
Individuality Corollary, each individual not only construes things differently but also
interprets other’s meanings in different ways. Therefore, resolving potential
disagreements may cause difficulties.
In solution, Honey (1979) has proposed the use of additional coders to resist
potential bias of the codes, i.e., categories developer. However, there is no evidence
in literature of how many coders should be involved. Moreover, Krippendorf (2004)
warned against additional coders because “they are more likely to be aware of the
researcher’s aims or general approach and are less likely to be truly independent”
what causes another kind of coders’ bias (cited in Green, 2004).
In general, data driven approach is favored by researchers working in business and
management field (Honey, 1979, Jankowicz, 2004, p. 169; Stewart et al., 1981;
Wright, 2004). It is recommended by these authors to cut up all respondents’ single
grids so that each elicited construct appears on a separate sheet and then to sort
constructs into groups all together revealing common patterns (Jankowicz, 2004, p.
151; Green, 2004).
In order to decide which approach should be employed it is important to consider the
respective advantages and limitations, as well as the range of available approaches.
It is also suggested to apply a combination of theory and data derived categories
4. Repertory grid technique within the personal construct psychology
107
(Stewart et al., 1981). A different hybrid approach proposed by Honey (1979) implies
an enclosure of supplied constructs which permit to sum individual’s perspective on
the topic in question.
In the current study the data driven approach of categorization is applied because the
main interest of research lies, first of all, in exploring peculiarities, commonalities and
utilized language of customers from another culture as well as their environment and
mentality aspects associated with the research topic.
4.6.1.3. Honey’s content analysis
Traditional content analysis approaches allow to categorize different meanings
expressed in a set of grids. However, traditional content analysis does not use all the
available information given in the grids. Thus, ratings of elements on constructs are
completely ignored in content analysis.
Honey (1979) developed a technique which permits to make use of individual
meanings being expressed by person’s ratings. It is based on the Kelly’s
assumptions that constructs are organized into a system (Organizational Corollary)
and that some constructs are more important and central to the person’s views and
knowledge about the topic while others even though relevant to the topic, are minor.
The personal construing system is not monolithic and it can be expected that while
there is certainly some kind of consistency among the constructs a person uses,
some constructs are particularly related to the discussed matter and carry more
meaning about the topic while others may bring secondary notes which do not reflect
“what the whole thing is really all about” (Jankowicz, 2004, p. 170).
Honey (1979) argued that the researcher is interested in exploring the individual
understanding of the topic in question and at the end all the elicited constructs are
more or less related to the “overall issue a person has in mind when thinking about
the topic”. He proposed at the end of the elicited list of constructs to supply an overall
assessment construct designed to sum up the interviewee’s individual attitude to the
topic as a whole. For instance, if a grid is made to explore customers’ preferences
among different brands of a certain product, the supplied construct could be “overall,
absolutely satisfied – overall, absolutely unsatisfied”.
Obtainment of the ratings on the overall construct gives an opportunity to compare
overall construct with all other constructs and see how close ratings on each
4. Repertory grid technique within the personal construct psychology
108
particular construct match the ratings on the overall construct. Constructs which
match highly with the overall construct represent what the person strongly felt and
thought about the topic, overall and constructs matching less highly also represent
what the person felt and thought about the topic, overall, however, somewhat less
strongly (Jankowicz, 2004, p. 170).
Two indices are used to reflect the extent of match between elicited and overall
constructs. The first index is called “percentage similarity score” (PSS). This index
reflects the actual similarity between constructs. Second, “H-I-L” index allows to
aggregate constructs across the sample taking into account the fact that people differ
in their typical percentage similarity scores (Jankowicz, 2004, p. 171).
1. The Percentage similarity score (PSS) index
The PSS index is based on counting the differences between the ratings along the
rows of the grid, where each row represents ratings on particular construct. At first,
absolute differences between each two ratings on the particular and overall
constructs corresponding to the same elements need to be found and summarized
(Formula 1). Each sum of differences indicates how small or high the difference is
and, therefore, how similar the constructs are. Some authors prefer the term
‘distance’ between the constructs rather than ‘difference’, however, the implied
meaning and the formula remain the same for a conventional grid (Bhatia and Yao,
1993).
If repertory grid consists of ratings on n elements and m constructs so that the sum
of differences is denoted by jjDD ′ and is given as
∑=
′−=n
i
jiijjj rrD1
', , (Formula 1)
where ijr is the rating assigned to element iE with respect to the construct jC .
The similarity then between the two constructs, i.e., the PSS which is denoted as
', jjσ can be computed as
nk
D jj
jj ×−
×−=
)1(
%200%100
',
,'σ , (Formula 2)
4. Repertory grid technique within the personal construct psychology
109
where k is the maximum possible rating. So ( k -1) is the largest possible difference
between ratings and ( k -1) × n is the largest possible sum of differences between
constructs in the whole grid.
The constructs are bipolar and each element can be rated on the construct in both
ways, running from ‘1’ to ‘5’ (on the 5-point scale) or running from ‘5’ to ‘1’ when
expressions at each pole of the construct are reversed. The expressed meanings
regarding personal attitude and preference to one of the poles remain the same,
however, relationships between meanings of reversed constructs with remain
constructs may change. Table 7 shows two constructs on which six elements are
being rated. The sum of difference computed according to Formula 1 is 16 (whereas
the largest possible sum of difference is (5-1)×6 = 24), which is rather large and
demonstrates weak relationships between the constructs.
Triad combination E1 E2 E3 E4 E5 E6 Sum of Difference,
1 5
E1 E3 E6 Construct A Speedy 2 4 5 5 3 1 Calm
E2 E4 E5 Construct B Large 4 2 2 1 5 4 Compact
2 2 3 4 2 3 16'ijij rr −
jjDD ′
Table 7: Relationship between two constructs about six elements
However, basic grid elicitation procedure implies that the way both poles are written
down in the grid depends on what has been expressed as similarity and what was
expressed as contrast during the triadic comparison. It is suggested as a convenient
way of recording the elicited information which helps to quickly see where the elicited
information is coming from. Triads are used just as combinations of elements in order
to encourage fresh constructs and are offered arbitrary. Most important is the
meaning expressed when person was comparing some triad but not the triad itself or
on which pole of the construct which meaning is placed (Jankowicz, 2004, p. 109).
In Table 7 for construct A a triad composed of elements E1, E3 and E6 was offered.
Elements E1 and E6 are defined as similar in that they are both “speedy” and, hence,
this emergent pole is written down on the left, indicating the ‘1’ end of the scale.
If the triad offered for eliciting construct B was E2, E4 and E5 and element E2 is
defined similar with E4 as more or less ‘large’, whereas E5 is considered as
‘compact’, the construct is to be written down as shown in Table 7.
4. Repertory grid technique within the personal construct psychology
110
However, if the same domain implicated in construct B was elicited from another
triad, for example, E4, E5, E6 where now E5 and E6 are construed similar as
‘compact’, then the pole ‘compact’ should be written on the left as a similarity
meaning defining the ‘1’ end of the scale. In this case the ratings on construct B
would be given as shown in Table 8.
Triad combination E1 E2 E3 E4 E5 E6 Sum of Difference,
1 5
E1 E3 E6 Construct A Speedy 2 4 5 5 3 1 Calm
E4 E5 E6 Construct B Compact 2 4 4 5 1 2 Large
0 0 1 0 2 1 4'ijij rr −
jjDD ′
Table 8: Relationship between two constructs showing a reversal
It can be seen that both ways of recording the construct B express the same
meanings, although, in Table 8 the sum of difference is 4 what already demonstrates
close relationship between two constructs and in particular between meanings of
emergent poles and between meanings of implicit poles respectively.
As exemplified, it is very important that construct analysis takes into account the
possible reversed relationship between constructs when calculating construct
similarity. The calculation should be conducted twice, once with each construct as it
stands and once with the ratings being the same of the first of compared constructs
and ‘reversed’ of the second of compared constructs. It is apparent that if ratings of
both constructs are reversed the sum of difference and, therefore, constructs’ PSS
would be the same as without construct reverse (Jankowicz, 2004, p. 108).
Coming back to the calculation of construct similarity (Formula 2), in order to take into
account possible reversed construct relationships the range of the percentage scale
should be spread over a 200%-point scale. Hence, the proportion is multiplied by
200%, but not by 100% (Jankowicz, 2004, p. 115; Bhatia and Yao, 1993).
Subtraction from 100% turns the percentage sum of difference into a percentage
similarity score. Similarity and difference are calculated in proportion Thereby, the
percentage permits to compare the extant to which pairs of constructs are matched
when grids to be compared have different amount of elements.
The PSS in the provided example is: %7,666)15(
4%200%100, =
×−
×−=BAσ
4. Repertory grid technique within the personal construct psychology
111
Constructs compared to demonstrate calculation of PSS are rather elicited according
to their meanings. However, the same way ‘overall assessment’ construct can be
compared with any elicited one.
2. The „H-I-L” index
The „H-I-L” index allows to analyze a group of individuals and to compare similarity
scores within a single grid with those of other grids and, thereby, to identify relevance
of each individual construct to general matter of the interview.
It is recognized that people have different construct similarity metrics. Their ranges of
PSSs for any topic may differ. For example, supposing in one interview the highest
PSS between given elicited constructs and overall construct is 100% and the lowest
is 80%, whereas in another interview where the person tends towards seeing many
different aspects when thinking about the topic, the highest PSS may be 85% and the
lowest, 55%. In this case, PSS of 82% indicates comparatively very similar meanings
between constructs for the second person and lies towards the bottom of the range
of similarity for the first person.
Therefore, the PSS index represents a relative value and Honey’s procedure is
aimed at assigning the actual percentage value of PSS index to either ‘high’,
‘intermediate’ or ‘low’ (‘H-I-L’) values for a particular individual. Then, for example,
when selecting constructs with similar meaning and with high similarity scores,
researcher should focus attention on constructs with individually high rather than
simply high PSSs, i.e., PSSs which are particularly high for the given individual who
provided that particular constructs (Jankowicz, 2004, p. 171). As a result, the
aggregated set of constructs, on the one hand, represents the categorized views, i.e.,
clusters of constructs on which there is a consensus across the sample of
interviewees and, on the other hand, preserves information about each individual’s
views in terms of personal relevance to the topic.
The steps for accomplishing Honey’s technique have the following consequence:
1. Ratings on a supplied overall construct are obtained.
2. Sums of differences for each construct against the overall construct are
calculated according to the formula provided in Formula 1. In order to check for
reversals, i.e., if the sum of differences is smaller when one of the two constructs is
4. Repertory grid technique within the personal construct psychology
112
reversed, at first, the sum of difference is computed between overall construct and
elicited construct. Secondly, the sum of differences is computed between reversed
overall construct and remaining unreversed elicited construct. The smaller sum is
then being noted.
3. Turn all computed sums of differences into PSSs (Formula 2), therefore,
ensuring comparability with other grids.
4. Within each grid the H-, I- and L-value intervals to be defined and each
construct is labeled according to the calculated score. The highest, intermediate and
lowest thirds defined by
ii lL =
3
iiii
lhlL
−+<
3
iiii
lhlI
−+=
3
)(2 iiii
lhlI
−+<
3
)(2 iiii
lhlH
−+=
ii hH =
Where:
ih and il – are the highest and lowest PSS between elicited construct ic and overall
construct respectively;
iL and iL – are the lower and the higher boundaries of the L-value interval;
iI and iI – are the lower and the higher boundaries of the I-value interval;
iH and iH – are the lower and the higher boundaries of the H-value interval.
4. Repertory grid technique within the personal construct psychology
113
When all the elicited constructs within the single grid are assigned with H-I-L indices,
those constructs that are labeled with ‘H’ index represent most relevant constructs to
the overall assessment construct and, thus, contain meanings that reflect what the
person particularly had in mind when thinking about the topic, even though, within
their category they may be different in meaning covering different aspect of the topic.
Constructs assigned with ‘L’ index appear less salient to the overall topic.
Procedure of Honey’s technique described in this section is demonstrated in Table 9.
The used RG represents the one shown in Table 3 as an example of elicited RG.
Similarity between elicited constructs with overall supplied construct is calculated.
The reversed ratings of the overall construct are shown at the bottom, below the
construct unreversed ratings. The unreversed sums of differences are shown below
each construct, on the left. The reversed sums of differences are shown below, on
the right. All sums of differences are turned into the PSSs. For each construct the
higher of the two values of PSS (for unreversed and reversed) are chosen and
shown in bold. The constructs are labeled (letters in bold below each construct, in the
middle) by H-, I- and L-values taking individual’s personal metric into account where
the highest PSS is 100% and the lowest PSS is 30%. Table 9 demonstrates how grid
data is complemented with Honey’s technique estimations. Interpretation and
inferences which can be derived from this will be discussed in 7.2.2.
When analyzing all the grids where constructs are ascribed to categories, a certain
consensus between the respondents can be revealed.
If most of the H-I-L indices are high within one category it signifies that the category,
i.e., attribute which the category represents is relevant and important for all or most
of the respondents in the sample and, therefore, important for the sample as a whole.
If most of the H-I-L indices are low within the category, it signifies that the sample as
a whole agree that the attribute does not relate particularly well to the topic in
general.
If the H-I-L indices are mixed within the category, especially if there are mostly either
‘H’ or ‘L’ indices, the researcher should pay special attention to the respondents of
the sample and the category because such a division may reveal a presence of latent
group among the sample for which the attribute presented by the category is
4. Repertory grid technique within the personal construct psychology
114
important and relevant to the topic in general despite the fact that for other
respondents of the sample it is less salient.
emergent pole (1) PC Mac G3 iMac G4 eMac Ideal implicit pole (5)
1 2 5 4 5
Unreversed Reversed
Sum of Diffs. 16 H Sum of Diffs. 0
PSS(%) -60 PSS(%) 100
1 2 4 2 1
Unreversed Reversed
Sum of Diffs. 9 L Sum of Diffs. 7
PSS(%) 10 PSS(%) 30
1 3 5 2 5
Unreversed Reversed
Sum of Diffs. 13 I Sum of Diffs. 3
PSS(%) -30 PSS(%) 70
5 1 1 2 1
Unreversed Reversed
Sum of Diffs. 3 I Sum of Diffs. 15
PSS(%) 70 PSS(%) -50
5 2 1 3 1
Unreversed Reversed
Sum of Diffs. 3 I Sum of Diffs. 13
PSS(%) 70 PSS(%) -30
2 3 1 1 1
Unreversed Reversed
Sum of Diffs. 5 L Sum of Diffs. 13
PSS(%) 50 PSS(%) -30
1 1 4 5 5
Unreversed Reversed
Sum of Diffs. 17 I Sum of Diffs. 3
PSS(%) -70 PSS(%) 70
5 4 1 2 1
1 2 5 4 5
The looks are to die for
Smaller range of software
Doesn't fit properly
Fast
Difficult to set up
Flimsy build
Upgrade is a dealer job
Transportable
Reversed
Unreversed
Looks boxy and 'standard'
Easy to set up
Slow performer
Large range of software
Difficult to move
Easy to upgrade
Good build quality
Appropriate for my work
Table 9: Repertory grid using Honey’s technique
In this work it is proposed that the relevance of the elicited constructs to the supplied
construct identified by application of Honey’s technique can also be applied regarding
not just overall assessment construct representing the general topic of the grid but
also regarding constructs representing subthemes of the survey. While the general
importance of the elicited constructs can be identified by other approaches, for
example, supplying the ‘ideal’ element, the relevance of constructs to a concrete
question which researcher is interested in within the general topic can be revealed by
4. Repertory grid technique within the personal construct psychology
115
supplying the construct which would reflect the question and comparing it with other
constructs by analogy with the described technique. This is demonstrated in the
empirical part of the work.
4.6.2. Quantitative methods for repertory grid analysis
When conducting an empirical study using the RGT analysis of several individual
grids aimed at exploration of similarities and differences in content and structure of
consumers’ individual and collective personal construct systems can be pursued. The
main methodological problem is to find a way of aggregating the information from all
participants without destroying the phenomena of individuality of each personal world
(Jankowicz, 1990).
It was George Kelly (1955), the founder of PCP theory who had first suggested to
analyze psychological space by means of statistical approaches. He invented a
nonparametric form of factor analysis for RG. Later two major contributions were
made to the history of RG quantitative analysis.
• The first contribution was made by Patrick Slater (1964, 1976, 1977) who
developed a computer program “Ingrid” based on application of “The Principal
Components of a repertory grid” (Slater, 1964).
• The second significant contribution was the adoption of cluster analysis for RG
data (Jankowicz and Thomas, 1982; Shaw, 1980).
There was also an attempt in suggesting multidimensional scaling within multivariate
methods for grid analysis, but it did not receive a wide use as well as adoption in
computer programs (Feixas and Alvarez, 1996). Most of computer programs
developed for analyzing grid data and most of conducted researches using
quantitative methods still use for RG analysis principal component analysis and
cluster analysis or derivatives of them (Fromm, o. J.).
For those researchers who are not familiar with statistics the quickest, simplest and
most effective method for interpreting grid data is counting simple frequencies.
However, for those who have got proper grounding in statistical procedures it is
suggested that repertory grids can be analyzed with standard computer statistics
packages (Bell, 1994).
4. Repertory grid technique within the personal construct psychology
116
Originally, the RG was devised as an interview-based or paper-and-pencil technique.
A number of computer packages have been developed for their elicitation and
analysis which are constantly being used by contemporary researchers. The thing is,
when using a RG in clinical purposes for understanding meanings about a certain
topic of a particular person, it is better to analyze constructs and ratings by means of
paper, pencil and professional look. However, since the technique is used in a variety
of disciplines where one of the aims of research is to aggregate the information from
a number of RGs or from a great grid consisting of a large number of constructs, it
became more helpful to conduct a comprehensive analysis by means of computer
program to recognize larger patterns. Computer software is also helpful as it provides
a graphical illustration so that some relationships between constructs or elements
can be seen at a first glance.
Despite the fact that there is a number of excellent programs which offer certain
analysis tools, repertory grid research deals with a large number of variables and
pursues different objectives to be reached and the programs may or may not match
the needs of a specific research project. None of the packages is a complete solution
to all the possible repertory grid designs and there is no one best way to interpret grid
data (Shaw and McKnight, 1981 cited in Hadley, 1996). Most important is to
remember as Easterby-Smith (1980, p. 17) says: “Qthe interpretation of grid data is
very much an art and not a technology”. So the researcher is relatively free to choose
the way of interpreting grid data “so long as it is well-informed and consistent with
repertory grid theory” (Hadley, 1996).
In this study the analysis are carried out by means of a popular WEBGRID IV
program developed by Brian R. Gaines and Mildred L. G. Shaw (Bell, 2003). It is a
Web version of REPGRID program (Shaw, 1989) available via the Internet which
offers “the capability to elicit and analyze grids over the Internet or a local network
with customized interactive dialogs and privacy of data” (Gaines and Shaw, 2005).
Some peculiarities of the RG form such as bipolarity of constructs and consideration
of construct reversing make application of software which is not developed specially
for RGT rather problematic and effort consuming. A variety of computer packages for
RGT analysis is provided by Bell (see 2003 for description of different software). The
program chosen for this study provides all the necessary analysis tools for proposed
4. Repertory grid technique within the personal construct psychology
117
methodology. Besides, it has friendly, clear design without any additional complexity
what makes its practical use suitable in business and managerial applications.
4.6.2.1. Hierarchical cluster analysis
Technique that is used for highlighting the relationships in a grid so that they become
visible at a glance is called Hierarchical cluster analysis (HCA). Originally, clustering
methods are based on analysis of correlations, however, for grid analysis the HCA of
distances, also called “Focus” or “focusing”, is commonly adopted (Fransella et al.,
2004, p. 87). This method sorts and rearranges rows (constructs) and columns
(elements) of the grid so that simultaneously similarly-rated elements are situated
together and similarly-rated constructs are situated together both being displayed in
same dendrogram. Such ‘focusing’ of the structure gives the technique its name.
In the WEBGRID IV package, used in the current study, one of the supplied tools is
called “Focus sorting and hierarchical clustering”. This tool was firstly designed and
presented in FOCUS program (Thomas and Shaw, 1976) and then, later, in
REPGRID (Shaw, 1989) which became one of the earliest computer programs
designed to carry out clustering of repertory grid data (Fransella et al., 2004, p. 88;
Thomas and Shaw, 1976). The program utilizes the city-block distance metric rather
than the common Euclidian distance metric (Shaw, 1980, p. 159) (see Appendix 2 for
more details on metrics).
In an output of HCA the PSS (in the software term ‘matching’ is used for PSS) of both
adjacent elements and adjacent constructs are presented in numeric form as well as
in a dendrogram like is shown further in Figure 11.
There are other general statistical packages such as SPSS, BMDP, CLUSTAN and
G-PACK which analyze the clusters. However, as above mentioned, the hierarchical
clustering presented in the FOCUS and REPGRID (WEBGRID) programs has
significant advantage over other statistical packages of taking into account the
bipolarity of the constructs. Thus, if two constructs are very different in their ratings a
distance coefficient is high indicating independence between the constructs.
However, as constructs represent bipolar dimensions, if one construct is reversed a
strong association between constructs may be revealed. Therefore, when distance
coefficient between constructs is high it is difficult to determine whether it indicates
inverted association or constructs independence. The “Focus” tool of WEBGRID
4. Repertory grid technique within the personal construct psychology
118
reverses the two poles of one of the constructs by replacing scores (e.g. a ‘1’
becomes a ‘5’ and a ‘2’ becomes a ‘4’, etc.) and, basically, establishes a single
direction of construing between each pair of constructs (Feixas and Alvarez, 1996).
As a result, those constructs that are more similar to others when being reversed, are
printed out so.
Reading of HCA outcome
This part is focused on reading and understanding of relationships between RG
components revealed by HCA produced by WEBGRID in a form of dendrogram and
supported figures (the term “Focus Cluster” is used for HCA outcome in the
WEBGRID software). The example of RG represents one of the elicited grids of the
current study. However, interpretation of underlying meanings is not presented in this
part and will be provided in the next chapter. The description here is stressed only on
the reading of quantitative analysis outcome where the RG form is used as an
example, therefore, reader should not pay attention to the expressed meanings
provided (therewith, expressed constructs are not fully provided but only bipolar
categories they were assigned to) but only bear in mind the bipolarity nature of
constructs elicited to describe elements set and rated on the 5-point scale according
to the described earlier in this work RG elicitation procedure.
The elicited individual grid is displayed in Figure 10 where elicited constructs form
grid’s rows and elicited elements form grid’s columns. To the left is the emergent pole
and to the right the implicit pole of a construct. The elements are printed out in turn
they were originally elicited (The numeric indications in front of emergent poles stand
for the number of construct, therefore, it can serve as numeration of constructs which
will be rearranged during analysis and also can serve as an indicator of emergent
pole which also may be placed to the other side in case of construct reverse).
Figure 10: Display of elicited repertory grid
4. Repertory grid technique within the personal construct psychology
119
In Figure 11 the elements are rearranged and constructs are rearranged and
reversed so that the construct poles to the left of the grid are the ‘1’ end of the scale,
and those poles to the right are the ‘5’ end. The rating values of ‘4’ and ‘5’ are
produced dark shaded, whereas ratings of ‘1’ and ‘2’ are left unshaded and rating
values of ‘3’ are given light shading. This reproduction of the grid is given in order to
highlight the clusters and make the identification of “blocks” within the “focused” grid
even more visualized and clear (McKnight, 2000).
The PSS of constructs/elements can be found if to follow adjacent lines from two
constructs/elements till they meet at a common apex and then further to the scale.
This “tree” structure which reflects the % similarity scores for adjacent constructs is
actually a construct dendrogram of the grid. Ditto for the adjacent elements and
element dendrogram (Jankowicz, 2004, p. 121). General dendrogram comprising
both construct and element dendrograms illustrates the strength of association
between elements and between constructs.
Figure 11: Graphic plot (dendrogram) of HCA of the grid
R - reversed PSS (%)
8 1 R 100
4 5 90
5 3 R 90
1 3 85
8 7 R 85
2 6 70
4 6 70
Constructs
Table 10: Links between constructs
4. Repertory grid technique within the personal construct psychology
120
PSS (%)
iMac G4 Ideal 87.5
eMac Ideal 71.9
PC Mac G3 59.4
Mac G3 eMac 59.4
Elements
Table 11: Links between elements
iMac G4 Ideal eMac Mac G3 PC
Sort of elements
Table 12: Element sort after cluster rearrangement
2 6 4 5 3R 1R 8 7R
Sort of constructs
Table 13: Construct sort after cluster rearrangement
High matches between constructs indicate that the relevant constructs distinguish
identically or similarly between the majority of elements. High matches between the
elements signify that the elements share identical or similar ratings on the majority of
constructs (Dillon and McKnight, 1990).
In Figure 11 if the cutoff point is 75%, i.e., if to break up any associations whose
matching score is less than 75%, 1 major cluster of constructs is identified consisting
of constructs 1, 3, 4, 5, 7 and 8; constructs 2 and 4 stand apart not referring to any
cluster.
The match of 100% between reversed construct 1 and construct 8 means that
elements that are seen as having ‘The looks are to die’ have a very strong tendency
to be seen as ‘Appropriate for my work’ and those that are seen as ‘Looks boxy and
'standard'’ are likely to be perceived as ‘Doesn't fit properly’. Rather strong
resemblance, 90%, is seen between meanings ‘Easy to set up - Difficult to set up’
and ‘Good build quality – Flimsy build’ (constructs 4 and 5) as well as between
meanings ‘Good build quality – Flimsy build’ and ‘Fast - Slow performer’ (construct 5
and reversed construct 3). Less strongly alike but still significantly, 85%, are
constructs ‘Looks boxy and 'standard' - The looks are to die’ and ‘Slow performer –
4. Repertory grid technique within the personal construct psychology
121
Fast’ (constructs 1 and 3); and ‘Appropriate for my work - Doesn't fit properly’ and
‘Transportable - Difficult to move’ (construct 8 and reversed construct 7). It is
important that similarity between construct meanings should be interpreted according
to the way poles are rearranged.
Similar inferences can be made considering elements. Thus, focusing on the
columns, i.e., the elements 1 cluster is sorted out if to take a 75% cutoff level. This
cluster includes elements ‘Ideal’ and ‘iMac G4’. Although, these elements are
comprised in a cluster their match is only 87% what indicates that ‘iMac G4’ still does
not entirely fit the person’s image of an ideal product. Elements ‘PC’ and ‘Mac G3’
appeared isolate and stand out. Element ‘eMac’ stands also out from the main
cluster, however, differs less strongly from the element ‘Ideal’ than other two isolate
elements. By using here the term ‘isolate’ it should be borne in mind that although
these elements do not match with other elements at 75% or higher, they still join a
general cluster by the 59% and 72% level respectively (McKnight, 2000).
The ‘Focus’ display of HCA allows a convenient illustration of variance graphically
what helps the researcher to get a first view of similarities or differences (Björklund,
2005).
4.6.2.2. Principal component analysis
Principal component analysis (PCA) permits, on the one hand, to analyze
relationships between constructs and elements within a single analysis, i.e., analysis
of one individual grid and, on the other hand, to group together the persons’
constructs and reveal the relationships between them (Easterby-Smith et al., 1996).
PCA indicates how different constructs group together and where different elements
are placed in relation to these. Besides, it indicates the statistical distance of each
element from any other (Senior, 1996).
When talking about PCA, two different calculation procedures can be implied, first,
calculation of the eigenvalue decomposition of a data covariance matrix and, second,
singular value decomposition of a data matrix, often after mean centering the data for
each attribute. The results of these two calculation procedures are expressed in
terms of component scores and loadings (Shaw, 2003). In this description both
procedures of decomposition are named as PCA, however some authors distinguish
the methods and apply the term PCA to the first procedure and the term singular
4. Repertory grid technique within the personal construct psychology
122
value decomposition (SVD) to the second method emphasizing its independency as
a statistical method.
It is important to note that PCA used for grid analysis differs from the original
statistical method of principal component analysis. One of the most significant
contributions to the advance in RG mathematical analysis was made by Patrick
Slater who introduced a most widely known method for representing both constructs
and elements. Slater proclaimed the method as “The Principal Components of a
repertory grid” (Slater, 1964), however, the mathematical procedure was, actually,
SVD and this fact caused later a lot of confusion in the literature.
The SVD is a technique which implies a linear transformation which is often used in
pattern recognition, i.e., to explore structures in multivariate data. A matrix such as
RG can be approximated by another matrix of a lower rank. As a mathematical
procedure it means that a number of variables (elements or constructs) are
translated into a lesser number of composite variables (principal components) which
explain the maximum possible variance (Feixas and Alvarez, 1996).
Further in this study the term PCA is adopted for the analysis name as most of
literature and computer packages devoted to the RG use the term PCA bearing in
mind SVD as an underpinning mathematical procedure. As the procedure is
developed and computerized for specialized software for a long time, its accuracy
and validity are assumed to be proved. The study is not aimed at providing a
comprehensive description of difference between calculation procedures of SVD and
PCA. At the end, both SVD and PCA are data compression techniques for
determining a set of weighted linear composites (principal components) of original
variables such that each principal component is uncorrelated with the others. In
marketing research it is rather important to understand an outcome produced by the
analysis tool provided by the software in order to make inferences and an accurate
interpretation of data.
Principal components identified through the analysis are used as the axes where the
constructs are plotted according to their factor loadings. Thereby, a certain
representation of a grid is produced which presents a useful map for further grid
interpretation (Feixas and Alvarez, 1996, Easterby-Smith et al., 1996). An essential
interpretation of this analysis is that the greater the amount of variance that an
4. Repertory grid technique within the personal construct psychology
123
element or construct accounts for within the overall RG data, the more important they
are (Metzler and Neimeyer, 1988).
Reading of the PCA map
The PCA map produced by PCA tool of the software (the term “PrinGrid map” is used
for PCA outcome in the WEBGRID software) contains the information about
relationships between constructs, between elements, and between the constructs
and the elements. The analysis tool provides a number of tables and graphs for the
data (Easterby-Smith et al., 1996).
First of all, the PCA tool provides a table with the percentage of variance accounted
for by each component, as shown in Table 14.
1 Component 2 Component 3 Component 4 Component
Percentage, % 76.38 15.50 5.09 3.03
Cumulative % 76.38 91.88 96.97 100.00
Table 14: Percentage of variance accounted for by each component
The information is then presented in a form of series of graphs. Each graph plots columns
(elements) and rows (constructs) according to their arrangement with respect to the principal
components as shown in a Figure 12.
Figure 12: Graphic plot of PCA of the grid
4. Repertory grid technique within the personal construct psychology
124
The vertical and horizontal axes stand for the first two components and plotted at
right angles to each other reflecting the maximal distinguish between two
components as patterns in the original grid data. One plot reproduces any two of all
revealed principal components.
The textual output of the data underlying the plot contains tables with constructs
(rows) and elements (components) loadings on each component. The absolute
values of the loadings produced are arbitrary and depend on how the grid data was
scaled. Only the relative values are meaningful. Therefore, in the resulting plot
distances between lines, points and poles are important and can be meaningfully
interpreted (Gaines and Shaw, 2005).
To produce the plot the grid is treated as if the elements were points plotted in an n-
dimensional space defined by the constructs as axes centered on the means of the
elements. The data is then rotated through principal components analysis to spread
the elements out as much as possible in a 2-dimensional plot. As a technical note,
PCA map carries out an analysis of the double-centered matrix of distances between
elements with all construct ranges scaled to be the same (because the software
supports grids with constructs having differing ranges) (Gaines and Shaw, 2005).
This analysis is equivalent to a principal components analysis of the construct
variance matrix which is the basis of most grid analysis packages emulating Slater's
(1976; 1977) “Ingrid” (Gower, 1966).
John Gower showed in his work (1966) that it is not always possible to recognize
different patterns (components) in the original sample of data and “these must be first
found by using an analysis based on similarities or distances which do not allow for
within data correlation”. He pointed that very often an interpretation of such methods
as PCA and factor analysis can be better understood by examining the distances
suitably defined between the variables rather than by correlations between them. He
suggested a method for finding co-ordinates for each variable referred to principal
axes which preserve these distances.
Gower’s work, published in 1966, where he described the technique based on the
concept of distance between variables has become a “bible” and a manual for Dr.
Brian Gaines, one of the developers of the RepGrid software who programmed the
PCA for RepGrid (and WEBGRID) software:
4. Repertory grid technique within the personal construct psychology
125
“Ny far the best ref on PCA is John Gower, "Some distance properties of latent root
and vector methods in multivariate analysis", Biometrika, 1966. 53, 325-338. John
sent me a draft of that paper in 1964 and I programmed PCA on the Cambridge
TITAN -- it has been my bible ever since and is a recognized citation classic.” (from
personal correspondents with Dr. Brian Gaines, 2008)
As mentioned before, the distances between variables are counted with Minkowski
metric where the default power of 1 defines the standard city-block metric (in contrast
to Euclidean metric which is defined by power of 2, see Appendix 2).
Components and constructs
Constructs are plotted as straight lines where the angle with respect to each
component reflects the extent to which the construct is represented by the
component.
The angle between any two construct lines reflects the extent to which the ratings of
elements on those constructs are correlated: the smaller the angle, the more similar
the ratings.
In a broader outlook, the angle between a group of construct lines and the
component line reflects the extent to which the component can be taken to represent
the grouping of constructs; the smaller the angle, the greater the extent (Jankowicz,
2004, p. 130; Gaines and Shaw, 2005).
The construct poles’ names are spread so it is also can be seen right away in which
direction the constructs are similarly rated.
In Figure 12, for example, the first, horizontal, component represents constructs 1
and 8 quite well (construct lines of 1 and 8 are coincided as their ratings are identical
when one of the constructs is reversed, i.e., PSS between them is 100% (Table 10)).
The second, vertical, component does not stand clearly for any construct. Lines of all
other constructs are disposed between the components being well represented by
neither of them. Interpretation of meanings underlying such plotting is presented in
section 7.3 of empirical part of the work.
Further, in this work, the term ‘dimension’ stands for a principal component which, in
its turn, represents a distinct pattern of variability in the grid, i.e., a certain direction in
4. Repertory grid technique within the personal construct psychology
126
which the values of the ratings vary. Thus, constructs with highest variance define
directions based on which respondent perceives a product. The PCA of a grid
reveals the dimensions in which a person perceives elements. When looking at the
constructs referred to one component usually a shared meaning of those constructs
can be identified. The shared meaning represents a common topic or direction of
meanings described by the constructs. Each person perceives elements in several
dimensions which can be named generally according to the meanings of the
constructs aligned to each dimension.
Components and elements
Elements are also placed on the plot. As each component stands for several
constructs, the elements can be positioned along each component in place of their
original position along each construct. It is like the position of a point on a graph far
along both axes. Itself, it does not bring a lot of sense and is not that useful for
interpretation until other elements are positioned on the plot (Jankowicz, 2004, p.
130).
The distance between any two elements gives an idea of the relationships between
these two elements because it reflects the ratings, each element was given on all the
constructs. Thus, if two elements are closely positioned on the graph it means they
received similar ratings on those constructs which are represented by the
components reflected on the plot.
Only two components can be reflected on the plot at a time. Therefore, when more
then two principal components are significant it is important to look how elements are
positioned on plots reflecting all pairs of components. It can be revealed that two
elements that are placed closely on the graphic plot, for example, of two first
components are placed apart on the plot of, e.g., first and third components, still
showing the similarity on the first component but real difference on the third one.
Such positioning observation reveals on which dimensions the elements are
considered differently by the respondent and on which dimensions perceptions of
elements are similar.
For example, in Figure 13 on the first plot reflecting first and second components,
i.e., dimensions, elements ‘Ideal’ and ‘iMac G4’ are placed identically on the first
dimension and appeared in different quadrants but also not very distantly on the
4. Repertory grid technique within the personal construct psychology
127
second dimension. However, on the second and third plots reflecting first and third;
and second and third components respectively these elements are positioned far
apart. This signifies that although element “iMac G4’ is considered similar to the
‘Ideal’ element on the first dimension, they are considered different on the second
and very different on the third dimension (construct lines are not shown in Figure 13
for visual convenience).
Figure 13: Element perception on different dimensions
This is a particularly useful property offered by grid analysis which allows to compare
how closely a person construes his perceptions towards certain elements on different
dimensions. Depending on how far the two elements are positioned on the plot it can
be concluded, for example, what might need to be changed to make element ‘iMac
G4’ to be perceived closer to or like ‘ideal’ element. Along which dimensions and,
4. Repertory grid technique within the personal construct psychology
128
hence, constructs the movement should occur so that the rating values of
perceptions on that constructs would be altered by the person (Jankowicz, 2004, p.
131).
4.6.2.3. Synergy of principal component and hierarchical cluster analysis
Analysis of similarities and differences between construct meanings allows to identify
groups of customers whose perceptions of elements, e.g., products or brands, are
similar or different in certain ways. These findings can help a company to appeal to
the customers in a proper way emphasizing and focusing on important attributes and
characteristics sought by each group taking into account customer’s preferences and
expectations.
The RG contains not only meanings but also ratings which allow to apply both
qualitative and quantitative approaches making the analysis of similarities and
differences more thorough. This study is aimed at proposing an application of PCA
and HCA in a way which would produce a synergy effect by structuring and
systemizing RG data into a certain form to simplify data visualization and to allow
customers' classification according to their perceptions and preferences.
PCA and HCA present pattern detection methods which purpose is to represent the
variation in a data set into a more manageable form by recognizing categories or
groups (Van Ooyen, 2001). The principle of synergy effect produced by combination
of PCA and HCA is illustrated in Figure 14.
At first, PCA allows to distinguish constructs according to the distinct patterns of
variability. Each dimension, i.e., component contains constructs that have the same
variation trend of the ratings (Jankowicz, 2004, p. 128).
Within each dimension by means of HCA the distances between constructs as well
as between elements are calculated and grouping according to PSSs in ratings is
performed.
Combination of PCA and HCA allows recognizing:
1. Common patterns of variation, i.e., common directions in which construct
ratings vary (where the first component accounts for the largest variance);
4. Repertory grid technique within the personal construct psychology
129
2. Similarities among rating absolute values which indicate possible similarity
between meanings underlying constructs.
Constructs with highest variance attributed to the principal components define
directions based on which a respondent perceives a product. And if there are
constructs that do not only vary in the same direction but also have the same or
similar rating values it may indicate strong resemblance between meanings
underlying constructs or, at least, strong correlation and complementarity of
perceptions.
PCA: distinct patterns of variability
HCA: similarity grouping by distances
1 Cluster
2 Cluster
1 Cluster
1 Component 2 Component
1 Component 2 Component
General data set
Figure 14: Synergy effect of PCA and HCA
It is very important to bear in mind that the nature of the research to be conducted by
means of described methodology is exploratory. Exploratory research, although
applying quantitative methods such as PCA and HCA, implies the biggest role of the
interpretation aspect comprised in methodology. Exploratory research is, first of all,
aimed at recognizing common patterns which can describe people, i.e., consumer
behavior and permits to classify people according to these patterns.
Therefore, although pattern detection is sometimes regarded as yet another form of
statistics, there are important conceptual differences which should be taken into
4. Repertory grid technique within the personal construct psychology
130
account when conducting quantitative analysis within the suggested methodology
(Hogeweg, 1976 cited in Dijkshoorn et al., 2001, p. 44):
1. In statistics deviations from randomness in the data set are looked for.
In pattern detection the structure in the data set is sought.
2. In statistics attempts are made to make sample-independent statements. The
data under consideration is assumed to be a random sample of the whole
population, and the objective is to make statements about the whole
population by looking at a representative sample of the population. Ideally,
these statements should not change if a different random sample is taken from
the population.
In pattern detection, the data set under study is not considered a sample from
a larger population but is considered all there is. A different structure may be
found if new data is added.
3. In statistics groups (and an underlying distribution) are presupposed and tests
are made to determine whether these groups differ significantly form each
other (i.e., more than can be expected on the basis of random fluctuations
alone). In other words, concepts are tested (i.e., attempts are made to answer
the question whether presupposed groups are different).
In pattern detection groups are generated per se. In other words, concepts
(i.e., groupings) are generated. Descriptive statistics, however, may be used in
pattern detection for characterizing the grouping obtained in cluster analysis.
PCA and HCA can best be seen as heuristic, rather than statistical, methods for
exploring the diversity in a data set by means of pattern generation. Thus, rather than
trying to find the “right” pattern or classification, the differences in the patterns as
revealed by analysis should be used to gain further understanding of the objects
under the study (Dijkshoorn et al., 2001, p. 45).
4.7. Reliability and validity of the repertory grid
When talking about the concepts of reliability and validity, first of all, the
determination needs to be differentiated according to whether it determines the
reliability and validity of the RG as a data collecting technique or it determines the
4. Repertory grid technique within the personal construct psychology
131
reliability and validity of measures applied for data analysis. It is also important how
these two concepts are construed from the PCT perspective underlying the RGT
(Fransella et al., 2004, p. 132).
Although a resulting grid contains a lot of numbers, it is important to remember that
the technique remains qualitative where the data represents the result of a sorting
process by a respondent. This data shows relations between the elements and is
rather susceptible to the selection of elements and the interview situation. The
elicited constructs contain very subjective individual meanings and labels which
require a great care of the researcher to be interpreted ‘correctly’. Therefore, an
excess of statistical processing may distort information derived from this ‘soft’ data
(Björklund, 2005).
It is important to remember that the RG represents a flexible and diverse
methodology but not a standardized test with the set procedure. The manuals written
about the RG outline the diversity of the technique, the variety of form design,
elicitation procedures and analyzing approaches (e.g., Fransella et al., 2004,
Jankowicz, 2004; Feixas and Alvarez, 1996). Unlike tests, where the administering
procedure is clearly specified and described and the ratings and scores are based on
an adequate standardization data, the RGT does not meet these requirements and
grids do not have norms (Beail, 1985, p. 22).
Given the wide variety of form, content and analysis available in the RG application it
makes no sense to discuss reliability and validity of RG in general. The same as
when talking about the reliability and validity of a questionnaire, the defining issues
comprise what grid, in what area administrated, to what kind of people, under what
kind of conditions, analyzed in what kind of manner and what kind of results are
sought (Fransella et al., 2004, p. 134). Therefore, the RG should be viewed as a
specific format which is used for collecting the data and the question of reliability and
validity should be referred to a particular case of RG application and its measurement
procedures.
When talking about the reliability of a measure, there are two various definitions of
the term used by researchers. The first definition of the reliability concept implies the
general capacity of a measure to ‘reliably’ assess a characteristic, whether or not the
‘amount’ of the characteristic is changing in the subject. As second, by ‘reliability’ is
4. Repertory grid technique within the personal construct psychology
132
meant the tendency of a test to produce exactly the same result for the same subject
at different times. There are the research contexts in which the second definition
could be regarded as a reasonable operational form of the first definition, for example
when it is assumed that the characteristic is relatively stable and unchanging for a
given subject such as the height of an adult or a temperature of a healthy person.
However, the universal application of the second definition as a requirement of a
measure becomes pointless when it is about things that change a priori.
In Kelly’s PCT, underlying the RGT, the person is a ‘form of motion’, not a static
object and, therefore, the consistence between person’s construct systems over time
could be low (Björklund, 2005). Hence, since much of life is about change and the
stability or the instability exists in what is measured, not in the measure, the aim is
not to produce stable measures but to assess as accurately as possible the
predictable stability and the predictable change (Fransella et al., 2004, p. 133).
The question arises under what circumstances a researcher would expect stability or
change from a personal construct system? This question can be answered only
regarding a particular form of the RG. It would depend on which part of the person’s
viewpoint, attitudes or knowledge is examined and re-examined. For example, a high
degree of stability can be expected when examining from time to time personal
notions of the rules of arithmetic. However, when examining, for example, personal
viewpoint about the political situation, more variation in attitude can be expected in
time (Fransella et al., 2004, p. 133). In marketing research, it can also be expected
that more superordinate constructs such as values or socio-psychological expected
consequences would appear more stable than consumer’s attitude or perception of a
particular product characteristic, such as high level of product quality. However, even
in this case a lot depends on what kind of product or product category is in question.
This exemplifies that the idea of the static mind represents a contradiction in terms.
Therefore, the application of the RG should not be done to repeat the same result but
to understand, when it shows change, what that change is signifying. In these terms
the reliability can be rather regarded as merely one aspect of validity (Fransella et al.,
2004, p. 134).
The discussion of the RG validity took place in a lot of studies again due to its flexible
format and a variety of applications (see Fransella et al., 2004, p. 143 for discussions
4. Repertory grid technique within the personal construct psychology
133
in different fields of research). However, at the end, as Fransella and colleagues
(2004, p. 146) point out that the grid became a problem for those researchers who
regarded it as a ready-made device or tool for their purposes rather than a broad
methodology which involves them in solving a series of experimental problems if it is
to be of any value.
The RG does not represent a test. It is a format in which data can be placed what
reveals if there is a pattern or meaning to the data. As it has no specific content, its
validity can be discussed in a sense that the researcher can question whether or not
it can effectively reveal patterns and relationships in certain types of data (Fransella
et al., 2004, p. 144).
The validity of the technique in terms of the PCT is its capacity to enable us to
elaborate our construing, our ability to anticipate (Björklund, 2005). Elaboration,
according to the PCT, occurs by the extension and definition of personal construing
system. In its turn, the extension implies an increase of the range of convenience of
personal constructs so that more events or elements are taken into consideration.
Elaboration of construing by definition implies tightening of the construing within an
investigated area so that the researcher attains a more precise, detailed and exact
comprehension of this area. Thus, Kelly equated the concept of validity with
usefulness and saw understanding as the most useful of enterprises. The validity
ultimately refers to the way in which a mode of understanding enables the researcher
or the respondent to take effective action (Fransella et al., 2004, p. 151).
Bannister and Bott (1973, p. 162) say: “Qif we substitute for validity the notion of
usefulness, or at least make usefulness the central feature of validity, we shall be
less concerned with the correlation between a test and some relatively arbitrary
criterion, and more concerned with the values which users of a test find in it”.
The interpretive paradigm and the PCT underpinning the RGT are congruent with
what Kelly argued that the validity of any theory is to be found in its usefulness. In
this respect the RGT has proven its utility by being applied and tested in widely
differing fields for several decades and remains a fruitful technique offering a number
of fundamental advantages to the researchers.
The above explanation and arguments confirm that the validity of a grid should be
questioned within a particular study applying a particular elicitation and administering
4. Repertory grid technique within the personal construct psychology
134
procedure. The fact that even slight procedural variations can produce a substantial
impact on the content and structure of the elicited personal constructs emphasizes
the role and responsibility of the researchers who have to understand their own
contributions to the grid outcomes that they interpret (Neimeyer, 2002). Thus, if it is
found that a particular grid does not have any predictive value and did not produce
the sought information, the researcher should look for the shortcomings in that
specific grid format and design rather than make general remarks regarding the
validity of the technique.
Although there are shortcomings and limitations of the method many of them can be
overcome due to the various ways of grid application, its flexibility. Within the each
area where the method is used researchers try to adopt and further expand the
methodology by suggesting new variations which can make the technique even more
suitable to a particular field.
In this work the general description of the RG includes discussions of reliability,
validity or, at least, reasonability of either option appropriateness depending on the
topic, objective and conditions of the research. Although almost in every aspect of
RGT various ways and approaches are described, there is still a lot left aside
because of its diversity especially in respect to different fields where the technique is
applicable. It is, therefore, a researcher’s task to consider the options and nuances
and design the grid which would meet the conditions and the requirements regarding
information of his study.
The market research study conducted in the frame of this work is aimed at exploring
the customers’ constructs related to a particular product category. Therefore, the
utility presents in the discovered information and inferences that can be made by
interpreting the data to be used further for strategic, positioning and advertising
objectives. The reliability estimates of the study are not tested as they need to be
based on re-test correlations replicated previous research finding which are not
available due to the specificity of the research topic, the conditions, the region and
the sample involved in the study.
4.8. Summary
♦ Repertory grid is a technique of structured interviewing designed to capture the
dimensions and the structure of personal meanings. It is aimed at eliciting
4. Repertory grid technique within the personal construct psychology
135
and describing the ways in which people give meanings to their experience in
their own terms, in their own language. Its methodology allows a new deeper
insight on how people think and perceive life experience and things around.
♦ The technique represents a qualitative approach for data collection designed so
that the gathered data is appropriate for both qualitative and quantitative
analysis methods.
♦ Repertory grid technique is developed within the personal construct theory
which maintains that a person’s understanding of the world is the result of an
active, constructive process rather than a passive reaction to some external
reality.
♦ This understanding is achieved through a construing, i.e., the process of
contrast and similarity. People construe situations and give meaning to events
by seeking to differentiate them from others and see them as similar to others.
♦ Similarity and contrast in events or objects that are observed and perceived by an
individual stimulate the creation of his personal classification standards –
constructs. Thus, subjective meanings take the form of bipolar constructs
which are organized into a system of superordinate and subordinate relationships.
♦ Therefore, not only the content but also the structure of meanings is considered
within the personal construct theory and the repertory grid technique.
♦ Personal construct theory is consistent with the interpretive paradigm which
focus lies on the exploring the psychological processes that people use to make
sense of their material and social environments.
♦ The resulting form of the repertory grid elicitation procedure in whole represents a
set of rating scales printed one above the other with the ratings arranged in
rows and columns into a table or a grid.
♦ A grid is always conducted about a particular topic with an intention of eliciting
just those constructs which a person uses in making sense of that particular field
of discourse, i.e., that particular piece of his experience.
4. Repertory grid technique within the personal construct psychology
136
♦ Repertory grid mainly consists of:
� a set of elements which are representative of the content area under the
study;
� a set of personal constructs that a person uses to compare and contrast the
elements;
� a rating system that evaluates the elements based on a bipolar arrangement
of each construct.
♦ Constructs represent individual scales that a person uses when thinking about
the elements, where poles of a construct represent scale limits. Constructs can be
elicited or supplied. Elicited constructs are more informative for an explorative
research. An enclosure of supplied constructs enables to expand the variety of
findings.
♦ Elements can also be elicited or supplied. Elicitation of elements based on
supplied categories represents an advantageous approach which, on the one
hand, implies consideration of elements a respondent is aware of and, on the
other hand, enables further comparison of individual grids within the sample.
♦ ‘Self’ elements such as ‘ideal’ or ‘current’ represent individual perceptions and
images of what a respondent actually has or wants when talking about things,
events or people enclosed in the element set. Supplying of such key elements
permits to significantly extend the findings from further analyses.
♦ When elements represent products or brands the ‘ideal’ element represents a
generalized image comprising mostly desired performance of a product or a
brand from the domain. This element appears helpful in any research in which a
grid is used to understand a respondent’s choice, i.e., preferences.
♦ Personal values represent desired end states which can be derived from
constructs by means of the laddering up technique and then prioritized and put
into a hierarchical order by means of the resistance-to-change technique.
4. Repertory grid technique within the personal construct psychology
137
♦ Resistance-to-change technique enables to identify values that are not sincere
and central for a person, but were expressed rather fictitious to seem more
meaningful.
♦ The higher the personal value in the hierarchy, the higher its resistant to change
and, thus, the representative subordinate constructs, including product attributes
are tend to be similarly resistant to change.
♦ In order to better understand what is being said in a grid a blend of both,
qualitative and quantitative approaches appear useful to be applied.
♦ Qualitative analysis methods that are traditionally used are frequency counts
and content analysis. These methods are concerned with analyzing contents of a
grid.
♦ Quantitative analysis methods most often used are hierarchical cluster analysis
and principal component analysis which also enable to analyze inter-
relationships contained in a grid.
♦ When applying traditional content analysis the ratings contained in a grid are
completely ignored. Honey’s content analysis technique permits to make use of
individual meanings expressed by person’s ratings.
♦ Honey’s content analysis suggests to supply an overall assessment construct
which is formulated to sum up the interviewee’s individual attitude to the topic as
a whole. Ratings on this construct are compared with the ratings of elicited
constructs to identify how closely the elicited meanings relate to the topic in
question.
♦ Honey’s content analysis allows aggregation of constructs across a sample
revealing the homogeneity in attributes’ relevance and importance. It takes
into account respondents’ personal metrics of construct similarity to the overall
construct.
4. Repertory grid technique within the personal construct psychology
138
♦ Honey’s content analysis also allows to discern a presence of latent groups for
which attributes that appeared less salient for the rest of a sample are commonly
important.
♦ It is proposed that the relevance of the elicited constructs to the supplied
construct identified by Honey’s technique can also be applied regarding
constructs representing subthemes of a survey rather than only its general topic.
♦ Hierarchical cluster analysis used for analyzing repertory grid data differs from
the traditional cluster analysis as:
� it sorts and rearranges rows (constructs) and columns (elements) of a grid so
that the resulting dendrogram comprises both construct and element
dendrograms illustrating the strength of association between elements and
between constructs;
� it takes into account the bipolarity of constructs by considering both reversed
and unreversed constructs.
♦ High matches between constructs indicate that the relevant constructs
distinguish identically or similarly between the majority of elements. High
matches between elements signify that the elements share identical or similar
ratings on the majority of constructs.
♦ Principal component analysis of a grid reveals the dimensions in which a person
perceives elements.
♦ The difference of the principal component analysis used for a grid data in
comparison to the original analysis approach is that it allows to compare how
closely a person construes his perceptions towards certain elements on
different dimensions. This is attained by the analysis tool capability to place
elements on the same plot that reflects constructs on different dimensions.
♦ It is suggested that application of the principal component analysis and
hierarchical component analysis on the complementary basis can produce a
synergy effect by structuring and systemizing repertory grid data into a certain
form which simplifies data visualization and allows consumers' classification
according to their perceptions and preferences.
4. Repertory grid technique within the personal construct psychology
139
� At first, principal component analysis allows to distinguish constructs
according to distinct patterns of the variability.
� Secondly, within the each dimension by means of the hierarchical cluster
analysis distances between the constructs as well as between the elements
are calculated and a grouping is performed.
♦ Combination of the principal component analysis and hierarchical cluster analysis
allows recognizing:
� Common patterns of variation, i.e., common directions in which construct
ratings vary (where the first component accounts for the largest variance);
� Similarities among rating absolute values which indicate possible similarity
between the meanings underlying constructs.
♦ Constructs with highest variance attributed to the principal components are
determinative for the directions based on which a respondent perceives a
product. Each person perceives elements in several dimensions which can be
generally named according to a common meaning of the constructs aligned to
each dimension.
♦ Constructs that do not only vary in the same direction but also have the same or
similar rating values may have a strong resemblance between their meanings
or, at least, there is a strong correlation and complementarity of these
perceptions.
♦ In an exploratory research after recognizing common patterns and similarities in
consumer perceptions by means of quantitative approaches the inferences and
conclusions are very much dependent on the interpretation of these findings.
♦ Concepts of reliability and validity of the repertory grid are built on the fact that
the technique represents a flexible and diverse methodology but not a
standardized test with the set procedure.
♦ In the personal construct theory a person is a ‘form of motion’, therefore, when
talking about technique’s reliability, application of the repertory grid should not
4. Repertory grid technique within the personal construct psychology
140
be done to repeat the same result but to understand, when it shows change,
what that change is signifying.
♦ The interpretive paradigm and the personal construct theory are congruent with
what Kelly argued that the validity of any theory is to be found in its usefulness.
The validity ultimately refers to the way in which a method of understanding
enables a researcher or respondent to take effective actions.
5. Modeling consumer motivational sphere
141
5. Modeling the consumer motivational sphere
In this chapter the individual consumer motivational sphere (CMS) described in the
theoretical part is further discussed with the focus on a procedure suggested to
identify the concepts and linkages represented in the CMS schema. The purpose of
the chapter is to define where each level of abstraction can be found when analyzing
RG data by means of analysis methods described in 4.6. After showing the initial
schema developed in 2.2 the model is further extended with its analysis sources
assigned to each level of the initial schema. Following parts of the chapter are
devoted to the description of each level presented in the model focusing on the
analysis approach suggested to reveal relevant information.
According to Trommsdorff (2009, p. 108) values represent too-general motives and
by themselves do not explain much about consumer buying motivation. It is important
to obtain the full sequence of motives starting from too-specific which explain a lot
about purchasing motivation of an individual, then moving to motives of the middle
range which allow revealing of commonalities within a group of consumers and
ending up with too-general motives in order to understand the full structure of
consumers’ mental representations regarding the investigated topic and the nature of
the consumer behavior underlined by their preferences and choices.
Product characteristics Product attributes Specific consequences Abstract consequences
"too specific motives" "motives of
middle range"
"too general motives"
MOTIVES
Expected consequences
(desirable/undesirable)Product
Values
Situational and Environmental incentives
Figure 15: Consumer motivational sphere
It is assumed that each level of abstraction presented in the schema can be
disclosed by analyzing data elicited by means of the RGT. In Figure 16 the schema is
extended with its sources, i.e., each level is assigned with that part of the
methodology by which the level can be identified and described.
5. Modeling consumer motivational sphere
142
Product characteristics Specific consequences Values
Personal constructs Laddering UP Laddering UP
WHICH Resistance-to-Change
Product attributes Abstract consequences
Content Analysis PCA Dimensions
WHAT Clusters
Sit. and Env. incentives
HOW+ Ideal ratings
Ana
lysi
s dr
iven
: re
lativ
ely
obj
ectiv
eP
erso
n dr
iven
: su
bjec
tive
Figure 16: The CMS schema extended with sources
Thus, it is suggested that by integrating both qualitative and quantitative approaches
in analyzing RG data a clear structure of an individual CMS schema can be obtained.
On the one hand, this structure contains very individual aspects which usually can be
useful when working with each customer individually or with very small group of
people. It is especially appropriate for companies producing high involvement
products usually at relatively high prices. On the other hand, the structure provides
more abstract information about the customer which permits comparison of different
individuals and revelation of similarities among people on different levels of
abstraction of their mental representations.
The suggested methodology integrates and extends existing qualitative and
quantitative approaches to investigate an individual construing system on two levels
according to the extent of meaning objectivity.
The first level of integrated method is based on producing a person driven outcome
which actually organizes the information elicited directly from an individual using his
own meanings. Findings made on this level are based on the individual expressed
meanings where the numeric part of the RG form is not considered. This method,
actually, represents the common method usually applied for RG analysis based on
the laddering technique developed on principles of the MEC theory. Within the
suggested methodology this ‘upper’ level represents a more subjective part of the
whole methodological chain.
The second, ‘bottom’ level of the method is based on producing the analysis driven
outcome which is aimed at revealing and structuring the underlying meanings
5. Modeling consumer motivational sphere
143
expressed, on the one hand, directly in a sense of elicited expressions and, on the
other hand, indirectly in a sense of meaning relationships implied by numeric
evaluations, i.e., given ratings. All the analyzed data is still elicited from an individual,
however, an attempt is pursued to uncover subconscious part of construing system
built on certain relationships between meanings which are hard to be recognized
when only interpreting individual’s directly expressed wording. Involving quantitative
analysis the ‘bottom’ level of methodology represents a relatively objective part of the
methodological chain.
The ‘bottom’ level of the methodology represents a major methodological contribution
to broadening the variety and complexity of approaches used for analysis of the RG
data. The proposed methodology does not contradict the traditional approaches to
analyze RG but rather complements them with an opportunity to use and produce
inferences from fuller capacity of the RG elicitation technique.
As for contribution to the theory of consumer motivation the suggested methodology
comprises a certain approach to define and assess motives of the middle range
building on the motives conception proposed by Trommsdorff.
In the Figure 16 the linkages within the chain show that the procedure comprised in
the upper level methodology is not directly connected to the procedure implied in the
bottom methodology level. It is assumed that although the levels of abstraction in the
scheme do not necessarily sequentially flow out from each other in terms of analysis
procedure, their linkages are preserved in terms of sense implied by revealed and
interpreted structured meanings in each level of abstraction. It means that the
methodology is developed to identify each level of abstraction according to its
definition where level distinction is based on the extent of its relation to the person’s
product or self knowledge.
Further each level of abstraction presented in the CMS schema is described in terms
of its disclosing procedure based on the elicited RG data.
5.1. Product characteristics
At first, elicited personal constructs represent product characteristics that are more or
less important for the consumer. Sometimes, personal construct is expressed in
relatively abstract way when it is not clear which specific product characteristic or
5. Modeling consumer motivational sphere
144
feature is implied, for example, “more comfortable versus less comfortable”. This
should be précised during the interview by means of laddering down technique, when
asking, for example, ‘what does it mean ‘more comfortable’ for you?’ Then person
states what is it that he considers ‘comfortable’ for him. In this case, one may say:
“there is enough space for me to stretch my legs and lay down”, another may say:
“the auto is better equipped with different technical devices that provide passenger’s
comfort”, while the third one points that: “comfortable means not to feel bumpy roads
at a high speed” and so on. Therefore, it is important to try to keep all the information
and expressions provided by the interviewee.
In order to detect which constructs should be specified during the interview it is
helpful to examine the constructs that distinguish between the elements on the
classificatory criteria employed by the respondents (Dillon and McKnight, 1990). To
do so, constructs can be classified virtually as referring to different aspects of the
elements’ description. Within the context of the current research elements which
represent different autos appear to be distinguishable on three types of elicited
constructs according to the direction of their meanings about the elements. Thus,
constructs refer to one of the following aspects of elements’ description:
• ‘Which’ characteristics the auto contains, e.g., large or compact in size, rigid or
soft suspension, attractive or unappealing design, capacity of high speed and
power or lack of the capacity, etc.
• ‘What’ the auto provides or emphasizes, e.g., comfortable or not, prestigious
or ordinary, provides a feel of safety or does not provide such a feel, enjoy
from driving or stressful driving, evokes respect on the road or imperceptible,
etc.
• ‘How’ the auto is being used, e.g., for work or for leisure time, in summer or
all-year around, in the city or in the countryside, with a driver or for self-driving,
for oneself or for the family use, etc.
Such a virtual classification of constructs helps the researcher to examine the
constructs right away during the interview. Thus, at first level of abstraction all
personal constructs are kept and those constructs that answer the question ‘which’
and ‘how’ are usually recorded as initially expressed while those constructs that refer
to the question ‘what’ often reflect more abstract meanings that rather stand for the
5. Modeling consumer motivational sphere
145
attributes or even consequences and, therefore, should be précised through the
laddering down technique.
The virtual classification of constructs is feasible in the studies regarding different
topics. It is only important for the researcher to accurately define the questions or the
aspects of elements to which the constructs refer. When aspects are accurately
defined, usually most constructs refer to one of them. Rarely, constructs that can be
referred to neither of aspects are either very specific to the individual or bear little
obvious semantic resemblance to any other construct with which they may match
(Dillon and McKnight, 1990).
The elicited constructs may appear at any level of abstraction, although it is usually
found that consumers make distinctions at relatively concrete level of abstraction
such as product physical characteristics or product perceived attributes. However,
some may discriminate elements at the higher level of abstraction such as
consequences and very rarely at the highest level, i.e., personal values (Olson and
Reynolds, 1983, p. 82). In this study it is found that in most cases by laddering down
procedure when asking the question “how, in what way?” or “what does it mean for
you?” interviewer can reveal chains of lower level of abstraction such as concrete
product characteristic or at least attribute (see also Jankowicz, 2004, p. 64).
5.2. Product attributes
Product attributes represent perceptible product characteristics and features.
Perceptible characteristics of a product often enclose several product physical
characteristics. As an example taken from one of the elicited grids of this study, a
person provides one construct “reliable vehicle which needs less service versus
unreliable, often breaks down” and another construct “producer’s proved image of
high quality production versus not proved”. As a result of content analysis and
categorization these constructs refer to the perception of quality of the vehicle. And
by looking at the ratings given by the interviewee (see Table 15) it is seen that most
ratings are similar and only on one element ratings differ at minimal value. The PSS
between these two constructs is 94%.
Product attributes are individually perceptible and, sometimes, presence and
performance of the same product physical characteristic may be associated with
5. Modeling consumer motivational sphere
146
different attributes by different people. As an example, in the conducted study several
interviewees mentioned “presence of rigid versus soft suspension” in vehicle.
1 2 3 4 5 6 7 8 9
reliable, needs less servicing 1 2 2 2 3 2 3 4 1 unreliable, breaks more often
producer's proved quality (e.g. German or Japanese producers)
1 3 2 2 3 2 3 4 1not proved production quality (e.g. Italian
proucers)
ElementEmergent pole (1) Implicit pole (5)
Table 15: Ratings given on constructs of the same product attribute.
For one person who purchases armored tough body vehicles, presence of a rigid
suspension is essential for a safe reliable exploitation and, therefore, he perceives
the presence and quality of rigid frame and suspension as one of the components of
a high quality vehicle.
Another person, who also pointed necessity of a rigid suspension, often utilizes his
auto out of the city area which is characterized by a very bad quality of roads. For
this person presence of rigid suspension defines the level of rough-terrain
performance provided by a vehicle implying assurance that the auto will not get stuck
in a ditch or mud.
However, for the third person who also expressed the construct “rigid versus soft
suspension”, it appears to be important to have soft suspension which provides a
comfortable ride. This person utilizes an auto only in the city where he does not face
problems with a bad quality of roads. For him it is important to experience a soft calm
ride without feeling any bumps on the roads.
This example proves the importance of understanding the underlying meanings
expressed in the constructs and importance of interpretation of expressions in
compliance with all elicited and even beforehand known information, if available,
about an interviewee.
Categorization and assignment of product characteristics to product attributes can be
highlighted later, as a researcher goes with analysis, by means of the HCA which
sorts and rearranges constructs so that similarly-rated constructs are situated
together. This method can, sometimes, help to prove an uncertain assignment,
however, it is important to always pay attention to the expressed meaning as similarly
rated constructs not necessarily imply reference to the same product attribute.
5. Modeling consumer motivational sphere
147
5.3. Specific expected consequences
Specific expected consequences represent the lowest of three levels of abstraction
which already represent the motives that stimulate and influence consumer decision
making process. Specific consequences are very personal and hard to approximate
in order to reach common set of expected benefits.
In suggested model specific consequences are identified through the laddering up
technique. Consumers are given laddering up task aimed at understanding of
consumers’ underlying personal motives with respect to a given set of elements, e.g.,
products or brands. Each unique pathway from product characteristic to a personal
value represents a possible perceptual orientation with respect to viewing the product
category (Reynolds and Gutman, 1988).
Laddering task is undertaken during the one-on-one interview to elicit individuals’
means-end knowledge structures linking product characteristics to related values or
end states. There are reasons to apply laddering technique only to some elicited
constructs rather than make a person provide ladders for each given construct.
One reason to be emphasized is time limit during an interview. This reason refers to
the conduction of the research, i.e., its practical accomplishment rather than scientific
or theoretical limitation. RGT requires relatively long time, usually not less than 1
hour for one interview. When the technique is to be used in clinical purposes or by
human resources department in organizations respondents usually may or ready to
devote as much time as needed because they are forced or self interested in it,
besides they are usually interested in the results received from analysis of their
personal repertory grids. In marketing it is researchers and managers who are
interested in the outcome and the findings rather than consumers who, indeed, make
a favor to a researcher by participating in survey. They are not interested in the
results of the study (as they rarely acquire access to them) and it is hard to ask them
to devote so much time as would be needed to elicit all possible cognitive meanings
and ladders related to the investigated topic. That is why it is vital to understand at
the beginning which information is useless and should not be asked in order to make
interview less time consuming.
Second reason refers to a common criticism of the original means-end chain
approach which claims that by asking “why is it important for you?” artificial abstract
5. Modeling consumer motivational sphere
148
levels may occur (Grunert and Grunert, 1995). This problem does not take place only
when all the characteristics from which ladders are elicited are considered as enough
important to produce benefits for an interviewee. Otherwise, respondents are
basically forced to find arguments for their buying behavior what makes them think in
a more strategic way and provide reasons that often do not explain their actual
preferences and choice (Botschen, et al., 1999).
The RGT does not imply elicitation of most important characteristics that respondents
seek for when choosing a concrete product. The triadic sort task described in section
4.5.2.1 is intended to elicit the salient distinctions respondents use to discriminate
among stimuli in the domain of interest (for example, brands in a product category)
(Olson and Reynolds, 1983, p. 82). These salient distinctions, i.e., elicited constructs
reflect cognitive meanings that respondents use when thinking about that domain,
however, not all of them appear to be important when making purchasing decisions.
The ‘ideal’ element which represents a generalized image comprising mostly desired
performance of the product or brand from the domain appears helpful in any research
in which a grid is used to understand the respondent’s choice, i.e., preferences
(Jankowicz, 2004, p. 58). The rating value assigned to the construct on the ‘ideal’
element points the preferred by the respondent construct pole. Rating value of ‘1’ or
‘2’ points that respondent prefers product characteristic performance as expressed in
the emergent pole and, correspondingly, rating value of ‘5’ or ‘4’ indicates that the
product characteristic performance of the implicit pole is preferred.
For this reason, in the proposed methodology it is suggested that laddering
procedure should be carried only with those constructs that are given extreme rating
value which is either ‘1’ or ‘5’ (in case of 5-point scale) on the ‘ideal’ element. It is
assumed that when respondent provides extreme rating on the ‘ideal’ element he is
consciously aware of why and what for he wants a product characteristic expressed
in the construct to be necessarily present as stated in the preferred pole. Therefore,
he can provide conscious consequences that he expects from obtaining the product
characteristic.
Depending on the field and research purpose ladders can be elicited either from
preferred pole or from both contrast poles. In the elicitation from both contrast poles
at each iteration when respondent answers the question “why is it important for you?”
5. Modeling consumer motivational sphere
149
the following question “what would the other end of that construct be?” is to be asked
next. Thus, the resulting ladder consists of a set of superordinate constructs on top
of which most superordinate one reflects personal value and its contrast for example
“life and hope versus death, hopelessness, and despair” (Jankowicz, 2004, p. 191).
In marketing, elicitation of both contrasts of personal values seems to be excessively
unlike. In condition of limited time asking for contrasts at each ladder doubles the
time and rarely brings significant information. Personal values are very general and
their contrasts are more or less clear unlike subordinate constructs, especially at the
lower levels of abstraction. If the procedure for eliciting personal values implied
straight asking for ‘your’ values, expressed ‘life and hope’ would not say much about
a person to the researcher, however, the elicited contrast ‘death, hopelessness, and
despair’ also would not be so helpful in explaining individual’s construing system of
his buying behavior. Sequence of ladders obtained only from the preferred pole
explains much more about personal mental representations which bridge product
characteristics and personal values what is of a greater interest for marketers.
Moreover, it saves time and does not let the respondent become bored with an
interview what is very important when the data is based on the meanings expressed
by respondent and his disinterest and unwillingness to support may lead to artificial
or scant data.
Specific expected consequences, being elicited from constructs which are given the
extreme rating value on the ideal element, ensure that revealed ladders are
consciously realized and, thus, meanings expressed in each ladder are scarcely
artificial. On the other hand, meanings implied by other constructs are not considered
on the higher levels of abstraction because of a doubt of the fact that a respondent
can actually produce consciously realized and, at the same time, not artificial ladders.
However, being elicited by the RGT these other constructs are assumed to be a part
of a customer’s construing system related to the domain and, therefore, need to be
considered, analyzed and interpreted in terms of their role within the CMS.
The term ‘specific’ used for defining the expected consequences elicited by means of
laddering technique emphasizes that these consequences are directly formulated by
the respondent based on his conscious mental representations and, thus, they are
individually peculiar.
5. Modeling consumer motivational sphere
150
Disclosing of abstract expected consequences, discussed next, implies application of
quantitative approaches in addition to the qualitative content interpretation in order to
consider all constructs, i.e., meanings associated with the topic in question, and
define their motivational role in the respondent’s cognitive structure related to the
topic.
5.4. Abstract expected consequences
Next level of abstraction in the proposed model concept represents consequences
that are identified by means of quantitative analysis and interpretation of elicited
product characteristics and attributes. In original MEC theory consequences of a
higher level of abstraction are obtained through the laddering and bridge product
characteristics, product attributes and consequences of a lower level of abstraction
with personal values. As described in 2.1.2 there is, however, a certain criticism in
literature concerning identification and distinguishing between consequences of
different levels of abstraction and values.
This work proposes different approach for disclosing abstract consequences, i.e.,
motives of the middle range that drive and direct consumers in their buying behavior.
The aim pursued is to suggest a method that would allow to identify and to formulate
consequences at a level of abstraction which permits to approximate and find
commonalities within a group of people but, at the same time, the identified
consequences would still be not too general and could be interpreted in terms of
product features and perceptions.
The virtue of identifying abstract consequences by means of suggested methodology
is that another type of human cognitive representation is additionally involved. In
laddering a person is actually using his ability for reasoning, that is, he expresses
underlying reasons for perceiving and preferring things in this or that way. In the
proposed method analysis of person’s comparing ability is involved. Therefore,
mental representations underlying person’s evaluative and comparative approach to
perceive things are analyzed and inferences are made not only from vertical
relationships between meanings, i.e., linkages between different levels of abstraction,
but also from horizontal interrelationships between meanings that are on the same
level of abstraction.
5. Modeling consumer motivational sphere
151
Slater’s PCA described in 4.6.2.2 reveals dimensions which represent directions
based on which a respondent perceives elements. By conducting PCA of elicited
constructs and ratings components, i.e., dimensions which represent distinct patterns
of variability in the grid can be identified. According to Jankowicz (2004, p. 128) and
as it was found in this study, in most cases constructs aligned to one component
share some common meaning and, hence, a common direction of characteristics
described by constructs can be defined.
Common meanings underlying each identified dimension represent abstract
expected consequences. Therefore, an abstract consequence is derived from
product characteristics and their perceptions by identification of sort of a bundle
comprised of product features which person construes in the same way and direction
and associates them with attainment of a certain benefit.
Sometimes, meanings of constructs within one dimension can be different but the
fact that their rating values vary in the same way indicates a strong correlation
between these constructs. For this reason after carrying the quantitative procedure of
the PCA it comes to the numerously discussed importance of the interpretive
process.
PCA as complex analysis requires the researcher to make assumptions when
interpreting the original grid and unlike, for example, cluster analysis these
assumptions are less visible (Jankowicz, 2004, p. 132).
An essential interpretation of this analysis is that the greater the amount of variance
that a construct accounts for within the component and within the overall grid data,
the more important the construct is (Metzler and Neimeyer, 1988). Therefore, this
construct should be thoroughly considered when identifying underlying component
meaning. Finding a common meaning between constructs which lie closest to the
component, i.e., account for greater variance within the component, and naming the
component requires cautious interpretation.
The name given to a component which, in turn, represents an abstract consequence
reflects researcher’s judgement. Therefore, it is suggested, if possible, to negotiate a
meaning with an interviewee and check whether the researcher’s judgement reflects
true meaning of an interviewee (Jankowicz, 2004, p. 132). However, collaborative
process of verifying components denotation with an interviewee, i.e., consumer, is
5. Modeling consumer motivational sphere
152
rarely feasible in marketing and it appears vital how good the researcher is qualified
and can understand respondent’s jargon and meanings. Anyway, quantitative
assistance in distinguishing the elicited meanings makes the qualitative procedure of
interpretation more straightforward, clear and scientifically grounded.
PCA also provides hierarchical importance of identified abstract consequences.
There are two indicators that can specify hierarchical relationship between
components, i.e., dimensions.
• First indicator is a percentage of variance that component accounts for within
the overall RG data. The greater the percentage of variance a component
accounts for the more important the component is.
• Second indicator is a loading of the ‘ideal’ element. It is assumed that the
greater the loading of the ideal element a component has got the more
important the component is. As each component represents an abstract
consequence a respondent seeks for, consequences can be arranged
according to their importance for the respondent.
An indicator to be used depends on the aim of the research. If a research focus lies
in pursuing better understanding of how people construe their meanings and what
aspects people consider when thinking about a domain such as, for example, product
category then first indicator is more appropriate. However, when research is aimed at
revealing consumer preferences and expectations regarding product consumption an
indicator based on the ideal element evaluation should be used.
When PCA of the RG data is performed and constructs are assigned to the
components, it is suggested to perform the HCA regarding each component. That is,
in order to reach a synergy effect described in 4.6.2.3 and to structure elicited
meanings more precisely, within the each identified dimension constructs are
rearranged so that similarly rated constructs are positioned as adjacent. Such
positioning indicates that the person perceives meanings underlying similar
constructs not only in the same direction but also in the same way. This strengthens
the credibility of inferences made regarding similarity of individual subjective
meanings. At the same time, meanings underlying constructs which appear in
different clusters within the same dimension are considered, although different but
5. Modeling consumer motivational sphere
153
highly correlated what signifies their complementary relationship within the personal
cognitive structure of meanings.
Moreover, elements are also clustered within the each dimension allowing revealing
on which dimensions elements are considered similar or different. This is important in
order to understand on which dimensions a particular product or brand is perceived,
for example, differently from the ‘ideal’ one. This would mean that this particular
product cannot cause the abstract expected consequences, implied by that
dimensions, in the desired manner and, therefore, product characteristics and
product attributes comprised in that dimensions need to be improved or repositioned
(see also 4.6.2.1).
5.5. Values
The highest level of abstraction stands for personal values. Values are “end benefits”
that are expressed by respondent during laddering procedure (Gutman, 1991). They
are obtained during an interview when respondent at first is given a laddering task
regarding only the constructs that are given extreme rating values and secondly, the
respondent is given a task to prioritize all named values by means of resistance-to-
change technique described in 4.5.4.2.
Process of laddering from different constructs may lead to the same personal values.
Since it is dealt with the hierarchy of meanings at different levels of abstraction a
single superordinate meaning can have several subordinate ones below it. That
means that an interviewee is likely to associate several of his original constructs with
the same value (Jankowicz, 2004, p. 189). In the conducted study it was found that
most respondents draw not more than three or four personal values associated with
consumption of a certain product.
Because values represent consciously desired end states, they are directly elicited
from a person. Although the process of identifying abstract expected consequences
is not directly linked to the process of values elicitation, they are still linked on the
principles of MEC where values represent end states and abstract consequences
represent psychosocial consequences provided by product characteristics and
attributes.
5. Modeling consumer motivational sphere
154
All constructs that are given extreme values on ‘ideal’ element are also ascribed to
the dimensions on which a person perceives elements. Therefore, values, laddered
from these constructs are also linked to the dimensions, i.e., abstract consequences.
However, now it can also be found how close are the other constructs, not rated by
extreme values, to the constructs from which values are elicited and, thereby, how
actually important is their performance in reaching consciously desired end states.
5.6. Situational and environmental incentives
As discussed in 2.2, in the proposed CMS schema situational and environmental
incentives reflect environmental systems and current life circumstances which a
consumer faces at a given time. They are relatively temporary and unstable, i.e., they
can be changed when person achieves another social level, moves to another place,
changes family status or work but they may significantly affect the perception of
desirable end states and choice criteria related to the product attributes at a given
time and circumstance.
It is almost impossible to reveal from an interviewee all the conditions related to the
product use, however, some important incentives appear to be available for the
researcher. As shown in section 5.1 elements appear to be distinguishable on
different types of constructs which represent product characteristics. Constructs can
be classified virtually as referring to one of aspects of the elements’ description
(Dillon and McKnight, 1990). In the current study three levels on which elements are
distinguishable were defined where one of the levels reflects an aspect ‘How’
element, i.e., product is being used. Constructs that pertain to this level basically
enclose meanings that demonstrate how and in which circumstances respondent
uses a product. Some examples of such constructs taken from the current study
where autos represent elements, are: ‘an auto can be used for (in)’: work – leisure
time, in the city – countryside, in summer – all the year round, business – family, with
driver – self-driving, only for me – for family, friends.
All these constructs reflect environmental incentives such as “in summer – all year
around” or “in the city – countryside” as well as life circumstances in which an auto is
to be used, for example “for business – for family” or “with driver – for self-driving”.
(Differentiation between usage of an auto in summer time or all year around reflects
climate peculiarities of the location the respondent is living; in this study conducted in
5. Modeling consumer motivational sphere
155
Russia the conditions of vehicle exploitation in winter time strongly differ from
conditions in summer time because of very low temperature degrees and
atmospheric precipitation. Differentiation between exploitation in the city or in the
countryside reflects the problem of quality of roads in the investigated area as well as
problem of clearance of roads, especially in winter time.)
Elicitation of such constructs does not itself indicate which pole reflects the way of
respondent’s exploitation. For example, construct “with driver – for self-driving”
expressed during triadic sort task does not show whether the respondent uses his
auto with driver or drives his auto himself. The pole referred to the type of
respondent’s use can be identified by looking at the rating value on the ‘current’
element. In the study it was found that the neutral rating value (‘3’ on a 5-point scale)
is given when respondent either uses his auto for (in) both kinds of exploitation, e.g.,
for family and for business purposes or that he does not face such choice and the
difference does not make sense in his current situation. For example, if a person has
moved to another location where climate factor is not actual anymore, however, he is
aware of difference from his previous experience and, therefore, might name such a
construct when differentiating elements in triads.
Ratings given on the ‘ideal’ element can also indicate how vehicle is being used.
However, this would represent useful information for marketer when sample
represents consumers that are likely to purchase a desired product from a product
category rather than consumers who are limited, e.g., by resources or available
product variety and are obliged to purchase not desired but available and affordable
products.
In the current study consumers of premium and luxury autos have been investigated.
These consumers are characterized by relatively high standard of living who can
afford products which they really like and which include desired attributes and
features. In this case ratings on both ‘ideal’ and ‘current’ elements should be taken
into account. It is likely that a person who currently drives his auto himself within the
next few months will higher personal driver and will buy a new auto to be a
passenger rather than a driver. Or, if a person is going to purchase a new house out
of the city and his ideal auto is to be appropriate for both, city and countryside.
5. Modeling consumer motivational sphere
156
It is useful to look at how similarly ‘ideal’ and ‘current’ elements are rated. If the
ratings on constructs highly differ, it may indicate that ideal auto represents some
kind of a dream auto which exists in respondent’s mind with respect to his drawn
ideal world and life. For example, respondent is a hard worker and mostly uses his
auto for business purposes, however, he dreams about calm life in a tropical country
where he drives an auto only for pleasure. In this case it is important to be aware of
the way the auto is being used currently because, probably, when he is going to
make his next choice he still will be living in a country of a temperate or frigid climate.
Overall, the model, proposed in this study, serves to make elicited mental
representations organized to a structural form which allows more thorough
understanding of consumers’ construing system and makes it possible to differentiate
consumers on different levels of abstraction and develop segments according to
consumers’ commonalities. In this chapter it was described how each level of
proposed chain can be identified from the data gathered by means of RG
interviewing technique.
The suggested approach does not claim to be a substitute to the traditional
qualitative procedure of laddering technique. The methodology represents an
integrated approach which is aimed at an attempt to get an insight into the individual
mind from another perspective. It is believed that quantitative analysis and related
interpretation of expressed meanings and evaluations provided by the RG form can
extend the existing approaches enabling more thorough and broad understanding of
individual’s construing system on different levels of abstraction and, thereby,
reaching closer insight to the consumer motivational sphere.
5.7. Summary and conclusion
♦ It is proposed that each level of abstraction presented in schema of consumer
motivational sphere can be derived from the data elicited by means of the
repertory grid technique.
♦ It is important to obtain the full sequence of motives starting from too-specific
which explain a lot about purchasing motivation of an individual, then moving to
motives of the middle range which allow approximation and revelation of
commonalities within the sample and ending up with too-general motives in order
5. Modeling consumer motivational sphere
157
to understand the full structure of consumers’ mental representations regarding
the investigated topic.
♦ Suggested methodology integrates and extends existing qualitative and
quantitative approaches to investigate individual construing system on two levels
according to the extent of meaning objectivity.
♦ The first level is based on producing the person driven outcome which
organizes information elicited directly from the person using his own meanings.
This level uses meanings elicited by laddering up technique and, therefore, the
numeric part of the repertory grid form is not considered. This level represents a
subjective part of the whole methodological chain.
♦ The second level is based on producing the analysis driven outcome which is
aimed at revealing and structuring the underlying meanings expressed, on the
one hand, directly in a sense of elicited expressions and, on the other hand,
indirectly in a sense of meanings relationships implied by numeric evaluations,
i.e., given ratings. This level of the methodology represents a relatively objective
part of the methodological chain.
♦ Analysis of meanings underlying evaluative and comparative approach to
cognitive representation uncovers subconscious part of construing system built
on interrelationships between meanings which are hard to be recognized when
only interpreting individual’s directly expressed wording.
♦ The suggested methodology comprises a certain approach to define and
assess motives of the middle range building on the motives conception
proposed by Trommsdorff.
♦ Product characteristics are expressed in the elicited constructs and should be
virtually categorized by the researcher during the interview according to their
extent of specification in order to precise the meanings if needed.
♦ Product attributes are identified by means of content analysis when all
constructs are assigned to the perceptible attributes.
♦ Sometimes, presence and performance of the same product physical
characteristic may be associated with different attributes by different people. It is
5. Modeling consumer motivational sphere
158
vital when interpreting the underlying meanings expressed in the constructs to
consider the expressions in compliance with all elicited, observed and even
beforehand known information, if available, about an interviewee.
♦ Specific expected consequences are identified through the laddering up
technique.
♦ Laddering up procedure is applied towards constructs that are given extreme
rating values on the ‘ideal’ element. In this case a person is consciously aware
of the reasons why he wants a product characteristic to be necessarily present as
stated in the preferred pole and, therefore, is likely to provide not artificial ladders.
♦ Constructs that are not given extreme ratings on the ‘ideal’ element, however,
being elicited by the repertory grid technique are assumed to be a part of a
customer’s construing system related to the domain and, therefore, need to be
considered, analyzed and interpreted in terms of their role within the consumer
motivational sphere.
♦ Disclosing of abstract expected consequences implies application of quantitative
approaches in addition to the qualitative content interpretation in order to consider
all the constructs, i.e., meanings associated with the topic in question
♦ The most apparent difference of identifying abstract consequences, i.e.,
motives of the middle range is that mental representations underlying person’s
evaluative and comparative approach to perceive things are analyzed and
inferences are made not only from vertical relationships between meanings but
also from horizontal interrelationships between meanings that are on the same
level of abstraction.
♦ Principal component analysis reveals dimensions which represent directions
based on which respondent perceives elements. Usually dimensions can be
generally named according to the common meaning or direction of the construct
meanings ascribed to each dimension, however, it requires cautious
interpretation.
♦ Common meanings underlying each identified dimension represent abstract
expected consequences. Therefore, an abstract consequence is derived from
5. Modeling consumer motivational sphere
159
product characteristics and their perceptions by identification of a bundle
consisting of product features which person construes in the same way and
direction and associates them with attainment of a certain benefit.
♦ The greater the amount of variance that a construct accounts for within the
dimension the more important the construct is to be thoroughly considered
when identifying underlying dimension meaning.
♦ Principal component analysis enables hierarchical order of identified abstract
consequences. Two indicators can be considered:
� First indicator is a percentage of variance a dimension accounts for within
the overall repertory grid data; the greater the percentage the more important
the dimension, i.e., abstract consequence is.
� Second indicator is a loading of the ‘ideal’ element a dimension has got; the
greater the loading the more important the dimension is.
♦ An indicator to be used depends on whether the aim is to better understand the
personal construing system regarding the domain or to reveal consumer
preferences.
♦ In order to reach a synergy effect and to structure elicited meanings more
precisely, within the each identified dimension constructs are rearranged by
means of hierarchical cluster analysis. It discloses the relationships between the
meanings which are either perceived in the same way or, at least, highly
correlated signifying complementary relationship within the personal cognitive
structure of meanings.
♦ Elements are also clustered within the each dimension allowing to understand on
which dimensions elements are considered similar or different. This is important in
order to understand on which dimensions a particular element is perceived, for
example, differently from the ‘ideal’ one, showing the weakness of the element in
providing certain abstract consequences in the desired manner.
♦ Values are elicited and prioritized by the respondent during the laddering up and
the resistance-to-change procedures. Constructs from which values are laddered
5. Modeling consumer motivational sphere
160
are also ascribed to the dimensions and, thereby, the abstract consequences are
linked to the values.
♦ The method for identifying abstract consequences exposes how close are other
constructs, not rated by extreme values, to the constructs from which values are
elicited and, thereby, how actually important the performance of other
constructs is in reaching the desired end states.
♦ Situational and environmental incentives are enclosed in the meanings and
ratings on the ‘ideal’ and ‘current’ elements of constructs that pertain to the virtual
category which answers the question ‘how and in which circumstances’
respondent uses a product.
♦ Overall, the model and the methodology proposed in this study serve to make
elicited mental representations organized to a structural form which allows more
thorough understanding of consumers’ construing system and makes it possible
to differentiate consumers on different levels of abstraction and develop segments
according to consumers’ commonalities.
♦ The proposed methodology does not contradict the traditional approaches to
analyze the repertory grid data and does not claim to be a substitute to the
traditional analysis approaches but rather extends them with an opportunity to use
and produce inferences from fuller capacity of the repertory grid elicitation
technique.
♦ An integrated method is aimed at getting an insight into the individual mind from
another perspective and extending the existing approaches allowing more
thorough and broad understanding of individual’s construing system on different
levels of abstraction and, thereby, reaching closer insight into the consumer
motivational sphere.
6. Research design
161
III. Field research
6. Research design
6.1. Research objectives
In the previous chapters issues of defining customer motivation and motives of
different ranges, modeling the CMS and assessing each level of the CMS schema
have been addressed using a number of theories, assumptions and analysis
approaches from different areas related to the fields of consumer behavior and
psychology. Apart from a systematic discussion of the many ways to design, proceed
and analyze RG data comprising various individual mental representations, the major
result of this work so far is the development of a conceptual model and supportive
methodology to assess each level of abstraction comprised in the CMS.
The aim of this chapter is to demonstrate the applicability of the suggested
methodology by conducting a comprehensive analysis of customers’ motivation
mechanism and preferences on the Russian automotive market with the focus on
customers of premium cars.
The full capacity of the RGT is hard to demonstrate as it allows numerous findings
and inferences depending on the design of the form and applied analysis. Therefore,
empirical part of the work is mainly concentrated on demonstration of the proposed
methodology, however, some additional inferences and managerial
recommendations for the company that ordered the study are provided as well.
The objectives of the empirical study presented in this thesis can be described as
follows:
• Design of a particular form of RG which would be appropriate according to the
research objectives and peculiarities;
• Elicitation of the individual consumer motivational sphere from an individual
grid;
• Performance of a customer segmentation based on the grid analyses of the
sample.
6. Research design
162
6.2. Research design
To demonstrate an applicability of the methodology proposed in this work a primary
data is used. The proposed methodology built on principles and design of a particular
interviewing technique implies that a research should be conducted from the very
beginning and the data acquisition has to be done according to the suggested
methodology. For this reason it was impossible to use secondary data, all analyses
and findings should be accomplished within the data gathered in a particular form.
Secondary data collected by the RGT implication would also not be appropriate for
this work because the interviewing technique is very flexible in its design and
included components (e.g., elicited or supplied elements and constructs, way of
rating and domain of elements). The ultimate form depends on the objectives of a
particular research.
The data set was generated in the context of a market research project conducted on
a behalf of a major German car manufacturer. For confidentiality reasons the
company name is not disclosed. The interest of the research practical outcome lies in
the exploration of perceptions, preferences and purchasing intentions of customers of
premium and luxury cars in the Russian car market. A particular interest also lies in
exploring customer attitudes towards auto safety and security measures as well as
the perceived confidence in a car.
6.2.1. Structure of the sample
The survey was conducted in two of three largest cities of Russia, Moscow and
Novosibirsk. A sample of 20 (18 male and 2 female) respondents was surveyed in
July-August 2007. Besides, 7 experts were asked to share their opinion on the
investigated topic in order to provide a more thorough insight into the behavior of
premium armored car customers. In this chapter the focus lies on the scientific part of
the research that is an application of the methodology suggested in this study.
Information and findings gathered for the practical outcome, such as expert opinions
and specific information about armoring and criminal situation is not provided in the
work because, on the one hand, it does not represent the scientific interest pursued
here and, on the other hand, most of the information was asked to be kept
confidential by the company.
6. Research design
163
Respondents represent current owners of premium or luxury autos who are as well
current or potential customers of armored cars. Design of the interview form required
a particular thoroughness and accuracy taking into consideration peculiarities of the
respondents who are:
• hardly accessible;
• continually busy and not ready to devote 1,5 hour for an interview which
doesn’t bring them any profit or benefit;
• not ready to share much about private and business issues of their life.
Selection criteria include objective and subjective criteria.
Objective criteria for respondents’ selection include two aspects:
1. class and price of a car in possession;
2. position and field of activity of a respondent.
Each respondent obtains at least one luxury auto starting from entry-level luxury cars
(i.e., premium according to European classification) and ending by high-end and ultra
luxury autos according to American classification of luxury autos (see Appendix 1).
Besides, all respondents spend their own means for an auto purchase.
In the initial stage of respondent selection it was assumed that people who spend
their own money are likely to make final purchasing decision and choice, although,
their decision may be influenced by opinions of opinion leaders from respondents’
surrounding or presented in different kinds of media. During the survey conduction
this assumption was proved. On the question: “Who makes decision on purchasing
an auto?” all respondents named themselves. On the question: “Who influences a
choice of technical equipment and other characteristics of an auto to be purchased?”
70% of respondents named themselves, 24% named “me and friends” and rest 6%
named “me and mass media”.
In addition to the current ownership of a premium or luxury car (further, in the study,
the notion ‘premium’ is used for both), all respondents represent experienced users
in a sense that currently used auto is not the first purchased auto.
6. Research design
164
Second objective criterion is respondent’s position and field of activity. This criterion
gives an insight into the social status and level of respondent’s welfare. The
specificity of the surveyed sample is that interviewer cannot ask direct question about
income or welfare. Most of the respondents either wouldn’t answer this question at all
or would give an answer which most probably does not reflect true state. Eighty
percent of respondents are owners or joint owners of private companies and the rest
represent chief executives and people working on top managerial positions.
Peculiarities of the sample do not allow using an objective criterion. For this reason,
subjective criterion appears supportive in sample selection.
Subjective criterion reflects respondents’ subjective estimation of personal income
and welfare. Personal income represents one of the most often asked data and can
be effectively used for predicting consumer behavior in marketing research.
However, this indicator when asked directly is not perfect. Consumers with low
income often pretend and behave as having higher income, especially when being
surveyed rather than observed during actual buying behavior, and vice versa. One of
the reasons for such contradiction is that consumers differ in their financial
commitments and abilities to manage own assets. Two people having the same
income, after fulfillment of necessary payments and purchasing essential goods may
perceive their discretionary income differently. One may perceive his discretionary
income as relatively low and therefore feel himself relatively poor while another one
perceives discretionary income as relatively high and feels himself as relatively rich
person. People with higher discretionary income have more opportunities to spend
money on special, luxury or comfortable goods and services or save money for future
expenses. Therefore, information about respondents’ subjective discretionary income
appears to be useful in selection, segmentation and prediction of consumers buying
behavior in marketing research (Churchill, 1996, p. 356).
Subjective discretionary income (SDI) is a measure of perceived spending power
originally conceptualized and developed by O’Guinn and Wells (1989). Specifically, it
is an estimate by the consumer of how much money he or she has to spend on
nonessentials. The SDI index is “considered as attitudinal rather than an objective
behavioral resource” and is based on the psychological approach allowing
assessment of a capacity to spend. This approach implies revealing subjective
estimation of personal well-being rather than asking respondents to provide complex
6. Research design
165
objective economic data. Analysis conducted in different fields provided evidence of
the SDI predictive validity (e.g., Rossiter, 1995; O’Guinn and Wells (1989) cited in
Bearden and Netemeyer, 1999).
SDI measurement approach allows to overcome criticism claiming that firstly,
consumers cannot objectively and precisely calculate and say about their
discretionary income. Secondly, understanding of “discretionary” and “necessary”
income is very subjective and depends on the individual perception.
Psychological approach of assessing respondents’ SDI implies estimation of three
statements in terms of how they reflect personal perception of his state of well-being.
The following statements were scored on 5-point scale starting from “absolutely
agree” (5) and ending up with “absolutely disagree” (1) (Bearden and Netemeyer,
1999):
1. “No matter how fast my income goes up, I never seem to get ahead.”
2. “I have more to spend on extras than most of my neighbors.”
3. “My income is high enough to satisfy nearly all of my important desires.”
The first statement reflects personal economic state and his ability to manage
personal assets. It shows the ability to earn and at the same time to spend less than
it is available. Second statement reflects availability of means and resources that are
not available for neighbors who represent respondent’s considerable reference
group. The third statement stands for an attitude towards money in terms of having
sufficient resources to purchase goods and services that are considered to be
desirable and, at the same time, represent rather important desires, because desires
in general are unlimited and there is always something unreachable at the moment
no matter how much a person can spend (Churchill, 1996).
Sum of individual scores given on each statement form an overall SDI index that can
range from 3 to 15. Respondents having higher sums on the scale show that they
can afford what they desire and even more. Respondents having smaller sums on
the scale show lack of means (Churchill, 1996). In this study four states of welfare
are proposed based on the sum of ratings (Table 16).
A majority of respondents estimate their income and welfare as relatively high and no
respondent estimates his income as “not enough”. This shows that the sample
6. Research design
166
comprises people who feel themselves as having high standard of well-being and at
the same time they can afford premium or luxury auto and, therefore, it can be
assumed that respondents represent consumers from upper-middle or higher class
who can afford almost everything. Although their purchasing behavior can be
sometimes perceived as irrational in terms of common understanding of main
reasons for buying behavior (usually pursued regarding volume brands), these
people still can be characterized as rational and adequate consumers who make a
choice instead of ‘buying everything up’.
State of welfare Sum of ratings Amount of respondents (%)
High income 13-15 25%
High income “for me” 10-12 65%
Not always high enough 7-9 10%
Not enough 3-6 0%
Table 16: Personal estimation of welfare
Measuring of SDI in the demonstrated way might also provoke some criticism
concerning measurement and data accuracy. However, in conditions of inability to
estimate respondents’ income directly, this approach was found most appropriate to
reach an insight to the respondents’ standard of living without asking them
undesirable and often provocative direct questions.
6.2.2. Survey instrument: repertory grid design
The survey instrument represents a questionnaire consisting of a repertory grid form
and additional questions concerning the car usage and demographic data. Some
complementary questions are included to get more detailed information about
respondents that is important for the company which is also interested in practical
findings. Within the context, these questions are only briefly addressed and the full
questionnaire form is provided in Appendix 3. For the empirical study to be
conducted in accordance with the proposed methodology the relevant part of the
questionnaire is a design of the RG form. This part is comprehensively described
further.
6. Research design
167
The final survey instrument consisted of an eight page questionnaire. Four pages
comprised repertory grid form, laddering down form, laddering up form and
resistance to change form. The rest pages included questions regarding:
♦ Utilized auto:
� Model and maker of an auto, car owner, driver, person who makes decisions
concerning the auto;
♦ Aim of purchase of a car (current or next);
♦ (open-ended) Importance and meaning of being confident about an auto;
♦ Usage of enhanced security devices and measures;
♦ (open-ended) Situations related to auto utilization and ownership which
respondent beware of:
� in general (what comes to mind at first);
� affecting:
• respondent’s image;
• respondent’s property (e.g. auto, notebook, briefcase, bag, etc.);
• respondent in physical or psychological way;
♦ Preference to either high-tech or physical security methods in terms of confidence
in case of need;
♦ Estimation of subjective discretionary income;
♦ Personal data:
• Field and sort of activity, education, family status, age, etc.
Next each component of the designed repertory grid form is described.
Topic
At first the topic has been agreed with representatives of the company which has
ordered the study in accordance with the aim of the scientific aspect of the work. The
topic of the repertory grid elicitation is “Exploration of similarities and differences in
the content and structure of customers’ subjective meaning systems related to a
decision making process of buying a car”.
Grids were elicited individually during face-to-face interviews conducted in places
convenient for respondents (home, office, restaurant) allowing for calm, undisturbed
6. Research design
168
1,5 hour (on average) conversation. Respondents were given a brief introduction into
the aims of the study and the nature of the repertory grid technique.
Set of elements
Two aspects need to be considered while selection of elements:
1. Decide on appropriate strategy for element selection (see 4.5.1 for strategies
description and their advantages and disadvantages);
2. Elements set should meet the requirements described in section 4.3.2 to
preserve the validity of a grid’s design.
In this study the third strategy described in 4.5.1 was chosen. Each respondent was
provided with general categories which cover the range of the topic and was asked
questions based on the categories, answers to which appeared as elements.
The aim of the RG application is to explore and build a model of a particular domain
of knowledge by enabling consumers to verbalize how they perceive certain
elements within the area of interest (McKnight, 2000). In the current study area of
interest covers questions such as how consumers of luxury and premium autos
perceive ownership and exploitation of an auto and what thoughts, feeling, priorities
and preferences do they have regarding auto purchase, ownership and usage.
To derive a set of elements so that the proposed method would be meaningful it is
necessary to identify a set of autos among respondent’s consideration set, i.e., autos
that are known to the respondent (McKnight, 2000). Therefore, at the end, elements
should be named by the respondent.
In order respondents’ grids to be comparable, each respondent names autos that
correspond to answers on the category based questions. All respondents are asked
the same questions so that the elements are elicited by using identical elicitation
categories (Jankowicz, 2004, p. 32). This approach, on the one hand, allows
respondents to consider their own elements which they are aware of and, on the
other hand, permits further analysis and comparison of all grids. Each respondent
names different autos which he can consider, that is, either he has got an experience
of exploitation or knows about it from, for example, friends and has got a certain
opinion and attitude towards named autos.
6. Research design
169
Each respondent was asked to name eight different cars as answers for the
questions listed in Table 17. To find most of individual’s constructs which also may
reflect undesired states and meanings, respondent had to provide a stratified set of
elements (Björklund, 2005), including some that he likes in a way and some that he
does not like or that are not suitable in a way. Provided questions make respondent
consider vehicles from different perspectives and, thereby, to think about different
autos according to the provided situation or condition. Provided categories cover
most important and common consideration aspects which arise when thinking about
auto ownership and exploitation.
Marking for analysis Categories questions for elements elicitation
1 (work) Which auto would you prefer to use mostly for work
2 (vacation) Which auto would you prefer to use mostly for vacation (weekend or holiday)
3 (evening) Which auto would you prefer to use mostly for leisure (evening) time
4 (in no way) Which car evokes negative emotions and you wouldn’t consider it as an alternative
5 (previous) Your previous auto
6 (doesn't fit) Which auto really attracts you, however it doesn’t fit your lifestyle or environment
7 (alternative) Which auto did you consider as an alternative to buy (during the last purchase)
8 (current) Your current auto (mostly used)
9 (ideal) Your generalized image of an ideal auto - supplied element, provided after all constructs were elicited
Table 17: Category based questions for eliciting a set of elements
First three categories cover most common auto exploitation situations:
• 1 (working): If a person is about to choose an auto which would be used
particularly for his working time, which auto would he prefer. His preferences
and choice would depend on the way the person exploits an auto during
working hours, for example, whether he has a driver or drives himself; whether
he uses time during the rides for work or for rest, etc.;
• 2 (vacation): An auto a person would prefer to use mainly during holidays and
weekends. Preferences would depend on the way the person spends his
vacation, for example, with family, with friends or by himself; whether he goes
out of the city to the mountains or suburbs or prefers to spend vacation in the
city or abroad, etc.;
• 3 (evening): An auto a person would prefer to use mainly for evening, i.e.,
leisure, after work time. Preferences would depend on the way the person
spends his free time, for example, whether he goes to fancy clubs or
6. Research design
170
restaurants or he prefers to spend evenings in cottage with family or closest
friends (what is very popular in Russia), etc.
Categories 5, 7 and 8 reveal autos with which respondents are good familiar and
experienced.
• 5 (previous): A respondent got an experience with previously owned auto,
therefore, he can surely express in the constructs what he did or did not like,
what disconfirmations were faced and reasons for purchasing next, i.e.,
“current” auto (the reasons, though, may not reflect negative experience);
• 7 (alternative): When making auto purchasing decision a respondent is likely
to choose among most desirable range (in case of this study, when people
can afford what they really like rather than for what they have enough money).
The final choice in favor of “current” auto is likely to leave behind at least one
auto which was carefully studied and examined, however, the choice has
fallen on another one;
• 8 (current): Auto which is currently used implies a certain experience and
opinion about it. When a respondent owns more than one auto it is up to him
which auto he names. The study shows, however, that respondents likely
name those autos which they use more often or which exploitation perceived
as primary.
Categories 4 and 6 derive elements regarding which respondents a likely to have
negative or not suitable judgements and perceptions:
• 4 (in no way): A respondent may have negative personal experience, be
aware of negative reference, have negative attitude towards car maker or
simply not like certain auto. There are certain meaning representations for
such an attitude which can be revealed by triadic differentiation;
• 6 (doesn’t fit): There are often situations, life circumstances or a way of
exploitation which entail possession of certain auto features or attributes.
However, desirable autos may often lack these features and, therefore,
appear not appropriate for a personal use in that particular circumstance or
situation. For example, in situation of a very cold winter time and bad quality of
roads it does not make sense to buy a sporty cabriolet unless to buy it only for
summer months, although, quality of roads is not a seasonable problem; or
6. Research design
171
when a person lives in downtown where problems with traffic jams and lack of
parking place are actual it becomes inconvenient to have a relatively big auto
such as, e.g., van or SUV.
During the elicitation of elements the last supplied ‘ideal’ element was not yet
provided. Only until thereafter it was included.
According to Yorke (1985), sampling of elements can actually rarely be
representative of their respective area. However, validity can be considered from the
notion of usefulness, that is, grids are valid in the sense that they allow accurate
capturing of respondent’s construing and “they tell the researchers what they don’t
know, thereby, contributing to the building of theory about individual concerned”
(Yorke,1985).
Elicited constructs
Once elements were elicited from the respondent, construct elicitation procedure
occurred. Respondent was presented with a subset of elements, i.e., triad and asked
to generate a construct which would meaningfully for him facilitate comparison and
discrimination between these elements. The aim is to elicit a set of bipolar constructs
which the respondent utilizes to perceive the elements (Dillon and McKnight, 1990).
Constructs were elicited by using the triadic difference method (see 4.5.2.1). This
method was selected because the produced constructs are able to better
discriminate between elements (Caputi and Reddy, 1999). It was important since
elements were rather similar in a sense that they represented autos mostly of a
higher class reflecting sample’s consideration sets. Besides, the method yields
constructs that are more cognitively complex (Fransella et al., 2004, p. 63)
Each respondent was presented with three elements at a time (triad) written on
different cards and asked to think of a way in which two of these three are similar
and, thereby, different from the third. The elements were always selected from the
set elicited from the respondent. Eight constructs were elicited by presenting different
triads composed according to a predefined sequence arranged so that no pairings of
elements are repeated and each element is presented 3 times among triads (Dillon
and McKnight, 1990; Jankowicz, 2004, p. 42). The sequence of triads allows future
comparison between the grids when certain categories are compared in terms of
6. Research design
172
individual constructs. For example, frequency of constructs with similar meanings
expressed towards same triad, i.e., same categories’ comparison. The combinations
were presented to each respondent in the same order:
128
427
137
368
265
534
578
164.
Depending on the research field and objective the sufficient amount of constructs
may vary. However, when it is not for a clinical purpose, it is agreed that eliciting
between 7 and 12 constructs during 1 hour session is usually sufficient and any
additional constructs provide little difference to the distribution of elements
(Jankowicz, 2004, p. 44). Easterby-Smith (1980) argues that even from eight to ten
constructs are quite adequate for most purposes.
Each respondent was expected to generate one construct per triad, however, in
practice several respondents produced more then one construct from a given triad. In
this case, these were noted and used in turn (Dillon and McKnight, 1990).
It is important to tell the person that he should voice that difference and similarity that
come at first moment when triad is presented. It is important that he uses the terms in
which he actually construes in mind, no matter if they are jargon or censorial.
Besides, it appeared to be useful to explain that each thought, each kind of
expression is important; there are no ‘right’ or ‘wrong’ answers and what is more,
there are no ‘clever’ or ‘stupid’ answers. What is important, is the person and his own
individual construing such as it is.
When a meaningful construct was generated the two poles were written in the cells of
the grid form, the emergent pole on the left and implicit pole on the right. Constructs
6. Research design
173
that were not enough precise were at first specified through laddering down
technique by asking the respondent “how/what do you mean; in what way?” (there is
a variety of ways the question can be asked depending on the context that can be
found in Jankowicz, 2004, p. 34) and the answer was used as a construct to be
noted in the grid form. Ladders were also recorded on a separate ‘laddering down’
form.
Elicited data was recorded on a recording sheet. The numbers of element cards
between which the similarity occurred were underlined to indicate which element pair
formed the construct pole.
There are plenty of pitfalls that may occur during interviewing concerning adequacy
of expressed meanings and ways to overcome uncertainties and misunderstandings.
Different examples and solutions are thoroughly described by Jankowicz (2004). This
study does not cover these details because most of the cases relate to application of
repertory grid in other fields rather than marketing such as clinics, human recourses,
education, etc.
Ratings
Once constructs on all given triads were elicited and noted in the grid form,
respondent was asked to rate all the elements, i.e., autos on each construct. Each
construct was presented as a rating scale with the contrast meaning on the left
(emergent pole) standing for the ‘1’ end of the scale and the contrast meaning on the
right (implicit pole) standing for the ‘5’ end of the scale. The 5-point scale was chosen
following recommendations and arguments provided in the literature (e.g.,
Jankowicz, 2004, p. 36; Marsden and Littler, 2000a; Dillon and McKnight, 1990).
Respondent rated each element based on how his perception of the element is
referred to one of the construct poles.
Supplied constructs
One of practical research objectives is to explore respondents’ attitude towards the
issue of safety related to the auto usage. Following aspects are of a special interest
in the research:
• What is meant by safety and feeling safe in a car;
6. Research design
174
• How is it important to feel safe in a car;
• How close are perception of feeling safe in a car and perception of feeling
confident about a car.
For this reason two overall assessment constructs were supplied:
• Feel confident about a car – Not feel confident about a car
• Feel safe in a car – Not feel safe in a car
In addition, beforehand, in the provided questionnaire an open end question “what
does it mean to you to feel confident about a car?” was asked to reveal how person
perceives ‘confidence’ within the context.
Respondent was asked to rate elements on the supplied constructs in the same way
as he did with elicited constructs.
Supplied element
After elicited elements, constructs and ratings were obtained the respondent was
asked to rate the supplied ‘ideal’ element on the same constructs. Ratings assigned
to the ‘ideal’ element represent the extent of preference for the characteristic
performance comprised in the corresponding pole. Ratings of ‘ideal’ element can also
point where respondent’s actual and desirable images are among the construed
elements.
Personal values
Once ratings on elicited and supplied elements were obtained, respondent was given
laddering up task regarding constructs which were assigned with the extreme rating
values on the ‘ideal’ element (see 5.3 for the reasoning). Laddering task was aimed
at eliciting individual’s means-end knowledge structures linking product
characteristics to associated values or ends.
Respondent was asked ‘why is this important for you’ and each given ladder was
noted on the separate ‘laddering up’ form (see Appendix 3). Thereby, interviewer
arrived at more abstract values associated with original constructs (Jankowicz, 2004,
p. 186; Marsden and Littler, 2000a).
6. Research design
175
After all personal values were elicited, respondent was asked to make choice
according to resistance-to-change technique described in section 4.5.4.2 aimed at
prioritizing personal values. He was asked to imagine the choice as an existential
one. For example, if the revealed values were ‘pleasure, enjoyment of life’ and
‘success, achievement’ the respondent was asked to choose one option either
“imagine, after some years you are going to live the rest of your life having everything
what you need to enjoy every day, but you will realize that you have not achieved
desired success and aspirations that you have been striving for” or, alternatively,
“imagine, you have gained everything you have been working for, absolute self-
realization, however, you realize that your achievements appeared to be not enough
to make you happy and enjoy the rest of your life”. Thereby, a person had to prefer
one end state at cost of another, i.e., to prefer one value over another. When all
respondent’s personal values were compared with each other the outcome record
represented a hierarchy of his personal values.
Example of the recorded individual RG is presented in Table 18. Such a full and
structured way of recording allowed by RGT has advantages which make the
technique even more convenient in application. Although the record is extremely full
of data, it is possible to reconstruct all the substantive points that occurred in the
interview by looking at the record. Another advantage is that the structured form
allows several interviewers working on the same problem where each researcher can
understand his colleagues’ interview records without making long discussions,
explanations and distorting reconstructions of what went on during their various
interviews (Stewart et al., 1981).
6. Research design
176
Table 18: Repertory grid record of the 14th respondent
of grid and construct
Combination
Em
erge
nt p
ole
(1)
Mercedes S221 Guard
Audi Q7
Porsche Carrera
Jaguar
Mercedes CLK
Aston Martin DB9 Coupe
Mercedes S220
Audi S8 Armored
IDEAL
Implic
it pol
e (5
)
(1)
wor
k(2
) va
catio
n(3
) ev
enin
g(4
) in
no
way
(5)
pre
viou
s(6
) d
oesn
't fit
(7)
alte
rnat
ive
(8)
curr
ent
(9)
idea
l
14.1
1 2 8
appro
pria
te for
arm
orin
g1
55
55
51
11
inex
pedie
ncy
of ar
mor
ing
14.2
4 2 7
beau
tiful
12
15
31
31
2uns
atis
fact
ory
des
ign
14.3
1 3 7
nov
elty
of a
mod
el1
11
53
13
22
old m
odel
14.4
3 6 8
for
leis
ure
time
51
13
11
55
5fo
r bus
ines
s, n
ot for
en
tert
ainm
ent
14.5
2 6 5
wor
se t
echnol
ogic
al e
quip
men
t in
sen
se o
f sa
fety
54
43
45
45
5 in
itial
ly b
ette
r te
chnol
ogic
al
equi
pm
ent
in s
ense
of sa
fety
14.6
5 3 4
ther
e is
no
late
r m
odel
:
tech
nol
ogic
ally
innov
ativ
e1
11
31
13
11
ther
e is
alrea
dy a
late
r m
odel
14.7
5 7 8
with
drive
r1
45
35
51
12
for
self-
driv
ing
14.8
1 6 4
appro
priat
e fo
r co
mfo
rt
elec
tron
ics
and t
echnol
ogie
s1
11
32
12
12
limite
d e
quip
men
t w
ith c
omfo
rt
elec
tron
ics
and t
echnol
ogie
s
14.9
sup
plie
dfe
el c
onfid
ent
abou
t th
is c
ar1
22
32
11
11
not
fee
l con
fiden
t ab
out
this
car
14.1
0su
pp
lied
feel
saf
e in
thi
s ca
r1
22
32
12
11
not
fee
l saf
e in
this
car
7. Analysis of repertory grids
177
7. Analysis of repertory grids
7.1. Preparation of data
The first step in analyzing the data is to turn it into a proper format. All interviews
were conducted in Russia and noted in Russian. Therefore, the first thing to do was
to translate all the grids and record them digitally. At this phase skills of the
researcher play an essential role in his ability to accurately translate and justify all the
expressed meanings. Thorough knowledge of both languages is required to be
aware not only of word meanings provided by dictionaries but also awareness of their
special underlying meanings which are used in certain cultural and social
surroundings.
Once all grids are translated and recorded an appropriate data format for the
computer package is required.
Constructs of all grids are coded with the first number denoting an interviewee and
the second number - the construct order within the grid. For example, the code 14.7
indicates that the construct in question is the seventh construct in the fourteenth
interviewee’s grid (Jankowicz, 2004, p. 151). After the constructs are coded and,
thereby, can be identified, a categorization procedure which represents the generic
procedure of content analysis is performed.
7.2. Content analysis
7.2.1. The generic content analysis procedure
Content analysis is a technique in which constructs of all the respondents are
grouped and categorized according to the meanings they express. As discussed in
4.6.1.2, categories can be theory driven or data driven. In this study categories are
derived from constructs themselves by identifying the various themes they express.
The reason for choosing data driven approach is explained by the research purpose.
The study is aimed at revelation of common meanings, perceptions and stereotypes
as well as jargon used by a certain kind of customers in considering a very specific
product category. Moreover, respondents represent residents of a different culture to
the one in which the producing company is founded.
To reveal categories and allocate constructs to them all grids are cut into pieces,
each piece containing one grid row, i.e., one construct. When all pieces are put on a
7. Analysis of repertory grids
178
large table they can be shuffled around until all of them are placed into piles, each
pile constituting a different category (Jankowicz, 2004, p. 151). Each item is to be
compared with another and if they are alike in some way they are placed together
under a single category created for them at that very moment or if meanings are
different they are put into separate piles. Each time remaining items are compared
with each of the already derived categories and allocated to the appropriate one or a
new category is created. Once a new category is created, other existing categories
are reconsidered and if needed redefined, i.e., combined or broken up with their
items reallocated accordingly (Jankowicz, 2004, p. 149).
It can happen that some number of items appears to be unclassifiable. In this case,
as suggested by Jankowicz (2004, p. 149), if the number of unclassified items is less
than 5% of the total they can be allocated to a single category labeled
‘miscellaneous’, otherwise one or more categories should be redefined so that at the
end not more than 5% of the total are categorized as ‘miscellaneous’.
In the current study all constructs are classified and allocated to a certain category.
However, it is found that some meanings of constructs actually enclose combination
of two (and in one case, three) meanings which belong to different categories. In this
case the construct is assigned to both categories. For example, the construct
“attractive, good design – bad, old (outdated) design” reflects respondent’s
judgement that “only autos of new models have attractive, good design as they are
not yet widespread”. This person considers only new car models as having good
design and he is ‘right’ as he is always right in his personal perceptions and attitudes.
In this case the construct was allocated to categories “design” and “novelty” as both
these aspects are equally important in the content of the construct meaning and need
to be satisfied to bring the desirable consequence.
Derived categories are also identified with respect to their position in the CMS
scheme. In the model conceptualization defined categories represent higher level of
abstraction of constructs. Therefore, categories are defined in such a way that they
represent product attributes reflecting perceived product characteristics rather than
product physical characteristics.
7. Analysis of repertory grids
179
Comfort 1.2 comfortable: driver's comfort in terms of less emotional and physical efforts 281.5 comfortable2: comfort electronics and technologies 14.2
2.3 comfortable: appropriate size, comfortable passenger seating arrangement2.11 comfortable2: comfort electronics and technologies 3.5 comfortable: don't feel hummocks, potholes and bumpy roads at driving Q
Prestige prestigious -- ordinary 4.4 prestigious: others will notice me 264.5 universal exploitation / large / prestigious -- narrow exploitation/compact/ordinary: universal-for business and family use, narrow-cool, only for me
12.7
5.1 prestigious / representative: not youth5.4 prestigious2: black color - spectacular, convincing5.8 prestigious3 -- ordinary3: super image, for men, not womenQ
Representativeness 1.8 representative -- unrepresentative: prestigious, which means reflects my status 232.8 representative: prestigious, which means reflects my status 11.7
3.4 representative -- unrepresentative: for business but not for entertainment, family or vacation3.6 representative2: not for self-drivingQ
of high technical quality -- of low technical quality
1.1 rough-terrain high -- low -performance/of high -- low technical quality: safe "anywhere" ride in tough climate conditions (mud, ice)
20 10.2
2.4 of high technical quality: time and money savings2.7 of high technical quality2: producer's proved high quality (producer reputation)3.2 of high technical quality: won't break down, time and money savings4.7 of high technical quality: safe in our climate (snowdrift, glare ice)Q4.3 narrow exploitation: sporty for pleasure, not city-universal 17
4.5 universal exploitation/large/prestigious -- narrow exploitation/compact/ordinary: for business and family use
8.6
6.5 universal exploitation: for year-round 6.8 universal exploitation2: highly functional7.5 universal exploitation -- narrow: for everyday life not only for leisure driving7.12 universal exploitation2 / unrepresentative2: can fit and seat my dog, when car is too representative, I can't seat in my dogQ
Size large -- compact 1.6 large: for several people or luggage 17
1.7 large2: comfortable luggage transportation (don't need to push in and stuff) 8.6
2.6 large: better observation (panoramic view) and feel safeQ
Design stylish -- unattractive 2.2 stylish: bright attractive appearance 17
3.3 stylish -- unattractive: stylish appearance 8.6
4.8 stylish: attractiveQ
Speed/Power 1.3 high-speed/powerful: adequacy between power and weight 13
4.12 high-speed/powerful -- limited: speedy 6.6
9.2 high-speed/powerful: powerful10.3 high-speed/powerful: sporty = speedy and powerfulQ1.1 rough-terrain high -- low -performance/of high -- low technical quality: safe "anywhere" ride in tough climate conditions (mud, ice)
12 6.1
2.1 rough-terrain high performance: city, countryside, mountains5.7 rough-terrain high performance: city, countryside, mountains7.1 rough-terrain high -- low -performance: for city/everywhere7.7 rough-terrain high performance2 / of high technical quality: rigid suspension, rugged auto will survive in fosseQ
Safety/Security safe/secure -- unsafe 1.4 safe/secure: appropriate for security measures (bodyguard, driver, armoring) 108.1 safe/secure -- unsafe: practical = not notable on the roads, less risk of criminal attempt 5.1
8.11 safe/secure: no risk of criminal attempt because of good car brand 9.5 safe/secure: maneuverabilityQ
Price 2.5 expensive -- affordable price: not affordable by anyone 8
7.8 expensive -- affordable price: rather expensive but without extra frills that make car overpriced
4.1
10.7 expensive -- affordable price11.8 affordable price -- expensive: affordable for majority - expensive16.4 expensive -- affordable price: need to work hard to afford this auto Q
Novelty new model -- old model 3.1 new model: new means technologically innovative, not outdated 7
5.3 new model -- old model: unique, uncommon 3.67.6 stylish2/new model: elegant because of novelty, uniqueness12.3 new model: new means technologically innovative, better steering, can't be burnt anymoreQ
Totals 197
100.0
Intended application (Universality)
Cross-country ability
Technical quality
universal exploitation -- narrow exploitation
Category Bipolar category's poles Construct
comfortable -- uncomfortable
expensive -- affordable price
high-speed/powerful -- limited
rough-terrain high
performance --
rough-terrain low performance
Sum and % in
general sample
representative -- unrepresentative
Table 19: Content analysis procedure
Twelve different categories are identified according to the content of constructs. It is
important to remember that the same attribute may imply different meanings for
7. Analysis of repertory grids
180
different respondents. For this reason the original construct phrases are kept and
their implied meanings are enclosed into the analysis and interpretation phases.
A total of 180 elicited constructs are content-analyzed and because, as mentioned
above, some can be attributed to two categories the total amount of constructs
comprised by categories is 197.
Table 19 provides all the categories and numbers of constructs categorized under
each heading. All constructs allocated to each category are provided and categories
are ordered according to the number of constructs allocated to them (see Appendix 4
for the full table). Construct poles are reflected according to preferences expressed
by ratings on ‘ideal’ element. Both poles are written when the ‘ideal’ element is given
the neutral rating, ‘3’ on the construct.
Performance of frequency counts by categorizing constructs discloses common
aspects which respondents consider when they construe meanings about the
discussed topic, i.e., about premium cars. Thus, derived categories and the amount
of constructs in each category reflect more or less common aspects being
considered. The top four categories which together cover 49% of all constructs are
‘comfort’, ‘prestige’, ‘representativeness’ and ‘technical quality’. It means that these
four categories represent product attributes that are most common among the
respondents and describe 49% of their expressed meanings.
Therefore, first aggregation of individuals’ meanings is done by recognizing and
categorizing similarities and differences in the respondents’ meanings.
Categorization technique, however, employs only the context presented in the
constructs and does not take into account information provided by ratings which
emphasize the ways in which respondents use those constructs.
Frequency counts conducted for all elicited constructs reflect the homogeneity of
aspects that respondents consider when thinking about the elements. It is also
meaningful to examine how important mentioned aspects are (see Table 20).
Relative importance of constructs is defined by rating values given to the ‘ideal’
element indicating the importance of the corresponding construct pole (Walker and
Olson, 1991). Thus, when constructs to be counted are only those that are given
rating value of ‘1’ of ‘5’, the extent of importance of each category can be assumed.
7. Analysis of repertory grids
181
In total, on 100 out of 197 elicited constructs (51%) the ‘ideal’ element is rated with
value of ‘1’ or ‘5’.
Category name
Sum in general sample
Homogeneity among all
elicited constructs, %
Rank of homogeneity
Sum among 'important' constructs
Homogeneity among
important constructs, %
Importance among
constructs in category, %
Rank of importance
Comfort 28 14 1 22 22 79 2
Prestige 25 13 2 10 10 40 7
Representativeness 23 12 3 8 8 35 8
Technical quality 20 10 4 16 16 80 1
Intended application (Universality) 17 9 5,6,7 5 5 29 10,11
Size 17 9 5,6,7 5 5 29 10,11
Design 17 9 5,6,7 12 12 71 4
Speed/Power 13 7 8 7 7 54 6
Cross-country ability 12 6 9 4 4 33 9
Safety/Security 10 5 10 6 6 60 5
Price 8 4 11 0 0 0 12
Novelty 7 4 12 5 5 71 3
Totals 197 100 100 100 51
Table 20: Homogeneity and the importance of the derived meanings
There are categories that are rather common and related meanings emerge very
often among respondents when thinking about a certain product category. Although
categories are common, these meanings can be of different importance for
respondents.
If categories are ranked according to the percentage of most important constructs the
order looks different (compare 3rd “rank of homogeneity” and 7th “rank of importance”
columns of Table 20). Category ‘technical quality’ is on the 4th place when looking at
the sum of constructs; however, 80% of these constructs are valued as very
important. Whereas category ‘prestige’ which is on the 2nd place according to the
aspect mentioned, when considering importance of the attribute this category drops
to the 7th place as only 40% of all constructs in this category are perceived as very
important.
Categories which are rare are likely to represent either very important aspect for
people or, on the contrary, may be not important and emerge within the expressed
meanings because respondents could not construe any other more actual for them
meaning.
Category ‘novelty’ which is not very common attribute among respondents is very
important for those who mentioned it. It means that there is a certain consumer
segment for which novelty of a car, its model or technical innovations play very
7. Analysis of repertory grids
182
important role in making decision about its purchase and this segment’s priority
should not be underestimated.
The second assumption can be demonstrated on the example of category ‘price’.
Only 4% of all elicited constructs are about price and none of them are considered to
be important. This is, actually, a peculiarity which characterises the sample that was
interviewed. For these customers price is not a determinant factor when making
purchase decision.
Overall, it can be noticed that the most important categories are those which reflect
attributes important for confident and convenient utilization such as ‘technical quality’,
‘comfort’, ‘novelty’, ‘design’, ‘safety’ and ‘speed/power’. Confident utilization implies
both confidence in a car’s quality and in its image.
Next two categories according to rank of importance rather relate to the external
perception by others which are ‘prestige’ and ‘representativeness’. These both
attributes imply ability of auto’s image and image of auto’s owner to reflect
customer’s social status.
Other categories represent attributes that allow proper auto usage depending on the
conditions and circumstances in which auto is used. These categories include ‘cross-
country ability’, ‘intended application (universality)’ and ‘size’.
7.2.2. Honey’s content analysis
Honey’s content analysis is applied in this research with the goal of identifying
constructs that are more related to a particular issue the survey is focused on. A
particular interest of the survey is to understand how people perceive and what their
attitudes towards the notion of confidence and safety when talking about autos are.
It was expected that for some people the perception of feeling safe and secure in an
auto may differ from the perception of being confident about an auto. For this reason
two overall assessment constructs were supplied to be rated by each respondent.
The first construct is “overall, feel confident about this car – overall, not feel confident
about this car” and the second construct is “overall, feel safe in this car – overall, not
feel safe in this car”.
7. Analysis of repertory grids
183
All single grids are analyzed according to the procedure described in section 4.6.1.3.
For each elicited construct the sum of differences and the percentage similarity score
(PSS) index against both supplied overall assessment constructs are calculated
(formulas were presented in the 4.6.1.3). The calculation regarding each elicited
construct against each supplied construct is conducted twice, once with the given
ratings and once with the reversed ratings of the supplied construct. After all PSS
indices in the grid are estimated each construct is labeled with either ‘H’, ‘I’ or ‘L’
index standing for high, intermediate or low relevance of the construct to the overall
assessment construct of the respondent. To do so the individual’s personal metric is
calculated by defining individual H-, I- and L-value intervals. Table 21 shows 14th grid
(presented in Table 18) analysis by Honey’s technique.
Emergent pole (1)
Mer
cede
s S
221
Gua
rd
Aud
i Q7
Por
sch
e C
arre
ra
Jagu
ar
Mer
cede
s C
LK
Ast
on M
artin
DB
9 C
oupe
Mer
cede
s S
220
Aud
i S8
Arm
ore
d
IDE
AL
Implicit pole (5)
(1) work (2) vacation (3) evening (4) in no way (5) previous (6) doesn't fit (7) alternative (8) current (9) ideal
1 5 5 5 5 5 1 1 1
15 16,67 21 -16,67
16 11,11 20 -11,11
1 2 1 5 3 1 3 1 2
7 61,11 25 -38,89
6 66,67 24 -33,33
1 1 1 5 3 1 3 2 2
9 50,00 25 -38,89
8 55,56 24 -33,33
5 1 1 3 1 1 5 5 5
19 -5,56 13 27,78
18 0,00 14 22,22
5 4 4 3 4 5 4 5 5
25 -38,89 1 94,44
24 -33,33 0 100,00
1 1 1 3 1 1 3 1 1
5 72,22 27 -50,00
4 77,78 26 -44,44
1 4 5 3 5 5 1 1 2
13 27,78 17 5,56
14 22,22 16 11,11
1 1 1 3 2 1 2 1 2
4 77,78 26 -44,44
3 83,33 25 -38,89
1 2 2 3 2 1 1 1 1
5 4 4 3 4 5 5 5 5
1 2 2 3 2 1 2 1 1
5 4 4 3 4 5 4 5 5
14.10 supplied
14.9 supplied
H
H
14.8
not feel safe in this car
not feel confident about this car
feel safe in this car
feel confident about this car
14.7
14.6
14.5
14.4
limited equipment with comfort electronics and
technologies
14.2
14.1
for business, not for entertainment
old model
unsatisfactory design
inexpediency of armoringappropriate for armoring
for leisure time
novelty of a model
L
L
initially better technological equipment in sense of
safety
there is already a later model
for self-driving
H
H
H
L
I
I
L
H
L
L
I
I
beautiful
appropriate for comfort electronics and technologies
with driver
there is no later model: technologically innovative
worse technological equipment in sense of
safety
14.3
Table 21: Honey’s technique calculations
7. Analysis of repertory grids
184
The sums of differences are shown below each construct followed by PSSs. The left
two numbers are calculations between the unreversed ratings and the right two
numbers are the calculations between ratings of the elicited construct as they are
and the reversed ratings of the supplied construct. The reversed ratings of the
supplied constructs are shown bellow the supplied constructs 14.9, 14.10. Two rows
bellow each construct show calculations, first, between the construct and the
supplied construct 14.9 “feel confident about this car – not feel confident about this
car” and, second, between the construct and the supplied construct 14.10 “feel safe
in this car – not feel safe in this car” placed one above another correspondingly. For
each construct the higher of the two values of PSS (unreversed and reversed) is
chosen and shown in bold with respect to each comparison with supplied constructs.
Thus, the sum of difference between unreversed constructs 14.1 and 14.9 is
159
1
', =−=∑=
′
i
jiijjj rrD and the PSS is then
%67,169)15(
15%200%100
)1(
%200%100
',
,' =
×−
×−=
×−
×−=
nk
D jj
jjσ .
The constructs are divided evenly into three sets, high, intermediate and low
according to the individual’s personal metric, i.e., based on the grid’s highest and
lowest PSSs. For this grid the intervals are shown in the Table 22.
Similarity with 14.9 Similarity with 14.10
Interval for PSS Interval for PSS
"H" 68,53 - 94,44 70,38 - 100,00
"I" 42,60 - 68,52 40,75 - 70,37
"L" 16,67 - 42,59 11,11 - 40,74
Label
Table 22: Intervals for H-I-L indices
All elicited constructs within the sample are allocated to the categories. When all the
constructs are assigned with the H-I-L indices according to the individual’s personal
metrics the consensus between all respondents can be defined. In Table 23 the
constructs within each category are shown in three groups according to their high,
intermediate and low relevance to the supplied constructs.
7. Analysis of repertory grids
185
Sum of constructs
% in the category
Sum of constructs
% in the category
Sum of constructs
% in the category
Sum of constructs
% in the category
Sum of constructs
% in the category
Sum of constructs
% in the category
Comfort 28 14 14 50% 8 29% 6 21% 18 64% 7 25% 3 11%
Prestige 25 13 12 48% 6 24% 7 28% 12 48% 3 12% 10 40%
Representativeness 23 12 3 13% 7 30% 13 57% 5 22% 8 35% 10 43%
Technical quality 20 10 11 55% 4 20% 5 25% 10 50% 5 25% 5 25%
Intended application (Universality) 17 9 1 6% 8 47% 8 47% 7 41% 6 35% 4 24%
Size 17 9 2 12% 5 29% 10 59% 4 24% 9 53% 4 24%
Design 17 9 8 47% 6 35% 3 18% 8 47% 5 29% 4 24%
Speed/Power 13 7 4 31% 4 31% 5 38% 3 23% 5 38% 5 38%
Cross-country ability 12 6 3 25% 3 25% 6 50% 3 25% 2 17% 7 58%
Safety/Security 10 5 1 10% 1 10% 8 80% 2 20% 1 10% 7 70%
Price 8 4 0 0% 2 25% 6 75% 1 13% 0 0% 7 88%
Novelty 7 4 4 57% 2 29% 1 14% 4 57% 2 29% 1 14%
Totals 197 100,0 63 56 78 77 53 67
Percentage of all (197) 32% 28% 40% 39% 27% 34%
Sum in general sample
Sum in general
sample, %
Similarity with construct 14.9 'Confidence' Similarity with construct 14.10 'Safety'
H LIHLI
Table 23: Honey’s content analysis of all grids
Category ‘comfort’ is not only most common but also has a high relevance to both
supplied constructs for the majority of respondents. Also constructs ascribed to the
category ‘technical quality’ are rated similarly with the supplied constructs indicating
high relevance of their meaning when respondents think about overall confidence
and safety of an auto.
Although category ‘novelty’ is least common among the respondents, for the majority
of those who expressed constructs ascribed to this category the constructs are also
highly relevant and important when considering the question of confidence and safety
of an auto. This group of respondents represents a latent group for which the fact
that the auto is new and embodies most recent technological achievements is an
important factor influencing the choice decision and, therefore, it can be effectively
used in positioning and advertising of a product targeting this latent group.
When thinking about confidence about an auto, categories ‘size’, ‘cross-country
ability’, ‘safety/security’ and ‘price’ have rather low relevance for the majority. At the
same time, categories ‘prestige’ and ‘design’ appear relevant and similarly rated for
the large part of the sample. This signifies that the notion of confidence is often
related to the image of an auto and the social perception of an auto as well as of its
owner rather than to the feeling of safety.
Moreover, when thinking about the safety of an auto or provided by an auto, the
category ‘safety/security’ has a low relevance for the majority of people expressed
the constructs ascribed to this category. However, categories ‘prestige’ and ‘design’
remain highly relevant for a large part of the sample.
7. Analysis of repertory grids
186
This confusion is likely to be explained by the specificity of the sample and its
common attitude towards questions reflected in the supplied constructs. The
respondents, on the one hand, represent consumers of premium autos who perceive
general safety measures provided by an auto as a must have in premium autos due
to their superiority in quality. Cars named for each element category present premium
autos which, when being distinguished between each other by triadic differentiation,
usually can not be distinguished by general safety measures. On the other hand, the
respondents represent people referred to the risk group who are, if really concerned,
have particular requirements about safety and security measures which often exceed
the ordinary measures provided by ‘average’ premium autos. For example, as can be
seen in the individual grid (Table 21) the first construct ascribed to the category
‘safety/security’ has a low relevance to the overall construct about auto safety
because the respondent considers auto’s appropriateness for armoring which implies
rather different requirements of an auto than those which are needed for general
safety measures. For example, Audi Q7, although possessing all or most of the
superior technologies and safety measures in the general notion, is perceived by the
respondent as inappropriate for armoring because of its relatively large size what
makes it easier to hit it with the bullet or projectile.
The common attitude towards the notion of auto safety is rather skeptical in a sense
that people, at least in Russia, believe that if there is a plan to commit crime against
a person in a car or aimed at stealing property from an auto or stealing or damaging
the auto itself it will most probably be accomplished. This belief is explained by
another one which is “there are no technologies smarter than a man”. That is why
most safety and security measures are perceived either against different kinds of
accidents and provided by premium autos as a must or against foolish, poorly
thought out crimes. More serious measures require both specific technological
achievements as well as physical protection involving the human factor.
The H-I-L indices regarding both supplied constructs do not differ much within each
category. That is because most respondents rated both supplied constructs rather
similarly and the general match between these constructs is 96% considering all the
grids. It can, therefore, be supposed that most of respondents when rating the last
supplied construct were actually thinking about each element as a package
consisting of technological achievements as well as of the social value provided by a
7. Analysis of repertory grids
187
car and the more appropriate meaning to such assessment is probably the
convenience about an auto.
Because of the relatively small number of respondents in the sample the
assumptions based on this data might need additional testing. However, this study is,
first of all, aimed at demonstration of the technique application and illustrates how
researcher can handle the estimations. It shows on which kind of differences and
similarities in calculations the researcher needs to pay a particular attention because
it can disclose unconsidered, unexpected or unknown aspects of the study. It is also
very important to be careful with the interpretations to produce accurate inferences.
7.3. Analysis of relationships within a single grid
RG analysis is aimed at systematization of data and presenting it in a structure which
allows to reveal commonalities and differences among consumers according to their
perceptions and preferences. For this purpose each grid is separately analyzed,
elicited meanings are systematized and the CMS schema of each respondent is
constructed and presented. Afterwards, all structured individual data is analyzed
together.
To conduct the analysis of relationships within a single grid the following steps need
to be accomplished:
1) Preparation and input of data into the used software package.
2) Systematization of the elicited meanings:
a) Perform Slater’s PCA and provide a convenient structure form for findings
representation;
b) Perform HCA within each defined PCA component and structure individual
constructs based on the interrelationships between their meanings;
c) Build an individual CMS.
3) Systematization of elements perceptions:
a) Analyze matches between elements perceptions with respect to the
revealed construct dimensions, i.e., abstract consequences.
7. Analysis of repertory grids
188
7.3.1. Preparation and input of data into the software package
In this study analysis are carried out by means of a web version of REPGRID
program (Shaw, 1989) namely WEBGRID IV. This is the latest version introduced in
2008. Description of the software was provided in section 4.6.2.
Table 18 in section 6.2.2.1 shows the grid elicited from 14th interviewee as it was
recorded. All grids are inputted into the software including elements, constructs and
ratings. Figure 17 shows the 14th grid as it is displayed in the WEBGRID package.
Elements are inputted as element categories instead of mentioned particular cars in
order to later compare the grids and to analyze a collective grid representing several
individual ones. This allows having different constructs but the same set of elements.
Constructs are inputted in a coded form so that each construct is represented by its
unique coding number and the category it is ascribed to. As well as constructs,
categories are also bipolar as shown in Table 24.
Figure 17: Display of the 14th repertory grid
7. Analysis of repertory grids
189
Construct category Elicited meaning of construct Construct category Elicited meaning of construct
14.1 Safety/Security safe/secure appropriate for armoring unsafe inexpediency of armoring
14.2 Design stylish beautiful unattractive unsatisfactory design
14.3 Novelty new model novelty of a model old model old model
14.4 Representativeness unrepresentative for leisure time representative for business, not for entertainment
14.5 Safety/Security unsafeworse technological equipment in
sense of safetysafe/secure
initially better technological
equipment in sense of safety
14.6 Novelty new modelthere is no later model:
technologically innovativeold model there is already a later model
14.7 Representativeness representative with driver unrepresentative for self-driving
14.8 Comfort comfortableappropriate for comfort electronics
and technologiesuncomfortable
limited equipment with comfort
electronics and technologies
14.9 Overall confidencefeel confident about this car
feel confident about this carnot feel confident about this car
not feel confident about this car
14.10 Overall safety feel safe in this car feel safe in this car not feel safe in this car not feel safe in this car
Emergent pole Implicit poleConstruct № Category name
Table 24: Correspondence between category poles and construct poles
After all individual grids are inputted, WEBGRID’s analysis tools are used.
7.3.2. Systematization of elicited meanings
a) Perform Slater’s PCA and provide a convenient structure form for findings
representation
Each individual grid is examined by means of PCA. The produced graphic plot is
presented in Figure 18. In addition, a table with percentage of variance accounted for
by each component is also provided as shown in Table 25. Description of reading the
plot was described in section 4.6.2.2.
1 Component 2 Component 3 Component 4 Component
Percentage, % 64.36 30.80 2.23 1.19
Cumulative % 64.36 95.17 97.39 98.59
Table 25: Percentage of variance accounted for by each component.
Interpretation of the graphical plot involves three steps.
• At first it is needed to determine how many components are significant and
should be examined.
In this graphical plot the first component accounts for 64% of the variance and
together with the second, 31%, it identifies 95% of the variance in the data. It means
7. Analysis of repertory grids
190
that to analyze this grid only one plot of first component against the second should be
examined.
Figure 18: Graphic plot of a PCA of the 14th grid
• Secondly, the positioning of the lines representing the constructs needs to be
examined to reveal the distinct patterns.
Around the first component, i.e., x-axis three constructs are grouped, including:
� construct ‘14.1 safe/secure – unsafe’,
� construct ‘14.4 unrepresentative – representative’,
� construct ‘14.7 representative 2 – unrepresentative 2’.
These constructs form a joint with small angles around the first principal component
and could all be considered as important for the grading process.
Second grouping consists of four constructs including:
� construct ‘14.2 stylish – unattractive’,
� construct ‘14.3 new model – old model’,
� construct ‘14.6 new model 2 – old model 2’,
� construct ‘14.8 comfortable – uncomfortable’.
These constructs have relatively small angles towards the second component
indicating another dimension in the assessment system of this respondent.
7. Analysis of repertory grids
191
Apart from the grouped constructs, three constructs 14.5, 14.9 and 14.10 are
disposed between the components being well represented by neither of them. Two of
these constructs are overall supplied constructs ‘14.9 feel confident about this car –
not feel confident about this car’ and ‘14.10 feel safe in this car – not feel safe in this
car’ and the third construct is elicited one ‘14.5 unsafe2 – safe/secure2’. Construct
lines of 14.5 and 14.10 are coincident as their ratings are identical when construct
14.5 is reversed, i.e., the PSS between them is 100%. The expressed meaning
underlying the elicited construct is ‘safe means initially better technological
equipment in sense of safety’ which has a close meaning to the overall construct
14.10.
Thus, constructs 14.5, 14.9 and 14.10 align neither with the first grouping nor the
second. It can be concluded that overall feeling of safety or confidence about a car
depends neither on attributes aligned to the first component nor on the attributes
aligned to the second component but more or less equally on both (Jankowicz, 2004,
p. 129).
The graphical plot allows interpretation of the data, however, it is not always easy to
attribute constructs to the components by visual inspection.
Together with graphical plot the tool provides an underlying textual output. It includes
tables of construct and element loadings on each principal component. To further
discuss the constructs a table of construct loadings is first presented and interpreted
(Table 26).
Construct Component 1 Component 2 Component 3 Component 4 Construct poles (categorized)
14.1 -2.95 0.29 -0.16 -0.11 safe/secure--unsafe
14.2 -0.32 1.87 0.12 0.00 stylish--unattractive
14.3 0.03 1.87 0.36 -0.06 new model--old model
14.4 2.77 0.48 0.07 -0.11 unrepresentative--representative
14.5 0.49 -0.79 0.33 0.03 unsafe2--safe/secure2
14.6 0.19 1.08 -0.32 0.49 new model2--old model2
14.7 -2.46 -0.43 0.42 0.16 representative2--unrepresentative2
14.8 -0.00 0.96 0.31 0.02 comfortable--uncomfortable
14.9 -0.67 0.64 -0.21 -0.38 feel confident about this car--not feel confident about this car
14.10 -0.49 0.79 -0.33 -0.03 feel safe in this car--not feel safe in this car Table 26: Construct loadings on each component
As described in section 4.6.2.2 construct loadings do not directly represent the
correlation between the component and its construct as in original principal
component analysis. Loadings represent distances which absolute values depend on
how the data has been scaled. Thus, for interpretation the relative values are
7. Analysis of repertory grids
192
meaningful (Gaines and Shaw, 2005). It means that the criteria for construct loading
value to be considered significant also depends on the relative loading values.
Therefore, the loadings are considered as follows:
� Constructs with highest loadings on one component and, at the same time,
with relatively low loadings on other components are considered indicative for
the component generalized meaning.
� Constructs with high loadings on two (or more) components are not indicative
for defining generalized meaning and name of each component. However,
these constructs are significant and demand deliberation bearing in mind that
their meanings take account of both components.
� Constructs with low loadings on all components are considered to be
insignificant for component definition.
In Table 26 constructs with the highest loadings are marked in bold. The constructs
that are attributed to the components and account for the largest variance repeat the
clusters identified through visual inspection of the plot. It is also seen that three
constructs that were not aligned to either of the components have similar loadings of
a middle level on both components.
• After lines in the plot are examined and groupings are recognized, the third
step is to identify a common meaning between constructs within each group.
At this phase a researcher examines construct meanings and looks if there is any
shared meaning. It is very important procedure of ‘naming components’ when
abstract desirable consequences are actually identified. However, in this study this
step is postponed and will be accomplished after reaching the synergy effect of
conducting both PCA and HCA. This will provide a more clear view on the meanings
of constructs in each component and their importance.
b) Perform HCA within each defined PCA component and structure individual
constructs based on the interrelationships between their meanings
Constructs within each dimension are clustered according to percentage of similarity
in their rating values. Thereby, each construct grouping is based on, firstly, common
pattern of variability and, secondly, close distances between rating absolute values.
7. Analysis of repertory grids
193
Dendrograms in Figures 19 and 20 illustrate ‘trees’ of both components identified by
PCA.
Figure 19: HCA: Component 1 including constructs 5, 9, 10
Links between constructs Links between elements
R - reversed PSS (%) PSS (%)10 5 R 100.0 1 (work) 8 (current) 100.09 5 R 97.2 3 (evening) 5 (previous) 100.04 1 R 94.4 1 (work) 9 (ideal) 95.87 4 R 94.4 2 (vacation) 3 (evening) 95.87 9 63.9 7 (alternative) 8 (current) 91.7
5 (previous) 6 (doesn't fit) 87.52 (vacation) 4 (in no way) 75.06 (doesn't fit) 9 (ideal) 54.2
Constructs Elements
Table 27: Links in the dendrogram of Figure 19
7. Analysis of repertory grids
194
Figure 20: HCA: Component 2 including constructs 5, 9, 10
Links between constructs Links between elements
R - reversed PSS (%) PSS (%)10 5 R 100.0 1 (work) 6 (doesn't fit) 100.09 5 R 97.2 1 (work) 8 (current) 96.42 3 94.4 2 (vacation) 3 (evening) 96.46 8 91.7 8 (current) 9 (ideal) 92.98 10 91.7 3 (evening) 6 (doesn't fit) 89.32 6 83.3 5 (previous) 7 (alternative) 89.3
2 (vacation) 5 (previous) 85.74 (in no way) 7 (alternative) 67.9
Constructs Elements
Table 28: Links in the dendrogram of Figure 20
Constructs 14.5, 14.9 and 14.10 are first included in both dendrograms as they have
similar loadings on both dimensions. In Figure 19 these three constructs compose
separate cluster, whereas, in Figure 20 they are embodied in the cluster with the rest
constructs aligned with second dimension. Thus, it is concluded that three constructs
have comparatively closer meanings with constructs from the second dimension and,
therefore, they are retained there in order to keep the completeness of information for
further interpretation.
c) Build an individual CMS
Table 29 presents the form developed in this study which comprises all the
attributions that were derived through consecutive conduction of PCA and HCA.
Columns of the table represent dimensions. In the upper part of rows constructs are
arranged according to their dimension ascription. Each construct is put down so that
7. Analysis of repertory grids
195
its number of interviewee (of RG), number of construct, category the construct is
ascribed, indicator of category repeat, and emergent and implicit poles are reflected.
Besides, preferred poles are written in bold.
Pole preference is identified by rating given on the ‘ideal’ element. In case of a
neutral rating on the ‘ideal’ element meaning of the pole which corresponds (i.e., is
on the same side of the dendrogram) to the preferred pole of the adjacent construct
is to be interpreted. If preferred pole of a construct appears on the other side
regarding preferred poles of adjacent constructs in cluster and percentage of
similarity between adjacent constructs is high, the researcher needs to pay more
attention to expressed meanings of each pole to interpret their similarity or difference
in meanings appropriately.
In this study such examples appeared very few times and the reason for
contradiction could be reasonably explained. For example, one interviewee (8th RG)
expressed the construct ‘high risk of criminal attempt because of good expensive car
brand’ versus ‘no risk of criminal attempt’. By rating the ‘ideal’ element on this
construct he gave ‘4’ meaning that he does not want to incur any criminal attempt.
However, by high risk he actually meant an ownership of a good expensive car which
is often a reason for criminal attempts. It is his personal perception and he has a full
right to judge so. Ratings of ‘ideal’ element on constructs reflecting prestigiousness
vary from ‘1’ to ‘3’ on cars including current and alternative as well as those that he
would prefer for work, vacation or leisure time. Hierarchical clustering arranged
respondent’s constructs so that pole ‘high risk of criminal attempt’ appeared to be on
the side among other preferred poles in favor of prestige and representativeness.
This can be interpreted as that the respondent prefers obtaining an expensive
prestigious car despite the fact that this kind of autos is incurred criminal attempts in
Russia. It would be ‘ideal’ for him when a good expensive car does not bring such
risk, although he realizes that it is not possible (in Russia) like it is not possible, for
instance, to purchase an ‘expensive’ auto at a cheap price.
In the bottom of Table 29 constructs are arranged according to percentage of match,
i.e., similarity between their ratings what is displayed in the right upper tree structure
of dendrograms (Figures 19, 20). This representation of data allows to see which
meanings are more or less similar. Constructs attributed to the first dimension all
match with each other between 90% and 100% (excluding 100%) where construct
7. Analysis of repertory grids
196
Table 29: Individual structure of construct meanings
1
dim
en
sio
n2
dim
en
sio
n
Loa
din
gs
Loa
din
gs
14
.1-2
.95
14
.21
.87
14
.31
.87
14
.4R
2
.77
14
.5R
-0.7
9
14
.61
.08
14
.7-2
.46
14
.80
.96
14
.90
.64
14
.10
0.7
9
M=
100
%1
00%
>M
>=
90
%9
0%
>M
>=
80%
80
%>
M>
=7
5%
M<
75
%M
=1
00%
10
0%
>M
>=
90%
90
%>
M>
=8
0%
80%
>M
>=
75
%M
<7
5%
14
.1
1
4.9
14
.4 (
R)
1
4.5
(R
)
14
.7
1
4.1
0
1
4.8
14
.6
14
.2
1
4.3
Con
stru
ctR
ever
sed
co
nst
ruct
sB
ipol
ar c
onst
ruct
sB
ipol
ar c
onst
ruct
s
14
.1 s
afe/
secu
re:
ap
pro
pri
ate
fo
r a
rmo
rin
g -
inex
ped
ien
cy o
f ar
mor
ing
14
.2 s
tylis
h:
be
au
tifu
l -
un
satis
fact
ory
des
ign
14
.3 n
ew m
odel
: n
ove
lty
of
mo
de
l -
old
mod
el
14
.4 r
epre
sen
tativ
e: fo
r le
isu
re t
ime
- fo
r b
usin
ess,
no
t fo
r
en
tert
ain
me
nt
14
.5 s
afe/
secu
re2
: w
orse
tec
hn
olog
ical
eq
uip
men
t in
sen
se o
f sa
fety
- i
nit
iall
y b
ett
er
tec
hn
olo
gic
al
eq
uip
me
nt
in s
en
se
of
sa
fety
14
.6 n
ew m
odel
2:
the
re i
s n
o l
ate
r m
od
el,
i.e
. te
ch
no
log
ica
lly
inn
ova
tive
- t
her
e is
a la
ter
mod
el
14
.7 r
epre
sen
tativ
e2:
wit
h d
rive
r -
for
self-
driv
ing
14
.8 c
omfo
rtab
le:
co
mfo
rt e
lec
tro
nic
s a
nd
te
ch
no
log
ies -
lim
ited
eq
uip
men
t w
ith c
omfo
rt e
lect
ron
ics
and
tec
hn
olog
ies
Mat
ches
be
twe
en c
ons
truc
ts (
%)
2 d
ime
nsio
n1
dim
en
sio
n
14
.9 f
ee
l c
on
fid
en
t a
bo
ut
this
ca
r--n
ot fe
el c
onfid
ent
abou
t th
is
car
14
.10
fe
el
sa
fe i
n t
his
ca
r--n
ot fe
el s
afe
in t
his
car
Va
r.:
64,3
6%
V
ar.
: 3
0,8
0%
7. Analysis of repertory grids
197
14.4 is reversed. This means that the absolute difference between their ratings is one
or two units and it is likely that constructs share very close meanings or are in some
sense complementary in respondent’s related perceptual structure. Hence, it is
important to look at the content of the expressed meanings.
Construct 14.7 ‘with driver – for self-driving’ and reversed construct 14.4 ‘for leisure
time – for business, not for entertainment’ basically both express personal perception
of a representative auto where he sits on a passenger seat and uses an auto for
business purposes also implying working process while being in the car (e.g.,
telephone calls, reading, working with notebook, etc.) and, therefore, requires
appropriate fulfillment. Although expressions ‘with driver’ and ‘for business purposes’
do not share the same meaning, they present complementary characteristics of
perceived representative, business auto.
At first sight construct 14.1 ‘appropriate for armoring – inexpediency of armoring’
does not share any common meaning with constructs 14.4 and 14.7. However,
appropriateness for armoring represents for the respondent another very important
characteristic of an owned auto. During the interview it seemed that appropriateness
for armoring is respondent’s prior requirement. This assumption is proved by the
produced quantitative analysis. The construct has got the highest loading (in absolute
value) on the first component, i.e., dimension.
Identifying similarities in the meanings, constructs of the first component seem to
relate to practical characteristics that are essential for the consumer’s car exploitation
taking into account his life circumstances and first purpose for the car use. According
to constructs answering the question how a car is used and ratings of ‘current’ and
‘ideal’ elements, the main purpose for car exploitation is business (‘for business, not
for entertainment’). An auto does not represent any working instrument (as, for
example, for taxi driver) by means of which the respondent earns money or receives
profit. The auto is to be used to provide proper conditions for its owner to lead his
business and spend time during the rides differently than when self driving a car
(‘with driver’). The respondent prefers autos that are appropriate for armoring and
estimates the characteristic as very important (rating value ‘1’ on ‘ideal’ and ‘current’
elements). Ratings of this construct vary in the same way as ratings on other two
constructs which refer to the preferences regarding working exploitation. Business is
the main purpose for the car exploitation. Therefore, it can be assumed that
7. Analysis of repertory grids
198
‘appropriateness for armoring’ is also perceived as an important requirement referred
to the main purpose of vehicle exploitation. Considering ratings on different elements,
it appears that those autos that respondent considers as representative and
appropriate for business exploitation are also considered as appropriate for armoring.
Based on this interpretation the resulting name of the first dimension which reflects
the shared meaning of included constructs and their preferred poles is formulated as
“Appropriate for my business use” and can also be precised as “Appropriate for my
business use: armored and representative”.
Here, it is useful to describe some details revealed during the interview in order to
provide a reader with a broader picture. The respondent is an entrepreneur who
leads several large businesses in Russia. A couple years ago he was shot five times
but survived. The reason behind armed assault was related to his business activities.
For this reason this person does not feel in safety so far and always moves in
armored auto with escort armored vehicles and thirteen bodyguards. These
circumstances and the fact that respondent still leads his businesses makes
construct pole ‘appropriateness for armoring’ likely to be called a requirement rather
than a preference.
Returning to Table 29, constructs attributed to the second dimension also share a
common meaning. First of all, indicative meanings of constructs which have high
loadings on the second component should be considered. Constructs 14.8 ‘comfort
electronics and technologies – limited equipment with comfort electronics and
technologies’ and 14.6 ‘there is no later model, i.e., technologically innovative – there
is already a later model’ match at 92%. They both relate to the technological
achievements which provide better comfort as well as general exploitation features.
According to the respondent’s perception the newer a model is the more innovative
technologies it possesses.
Constructs 14.2 ‘beautiful - unsatisfactory design’ and 14.3 ‘novelty of model - old
model’ match at 94%. During laddering down specification the respondent mentioned
that beautiful design is only provided by new models which are not yet often seen on
the streets. He basically perceives auto’s beauty depending on its uniqueness.
Quantitative analysis proves this judgement and shows this strong interrelation in
7. Analysis of repertory grids
199
meanings in form of two adjacent constructs which highly match and have high
loadings on the second important dimension.
Besides, constructs 14.6 and 14.2 match at 83% in rating values. The double line
between constructs 14.6 and 14.2 indicates that their matching score lies in the
column where percentage of match is between 80% and 90%, while they both match
with adjacent constructs, 14.8 and 14.3 correspondingly, at score between 90% and
100%.
Four constructs which have higher loadings on the second dimension appear to
share respondents attitude and preference to novelty of auto model which, in turn,
means (for the respondent) representation of most updated technologies in
automotive industry. As these four constructs appeared to be indicative for
determining the second dimension, the resulting name of second dimension is
“Novelty of a model with recent technological achievements”.
Three constructs 14.5, 14.9 and 14.10 form a cluster within the second dimension.
Overall assessment construct 14.10 ‘feel safe in this car – not feel safe in this car’
matches at 100% with reversed elicited construct 14.5 ‘worse technological
equipment in sense of safety - initially better technological equipment in sense of
safety’. It can be interpreted that for respondent a question of safety when talking
about car features relates to the technical equipment and level of quality which can
increase security and protection measures. If the respondent felt safe in an auto
which meets certain attribute requirements he probably would not need other security
measures such as bodyguards and escort. The person feels safe only when all
measures are presented in complex which, in turn, involves a maximum technical
quality and equipment of the used vehicle, besides measures not related to the auto
itself. For this reason the respondent when talking about feeling safe and confident
about an auto, is likely to consider its provided technical quality and features enabling
its appropriateness for armoring. Therefore, construct loadings laying between two
dimensions show that overall feeling of confidence and safety when talking about an
auto depends more or less equally on both dimensions. However, meanings of three
constructs are closer to the meanings expressed in constructs of the second
dimension.
7. Analysis of repertory grids
200
The resulting schema of individual CMS is presented in Table 30. The schema
accumulates all the elicited data which was analyzed and structured according to
different levels of abstraction of respondent’s personal meanings.
In the bottom of the hierarchy (on the left of the schema) product characteristics are
put into a tabular form reflecting the extent of similarity between personal meanings.
Constructs are arranged according to the dimensions and clusters they have been
attributed to. It repeats the Table 29 but dimensions are located one on another
instead of side by side.
Product attributes reflect categories that were assigned to the elicited constructs.
They are ordered according to their relation to one of the dimensions. Thus, it can be
seen which attributes are sought to satisfy each of identified abstract expected
consequences.
Specific expected consequences which represent too-specific motives include
individual expressions revealed during laddering technique procedure carried out
with mostly preferred constructs according to their given ratings on the ‘ideal’
element. Resulting expected consequences, actually, reflect desirable consequences
as they were derived from the preferred construct poles. It was proved that people do
not necessarily provide ladders that link product characteristics with personal values
through all levels of abstraction, especially, in terms of six level chain proposed by
Olson and Reynolds (1983, p. 81). For example, ladders from constructs 14.5 and
14.6 have only one meaning linking product attribute with personal value. Actually, it
does not mean that the respondent could not produce and explain details
corresponding to each level of abstraction. In contrast to clinical purposes,
application of laddering in marketing research targets relatively adequate people who
usually can clearly express their thoughts especially about ‘obvious’ things. During
the current study it was found that asking for more details often does not provide
additional information, however, can evoke respondent’s irritation and, thereby,
endanger the remained interviewing procedure. It depends on the researched topic,
type and complexity of product, sample and conditions of the survey that makes
interviewer either feel comfortable and free to ask for more details even regarding
obvious things or be cautious and realize the situation going on to adjust his
scrupulosity.
7. Analysis of repertory grids
201
Table 30: Individual schema of a consumer motivational sphere
M =
10
0%
10
0%
> M
>=
90
%9
0%
> M
>=
80
%8
0%
> M
>=
75
%M
< 7
5%
14
.1 s
afe/
secu
re:
app
rop
riate
fo
r ar
mor
ing
Security
(14
.1)
Pro
vid
e m
axim
um
se
curit
y m
easu
res
Sec
urit
y
14
.4 r
epre
sen
tativ
e: fo
r b
usi
nes
s, n
ot fo
r en
tert
ain
men
tRepresentativeness
Be
con
fiden
t ab
out
surr
oun
din
g
peo
ple
an
d t
hin
gs
14
.7 r
epre
sen
tativ
e2:
not
for
self-
driv
ing
Sta
y al
ive
as lo
ng
as
pos
sib
le
14
.9 f
ee
l co
nfid
en
t a
bo
ut
this
ca
rNovelty
14
.5 s
afe/
secu
re2
: in
itial
ly
bet
ter
tech
nol
ogic
al e
qu
ipm
ent
in s
ense
of s
afet
yDesign
(14
.4)
Fee
l com
fort
able
as
a p
asse
ng
er
14
.10
fe
el sa
fe in
th
is c
ar
Comfort
Sav
e tim
e fo
r w
orki
ng
pu
rpos
es
14
.8 c
omfo
rtab
le:
com
fort
el
ectr
onic
s an
d t
ech
nol
ogie
sSecurity
14
.6 n
ew m
odel
2:
new
mea
ns
tech
nol
ogic
ally
inn
ovat
ive
(14
.5)
Fee
l saf
er d
ue
to la
st h
i-te
ch a
chie
vem
ents
14
.2 s
tylis
h:
bea
utif
ul
14
.3 n
ew m
odel
: n
ovel
ty(1
4.6
) E
njo
y rid
ing
du
e to
last
hi-
tech
ach
ieve
men
ts
Bu
sin
ess
With
driv
er
Sec
urit
y m
easu
res
Pro
du
ct
ch
ara
cte
risti
cs
Ach
ieve
men
t
Pro
du
ct
att
rib
ute
sS
pe
cif
ic e
xp
ec
ted
co
nse
qu
en
ce
s
Ab
str
ac
t e
xp
ec
ted
co
nse
qu
en
ce
s
Dim
. 1
: Appropriate for
my business use
(arm
ore
d a
nd
rep
rese
nta
tive
)
Inc
en
tive
s
Dim
. 2
: Novelty of a
model with most recent
technological
achievements
Va
lue
s
Pro
du
ct
ch
ara
cte
risti
cs
Sp
ec
ific
co
nse
qu
en
ce
sV
alu
es
Pe
rso
nal c
onst
ruct
s L
ad
deri
ng
UP
Lad
der
ing
UP
WH
ICH
Re
sist
an
ce-t
o-C
han
ge
Pro
du
ct
att
rib
ute
sA
bs
tra
ct
co
ns
eq
ue
nce
s
Co
nte
nt
Ana
lysi
sP
CA
Dim
en
sio
ns
WH
AT
Clu
ste
rs
Sit.
and
Env
. in
cen
tives
HO
W+
Ide
al r
atin
gs
Analysis driven: relatively objective
Person driven: subjective
7. Analysis of repertory grids
202
Abstract expected consequences representing individual motives of the middle range
consist of defined dimensions. They aggregate individual meanings which are not
only put into words but also expressed in individual estimations. These abstract
consequences are more general and can be compared with preferred abstract
consequences of other individuals, although they are still expressed in terms of
product features.
The incentives are revealed from the content and preferred construct poles of
constructs which explain how and in which circumstances the auto is used.
Values represent motives at the highest level of abstraction that are too-general and
do not directly reflect product preferences and requirements which help come closer
or reach the value state. Nevertheless, values represent the final states which person
is trying to achieve by living his life and doing as well as consuming what he does.
Values associated with investigated product consumption can be partly or fully
attained through benefits identified on the lower levels of abstraction.
7.3.3. Systematization of elements perceptions
a) Analyze matches between elements perceptions with respect to the revealed
construct dimensions, i.e., abstract consequences
Elements, i.e., autos named for each supplied category can also be analyzed and
differences in perceptions can be revealed by the suggested methodology.
Figure 21 displays the PCA plot reproducing only the elements. PCA of the grid data
rely on one plot reproducing two first principal components which account for 95% of
the variance in the data.
Looking at the distances between the elements, elements 2 (vacation), 3 (evening)
and 6 (doesn’t fit) are plotted close to each other indicating that they received similar
ratings.
The same similarity can be seen within another group including elements 1(work), 8
(current) and 9 (ideal).
Element 4 (in no way) is plotted far apart from other elements which tends to show
rather different ratings given on this element.
7. Analysis of repertory grids
203
Figure 21: Relationships between elements
Positions of elements 5 (previous) and 7 (alternative) provide less evidence of their
interrelations to other elements, although with respect to the first dimension element
5 (previous) seems to be similar with the first mentioned group of elements and
element 7 (alternative) is similar to the second group of elements. Positions of
elements with respect to the second dimension show strong similarities in ratings
between both groups mentioned above, however, when looking at elements 9 (ideal)
and 6 (doesn’t fit) they are plotted already far enough from each other.
This exemplifies how visible relations can be recognized from the plot, however,
more precise and comprehensive interpretation of element relationships requires
additional procedures. Moreover, it becomes harder to interpret and easier to
misinterpret the relationships between elements when more than two components
appear significant for data analysis.
Relationships between cars based on the constructs, i.e., perceptions are presented
in dendrogram of Figure 22. Table 31 shows links between elements, i.e., the
percentage of match between adjacent elements. Dendrogram displays the result of
HCA conducted with all respondent’s constructs. An advantage of such
representation offered by the computer package is that clustering of both constructs
7. Analysis of repertory grids
204
and elements is provided simultaneously and both ‘trees’ are displayed in the
analysis outcome.
Figure 22: HCA producing element and construct dendrograms
PSS (%)1 (work) 8 (current) 97.52 (vacation) 3 (evening) 95.03 (evening) 6 (doesn't fit) 92.58 (current) 9 (ideal) 92.52 (vacation) 5 (previous) 87.57 (alternative) 9 (ideal) 82.51 (work) 6 (doesn't fit) 70.04 (in no way) 5 (previous) 65.0
Elements
Table 31: Links between adjacent elements in the dendrogram of Figure 22
Examining the element dendrogram presented in Figure 22 and its supportive Table
31, two major clusters are sorted out by 75% cutoff level. First cluster includes
elements 1, 8, 9 and 7, second cluster includes 2, 3, 6 and 5 elements; and there is
also one “isolate” element 4 which stands out. By using the term “isolate” it should be
borne in mind that although this element does not match with any other elements at
75% or higher, it still joins a general cluster by the 65% level (McKnight, 2000).
This illustration replicates what was assumed from the first view on PCA plot
considering elements interrelations.
7. Analysis of repertory grids
205
Additional inferences can be made when elements are clustered with respect to
constructs within each dimension. Figures 23 and 24 show dendrograms displayed
for each PCA dimension.
Figure 23: HCA conducted within the first PCA dimension
Figure 24: HCA conducted within the second PCA dimension
In Table 32 clusters are summarized and elements are arranged according to their
matching scores. The algorithm is similar to the one shown in Table 29 but this time
examining elements rather than constructs.
• In the first column which reflects clusters considering all the constructs the first
cluster consists of elements 1, 7, 8 and 9.
7. Analysis of repertory grids
206
Elements 1(work) and 8(current) have the highest match of 97%. Elements 8
(current) and 9 (ideal) have also relatively high match of 92%. In turn,
elements 9(ideal) and 7(alternative) match at 82%.
It can be assumed that the respondent when choosing his current auto was
looking for a one that is mostly appropriate for his working purposes. He draws
his ideal car very close to the current one. Besides, he considered as
alternative purchasing an auto which also was alike with his ideal image.
Second cluster of the first column illustrates how similar the ratings between
elements 6 (doesn’t fit), 3 (evening), 2 (vacation) and 5 (previous) are. In other
words, autos that are more suitable for free time and an auto that this person
really likes but it doesn’t fit to his surrounding, environment or life
circumstances are perceived relatively similar to each other and different from
those that are practical and to be used for work.
• In the second column it is shown how elements are similar on the constructs
ascribed to the first dimension. There are no cardinal changes in element
groupings. Moreover, it can be seen that elements 8, 1 and 7 as well as 6, 3
and 5 match now at 100% correspondingly. This is not surprising as first
dimension “Appropriate for my business use” accounts for the largest
percentage of the variance in the grid data and, basically, is most
determinative in the general positions and distances between the elements.
• In the third column elements grouping appears different from those in previous
two columns. There is only one major cluster and an isolate element 4 (in no
way). Here, elements are compared regarding constructs attributed to the
second dimension “Novelty of a model with recent technological
achievements”. The respondent named autos that are relatively new and
correspond to recent technological achievements in automotive industry (see
in the bottom of Table 32). For this reason the ratings on these constructs do
not differ so much to build different clusters. However, the extent of similarity
between elements is different.
7. Analysis of repertory grids
207
Table 32: Interrelationships between element perceptions
PS
S (
%)
PS
S (
%)
PS
S (
%)
1 (
wor
k)8
(cu
rren
t)9
7.5
1 (
wor
k)7
(al
tern
ativ
e)1
00
.01
(w
ork)
6 (
doe
sn't
fit)
10
0.0
2 (
vaca
tion
)3
(ev
enin
g)
95
.01
(w
ork)
8 (
curr
ent)
10
0.0
1 (
wor
k)8
(cu
rren
t)9
6.4
3 (
even
ing
)6
(d
oesn
't fit
)9
2.5
3 (
even
ing
)5
(p
revi
ous)
10
0.0
2 (
vaca
tion
)3
(ev
enin
g)
96
.48
(cu
rren
t)9
(id
eal)
92
.53
(ev
enin
g)
6 (
doe
sn't
fit)
10
0.0
8 (
curr
ent)
9 (
idea
l)9
2.9
2 (
vaca
tion
)5
(p
revi
ous)
87
.52
(va
catio
n)
5 (
pre
viou
s)9
1.7
3 (
even
ing
)6
(d
oesn
't fit
)8
9.3
7 (
alte
rnat
ive)
9 (
idea
l)8
2.5
7 (
alte
rnat
ive)
9 (
idea
l)9
1.7
5 (
pre
viou
s)7
(al
tern
ativ
e)8
9.3
1 (
wor
k)6
(d
oesn
't fit
)7
0.0
2 (
vaca
tion
)4
(in
no
way
)7
5.0
2 (
vaca
tion
)5
(p
revi
ous)
85
.74
(in
no
way
)5
(p
revi
ous)
65
.04
(in
no
way
)9
(id
eal)
41
.74
(in
no
way
)7
(al
tern
ativ
e)6
7.9
Gen
eral
mat
ch
1 D
im2
Dim
Inte
rval
of
PS
SM
= 1
00%
100%
> M
≥ 9
0%
90%
> M
≥ 8
0%
80%
> M
≥ 7
5%
M <
75
%M
= 1
00%
100%
> M
≥ 9
0%
90%
> M
≥ 8
0%
80%
> M
≥ 7
5%
M <
75
%M
= 1
00%
10
0%
> M
≥ 9
0%
90
% >
M ≥
80%
80
% >
M ≥
75%
M <
75%
Clu
ster
17
(al
tern
ativ
e)8
(cu
rren
t)9
(id
eal)
9 (
idea
l)1
(w
ork)
8 (
curr
ent)
8 (
curr
ent)
7 (
alte
rnat
ive)
1 (
wor
k)
1 (
wor
k)9
(id
eal)
6 (
doe
sn't
fit)
3 (
even
ing
)
2 (
vaca
tion
)
5 (
pre
viou
s)
7 (
alte
rnat
ive)
4 (
in n
o w
ay)
Clu
ster
26
(d
oesn
't fit
)6
(d
oesn
't fit
)
3 (
even
ing
)3
(ev
enin
g)
2 (
vaca
tion
)5
(p
revi
ous)
5 (
pre
viou
s)2
(va
catio
n)
4 (
in n
o w
ay)
4 (
in n
o w
ay)
(1)w
ork
(2)v
acat
ion
(3)e
ven
ing
(4)in
no
way
(5)p
revi
ous
(6)d
oesn
't fit
(7)a
ltern
ativ
e(8
)cu
rren
t
Mer
ced
es
S2
21
Gu
ard
Au
di Q
7P
orsc
he
Car
rera
Jag
uar
(an
y)M
erce
des
CL
KA
ston
Mar
tin
DB
9 C
oup
eM
erce
des
S
22
0A
ud
i S8
A
rmor
edM
erce
des
C
14
0 G
uar
dM
erce
des
GW
G
uar
d
E
lem
ents
E
lem
ents
E
lem
ents
Nam
ed a
uto
s
Oth
er o
wn
ed a
uto
s
7. Analysis of repertory grids
208
Element 6 (doesn’t fit) is closer to elements 1 (work), 8 (current) and 9 (ideal)
rather than to elements 3, 2 and 5 as in previous columns (the double line
stands for the match score between 80%-90%). Moreover, element 6 matches
element 1 at 100%. It indicates that the car named as an auto that ‘I really like
but it doesn’t fit’ mostly meets all individual preferences regarding
technological equipment, design and comfort. Therefore, the main reason why
this auto would not probably be purchased by the respondent lays in lack of
correspondence to the preferences presented in the first dimension. During
the interview the respondent has mentioned that he is delighted with Aston
Martin DB9 Coupe, auto that he named as element 6 (doesn’t fit), and he
would love to have it. However, his current life circumstances prevent this
purchase.
Besides, element 7 (alternative) is now positioned relatively far from elements
1 (work), 8 (current) and, especially, from 9 (ideal). Respondent didn’t choose
the ‘alternative’ auto over the ‘current’ one because its technological and
design features did not enough meet his preferences, although this auto fully
corresponded to most important preferences related to its appropriateness for
security measures and representativeness.
Such element comparison is very useful when differences between elements carry
special meaning for researcher and, especially, for the manufacturer. It allows to
draw inferences regarding what needs to be changed or improved to make a certain
product be perceived closer to the ‘ideal’ or at least ‘current’ product. It gives a clue
along which constructs and in which direction product improvement or its
repositioning should occur to make ratings of the improved product move towards
customer’s ‘ideal’ image.
Representation of dimensions and clusters as in Table 32 permits to compare
element interrelationships and to recognize how respondent perceives each element
with respect to his perceptual dimensions, i.e., abstract expected consequences he
seeks for. It can be useful for managers who want to reveal disadvantages of their
product or brand from perspective of a consumer.
To sum up, the suggested methodology for analyzing individual grid data by means
of subsequent conduction of PCA and HCA permits to reveal relationships,
7. Analysis of repertory grids
209
similarities and differences between individual perceptions and between products
within the product category which often can be already assumed from observational
inspection of resulting graphs.
However, inferences made from observational inspection often bear doubt and
should be supported by other evidence to be persuasive. Otherwise, there is often a
risk to be claimed in making conjectures. The methodology implies both qualitative
and quantitative analyzing procedures which on their own represent solid and
recognized techniques and the findings made from their application are considered
reliable. It has also proved to reflect all assumptions and notices recognized during
interviewing. Therefore, it is believed that if interviewer and analyst appear to be two
different people the proper inferences could still be drawn revealing different aspects
of consumer’s life. The variety of aspects includes life circumstances, environment
and some features of character which influence the way a product is exploited. Last
but not least, personal perceptions, preferences and choice depending on the whole
complex of internal and external incentives can be drawn.
7.4. Segmentation based on the comparison of the findings from the single grids analyses
All twenty grids were analyzed according to the suggested method and for all
respondents CMS schemas were produced similarly to the provided example of 14th
respondent’s grid. Such representation of individual meanings associated with the
topic allows researcher a comparison of respondents on different levels of
abstraction depending on the aim in research and application of findings.
In this study an attempt to identify groups of customers is pursued to differentiate
them on higher levels of abstraction in order to appeal to the customer’s motives in
positioning products which are very similar to the competitors’ products in sense of
technological, functional and quality performance and the distinction between
products and brands is reached mostly on the emotional component of a high
involvement product.
The selection of appropriate segmentation type depends on the purpose of the
research, e.g., understanding markets, product positioning or new product
development (Wind, 1978; Botschen et al., 1999). When the main purpose is to use
findings in effective new product development differences of rather specific meanings
7. Analysis of repertory grids
210
are of an interest and segmentation that works well statistically based on product
characteristics and attributes as well as psychographic and general attitudinal
approaches is reasonable (Young et al., 1978). However, such approaches to
segmentation are less useful when the research is aimed at deriving effective
positioning and marketing strategy. For these purposes segmentation on higher
levels of abstraction such as benefits and values is preferred (Wind, 1978).
Human values as a basis for market segmentation have attracted much attention
from consumer researchers driven by the view that values are more closely related to
behavior than are personality traits (Kamakura and Novak, 1992). Besides, values
are less numerous, more central and stable determinants of behavior and attitudes
(Vallette-Florence, 1988 cited in Kamakura and Novak, 1992).
Conducted value-segmentation studies applied different approaches, some have
relied on a single, and most highly-ranked value from each individual to define
segments, whereas, other suggested segmentation based on value systems shared
by groups of customers (Kahle et al., 1986; Novak and MacEvoy, 1990; Kamakura
and Novak, 1992). Value construct is used both as criteria for segmentation and as a
way to enrich the description of segments identified through other criteria.
Researchers agree that segmentation based on higher levels of abstraction such as
values is more meaningful and stable (Gutman, 1982; Reynolds and Gutman, 1988).
However, it is also argued that any situation in life including product purchasing or
consumption is likely to activate several values rather than only one from a whole set
of personal values (Rokeach, 1973) and it is reasonable to consider the priority of
values associated with particular situation or, in this context, with particular
purchasing decision (Kamakura and Novak, 1992).
Value system of a person consists of values that are most central and general
determinants of consumer behavior and, therefore, fairly remote from each particular
decision, which can also be influenced by many other more immediate but less stable
situational and environmental affects. That is why segmentation on particular product
market based on general value-systems doesn’t seem reliable and meaningful. Other
more immediate and closer influences such as product attributes, product
consequences and consumer preferences must be taken into account when defining
segments for a particular market (Kamakura and Novak, 1992).
7. Analysis of repertory grids
211
Maslow’s (1954) analysis of “deficiency” versus “growth” needs suggests that
deficiency needs (e.g., health, safety) are inactive or functionally absent once the gap
between desired standard or satisfaction and the person’s perceived current state is
eliminated. Satisfaction of growth needs (e.g., self-actualization, curiosity), on the
contrast, is pursued even after high levels of satisfaction are reached and there is no
stable external standard that when is attained turns the need into inactive. From
psychological perspective behavior patterns aimed at satisfying growth needs are
different from those aimed at satisfying deficiency needs and, hence, people are
likely to attribute high importance to different values depending on the pursued
behavior pattern (Bilsky and Schwartz, 1994). This conception can be helpful in
marketing implementation when deciding from which perspective communications
and positioning are to be provided.
In this study segments are formed according to individual differences in their
consumer behavior patterns. Values have long been suggested as means for
understanding consumers’ underlying motivations and groups identified accordingly
are more likely to engage in a common pattern of attitudes, beliefs and behavior
(Kamakura and Novak, 1992).
Botschen and his colleagues (1999) provided a study aimed at clarifying the
distinction between attributes and benefits sought and suggesting modified laddering
technique to be applied to form benefit segments. It is not an aim of this work to
prove the appropriateness of a certain segmentation approach in different situations.
Relying on the statements and inferences made by authors it is agreed that the MEC
concept is ideally suited to form market segments according to different levels of
abstraction, such as attributes, desirable consequences (benefits), values or linkages
between the meanings. In this work it is demonstrated how segmentation can be
performed based on the systemized structure of personal meanings represented in
individual motivational schemas.
Expected consequences and values being motives of different ranges represent
actual drivers for purchase and choice of certain set of product attributes. Because
same product attribute may lead to different consequences and single consequence
may be reached by combination of attributes (Pieters, 1993), segments based on
attributes may differ from segments based on consequences sought and derived
segmentation may appear grossly misleading.
7. Analysis of repertory grids
212
In this study individual analysis of respondents’ motivational schemas allowed to
conduct a two-stage segmentation considering at first stage the most general
motives, i.e., values and on the second stage the revealed abstract consequences
which represent the motives of the middle range (Table 33).
The first stage of segmentation considers values which are derived by laddering
technique related to the subjective part of the proposed integrated method. To
remind the procedure, from all elicited constructs that were given rating ‘1’ or ‘5’
laddering technique was applied and values associated with the product category
were derived. Elicited personal values were then prioritized by respondent according
to the resistance-to-change technique described in section 4.5.4.2.
Respondents are compared in terms of frequency and priority of elicited values. It
was found that most common associated values named by respondents were ‘self-
actualization, achievement’ and ‘pleasure and enjoyment of life’. All respondents
named either both values or at least one of them.
For this reason, at first, participants are distinguished according to whether their first
priority value implies ‘achievement and self-actualization’ or whether it implies
‘pleasure and enjoyment of life’.
Thus two main groups are identified according their preference between these two
general values.
• The first group includes respondents who prefer ‘self-actualization and
achievement’ over ‘pleasure and enjoyment of life’ or who did not mention the
second at all. Among these respondents two have also named value “power
and social recognition” and prioritized it over the ‘self-actualization and
achievement’. Power and social recognition imply social influence and
achievements recognized not only by individual himself but also by social
surrounding. Power and achievement are compatible in that they both
emphasize superiority and esteem (Schwartz, 1992 cited in Paulssen, 1999).
• The second group includes respondents who either prefer ‘pleasure and
enjoyment of life’ over ‘self-actualization and achievement’ or named only
‘pleasure and enjoyment of life’ as an only value or among other values
excluding ‘self-actualization and achievement’ associated with car use.
7. Analysis of repertory grids
213
On the second stage commonalities among important individual abstract
consequences are explored. For this purpose from each individual CMS attributes,
i.e. categories the constructs are ascribed to, are extracted that appeared to be
indicative for dimensions on which ‘ideal’ element has got highest loading(s).
Indicative attributes are considered those that have high loadings on the revealed
dimensions of each single grid. Dimensions that are significant by variance and on
which ‘ideal’ element has highest loading(s) are considered in segmentation.
Table provided in Appendix 5 shows respondents’ most important attributes (on the
left hand side) and prioritized personal values derived from individual ladders (in the
right part). Respondents are divided into two groups according to the presence of
attribute ‘prestige’ among three most preferred attributes.
Five groups are identified from the investigated sample which description is
presented in the form of matrix shown in Table 33. Testing of representativeness and
validity of segmentation is not pursued in this study because the aim is to derive
common patterns of consumer behavior in the surveyed sample rather than to claim
to be a statistically representative (see 2.6.2.3).
Horizontally the groups are divided into two types of customers according to their
priority regarding aspiration for self-actualization and professional success or
aspiration for pleasure and enjoyment of life ensured by professional success in
terms of financial and status security. Thus, groups and divided into ‘careerists’ and
‘hedonists’.
Groups of careerists refer to people for whom “success and self-realization means
pleasure and enjoyment of life”, whereas groups of hedonists refer to people for
whom “self-realization serves as means for pleasure and enjoyment of life”. In the
first definition the term ‘mean(s)’ is used as a verb to emphasize that business or
professional success and self-realization make a person feel enjoyment of life. In the
second definition the term ‘means’ is used as a noun to emphasize that self-
realization in terms of professional success does not bring enjoyment itself but
serves as a resource to obtain status and to afford things that make a person enjoy
his life.
Vertically the groups are divided into two types according to social orientation and
conspicuity of product consumption. The first type represents people who are self-
7. Analysis of repertory grids
214
oriented regarding products consumption and consume premium goods to benefit,
first of all, from their quality and functional superiority. Attributes like image and
design are also important for these people but for own enjoyment rather than to make
sure to be perceived properly by others. The second type represents people whose
consumption of premium goods is more socially oriented and they pursue social
prestige and acknowledgement by such a conspicuous consumption. Thus, vertically
groups are divided into ‘introverts’ and ‘extroverts’.
Prioritized values Common important attributes Prioritized values Common important attributes
Self-actualization, Achievement Representativeness Pleasure and enjoyment of life Comfort
Pleasure, enjoyment of life Comfort Technical quality
Technical quality Cross-country ability
Prioritized values Common important attributes Prioritized values Common important attributes
Power, Social Recognition Prestige Pleasure and enjoyment of life Prestige
Self-actualization, Achievement Speed/Power Self-actualization, Achievement Design
Novelty / Comfort
Prioritized values Common important attributes
Pleasure and enjoyment of life Prestige
Affiliation, Recognition Speed/Power
Design
Se
lf-or
ien
ted
Careerist-introvert: "I know how successful I am" Hedonist-introvert: "I enjoy my life"
Orientation: professional success Orientation: enjoyment of life
Consume prestigious brands to reflect achieved successConsume brands that proved to be qualitative and
appropriate for personalized use
Orientation: be unique among others
Consume prestigious brands to demonstrate themselves
Success and self-realization means pleasure and enjoyment of life
Self-realization as means for pleasure and enjoyment of life
So
cial
ly o
rient
ed
Careerist-extrovert: "Others know how successful I
am"
Hedonistic careerist-extrovert: "Others should know
how successful I am"
Orientation: strong leadership Orientation: become successful and recognized
Consume prestigious brands to demonstrate success and
power
Consume prestigious brands to demonstrate achieved
success
Hedonist-extrovert: "Others notice me"
Table 33: Customer segmentation
The terms ‘introvert’ and ‘extrovert’ are used to define people with an ‘introverted
type of consumption’ and an ‘extroverted type of consumption’ respectively. In this
work an introverted type of consumption implies a consumer behavior oriented first of
all at person’s own opinion and judgement. In contrast, an extroverted type of
consumption implies a consumer behavior strongly influenced by stereotypes,
attitudes and perceptions generated by the society. Thus a consumer with an
extroverted type of consumption is predominantly concerned with other’s opinion and
7. Analysis of repertory grids
215
tends to make his choice towards products or brands that are first of all recognizable
by others. Two types of behavior represent opposite dispositions, however,
consumers fluctuate in their behavior and, therefore, their belonging to one of the
types is relative and based actually on the tendency towards one of the behavior
types in particular situation.
Customers who belong to introverts do not have the attribute ‘prestige’ among three
most important attributes. For these people the social function of premium brand
autos provides first of all a sense of life they possess or aspire to. In contrast, three
remaining groups related to extroverts all have a common most important attribute
‘prestige’ among first three. For these people the social function of premium brand
autos provides social prestige and recognition. Next each segment is described in
details.
Careerist-introvert: "I know how successful I am"
This group includes respondents oriented at professional success. They are
motivated by achievement and self-realization and exhibit their professional success.
This people refer to self-oriented individuals who put own opinion and self-esteem
over the social opinion and social recognition.
Members of this group prefer premium brand cars first of all to profit from their
technological and aesthetical advantages. In the first place they value the practical
use and an appropriateness of a car for their individual way of its exploitation.
Through a social prestige provided by the social function of premium brand these
people reflect their maturity and achieved success. They are likely to seek for a
symbol of well-being and wealth, taste and connoisseurship. Therefore, they are
especially interested in the representativeness of a car which reflects their social
status of mature, successful and respectable people.
Common most important attributes that contribute to the perceived product utility are
comfort, representativeness and technical quality of an auto.
Careerist-extrovert: “Others know how successful I am”
This group comprises people oriented towards strong leadership. As a car
exploitation associated value they name ‘self-actualization and achievement’ but in
7. Analysis of repertory grids
216
the first place they put ‘power’ as a prior value. They are motivated not only by
achievement but they perceive self-realization in attaining social status that enables
strong influence, authority, control and dominance over people and resources.
Although for these people own opinion is primary factor, the desire to influence other
people requires social recognition and preserving public image. Their buying
behavior regarding remarkable products is likely to be socially-oriented by which they
exhibit their strength and power.
For this group social function of premium or luxury brands comes to the foreground
although they also require sufficient level of functional fulfillment. To demonstrate
success and power they seek for attribute ‘prestige’ which reflects their status and
attribute ‘speed/power’ that underlines strength and risk-loving. As already discussed
rather functional attribute ‘speed/power’ significantly contribute to the social function
of car brand in customers’ perception in Russia.
Hedonist-introvert: “I enjoy my life”
This group includes people oriented at enjoyment of life. They are motivated by life
gratification and tend to enjoy an active life full of excitement, entertainment, and
socializing. For this group achievement and professional success present means to
attain sufficient social level when they can afford things they need to enjoy time and
life in general. Relative to the previous two groups, members of this group are likely
to spend more time for leisure and fun-filled, social activities. They also spend more
time with family and friends and are likely to try new things in life.
Premium brand car is valued by these people first of all due to its technical and
functional advantages over volume brands. They prefer products that proved to be
highly qualitative and appropriate for their personal way of exploitation.
Members of this group are self-oriented in terms of car perception. Social perception
and attitude play secondary role in their purchasing decision. Common most
important attributes that contribute to the perceived product utility are ‘comfort’ and
‘technical quality’. Also such attributes as ‘size’ and ‘cross-country ability’ appeared
important for majority of members of the group according to their individual favors for
outdoor activities.
7. Analysis of repertory grids
217
Hedonistic careerist-extrovert: “Others should know how successful I am”
Members of this group are oriented at achieving success and, thereby, are being
recognized by others. They are motivated by social recognition and respect because
of achieved success. If compare with the first group: “I know how successful I am”,
members of the first group are also oriented at achievement, however, they are
interested first of all in professional success i.e., to make a career in particular field,
to master every aspect of their professional activities.
In contrast, members of this group want to gain success to prove others what they
can. It may not be so important in what field and how interesting their work is but
what is important is an attained social status through earned wealth. At the end this
people are motivated by enjoyment and pleasure in life. They believe that to reach
this state they need to earn enough money and show or prove others that they are
successful. Thereby they build up a sense of self-esteem which is necessary for
further full enjoyment of life. This group represents some kind of a gold middle
between full hedonists and full careerists for whom pleasure and enjoyment of life is
a prior value (associated with product consumption), however, not possible without
achievement and success.
Because they pursue to let others recognize their achievements, social function of
premium brand car comes to the forefront. It can be expected that these people tend
to change cars and purchase new ones with stronger image and brighter appearance
every time they feel they have reached next level within society and in well-being.
Most important common attributes for them are ‘prestige’ and ‘design’ and often also
considered ‘novelty’ and ‘comfort’ which reflect pacing in step with trend and
progress and enjoying correspondingly.
Hedonist-extrovert: “Others notice me (and my success)”
Members of this group present highly hedonistic group oriented at becoming noted
by others. They are extroverts motivated by pleasure and enjoyment within social
environment. This people usually have already attained a status which they consider
sufficient for their wished state of well-being and rather maintain this position than
pursue new challenges. They enjoy life and like being recognized by others and
being affiliated with subjectively perceived highest strata of society.
7. Analysis of repertory grids
218
This group is tended to purchase cars which already belong to high-end or ultra-
luxury cars as they impress first of all with their sophistication and excess. They
purchase such cars to demonstrate themselves, to make others notice them on the
road or in general. Therefore, this type of product consumption is also regarded as
socially-oriented.
They are likely to seek for a symbol of wealth, omnipotence and lifestyle opened to
enjoyment and excitement. Such image is brought by sufficient performance of
attributes: ‘prestige’, ‘design’ and ‘speed/power’. According to the conducted survey,
this combination is usually found in luxury sport cars e.g., Bentley Continental GT
Speed or Jaguar S-type.
As a result, people that were surveyed for this research work are distinguished.
These respondents all together already represent relatively homogeneous group of
customers who all belong to an upper-middle or a higher class characterized by a
high level of well-being able to purchase almost everything including property and
assets. They all are successful professionals or entrepreneurs who dispose self-
earned means. However, provided segmentation permits to differentiate these people
according to very stable meanings and to define common patterns in consumer
behavior based on desired and pursued end states of well-being and on directed
intervening states which facilitate approaching of desired end states.
Vershofen and Belz (Belz, 1994) argued that premium products provide, besides
their intended use, additional utilities of satisfying needs for prestige and self-esteem.
According to the findings of this study, both these utilities should be presented.
However, the customer’s prioritized utility among these two significantly defines his
behavior, his choice for particular brand and product and combination of important
attributes, sought to provide desirable consequences.
7.5. Multi-grid analysis
In this part of the work elicited constructs from all respondents are to be aggregated
in order to draw inferences about commonalities and differences in customers’
perceptions, i.e., meanings. This is another opportunity provided by RGT which is
helpful when analysis of single grids is time consuming in a sense of big amount of
grids.
7. Analysis of repertory grids
219
Multi-grid analysis implies analysis of respondents’ constructs by analogy with the
proposed methodology regarding analysis of a single grid. However, in a multi or
collective grid each construct represents a collective construct developed from
respondents’ constructs ascribed to a particular category (Wright and Cheung, 2007).
Analysis and comparison of single grids allow to compare customers, their behavior
and attitudes, whereas, multi-grid analysis allows comparison of meanings of
different customers.
Twenty face-to-face in-depth RG interviews produced 180 bipolar personal
constructs of how customers perceive ownership and exploitation of cars. In addition
40 supplied constructs were rated by respondents to reveal their preference on the
overall assessment constructs and its relation to elicited perceptions. The
categorization procedure of content analysis allowed to aggregate and to reduce the
original 180 elicited constructs into 12 major collective bipolar construct categories
(see Appendix 4). Two more categories enclose supplied constructs, that is, each
additional category includes supplied overall assessment constructs of all
respondents. Therefore, each additional category consists of 20 constructs. As
described in section 7.2 constructs that actually enclose meanings of two (and in one
case, three) categories are allocated to both related categories. Therefore, the sum
of constructs allocated to twelve identified categories is 197.
Table 34 provides an example of how one collective construct ‘comfortable-
uncomfortable’ which represents a bipolar category ‘comfort’ was generated from 28
individually elicited constructs. Codes on the left-hand side of the table stand for the
number of a construct and number of respondent the construct is elicited from. In this
category six constructs were reversed so that all preferred poles were aligned on the
left-hand side to provide easier grouping of the constructs.
Table 35 shows all collective constructs with the numbers of constructs comprised in
each collective construct. In some categories the preferred poles of customers do not
relate to one of the poles of collective construct as in category ‘comfort’. For
example, some respondents may prefer having large cars, whereas, others prefer
rather compact autos or some prefer autos of universal exploitation and others prefer
more specialized cars such as sport cars. Therefore, constructs in such categories
were reversed so that meanings of their poles correspond to the poles of collective
constructs.
7. Analysis of repertory grids
220
Coded elicited
construct Construct pole: 'comfortable' Construct pole: 'uncomfortable'
1.2 Driver's comfort in terms of less emotional and physical efforts More stress at driving
1.5 Comfort electronics and technologies Limited technological features
2.3 Appropriate size, comfortable passenger seating arrangement Uncomfortable
2.11 Comfort electronics and technologies Lack of comfort electronics and technologies
3.5 R Don't feel hummocks, potholes and bumpy roads at driving Feel hummocks, potholes and bumpy roads at driving
3.7 Feel comfortable on the passenger seat (don't feel bumpy roads and potholes) Feel bumpy roads and potholes on the passenger seat
3.8 R Enough space to lie (stretch legs) Only sitting position
4.6 Appropriate size for tall people Not appropriate size for tall people
5.6 R Comfort electronics and technologies Lack of comfort electronics and technologies
5.11 R Don't feel hummocks, potholes and bumpy roads at driving Feel hummocks, potholes and bumpy roads
6.3 Smooth running, feel comfortable on the passenger seat Not smooth running
6.6 Good ergonomics of saloon Bad ergonomics of saloon
6.11 Feel comfortable on the passenger seat Feel uncomfortable on the passenger seat
8.5 R Comfort electronics and technologies, feel my own space Lack of comfort electronics and technologies, lack of space
9.7 Don't feel hummocks, potholes and bumpy roads (soft suspension) Feel hummocks, potholes and bumpy roads (rigid suspension)
10.1 Driver's comfort, relatively large Not comfortable at driving
12.4 Large and high for tall people Not appropriate size for tall people
12.8 Comfortable for tall people to get in Uncomfortable to get in
13.1 Driver's comfort at driving Not comfortable at driving
13.8 Attractive, comfortable and not harsh interior Uncomfortable and harsh interior
14.8 Comfort electronics and technologies Lack of comfort electronics and technologies
15.2 Driver's comfort in car handling Not comfortable handling
16.3 R No bumpy ride at high speed Bumpy ride at high speed
17.2 Comfortable enjoying steering Not enjoying steering
17.13 Don't feel hummocks, potholes and bumpy roads Feel hummocks, potholes and bumpy roads
18.1 For self-driving at high speed: steering, not noisy, good view Noisy, bad view at a high speed
19.5 Comfortable, good noise isolation Bad noise isolation
19.11 Soft car suspension, not sporty car suspension Rigid, sporty car suspension
Table 34: Collective construct ‘comfortable – uncomfortable’
Type of construct
№ CategoryNumber of individual
constructs representing collective construct
1 Comfort Comfortable Uncomfortable 28
2 Prestige Prestigious Ordinary 25
3 Representativeness Representative Unrepresentative 23
4 Technical quality Of high technical quality Of low technical quality 20
5 Intended application (Universality) Universal exploitation Narrow exploitation 17
6 Size Large Compact 17
7 Design Stylish Unattractive 17
8 Speed/Power High-speed/powerful Limited 13
9 Cross-country ability Rough-terrain high-performance Rough-terrain low-performance 12
10 Safety/Security Safe/secure Unsafe 10
11 Price Expensive Affordable price 8
12 Novelty New model Old model 7
13 Overall confidence Feel confident about a car Not feel confident about a car 20
14 Overall safety Feel safe in a car Not feel safe in a car 20
Total 237
Elicited
Supplied
Collective grid bi-polar construct
Table 35: Constructs of the collective grid
For further analysis a collective aggregated grid incorporating average ratings and
collective constructs of the sample group was produced (Wright and Cheung, 2007).
Display of the collective grid is shown in Table 36.
7. Analysis of repertory grids
221
Table 36: Collective RG with average ratings
(1)
wo
rk(2
) va
catio
n(3
) ev
enin
g(4
) in
no
wa
y(5
) p
revi
ous
(6)
doe
sn't
fit(7
) a
ltern
ativ
e(8
) cu
rren
t(9
) id
ea
l
C1
Com
fort
able
1.6
81
.89
2.2
53
.82
3.0
42
.64
2.5
02
.04
1.2
9U
nco
mfo
rtab
le
C2
Pre
stig
iou
s2
.28
2.4
41
.96
4.0
83
.48
1.9
21
.88
2.4
01
.80
Ord
inar
y
C3
Rep
rese
nta
tive
1.7
42
.35
3.5
23
.43
3.8
33
.09
2.7
82
.78
2.3
0U
nre
pre
sen
tativ
e
C4
Of h
igh
tec
hn
ical
qu
ality
1.8
01
.70
2.1
53
.75
2.3
02
.75
2.2
01
.90
1.2
5O
f low
tec
hn
ical
qu
ality
C5
Un
iver
sal e
xplo
itatio
n2
.06
2.0
63
.59
2.3
51
.76
3.4
72
.65
2.1
22
.35
Nar
row
exp
loita
tion
C6
Lar
ge
2.3
52
.06
4.1
23
.24
2.8
23
.06
2.6
52
.24
2.2
9C
omp
act
C7
Sty
lish
2.2
42
.35
1.9
44
.18
2.7
11
.71
2.4
12
.76
2.1
2U
nat
trac
tive
C8
Hig
h-s
pee
d/p
ower
ful
1.6
92
.08
1.3
84
.00
2.6
21
.77
1.8
52
.00
1.8
5L
imite
d
C9
Rou
gh
-ter
rain
hig
h-p
erfo
rman
ce2
.50
1.8
32
.83
3.4
21
.75
4.1
72
.08
2.4
22
.08
Rou
gh
-ter
rain
low
-per
form
ance
C1
0S
afe/
secu
re2
.50
3.1
03
.50
3.8
02
.70
2.9
02
.20
2.3
01
.60
Un
safe
C1
1E
xpen
sive
2.8
83
.25
2.3
84
.50
3.8
81
.75
3.5
03
.00
2.7
5A
fford
able
pric
e
C1
2N
ew m
odel
1.2
91
.57
1.1
44
.14
2.8
61
.57
2.7
12
.86
1.4
3O
ld m
odel
C1
3F
eel c
onfid
ent
abou
t a
car
1.5
51
.75
2.1
03
.85
2.5
02
.50
2.3
01
.70
1.0
0N
ot fe
el c
onfid
ent
abou
t a
car
C1
4F
eel s
afe
in a
car
1.6
51
.75
2.6
03
.25
2.6
02
.50
2.2
01
.75
1.1
0N
ot fe
el s
afe
in a
car
Ele
men
t ca
teg
ory
Col
lect
ive
con
stru
ct p
ole
(1)
Col
lect
ive
con
stru
ct p
ole
(5)
7. Analysis of repertory grids
222
PCA was performed with the collective grid (the produced PCA map is shown further
in Figure 26). Table 37 shows construct loadings on two principal components, i.e.,
on two dimensions in psychological space (Wright and Cheung, 2007). Component 1
and component 2 account for 85.58% of the total variance.
Dimension 1 Dimension 2
C1 Comfortable Uncomfortable 1,03 0,22
C2 Prestigious Ordinary 1,00 -0,33
C3 Representative Unrepresentative 0,62 0,36
C4 Of high technical quality Of low technical quality 0,92 0,37
C5 Universal exploitation Narrow exploitation -0,12 0,85
C6 Large Compact 0,33 0,71
C7 Stylish Unattractive 0,88 -0,40
C8 High-speed/powerful Limited 1,00 -0,31
C9 Rough-terrain high-performance Rough-terrain low-performance 0,34 0,87
C10 Safe/secure Unsafe 0,62 0,46
C11 Expensive Affordable price 0,84 -0,72
C12 New model Old model 1,26 -0,48
C13 Feel confident about a car Not feel confident about a car 1,08 0,31
C14 Feel safe in a car Not feel safe in a car 0,79 0,43
Percentage of variance for each component 61.09% 24.49%
Collective grid bi-polar construct
Table 37: Fundamental core perceptual dimensions based on construct loadings
Given the constructs with heaviest loadings, dimension 1 includes attributes which
define the class and the level of a car interior design and equipment. Thus, for
example, the more comfortable, technologically equipped, well-designed, prestigious
and powerful a vehicle is, the higher class it belongs to.
Dimension 2 comprises attributes which define level of practicality of auto
exploitation. Thus, large, multifunctional auto with rough-terrain high performance
implies different kind of exploitation than rather compact, specific, e.g., sport car.
However, these attributes do not define how premium a car is as it can be premium
mid-size SUV as Porsche Cayenne or BMW X5 as well as ordinary mid-size SUV
such as Toyota Highlander or Honda Pilot which all satisfy the same characteristics
regarding size, cross-country ability and multifunctional exploitation.
7. Analysis of repertory grids
223
Therefore, dimension 1 is labeled as “Class and image defining” and dimension 2 is
labeled as “Practicality defining”.
The findings indicate that customers’ preferences about a car are built on, on the one
hand, desirable and affordable social aspect which defines the price and
corresponding level of car interior design, equipment, quality and image, and on the
other hand, desirable and needed utility aspect which defines the size and range of
exploitation.
Construct correlations of car perceptions
Table 38 shows correlation matrix of collective constructs. Of particular interest of this
study are supplied overall assessment constructs C13 and C14. The purpose for
supply was to reveal perceptions which are highly correlated with perception of
overall confidence about an auto and overall safety feeling in a car. Both overall
assessment constructs are highly correlated with each other signifying the
importance of perception of feeling safe in general perception of confidence about an
auto. Besides, both supplied constructs are highly correlated with construct C1 and
construct C4 which represent comfort and technical quality of an auto. It means that
for the interviewed sample of customers comfort and technical quality represent
attributes that are determinative when considering general confidence and safety of
an auto. Furthermore, correlation between constructs C1 and C4 is also very high
signifying the interdependence of perceptions of technical quality and comfort.
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
Comfort C1 1.00
Prestige C2 0.78 1.00
Representativeness C3 0.76 0.49 1.00
Technical quality C4 0.94 0.68 0.61 1.00
Intended application (Universality) C5 0.06 -0.48 0.28 0.22 1.00
Size C6 0.49 0.12 0.70 0.53 0.73 1.00
Design C7 0.69 0.88 0.31 0.65 -0.45 0.02 1.00
Speed/Power C8 0.78 0.92 0.40 0.74 -0.39 0.06 0.94 1.00
Cross-country ability C9 0.41 0.04 0.24 0.66 0.67 0.51 0.04 0.16 1.00
Safety/Security C10 0.67 0.54 0.52 0.72 0.30 0.64 0.41 0.46 0.47 1.00
Price C11 0.57 0.82 0.26 0.41 -0.65 -0.12 0.88 0.82 -0.37 0.23 1.00
Novelty C12 0.78 0.78 0.47 0.69 -0.37 0.02 0.90 0.86 0.06 0.26 0.81 1.00
Overall confidence C13 0.98 0.74 0.67 0.98 0.14 0.52 0.69 0.78 0.51 0.74 0.53 0.74 1.00
Overall safety C14 0.94 0.63 0.81 0.92 0.31 0.73 0.51 0.59 0.51 0.81 0.37 0.57 0.95 1.00
Category
Table 38: Correlation matrix of collective constructs
7. Analysis of repertory grids
224
It is also interesting that construct ‘prestige’, C2, is very high correlated with construct
‘speed/power’, C8, showing the importance of attribute performance ‘speed/power’ in
perceiving prestigiousness of an auto. Perception of ‘prestige’, C2, is also highly
influenced by ‘design’, C7, and ‘price’, C11. It is rather evident that prestige of an
auto is reflected in its relatively high price.
Construct ‘design’, C7, is also highly correlated with construct ‘speed/power’, C8, as
well as with construct ‘novelty’, C12. During interviews it was found that very often
customers consider design as good and attractive when car model is new and not
often seen on the roads yet.
For Russian customers of premium autos ‘speed/power’ attribute represents not just
physical characteristics of vehicle capacity. Speed and power are indirect indicators
of consumer character and ambition. The more risky, ambitious and, sometimes,
even aggressive a person is the higher potential speed as well as more horsepower
the person is likely to ask for. People who are also ambitious but prefer rather safe
business or activity in a sense of, for example, legitimacy, possible loss or
competition rules are likely to choose less speedy, rather calm and gentle types of
autos.
Construct ‘price’, C11, in turn, stronger correlates with constructs ‘design’, C7,
‘speed/power’, C8 and ‘prestige’, C2. It turns out that three components are important
for a car to be perceived as reasonably expensive: appearance, physical potential
and social acknowledgement. At the same time, construct ‘price’ is low correlated
with ‘comfort’ and ‘technical quality’. It signifies that for customers higher price is
rather an indicator of status and social acknowledgment rather than good quality and
comfortableness. Probably, such an attitude gives rise to a rather common notion
between customers of ‘adequacy between quality and price’.
HCA of the collective grid presented in Figure 25 confirms the results generated from
the construct correlations. It can be seen that the smaller the clusters, the higher the
significant matches between the ratings and, hence, the stronger the matching.
Constructs C2, C13, C14 and C1 build a tight cluster and when this cluster is
extended it is linked to another tight cluster of C2, C7 and C8 and then to C12.
Another distinct cluster links less correlated constructs and, therefore, it was not seen
evidently in the correlation matrix. This cluster ties constructs C5, C6, C3, C9 and
7. Analysis of repertory grids
225
C10 further confirming the identified by PCA dimensions, the second of which named
“Practicality defining” comprises constructs which reflect the exploitation and practical
appropriation rather than image and class component of an auto.
Figure 25: HCA of the collective grid
Correlation matrix of element categories presented in Table 39 shows no strong
correlation between any element categories. It signifies that consideration set used
for construct elicitation comprises diversified auto categories which allows thorough
distinction between elements and, thereby, increases the validity of findings.
Category E1 E2 E3 E4 E5 E6 E7 E8 E9
Work E1 1.00
Vacation E2 0.75 1.00
Evening E3 0.43 0.28 1.00
In no way E4 0.09 0.33 -0.68 1.00
Previous E5 0.12 0.53 -0.08 0.56 1.00
Doesn't fit E6 0.20 -0.18 0.73 -0.75 -0.52 1.00
Alternative E7 0.25 0.37 0.24 0.08 0.46 -0.11 1.00
Current E8 0.40 0.50 -0.04 0.41 0.53 -0.27 0.59 1.00
Ideal E9 0.68 0.57 0.41 -0.09 0.26 0.12 0.53 0.67 1.00
Table 39: Correlation matrix of element categories
Collective cognitive map of all customers’ perceptions of autos
Figures 26, 27 show the collective cognitive map (Bougon, 1992) of all twenty
respondents. In Figure 26 cognitive map displays the relationships between
7. Analysis of repertory grids
226
elements, the relationships between constructs and how elements relate to
constructs, i.e., perceptions in psychological space on two major perceptual
dimensions. Figure 27 displays the cognitive map without vector lines, i.e., construct
lines. This allows to easily distinguish the location of each element with respect to
others.
Figure 26: PCA of the collective grid
Figure 27: Relationships between elements
The ‘ideal’ element is located in the same quadrant with elements ‘work’, ‘vacation’,
‘current’ and ‘alternative’. Given the labeling of two core perceptual dimensions, it
can be analytically defined in which directions elements categories are distinguished.
7. Analysis of repertory grids
227
It is apparent that elements ‘evening’ and ‘doesn’t fit’ differ from the ‘ideal’ and
‘current’ as well as from other elements located in the bottom quadrants on the
second dimension “Practicality defining”. Elements ‘evening’ and ‘doesn’t fit’ are
perceived more as autos of a narrow exploitation which are likely to be compact,
unrepresentative and with low cross-country ability (rough-terrain low-performance).
At the same time these elements represent cars that have strong image and mostly
relate to premium or luxury cars which are prestigious, expensive, stylish, novel,
high-speedy and powerful. Such set of attributes including both perceptual
dimensions is likely to describe sports cars. Most brands and car models named by
respondents for categories ‘evening’ and ‘doesn’t fit’ relate to sports cars, supercars
or luxury speedy cars, for example, Porsche 911 Carrera, Porsche Cayman, Ferrari,
Mercedes-Benz S-Class, etc. (see Appendix 6 for cars mentioned). These cars are
very attractive for respondents, however, according to the findings most of the
customers use auto first of all for business purposes and, hence, sport cars are
usually not appropriate. Other reasons for being not suitable are climate conditions
and the quality of roads which are significant regional distinctions of the most parts of
Russia.
Element 5 ‘previous’ represents car category which includes cars used directly before
the ‘current’ ones. From plots it can be seen that ‘previous’ differs from elements in
the bottom, left-hand quadrant mainly on the first dimension “Class and image
defining”. The inference can be made that participants are experienced and
determined concerning what kind of auto they need in terms of their practical
exploitation. They are not likely to change an auto type but rather look for an updated
quality and comfort features as well as new, stronger or improved image component
and higher class.
This inference is also confirmed by the main difference between ‘current’ and ‘ideal’
elements. They are very close, almost the same on the second dimension and
distinct on the first dimension although both already being plotted in the same
quadrant. The same tendency occurs between elements ‘previous’ and ‘current’.
However, ‘current’ element is already closer to the ‘ideal’ on the first dimension
appearing in the same quadrant.
7. Analysis of repertory grids
228
Similar analysis can be performed regarding other elements using two major
perceptual dimensions and the construct lines to describe how each auto category is
perceived by respondents.
In the current study elements were elicited from respondents according to the
supplied categories in order to compare single grids and carry out collective
aggregate analysis of the grids. However, each respondent was considering his own
category set comprised of particular auto models associated with supplied element
categories what enabled each respondent to express constructs comparing subjects
he is familiar with. Therefore, in aggregated collective grid customers’ perceptions of
supplied auto categories rather than certain auto models or brands are analyzed. The
study is aimed, first of all, at deriving cognitive meanings and mental representations
as well as preferences associated with auto purchase and exploitation. The study
was not aimed at revealing perceived differences and preferences in consumers’
mental structures regarding particular autos. When this goal is pursued in the
research it is recommended to supply a set of elements consisting of interested
brands or particular products, however, the researcher should then conduct the
survey with respondents who are familiar or, at least, aware of all supplied elements.
7.6. Additional analysis and findings for managerial implications
7.6.1. Segmentation according to aims for purchasing an auto
In the customer segmentation according to associated values and benefits sought
(described in 7.4) the revealed abstract consequences are considered as bundles of
common important attributes defined for each identified group. Thus, commonly
important attributes are recognized within the each group. This kind of segmentation
provides useful information first of all for the company and can be used for internal
processes such as product development and positioning strategies.
Abstract consequence, however, also provides a common meaning of product
characteristics and attributes comprised in each bundle. This shared meaning if
accurately defined can be effectively used in advertising and positioning as a
message for an appeal to customers.
One of the questions of the questionnaire offered to respondents after RG elicitation
was “What was (is) the aim of purchase of the current or next car?” The respondents
7. Analysis of repertory grids
229
were asked to select from a multiple choice or to provide own answer in case none of
the options appeared suitable. The options included:
1. Need for transportation during leisure and free time
2. Need for transportation at work
3. To maintain and fit a social status
4. To enjoy driving (ride)
All respondents chose one of the provided options and none of them chose the first
option. The resulting distribution of answers is presented in the diagram of Figure 28.
To enjoy driving (ride)50%
To maintain and f it a
social status
30%
Need for transportat ion
at work
20%
Figure 28: Aims of purchasing the current or next auto
Respondents are divided into three groups according to the aim of purchasing they
named. Constructs of all respondents within one group are put together and analyzed
by means of PCA. It is permitted by the RG form designed for this study where
elements are elicited by common categories.
The revealed dimensions are ordered according to the percentage of variance each
dimension accounts for within the overall number of constructs in each group.
Constructs in each dimension are considered and common meanings are suggested
for each dimension. Names reflecting meanings of abstract consequences that
allowed to identify commonalities within each group are presented in Table 40.
In each group along with common significant dimensions there are also significant
dimensions which reflect very personal requirements and, therefore, they are
reflected in the table as individual requirements. For example, in the first group the
fifth dimension consists only of two constructs which both are expressed by the same
respondent, construct 3.1 ‘new model: new means technologically innovative, not
7. Analysis of repertory grids
230
outdated’ versus ‘old model’ and construct 3.8 ‘comfortable: enough space to lie
(stretch legs) so I can take a rest during a ride’ versus ‘uncomfortable’. These
constructs are rather specific and do not have common variation pattern of ratings.
However, when a customer is to be considered individually these requirements are
very important to take into account.
Aim Dimension Variance % Cumulative %Loadings of
'ideal' element
1. Most important features of a car to be used for my working purpose 48,43 48,43 -3,35
2. Reflecting my status 24,37 72,8 -0,37
3 Individual requirements 9,26 82,07 -0,97
4 Individual requirements 7,04 89,1 -0,62
5 Individual requirements 5,08 94,18 -1,32
1. "Auto - as my mirror" - reflects my character 36,05 36,05 -3
2. Perception of ME in an auto - everybody should know who I am 17,9 53,96 0,11
3. Purpose of use especially during leisure time - to enjoy the ride (for country side / for city)
15,24 69,2 -0,64
4. Image and perception of my AUTO 11,36 80,55 -1,69
5. Individual requirements 8,06 88,61 -0,32
6 Individual requirements 4,45 93,06 1,51
7 Individual requirements 3,89 96,95 2,11
1. Auto for me - fits my character and driving style 32,45 32,45 -3,10
2. Practical use (e.g., for work but not for enjoyment) 24,72 57,17 -3,02
3. Use for long, far trips (camping, country side) 11,67 68,83 1,73
4. What means for ME to "feel confident (safe)", i.e., a 'must' which defines the type of my car
9,47 78,30 0,69
5 Individual requirements 8,12 86,42 -0,24
6 Individual requirements 5,92 92,34 1,63
7 Individual requirements 3,95 96,29 0,10
8 Individual requirements 3,71 100 2,54
Need for transportation at work
To maintain and fit a social status
To enjoy driving (ride)
Table 40: Common customer abstract consequences
Although, in each customer group individual requirements are presented, first
dimensions within each group can be defined according to a shared meaning of
comprised constructs.
In the first group of customers whose aim of purchase of the current or next auto is
associated with business purposes the first dimension (accounts for 48% of variance)
7. Analysis of repertory grids
231
reflects constructs mostly referred to the requirements related to business
exploitation of an auto. Most of the ascribed attributes are technical quality, comfort,
cross-country ability and speed/power. The supplied constructs regarding feeling
safe and confident are also mostly ascribed to the first dimension.
The second dimension comprises constructs that are mostly associated with the
status and reflect the attribute “representativeness”.
In the second group where consumers pursue to maintain and fit the social status the
first dimension is named “auto as my mirror which reflects my character”. Attributes
ascribed to this dimension are very different such as prestige, design, comfort,
speed/power and size, however, they all reflect characters of respondents who
named the constructs. For example, some respondents who are aimed at strong
leadership and power have prestige and speed/power ascribed to this dimension,
others who refer to the type of ‘hedonistic extroverts’ have comfort, design,
speed/power, size and cross-country ability which reflect their way of driving or
exploitation of an auto as well as aesthetic preferences. Defining general meaning of
this dimension especially required a thorough consideration of all the observed and
known data about the respondents obtained by the researcher.
The second dimension “perception of me in an auto” contains attributes which can
reflect and project auto peculiarities on its owner. Attributes such prestige reflecting
tough image, new model emphasizing uniqueness, limited exploitation in a sense of a
speedy car, pompous auto, unsafety in a sense of being notable on the road, etc. are
ascribed to this dimension.
The third dimension contains attributes related to the practical use of a car,
especially, during leisure time. Thus, it comprises size, cross-country ability, intended
application, comfort as well as technical quality and safety/security.
The fourth common dimension is “image and perception of my auto”. This dimension
reflects attributes and underlying characteristics that can be perceived by others and
emphasize a certain image of an auto. For example, the construct (construct 17.14,
see Appendix 4) related to the attribute safety/security is ascribed and respondent’s
preferred pole is reversed, showing that he actually prefers an unsafe auto. This is
because by an ‘unsafe auto’ the respondent implies a car of a good brand which has
7. Analysis of repertory grids
232
a high risk of being hijacked. Therefore, this construct actually emphasizes a brand of
the auto. Other attributes are unrepresentativeness emphasizing sports type of a
vehicle, prestige, novelty, price and design.
The third group comprises people who purchase the next auto to enjoy driving or
ride. Similarly to the first dimension of the second group, the first dimension of this
group “auto for me, e.g., fits my character and driving style” (accounts for 32% of
variance) contains a variety of attributes which actually reflect car features that
characterize the way of exploitation and the style of driving. Most common attributes
in this dimension are comfort, speed/power and intended application. However,
regarding some respondents this dimension also reflects prestige and design.
The second dimension “practical use, e.g., for work but not for enjoyment” consists of
constructs which mostly reflect auto characteristics and attributes necessary for
business purposes. Although the respondents enjoy driving, they still devote much of
the time to working and business purposes and, therefore, their preferences
concerning car attributes relate to a considerable degree to the practical, business
use. Thus, attributes such as representativeness, prestige, quality and intended
application are most common for this dimension.
The third dimension “use for long, far trips (camping, country-side)” reflects the group
peculiarity of enjoying driving and ride. It contains constructs which show
requirements of a car that make it appropriate for long distances and driving out of
the city. Prevailed attributes are size, intended use, technical quality and cross-
country ability.
The last dimension “what means for me to feel confident (safe), i.e., a ‘must’ which
defines the type of my car” reflects attributes which define individual perception of
safety and confidence associated with car exploitation. Attributes such as intended
application emphasizing auto universal exploitation, cross-country ability reflecting
rough-terrain high performance and comfort are presented in this dimension. Also
attributes prestige and representativeness are presented as reversed describing
ordinary, not prestigious and unrepresentative types of autos.
From the last column of Table 40 it is seen that in all groups first dimensions have
the highest loadings of the ‘ideal’ element. It emphasizes not only the homogeneity
but also the preference and importance of the identified abstract consequences.
7. Analysis of repertory grids
233
It is also shown that in the second group of customers whose aim is to maintain their
social status the second highest loading of the ‘ideal’ element is presented by the
fourth dimension reflecting attributes which allow to perceive an auto by others. This
finding supports theoretical statements about the conspicuous consumption of people
who strive for social status, prestige and recognition.
In the third group the second and the third dimensions have second and third highest
loadings of ‘ideal’ element emphasizing customers’ enjoyment from using a car in
their free time. However, the peculiarity of the whole sample is also taken into
account and customers’ need to devote a considerable part of auto exploitation to
business purposes is reflected.
This type of segmentation illustrates how common abstract consequences which are
pursued by customers can be recognized and formulated as a message in which
people can recognize their own benefits sought. When the common meaning or
direction of a consequence is defined it is further a wording art to make a strong
appeal but it is clear which meaning the message should communicate.
7.6.2. Stereotypes of car attribute perceptions
Table 41 provided in this section shows most common consumers’ perceptions of
product attributes. Although these meanings are common among consumers of the
sample, some of them may appear new for the researcher and for company
managers.
RGT allows to reveal meanings that are common among a particular group or type of
consumers. For example, perceptions can reflect or be based on regional
peculiarities from where respondents come from as well as cultural peculiarities and
generally adopted attitudes towards product consumption peculiar for the sample.
Especially when conducting a survey in a foreign country or in an uninvestigated
region, the RGT enables to reveal perceptions that are not self-evident and are hard
to think of when developing a questionnaire in office. Often, important aspects of
consumers’ construing systems associated with the domain are not considered in
surveys based on questionnaires with supplied questions regarding individual
perceptions (Stewart et al., 1981, p. 49).
7. Analysis of repertory grids
234
Comfort
Comfort electronics and technologies
Not feel hummocks, potholes and bumpy roads
Comfortable passenger seating arrangement
Comfortable enjoying steering
Appropriate size
Feel my own space being far from other people
For self-driving at high speed: steering, not noisy, good view
Prestige
Reflects my status
Respect of other drivers on roads who give me a way
For mature experienced users who are good judges of autos
Expensive, immodest image
Image of brand and manufacturer
Not youth
Super image, for men, not women
Others will notice me
Representative: Prestigious, which means reflects my status
With driver / not for self-driving
For business, not for family or entertainment
For experienced users of premium cars who are good judges
Not youth
Unrepresentative: For self-driving
For youth who are more risky (not for mature people)
Speedy self-driving of a cool auto
Technical quality
Won't break down, time and money savings
Producer's proved high quality (producer reputation)
Available good service
Safe in our climate (snowdrift, glare ice)
Rigid frame and suspension
Universal exploitation: For work and leasure
All-year round, for climate conditions
Can be only one in garage
Not pompous
Highly functional
For everyday life not only for entertainment
Not sporty
Narrow exploitation: With strongly marked peculiarities (e.g. high-speed or powerful)
Sports car for pleasure
Size
Large: Better observation (panoramic view) and feel safeCan fit people and luggageSteady, not short: not to go off from rut For more than 2 people
Compact: Others will notice me, aesthetically acceptableBetter to move on in traffic jam or narrow roads
Design
Attractive appearanceStylish: beautifulCan be expected uncommon, not boring coloring Demonstrative, bright designModern, not datedSharp, aggressive forms
Representativeness
Intended application (Universality)
7. Analysis of repertory grids
235
Speed/Power
Speed: SpeedyAllowable high speedCombination of allowable high speed and cross-country abilityLimited in speed: for comfortable, calm driving
Power: Powerful: not girlyAdequacy between power and weightCombination of power and economy
Cross-country ability
For city, countryside and mountainsSafe "anywhere" ride in tough climate conditions (mud, ice) No worry about potholesRigid suspension
Safety-Secure
No risk of criminal attempt because of good car brand Appropriate for bodyguard, driver, armoring
Price
Not affordable for majority Value for moneyOverpriced to make a splash
Novelty
Technically innovativeModern, not datedUnique, uncommonElegant because of novelty, uniqueness
Table 41: Stereotypes of car attribute perceptions among the respondents
7.7. Summary
♦ Twelve categories are identified according to the content of elicited constructs.
♦ The top four categories which together cover 49% of all constructs are ‘comfort’,
‘prestige’, ‘representativeness’ and ‘technical quality’. It means that these
four categories represent product attributes that are most common among the
respondents.
♦ Category ‘novelty’ is the least common attribute among respondents, however, it
is very important for those who mentioned it. These respondents form a latent
group for which the fact that a car is new and embodies most recent
technological achievements is an important factor influencing the choice decision.
♦ The notion and feeling of confidence about an auto is often associated with the
image and social perception of an auto as well as with the perception of its owner
rather than with the feeling of safety.
7. Analysis of repertory grids
236
♦ For the majority of customers who expressed constructs ascribed to the category
‘safety/security’ these constructs have a low relevance to the perception of safety
of an auto or provided by an auto. The respondents represent people referred to
a risk group who perceive general safety measures provided by an auto as a
must have in premium autos due to their superiority in quality. However, when
really concerned in a sense of criminal threat, they have particular requirements
about safety and security measures which often exceed ordinary measures
provided by ‘average’ premium autos. Respondents of the sample share belief
that “there are no technologies smarter than a man”. Therefore, the common
attitude towards the notion of auto safety is rather skeptical in a sense that people
believe that if there is a plan to commit a crime it will most probably be
accomplished.
Single grid analysis
♦ Individual schema of consumer motivational sphere is built. It accumulates all
the elicited personal data which was analyzed and structured according to
different levels of abstraction of the respondent’s personal meanings.
♦ Abstract expected consequences representing individual motives of the
middle range aggregate individual meanings which are not only put into words
but also expressed in individual estimations.
� Respondent’s first abstract consequence is defined as “Appropriate for my
business use: armored and representative”. The shared meaning between
constructs attributed to the first dimension, i.e. principal component is that they
relate to practical characteristics that are essential for the consumer’s car
exploitation taking into account his life circumstances and business as the
main purpose for car exploitation.
� Respondent’s second abstract consequence is defined as “Novelty of a
model with recent technological achievements”. Indicative constructs of
the second dimension reflect respondent’s attitude and preference to novelty
of an auto model which, in turn, means for him a representation of most
updated technologies in automotive industry.
7. Analysis of repertory grids
237
♦ Structuring of elements relations showed that the respondent chooses cars that
are mostly appropriate for his working purposes. He draws his ideal car very
close to the currently possessed auto as well as to the one that he considered as
an alternative.
♦ The respondent did not choose the ‘alternative’ auto over the ‘current’ one
because its technological and design features did not enough meet his
preferences, although this car was considered as appropriate for security
measures and representativeness.
♦ It is also found that the car (Aston Martin DB9) named as an auto that ‘I really like
but it doesn’t fit’ meets all individual preferences regarding technological
equipment, design and comfort. That is, it is very close to the ideal image on
the second dimension which reflects the second abstract consequence. The
main reason why this auto ‘does not fit’ (and is not purchased) lays in a lack of
its correspondence to the consumer preferences regarding auto
appropriateness for his working purposes reflected in the first dimension.
♦ Analysis of single grids has shown that the suggested methodology reflects all
assumptions and notices recognized during the interview. Therefore, it is
believed that if an interviewer and an analyst appear to be two different people
the proper inferences can still be drawn revealing different aspects of
respondent’s life influencing his consumer behavior.
Segmentation based on the comparison of the findings from the single grids analyses
♦ Individual analysis of respondents’ motivational schemas allowed to conduct a
two-stage segmentation. At first stage consumers are divided according to the
most general motives, i.e., values and on the second stage – according to the
revealed abstract consequences which represent the motives of the middle
range. As a result, customers are divided into five groups. Description of groups
is presented in a form of matrix.
♦ Horizontally the groups are divided into two types of consumers ─ ‘careerists’
and ‘hedonists’:
� Careerists represent people who aspire to self-actualization and
professional success
7. Analysis of repertory grids
238
� Hedonists represent people who in the first place aspire to pleasure and
enjoyment of life. For these people professional success represents means
by which they ensure the aspiration in terms of financial and status security.
♦ Vertically the groups are divided into two types, ‘introverts’ and extroverts’,
based on the social orientation and the extent of conspicuity of product
consumption:
� Introverts represent people with an introverted type of consumption. These
people are self-oriented in consumption and purchase premium goods to
benefit from its quality and functional superiority in the first place. Attributes
like image and design are also important for these people but for own
enjoyment rather than to make sure to be perceived properly by others. For
these people the social function of premium brand autos provides first of all
a sense of life they possess or aspire to.
� Extroverts represent people whose consumption of premium goods is more
socially oriented and they pursue social prestige and acknowledgement
by such a conspicuous consumption.
♦ It is found that additional utilities of satisfying needs for prestige and self-
esteem provided by premium products both should be present. However, a
customer’s prioritized utility among these two significantly defines his behavior,
his preference for a particular brand or a product and combination of important
attributes sought to provide desirable consequences.
Multi-grid analysis
♦ Analysis of the collective grid which incorporates average ratings and collective
constructs of the sample group revealed two major directions in which
customers’ perceptions are built.
� The first direction is called “Class and image defining”. It represents the first
dimension where attributes define the price and corresponding level of car
interior design and equipment, technical quality and image.
� The second direction is called “Practicality defining”. It represents the
second dimension which comprises attributes defining the level of practicality
7. Analysis of repertory grids
239
of car exploitation. These attributes define the size and range of vehicle use.
That is, they define the type of a vehicle but they do not define how premium
the car is.
♦ The collective construct ‘design’ is highly correlated with the construct ‘novelty’. It
is found that very often customers consider design as good and attractive when
car model is new and not often seen on the roads yet.
♦ The collective construct ‘speed/power’ is attributed to the first dimension
which defines a car class and image rather than its practicality. For Russian
customers in the premium car market the ‘speed/power’ attribute represents not
just physical characteristics of a vehicle capacity. Speed and power are indirect
indicators of consumer character and ambition.
♦ It is revealed that when purchasing a next car to replace the previous one the
participants are not likely to change a car type. They are experienced and
determined concerning what kind of auto they need in terms of their practical
exploitation. So when purchasing a new car they rather look for an updated
quality and comfort features as well as a new, stronger or improved image
component and higher class.
Additional analysis and findings for managerial implications
♦ Customer segmentation according to customers’ aim of purchasing a next
auto is presented. Common abstract consequences pursued by customers within
each segment are recognized and formulated as a message in which people can
then recognize their own benefits sought. When the common meaning or direction
of an expected consequence is defined marketers can produce a strong appeal
by communicating this meaning.
♦ The repertory grid technique allows to reveal meanings or stereotypes that are
common among a particular group or type of consumers, but not necessarily
known before to a researcher and company managers.
Summary and conclusion
240
Conclusion
1. Theoretical contribution
The main goal of this dissertation was to contribute to the understanding of consumer
motivation which explains the psychological dynamics determining consumer
behavior.
In literature, concepts of consumer motivation and motives have been often
approached, however, there are certain limitations regarding the explanation and
disclosure of these notions. Most concepts of consumer motivation suggest universal
sets or categories of motives that are presumed to apply universally across contexts
and, therefore, they appear unable to account for specific actions and to recognize
situational variability of consumer behavior. Most theoretical concepts of consumer
motivation do not provide certain supportive methodology which would enable
sampling it in a particular investigation of consumers. Besides, in consumer research
a long-standing dispute between qualitative and quantitative approaches to
motivation research resulted in attempts to develop a compromise approach based
on the synthesis of motivation research with quantitative practice, however, an
accomplishment of this task has remained vague.
Lack of a certain theoretical model of consumer motivation and causality which would
provide an active guidance for analysis algorithms results in the conduction of most
of the research and analysis of consumer motives as purely atheoretical predictive
exercises producing an ad hoc inductive explanation.
Considering the aforementioned limitations and dilemmas in the field of consumer
motivation research, the main theoretical contribution of the thesis is the suggested
conceptualization of the consumer motivational sphere that provides a certain
representation of a personal cognitive structure. The conceptualized schematic
model is:
• a theory based conception of content and structure of consumers mental
representations and their interrelations;
Summary and conclusion
241
• constructed with respect to the methodology supporting theoretical
assumptions and, thereby, allowing for further practical application by
marketers and researchers.
Consumer motivation is goal-oriented. It consists of the anticipation of possible
actions expected to result in certain consequences which can bring a person closer
to his goal. The reasons for choosing a particular goal are represented in a
hierarchical network of motives.
The suggested taxonomy of the individual consumer motivational sphere links an
individual’s product related knowledge with person related self-knowledge where
representations such as product characteristics, attributes, consequences and values
are ordered hierarchically according to their levels of abstraction and connected
through means-end linkages. The schema constitutes the relationships between
purchasing decision and its associated motives, where the motives are interrelated
according to their levels of abstraction.
Consumer motives are represented in the person related self-knowledge part of the
schema where specific and abstract consequences represent “too-specific motives”
and “motives of the middle range”, and personal values represent “too-general
motives”.
The schematic model is based on the assumptions of the means-end theory and
some of its modifications and on the concept of consumers’ motives implying three
levels of motivational abstraction. Values represent too general ends which
themselves cannot show what is it expected from the product consumption that
would bring a person or a group of people to the desired state of being. Too-specific
motives explain very specific purchasing behavior which is hard to approximate in
order to reach a common set of expected consequences of a group of people. As a
result, motives of the middle range represent a theoretical challenge in defining it and
providing a measurement instrument which would allow a balance between too-
specific and too-general motives.
Built on that, another contribution of the work is the definition and assessment of
motives of the middle range comprised in the consumer motivational sphere. Since
the definition of this type of motives is tied to its assessment, this contribution is
considered as methodological and discussed in the next section.
Summary and conclusion
242
The proposed schema of consumer motivational sphere serves as a theoretical
framework underpinning the suggested methodology which enables the
conceptualization to be appropriate for practical implication.
2. Methodological contribution
Extensions and modifications adopted in the theoretical model of consumer
motivational sphere are made to overcome some limitations of traditional approaches
in construing and measuring consumers’ cognitive structures. It is aimed at
suggesting a congruent methodology which gives an insight into other aspects of
consumers’ mind rather than the ones that are usually explored within consumer
motivation research.
The major methodological contribution of this work is the proposed integrated method
for assessing each level of abstraction in the consumer motivational sphere. The
methodology shows how each level of abstraction represented in the schema can be
derived from the data elicited by means of the repertory grid technique and analyzed
by qualitative and quantitative approaches.
Repertory grid is a cognitive mapping instrument designed to capture the dimensions
and the structure of personal meanings allowing the researcher to attain new, deeper
insight on how people think and perceive life experiences and the world around.
Based on the personal construct theory, the repertory grid technique provides an
interpretive research framework for exploring similarities and differences in the
content and structure of consumers’ subjective meaning systems.
The integrated method represents an algorithm built on existing qualitative and
quantitative approaches that produces two types of analysis outcomes. The first type
is based on producing a ‘person driven’ outcome which organizes information elicited
directly from the person using his own meanings expressed in wording and,
therefore, the numeric part of the data is not considered. This qualitative approach is
based on the laddering technique which is often used for qualitative analysis of the
repertory grid data. The proposed methodology does not contradict traditional
approaches to analyze repertory grid data but is rather aimed at further
methodological elaboration by involving quantitative analysis to benefit from the fuller
capacity of the repertory grid technique.
Summary and conclusion
243
Hence, the second type of analysis produces an ‘analysis driven’ outcome whose
goal is to reveal and structure the underlying meanings expressed in both direct
wording and numeric evaluation and comparison. Analysis of meanings that are
construed by an evaluative and comparative approach to cognitive representation
permits one to uncover some part of subconscious or even unconscious construing
system. Thus, mental representations built on the interrelationships between
meanings are considered. Such interrelationships are hard to recognize when only
interpreting an individual’s directly expressed wording.
As mentioned above, another contribution of this dissertation is that the suggested
methodology comprises of a certain approach to define and assess motives of the
middle range. To assess motives of the middle range a quantitative approach is
suggested which structures personal meanings so that they can be approximated
and commonalities within a group of people can be revealed.
Motives of the middle range are reflected in abstract consequences which are
positioned in the chain of levels of the consumer motivational sphere between
specific consequences and values representing too-specific and too-general motives
respectively. Principal component analysis of a grid data reveals dimensions which
represent directions based on which a person perceives objects. By cautious
interpretation, the dimensions can be generally defined according to the common
meaning, or the direction of construct meanings, ascribed to each dimension.
Common meanings underlying each identified dimension represent abstract
expected consequences. Therefore, an abstract consequence is derived from
product characteristics and their perceptions by identification of a bundle consisting
of product features, each of which a person construes in a similar way and direction
and associates them with an attainment of a certain benefit.
The virtue of identifying abstract consequences by means of the suggested
methodology is that another type of human cognitive representation is additionally
involved. In laddering a person actually uses his ability to reason, that is, he
expresses underlying reasons for perceiving and preferring things in this or that way.
In the proposed method, analysis of a person’s comparing ability is involved.
Therefore, mental representations generated by a person’s evaluative and
comparative approach in perceiving things are analyzed and inferences are made not
only from vertical relationships between meanings, i.e., linkages between different
Summary and conclusion
244
levels of abstraction, but also from horizontal interrelationships between meanings
that are on the same level of abstraction.
The quantitative methods applied together with the qualitative content interpretation
allow to consider respondent’s less salient meanings which are still presented in his
construing systems related to the topic in question and may influence person’s
choice decision. These meanings reflect consumer preferences regarding product
characteristics which may not be very important for the consumer in terms of benefits
sought as long as these characteristics are appropriately provided. However,
deficiency of appropriate presence of these characteristics may significantly impact a
consumer’s choice and, therefore, it is important to consider them and define their
role in the person’s cognitive structures related to the domain.
The proposed methodology approaches some concepts and analysis which are
rarely or never have been used by researchers for analysis of repertory grid data,
although their application may significantly extend the findings during further
analysis.
♦ The application of principal component analysis and hierarchical cluster analysis
on the complementary basis allows one to reach a synergy effect by structuring
and systematization of the repertory grid data. The subsequent implementation of
these analysis methods enables a more precise and reliable recognition of the
similarities and differences between meanings. It also permits the presentation of
the data and interpreted inferences in a form which simplifies data visualization
and allows customer segmentation according to customers’ perceptions and
preferences.
♦ Supplying of ‘self’ elements appears helpful in any research in which a grid is
used to understand the respondent’s choices, i.e., preferences. ‘Self’ elements
such as ‘ideal’ or ‘current’ represent individual perceptions and images of what
the respondent actually has or wants when talking about things, events or people
enclosed in the elements set. When elements represent products or brands the
‘ideal’ element represents a generalized image comprising mostly desired
performance of a product or a brand from the domain.
♦ Honey’s content analysis, in contrast to the traditional content analysis in which
the numeric part of the grid data is ignored, permits the researcher to make use of
Summary and conclusion
245
individual meanings expressed in ratings. This analysis method identifies how
closely elicited meanings relate to the overall issue about the topic. Honey’s
content analysis also allows the discernment of a presence of latent groups of
customers for whom attributes that appeared less salient for the rest of the
sample are commonly important.
♦ Application of resistance-to-change technique enables prioritizing personal
values. An identified, hierarchical order of values permits the researcher to reveal
values that are not sincere and central for a person, but being rather expressed
fictitiously to seem more meaningful.
Overall, the development of the methodology was aimed at providing approaches
which would allow the researcher to overcome some limitations of concepts of
consumer motives and motivation described in the beginning. As a result, the
methodology:
• is supportive and congruent with the theoretical conceptualization of consumer
motivation;
• is appropriate to reveal motives which define particular consumer behavior in a
given situation and circumstance;
• integrates qualitative and quantitative analysis methods that enables an
insight into different aspects of consumers’ cognition;
• allows one to take into account product characteristics and perceived
attributes that are less salient but, however, may significantly impact the
consumer behavior.
Moreover, the suggested approach to define and assess motives of the middle range
provides probably the first attempt to build upon Trommsdorff’s concept of motives of
the middle range by extending it with a methodized assessment approach which
overcomes the limitations of given universal sets of motives.
3. Practical contribution and implication
The following quotation clearly paraphrases the inspiration on which a contribution of
this dissertation to the managerial implication was focused.
Summary and conclusion
246
“The danger is not of getting too close to customers; the real danger is not getting
close enough.” (Johnson, 1998, p. 27)
Nowadays highly competitive markets and increasing complexity of consumer
behavior influenced by their social environment make companies strive for each
individual customer by providing benefits far beyond just primary functions of
products. The act of purchase and consumption is no longer neutral but a particular
form of social behavior, especially, regarding products which have public resonance.
In the markets of premium and luxury goods a social function of a product, to a large
extent, defines the product success. Thus, together with the instrumental function
premium brand cars have a social function which can ensure a consumer the social
prestige and sense of life he aspires to.
In order to offer a consumer product which would generate symbols emphasizing
exactly the aspects on which the consumer puts an emphasis, a company needs to
know the consumer much more deeply than just his buying habits and main
preferences. The consumer needs to be considered as a bearer of a certain ideology
or world perception. A consumer’s mental representations comprised in certain
cognitive structures, his goals and values have to be investigated in order to meet his
personal dispositions in life which in turn can be somehow reflected by or associated
with the offered product.
Based on the theoretical model an assessment approach is developed in this thesis
within the interpretive research framework underlying the repertory grid technique.
This technique is not only advantageous because of its capability to gather a variety
of personal data expressed in the person’s own language and to organize it into a
convenient form, but also due to its flexibility and adaptability in different fields and
areas of research. In this work, the description of the repertory grid technique is
presented rather thoroughly in order to demonstrate how versatile this tool is and
how its design can be formed according to the field, objectives and circumstances of
a research. The repertory grid is widely applied in psychology along with other
numerous fields of application, it has also been proved to be potentially useful
approach for conducting qualitative market research, however, up to now this
technique is rarely used with its full potential by marketers. The repertory grid
technique can be very effective for exploratory research when consumers’ opinions
on the researched topic are to be disclosed and investigated.
Summary and conclusion
247
The suggested methodology allows the structuring of personal meanings according
to different levels of abstraction that, in turn, allows the approximation and revelation
of commonalities in personal representations within a group of people. Thus, the
proposed format of the findings structure enables customer segmentation based on
different levels of abstraction giving marketers more flexibility in deciding on an
appropriate appeal to customers with respect to, for example, the size of the sample
and derived segments.
The method can also appear useful for advertising and positioning as it allows the
researcher to reveal stereotypes associated with the topic and relative products or
brands. The meanings are elicited in the customers’ own language and jargon that
permits the use of common expressions and underlying meanings to create
messages which would be understandable and more personal to the people.
Another perspective of the methodology's application is to reveal customers’ attitudes
towards both their own and their rivals’ products or brands. Elements representing
different products or brands can be compared according to their ability to provide
abstract expected consequences. By clustering the elements within each dimension,
i.e., principal component which represents an abstract consequence, marketers can
make conclusions about how products can be improved. For example, if a product
has a high match with its ideal, generalized image on the first dimension but appears
rather distant on the second, it signifies that this particular product cannot provide or
cause the consequence implied by the second dimension in the desired manner.
Therefore, product characteristics and product attributes comprised in the second
dimension need to be improved or repositioned.
The addition of the repertory grid form with supplied constructs, which reflect a
general topic or sub-theme of a research, allows one to discover latent groups of
customers with specific requirements and preferences. This feature is especially
useful for studies conducted in the markets where products are almost individually
customized.
Segmentation, from the empirical point of view, is done according to the values and
abstract consequences. It permits the differentiation of customers according to
relatively stable meanings in life. Common patterns in the studied consumer behavior
Summary and conclusion
248
are defined based on desired and pursued end states of well-being and on
intermediary pursuance which facilitates an attainment of desired end states.
As a result respondents, who together already represent relatively homogeneous
group of customers of premium cars, being professionals and self-made
entrepreneurs able to purchase almost everything including desired cars, property
and assets, are divided into five groups.
The whole segmentation has a form of a matrix where the horizontal dimension
differentiates between careerists and hedonists and the vertical dimension – between
introverts and extroverts.
Horizontal division is based on their priority regarding aspiration for self-actualization
and professional success or aspiration for pleasure and enjoyment of life ensured by
professional success in terms of financial and status security.
Vertical division is based on the respondents’ social orientation and the extent of
conspicuity of product consumption.
Thus, customers who relate to introverts do not have an attribute ‘prestige’ among
their three most important common attributes. For these people the social function of
premium brand autos provides, first of all, a sense of life they possess or aspire to. In
contrast, extroverts have a common most important attribute ‘prestige’ among their
first three. The social function of premium brand cars provides extroverts with social
prestige and recognition.
Findings of this dissertation show that additional utilities of satisfying needs for
prestige and self-esteem provided by premium products should both be present.
However, the customer’s prioritized utility among these two significantly defines his
behavior, his preference for a particular brand or a product and combination of the
important attributes sought to provide the desirable consequences.
Based on the differences described above and considering other revealed
peculiarities of the identified groups, fine borderlines can be drawn between relatively
homogenous customers that marketers can use for effective communication to
appeal to deep and rather stable dispositions and values.
Summary and conclusion
249
Multi-grid analysis has shown that customers’ preferences about a car are built in two
major directions. The first direction reflects desirable and affordable social aspect
which basically determines a car class. Preferences representing the direction define
price and corresponding level of car interior design and equipment, technical quality
and image. The second direction reflects desirable and needed utility aspect which
determines type of a car. Respective preferences define the size and range of
vehicle use.
Analysis of elements relations has shown that when purchasing a next car to replace
the previous one customers are not likely to change an auto type but rather look for
an updated quality and comfort features as well as for a new, stronger or an
improved image component and a higher class.
Customer segmentation according to customers’ aims of purchasing the next auto
illustrates how common abstract consequences pursued by customers within each
segment can be recognized and formulated as a message in which people can then
recognize their own sought benefits.
When common meaning or direction of an expected consequence is defined it is
further a wording art to make a strong appeal but it is clear which meaning the
message should communicate.
All findings and inferences made during the empirical verification of the suggested
conceptualization of the consumer motivational sphere and supportive methodology
provided a detailed picture of relevant buyer segments. The company has got a clear
view on the portrait of relevant buyers of premium cars in Russia including rather
specific and hardly assumed characteristics of the market and its customers.
Russian customers of premium cars and premium goods in general are rather
peculiar. Their consumer behavior is impacted by their attitudes and mental
representations built on historical and cultural peculiarities as well as on personal
experiences which took place during the last two decades. The “turbulent nineties”
during which Russia has gone through drastic alterations of the whole national
system, strongly influenced people’s system of values, aspirations, attitudes and
behavior including their purchasing behavior. That is why it is important for
companies that enter Russian markets, especially the markets of premium goods, to
Summary and conclusion
250
investigate consumers born after the seventies, whose values, interests, notions and
world-view significantly differ from those of older generations.
4. Limitations
Several limitations should be mentioned regarding the methodology and empirical
study.
The suggested methodology based on the repertory grid technique has its limited
applicability to a relatively small sample of participants.
It is appropriate to apply it in exploratory comprehensive research aimed at
revealing new information about customers or products. “The grid is par
excellence a technique for measuring individual perceptions and any attempts
to generalize to collective perceptions begin to lose the advantages peculiar to
it” (Easterby-Smith et al., 1996). Therefore, when new things are discovered
they can be used further in quantitative analysis aimed at testing hypotheses
regarding the population.
Considering the empirical study, the major critique might be addressed to the
question of reliability of performed content analysis. The procedure of categorization
did not involve additional coders and, therefore, its accuracy might be perceived as
doubtful.
The peculiarity of the repertory grid interviewing and, especially, of the survey
conducted for this thesis is that the person who has carried out the interviews,
besides the information recorded in the grid form, often bears additional
observed or noticed information. According to Jankowicz (2004, p. 163),
agreement on the content with a collaborator may cause prevalence of the
collaborator’s reality and perception of the content. If the collaborator did not
participate in the design of the study, design of the grid form and the
interviewing it is recommended to use the researcher’s content analysis rather
than the collaborator’s.
The reliability check would of course be desirable but the fact is that the survey
was conducted in Russia, in the Russian language, with a very specific group of
customers, by a single Russian researcher and applying a rarely used
technique and, finally, the data was analyzed and presented in Germany for a
Summary and conclusion
251
German automaker, making the reliability check rather difficult. For this study it
would make sense if there was a second coder who is familiar with the
language, jargon, the cultural and mental peculiarities of the participants and
region as well as being familiar with the elicitation procedure technique and the
topic in question.
As a practical outcome presented for the company all interpretations and
findings appeared informative, useful and interesting due to the fact that the
researcher comes from the country the survey was conducted in and can,
therefore, understand and communicate much better the implied meanings
expressed by participants. From the scientific perspective, on the one hand, the
empirical study is first of all aimed at demonstration of the applicability of the
proposed methodology and its practical findings and inferences regarding
meanings remain secondary. On the other hand, the concept of the reliability of
the repertory grid technique is regarded as merely one aspect of validity in a
sense that the repertory grid should be done to explain causality and signify
changes occurring in a person’s construing systems but not to repeat the same
result (Fransella et al., 2004, p. 134). In turn, the validity, according to Kelly,
ultimately refers to the way in which a mode of understanding enables the
researcher or the respondent to take effective action.
The design of the repertory grid form developed for this study can evoke some
questions concerning, for example, the representativeness of the set of elements and
categories developed for elements elicitation or concerning design of other
components comprised in the form.
The repertory grid technique represents a very flexible tool which has almost no
rules to be followed, only recommendations which can help the researcher to
design the form which would be appropriate for his particular research. The
variety of procedures which can be chosen and the reasons and conditions for
their appropriateness are thoroughly described in this work and reasons for
choices made for the repertory grid design applied in this study are explained to
the extent it is permitted within the underlying personal construct theory.
Summary and conclusion
252
5. Potential for future research
The potential of the repertory grid technique and methods for analyzing its data is
great within market research, especially in the area of consumer motivation, and it
provides a large number of actions for its further development.
The proposed methodology applies analysis approaches which are not new and have
proved their appropriateness and usefulness. Therefore, apart from the aspect of
interpretation, research results strongly depend on the elicitation procedure. It is
important to investigate whether the potential changes in the findings in case of
repeated interviewing would be caused by the fact that a person represents a
permanent ‘form of motion’ and the consistence of a person’s construct systems over
time could be low or the changes are caused by the application of different elicitation
procedures. Thus, the question should be investigated whether the elicited set of
meanings, i.e., constructs and corresponding results, would differ depending on the
elicitation approach used. And if it differs, then which set of meanings should be
considered as the correct one?
One of the suggestions that could make interpretation and findings more accurate is
to perform quantitative analysis taking into account the weight of constructs
according to their relevance and importance. This would enable better differentiation
between personal meanings and, thereby, a more accurately structured personal
construing systems. Thus, an estimation measure for a construct's weight could be
suggested. However, it is important to bear in mind that quantitative analysis without
its clear understanding may cause a misleading interpretation. Quantitative analysis
has to be carefully approached, making sure that the researcher understands it and
can confidently explain it.
Further research can also address practical applicability of the repertory grid and
supportive methodology. Procedures and forms to interpret and perform information
regarding elements can be further suggested that would provide more findings about
consumers’ consideration sets and their attitudes towards particular products or
brands. This implies a different objective of the research and needs a different design
of the repertory grid form.
Besides, researchers can make use of triads to derive information about elements.
For example, constructs elicited by the triads containing particular elements can be
Summary and conclusion
253
compared or constructs elicited by the same triads can be compared within the group
of people.
These are several ideas to further extend the use of the repertory grid technique,
methods for its analysis and types of resulting outcome, leveraging the technique’s
flexibility and adaptability. Further research should take a closer look at the potential
for using repertory grids in the research of different markets and product categories,
attaining an insight into consumers’ related cognitive construing systems and
purchase motivation.
254
References
Adlwarth, W. (1983), Formen und Bestimmungsgründe prestigegeleiteten
Konsumverhaltens. München
Alexander, P.M. and van Loggerenberg, J.J. (2005), The repertory grid: "discovering"
a 50-year-old research technique, SAICSIT '05: Proceedings of the 2005 annual
research conference of the South African institute of computer scientists and
information technologists on IT research in developing countries
Antaki, C. (1988), Structures of belief and justification, in C. Antaki (Ed.), Analyzing
everyday explanation, London: Sage, pp. 60–73
Antaki, C. (1989), Structured causal beliefs and their defense in accounts of student
political action, Journal of language and social psychology, 8 (January), pp. 39–
48.
Assael, H. (1995), Consumer behavior and marketing action. South-Western College,
Publishing, Cincinnati, OH
Atherton, J.S. (2007), Learning and Teaching: Personal Construct Theory,
http://www.learningandteaching.info/learning/personal.htm#Formal, accessed:
25 March 2008
Austin, J.T. and Vancouver, J.B. (1996), Goal constructs in psychology: structure,
process and content, Psychological Bulletin, 120, pp. 338-375
Bagozzi, R.P. and Dabholkar, P.A. (1994), Customer recycling goals and their effect
on decisions to recycle: A means-end chain analysis, Psychology and
Marketing, 11, pp. 313–340
Bagozzi, R.P. and Dabholkar, P.A. (2000), Discursive psychology: An alternative
conceptual foundation to means-end chain theory, Psychology and Marketing,
17, pp. 535–586
Bagozzi, R.P., Bergami, M. and Leone, L. (2003), Hierarchical Representation of
Motives in Goal-Setting, Journal of Applied Psychology, 88, pp. 915-943
Baillie-Grohman, R. (1975), The use of a modified form of repertory grid technique to
investigate the extent to which deaf school leavers tend to use stereotypes,
Unpublished MSc dissertation, University of London, London.
Banister, P., Burman, E., Parker, I., Taylor, M. and Tindall, C. (1994), Qualitative
Methods in Psychology: A Research Guide. Milton Keynes: Open University
Press.
Bannister, D. and Bott, M. (1973), Evaluating the person, in P. Klein (Ed), New
Approaches to Psychological Medicine, Chichester: John Wiley
255
Barton, E.S., Walton, T. and Rowe, D. (1976), Using grid technique with the mentally
handicapped, in Slater, P. (Ed.), The Measurement of Intrapersonal Space by
Grid Technique. Volume 1. Explorations of Intrapersonal Space, London: John
Wiley & Sons
Bauer, H.H., Huber, F. and Keller, T. (1998), Using product-policy variants to increase
customer retention in the automotive industry - Results of an empirical study, in
Proceedings of the Australian and New Zealand Marketing Academy
Conference
Beail, N. (1985), Repertory grid technique and personal constructs: Applications in
clinical and educational settings. London: Croom Helm
Bearden, W.O. and Netemeyer, R.G. (1999), Handbook of Marketing Scales: Multi-
item Measures for Marketing and Consumer Behavior Research, Inc; 2 Sub.
Ed., Sage Publications
Beckmeier-Feuerhahn, S. (1998), Marktorientierte Markenbewertung. Eine
Konsumenten und unternehmensbezogene Betrachtung. Wiesbaden: Gabler
Bell R.C. (2003), http://www.pcp-net.org/encyclopaedia/editors.html, accessed: 24
November 2008
Bell, R.C. (1994), Using SPSS to analyse repertory grid data, Paper presented at the
29th Annual Conference of the Australian Psychological Society, Wollongong
Belz, O. (1994), Luxusmarkenstrategie, in Bruhn, M. (Ed.), Handbuch Markenartikel,
Band 1, Stuttgart, pp. 646-651
Berkmann, H.W. and Gilson, C. (1986), Consumer Behavior - Concepts and
Strategies. Boston: Kent Publishing
Bieri, J., Atkins, A.L., Briar, S., Leaman, R.L., Miller H. and Tripodi T. (1966), Clinical
and social judgment. New York: John Wiley & Sons
Björklund, L. (2005), The Repertory Grid Technique: Making Tacit Knowledge Explicit:
Assessing Creative work and Problem solving skills, in H. Middleton (Ed.),
Researching Technology Education: Methods and techniques, Sense
Publishers, Netherlands
Boeree G. (o. J.), Personality theories: George Kelly 1905-1967, in
http://webspace.ship.edu/cgboer/kelly.html, accessed: 16 January 2009
Bonarius, H. (1977), The interaction model of communication: Through experimental
research towards existential relevance, in A.W. Landfield, and J.K. Cole (Eds.),
Nebraska symposium on motivation 1976: Personal Construct Psychology, Vol.
26, Lincoln, NE/London: University of Nebraska Press
256
Botschen, G., Thelen, E.M. and Pieters, R. (1999), Using means-end structures for
benefit segmentation, European Journal of Marketing, 33, (1/2), pp. 38-58
Bougon, M.G. (1992), Congregate cognitive maps: a unified dynamic theory of
organization and strategy, Journal of Management Studies, Vol. 29, No. 3, pp.
367-389
Bourdieu, P. (1982), Die feinen Unterschiede: Kritik der gesellschaftlichen Urteilskraft.
Frankfurt am Main.
Brinkman, W.-P. and Love, S. (2006), Developing an instrument to assess the impact
of attitude and social norms on user selection of an interface design: a repertory
grid approach, ECCE '06: Proceedings of the 13th Eurpoean conference on
Cognitive ergonomics: trust and control in complex socio-technical systems
Burkitt, I. (1991), Social Selves: Theories of the Social Formation of Personality.
London: Sage
Caputi, P. and Keynes, N. (2001), A note on the stability of structural measures based
on Repertory Grids, Journal of Constructivist Psychology, 14, pp. 51-55
Caputi, P. and Reddy, P. (1999), A Comparison of Triadic and Dyadic Methods of
Personal Construct Elicitation, Journal of Constructivist Psychology, 12, pp.
253-264
Cassell, C., Duberley, J., Close, P. and Johnson P. (2000), Surfacing Embedded
Assumptions: Using Repertory Grid Methodology to Facilitate Organizational
Change, European Journal of Work and Organizational Psychology, 9(4), pp.
561-573
Chisnall, P.M., (1992), Marketing Research, 4th Ed., McGraw-Hill, London
Churchill, G.A.Jr. (1996), Basic Marketing Research, 3rd Ed., Chicago: The Dryden
Press
Clancey, W.J. (1989), Viewing knowledge bases as qualitative models, IEEE Expert,
Summer, pp. 9-23
Cohen, J.B. (1979), The structure of product attributes: defining attribute dimensions
for planning and evaluation, in Shocker, A.D. (Eds), Analytic Approaches to
Product and Marketing Planning, Marketing Science Institute, Cambridge, MA,
pp. 54-86
Coltheart, V. and Evans, J. (1982), An investigation of semantic memory in individuals,
Memory and Cognition, 9(5), pp. 524-532
257
Coombe G. (2006), An introduction to Principal Component Analysis and Online
Singular Value Decomposition, PhD Dissertation Draft,
http://www.cs.unc.edu/~coombe/research/phd/, accessed: 26 September 2008
DeCharms, R. and Shea, D.J. (1976), Enhancing Motivation: Change in the
classroom. New York.
Dholakia, N. and Arndt, J. (1985), Changing the Course of Marketing: Alternative
Paradigms for Widening Marketing Theory. JAI Press, Greenwich, CT
Dichter, E. (1964), Handbook of Consumer Motivation - The Psychology of the World
of Objects. New York: McGraw
Diez, W. (2001), Herausforderungen und Perspektiven im Premiummarkt für
Automobile, Forschungsbericht Nr. 22 des Instituts für Automobilwirtschaft (IFA)
an der Hochschule Nürtingen-Geislingen
Dijkshoorn, L., Towner, K.J. and Struelens, M. (2001), New approaches for the
generation and analysis of microbial typing data. Elsevier
Dillon, A. and McKnight, C. (1990), Towards a classification of text types: a repertory
grid approach, International Journal of Man-Machine Studies, 33, pp. 623-636
Dröge, C. and Calantone, R. (1984), Assumptions underlying the metatheoretical
debates regarding methods and scientific theory construction, Advances in
Consumer Research, Vol. 11, pp. 5-9
Duck, S.W. (1994), Meaningful Relationships. Sage, London
Earl, P.E (1986), Lifestyle Economics: Consumer Behaviour in a Turbulent, World,
Wheatsheaf: Brighton
Easterby-Smith, M. (1980), How to Use Repertory Grids in HRD, Journal of European
Industrial Training, 4/2, pp. 2-32
Easterby-Smith, M., Thorpe R. and Holman D. (1996), Using repertory grids in
management, Journal of European Industrial Training, 20/3, pp. 3-30
Eden, C. and Jones, S. (1984), Using Repertory Grids for Problem Construction,
Journal of the Operational Research Society, Vol. 35, No. 9, pp. 779-790
Epting, F.R., Probert, J.S. and Pittman, S.T. (1993), Alternative strategies for construct
elicitation: Experimenting with experience, International Journal of Personal
Construct Psychology, 6, pp. 79-98
Esch, F.-R., Redler, J. and Langner, T. (2003), Promotional Efficiency and the
Interaction between Buying Behavior Type and Product Presentation Format -
Evidence form an Exploratory Study, in ANZMAC 2003 Conference
Proceedings, pp. 1838 - 1845
258
Feemers, M. (1992), Der demonstrativ aufwendige Konsum - Eine theoretisch-
empirische Untersuchung, Europäische Hochschulschriften, Reihe 6:
Psychologie Bd. 386, Frankfurt am Main: Lang
Feixas, G. and Cornejo-Alvarez, J.M. (1996), A Manual for the Repertory Grid: Using
GRIDCOR Programme (Version 2.0), Barcelona: Paidos
Fennell, G. (1975), Motivation Research Revisited, Journal of Advertising Research,
Vol. 15 (June), pp. 23-28
Fennell, G. (1978), Consumers' perceptions of the product-use situation, Journal of
Marketing, (April), pp. 38-47
Fewtrell, D. and O'Connor, K. (1995), Clinical Phenomenology and Cognitive
Psychology. Routledge, London
Fichter, J.-H. (1968), Grundbegriffe der Soziologie. Wien
Fisher, J. and Savage, D. (1999), Introduction to Personal Construct Psychology,
http://www.businessballs.com/personalchangeprocess.htm, reprinted from
Fisher, J. and Savage, D. (Eds), Beyond Experimentation into Meaning, EPCA
Publications, Farnborough
Ford, K. and Petry, F. (1989), Knowledge acquisition from repertory grids using a logic
of confirmation, ACM SIGART Bulletin, Issue 108
Foxall, G.R. (1998), Intention versus context in consumer psychology, Journal of
Marketing Management, 14, pp. 29-62
Fransella, F. and Bannister, D. (1977), A manual for repertory grid technique. London:
Academic Press
Fransella, F., Bell, R. and Bannister, D. (2004), A Manual for Repertory Grid
Technique. John Wiley & Sons, Ltd., 2nd Edition
Fransella, F., Jones, H. and Watson, J. (1988), A range of applications of PCP within
business and industry, in Fransella, F. and Thomas, L. (Eds), Experimenting
with Personal Construct Psychology, Routledge, London, pp. 405-417.
Fromm, M (o. J.), Grid Software, European Personal Construct Association,
http://www.epca-net.org/Default.aspx?pageId=42152, accessed: 22 October
2008
Frost, W.A. (1982), Book review: Business Applications of Repertory Grid, Journal of
the Market Research Society, Vol. 24, No.1, pp. 81-83
Gaines B.R. and Shaw M.L.G. (2005), Rep IV: Manual for Research Version 1.12,
http://repgrid.com/RepIV/RepIVManual/, accessed: 23 October 2007
259
Geistfeld, L.V., Sproles, G.B. and Badenhop, S.B. (1977), The Concept and
Measurement of a Hierarchy of Product Characteristics, in Perreault W.D.Jr.
(Ed.), Advances in Consumer Research, Vol. 4, Atlanta: Association for
Consumer Research, pp. 302-307.
Gengler, C., Mulvey, M. and Oglethorpe, J. (1999), A Means-End Analysis of Mothers'
Infant Feeding Choices, Journal of Public Policy and Marketing, 18(2), pp. 172-
188
Gordon, W. and Langmaid, R. (1988), Qualitative Market Research: A practitioner's
and buyer's guide. Gower
Green, B. (2004), Personal construct psychology and content analysis, Personal
Construct Theory and Practice, 1, pp. 82-91
Gries, R., (2005), about Ernest Dichter,
http://www.vmoe.at/show_content.php?sid=198&language=en, accessed: 15
May 2008
Gröppel-Klein, A. and Königstorfer, J. (2007), Projektive Verfahren, in Buber, R.,
Holzmüller, H. (Eds.), Qualitative Marktforschung - Theorie, Methode, Analyse,
Wiesbaden: Gabler, pp. 537-553
Grunert, K.G. and Grunert, S.C. (1995), Measuring subjective meaning structures by
laddering method: theoretical considerations and methodological problems,
International Journal of Research in Marketing, Vol. 12, pp. 209-225
Grunert, K.G., Sorensen, E., Johansen, L.B. and Nielsen, N.A., (1995), Analysing food
choice from a means-end perspective, European Advances in Consumer
Research, Vol.2, pp. 366-371
Gutman, J. (1982), A means-end chain model based on consumer categorization
processes, Journal of marketing, 46, pp. 60-72
Gutman, J. (1991), Exploring the nature of linkages: consequences and values,
Journal of Business Research, Vol. 22, pp.143-148
Gutman, J. and Reynolds, T.J. (1979), An investigation of the levels of cognitive
abstraction utilized by consumers in product differentiation, in J. Eighmy (Ed.),
Attitude research under the sun, Chicago: American Marketing Association, pp.
128-152
Hadley G.S. (1996), Bridging the cross-cultural gap with personal construct repertory
grids, Dissertation, University of Birmingham
260
Hagans, C.L., Neimeyer, G.J. and Goodholm, C.R. (2000), The Effect of Elicitation
Methods on Personal Construct Differentiation and Valence, Journal of
Constructivist Psychology, 13, pp. 155-173
Hallsworth, A.G. (1988), Repertory grid methodology and the analysis of group
perceptions in retailing, International Journal of Retailing, Vol. 3, No. 4, pp. 43-
54
Harper, D. (2001), The Online Etymology Dictionary, www.etymonline.com, accessed:
24 May 2008
Hassard, J. (1987), FOCUS as a phenomenological technique for job analysis: its use
in multiple paradigm research, International Journal of Man-Machine Studies,
27, pp. 413-433
Hassenzahl, M. and Trautmann, T. (2001), Analysis of web sites with the repertory grid
technique, in Proceedings of the CHI 2001 Conference on Human Factors in
Computing. Extended abstracts on Human factors in computing systems, pp.
167-168, New York: ACM Press, Addison-Wesley
Haugtvedt, C.P., Herr, P.M. and Kardes, F.R. (2008), Handbook of Consumer
Psychology. CRC Press
Haury, K. (2007), Geltungsmotivierte Konsum: Befunde und Empfehlungen für eine
kundenorientierte, bedarfsgerechte Produktkonzeption, Vol. 748 der
Europäischen Hochschulschriften, Dissertation, Frankfurt am Main
Hausruckinger G. and Helm R. (1996), Die Bedeutung des Country-of-Origin Effektes
vor dem Hintergrund der Internationalisierung von Unternehmen, Marketing
ZFP, Nr. 4, pp. 267-278
Heath, T. (1992), The reconciliation of humanism and positivism in the practice of
consumer research, Journal of the Academy of Marketing Science, 20(2), pp.
107-118
Heckhausen, H., Schmalt, H. -D. and Schneider, K. (1985), Achievement motivation in
perspective. New York: Academic Press
Herm, S. and Gall, K. (2008), Zur Messung markenspezifischer Risikowahrnehmungen
mit der Means-End-Analyse, Marketing ZFP, Nr. 3 (30. Jg.), pp. 161-174
Hessenzahl, M. and Wessler, R. (2000), Capturing Design Space From a User
Perspective: The Repertory Grid Technique Revisited, International Journal of
Human-Computer Interaction, 12 (3 and 4), pp. 441-459
261
Highlen, P.S. and Finley, H.C. (1996), Doing qualitative analysis, in F. T. L. Leong and
T. A. Austin (Eds.), The Psychology Research Handbook: A Guide for Graduate
Students and Research Assistants, Sage Publications, pp. 177-192
Hinkle, D. (1970), The game of personal constructs, in D. Bannister (Ed.),
Perspectives in personal construct theory, London: Academic Press, pp. 91-110
Hinkle, D.N. (1965), The change of personal constructs from the viewpoint of a theory
of construct implications, Unpublished PhD thesis, Ohio State University,
Columbus, OH
Hogeweg, P. (1976), Topics in biological pattern analysis, PhD Thesis, University of
Utrecht
Honey, P. (1979), The repertory grid in action, Industrial and Commercial Training,
Nov 79, 11(11): pp. 358--369; 407--414; 452--459
Huber, F., Beckmann, S.C. and Hermann, A. (2004), Means-end analysis: does the
affective state influence information processing style?, Psychology and
Marketing, Vol. 21, No.9, pp.715-37
Hunt, S.D., (1994), On Rethinking Marketing: Our Discipline, Our Practice, Our
Methods, European Journal of Marketing, Vol. 28 (3), pp. 13-25
Jacob F. (2009), Marketing: Eine Einführung für das Master-Studium. Kohlhammer,
Stuttgart
Jankowicz, A.D. (1990), Applications of personal construct psychology in business
practice, Advances in Personal Construct Psychology, Vol. 1, pp. 257-87
Jankowicz, A.D. (2001), Why does subjectivity makes au nervous?: Making the tacit
explicit, Journal of Intellectual Capital, Vol. 2, No. 1, pp. 61-73
Jankowicz, A.D. (2004), The Easy Guide to the Repertory Grid. Wiley, England
Jankowicz, D. and Thomas, L.F. (1982), An Algorithm for the Cluster Analysis of
Repertory Grids in Human Resource Development, Personell Review, No. 4
(Vol. 11), p. 15-22
Kahle, L.R., Beatty, S.E. and Homer, P.M. (1986), Alternative Measurement
Approaches to Consumer Values - the List of Values (Lov) and Values and Life-
Style (Vals), Journal of Consumer Research, 13 (3), pp. 405-409
Kamakura, W.A. and Novak, T.P. (1992), Value-System Segmentation: Exploring the
Meaning of LOV, Journal of Consumer Research, 19 (June), pp. 119-132
Karmasin, H. (1998), Produkte als Botschaften, 2. Aufl., Wien
262
Kassarjian, H.H. (1994), Scholarly Traditions and European Roots of American
Consumer Research, in Laurent, G. Lilien, G.L. and Pras, B. (Eds), Research
Traditions in Marketing, Kluwer Academic, Boston, MA, pp. 265-279
Kavanagh, D. (1994), Hunt versus Anderson: Round 16, European Journal of
Marketing, Vol. 28, No. 3, pp. 26-41
Kelly, G.A. (1955/1991), The Psychology of Personal Constructs, Vol. 1, 2. Norton,
New York (2nd printing: 1991, Routledge, London, New York)
Kelly, G.A. (1969), A mathematical approach to psychology, in Maher, B. (Ed.), Clinical
Psychology and Personality: The Selected Papers of George Kelly, New York:
Wiley, pp. 94-113
Kiefer, M. (2008), Bewusstsein, in Müsseler, J. (Hrsg.), Allgemeine Psychologie. 2.
Aufl. Berlin, pp. 154-188
Kihlstrom, J.F. and Cantor, N. (1984), Mental representations of the self, Advances in
experimental social psychology, Vol. 17, pp. 1-47
Kluth, H. (1957), Sozialprestige und Sozialer Status. Stuttgart: Enke
Krippendorf, K. (2004), Content analysis: An introduction to its methodology, 2nd Ed.,
Thousand Oaks, CA: Sage Publications
Kroeber-Riel, W., Weinberg, P. and Gröppel-Klein, A. (2009), Konsumentenverhalten,
9. Aufl., München: Vahlen
Lass, D. (2002), Kundenwünsche analysieren und verstehen: Theorien, Methoden und
Anwendungsbeispiele für Wissenschaft, Marktforschung- und Management-
praxis, Berlin: Berliner Beiträge zur Marketingforschung
Lin, C. (2002), Attribute-consequence-value linkages: A new technique for
understanding customer's product knowledge, Journal of Targeting
Measurement and Analysis for Marketing, 10, pp. 339-352
Lucerna, C. (1999), Das Ritual um die Marke, in Brandmeyer, K. and Deichsel, A.
(Hg.), Jahrbuch Markentechnik 2000/2001, Frankfurt am Main, pp. 359-387
Mahoney, M.J. and Gabriel, T. (1987), Psychotherapy and the cognitive sciences: An
evolving alliance, Journal of Cognitive Psychotherapy, 1, pp. 39-59
Marsden, D. (1997), Product construct systems: a personal construct psychology of
consumer behaviour, Manchester School of Management, UMIST, Manchester,
unpublished doctoral thesis
Marsden, D. and Littler, D. (2000a), Repertory grid technique: An Interpretive research
framework, European Journal of Marketing, Vol.34, No.7, pp. 816-834
263
Marsden, D. and Littler, D. (2000b), Exploring consumer product construct systems
with the repertory grid, Qualitative Market Research: An International Journal,
Vol.3, No.3, pp.127-144
Maslow, A. (1954), Motivation and Personality. Harper and Row, New York
Mason, R.S. (1981), Conspicuous Consumption: A Study of Exceptional Consumer
Behavior. St Martin’s Press, New York
McClelland, D.C. (1987), Human motivation. New York: Cambridge University Press
McKnight, C. (1981), Subjectivity in sentencing. Law and Human Behavior, 5(2/3), pp.
141-147
McKnight, C. (2000), The Personal Construction of Information Space, Journal of the
American Society for Information Science, Vol. 51(8), pp. 730-733
Metzler, A.E. and Neimeyer, G.J. (1988), Vocational hierarchies: how do we count the
ways?, International Journal of Personal Psychology, Vol. 1, pp. 205-17
Mick, D.G. and Buhl, C. (1992), A meaning-based model of advertising experiences,
Journal of Consumer Research, Vol. 19, pp. 317-338
Mook, D.G. (1987), Motivation. New York: W.W. Norton
Mulvey, M., Olson, J., Celsi, R. and Walker, B. (1994), Exploring the Relationship
Between Means-End Knowledge and Involvement, Advances in Consumer
Research, Vol. 21, pp. 51-57
Murray, J.B. and Evers D.J. (1989), Theory Borrowing and Reflectivity in
Interdisciplinary Fields, Advances in Consumer Research, Vol.16, pp. 647-652
Neimeyer, G.J. (1992), Personal constructs in vocational structure: A critique of poor
reasons, in R. Neimeyer and G. Neimeyer (Eds.), Advances in Personal
Construct Psychology, Vol. 2. Greenwich, CT: JAI Press
Neimeyer, G.J. (1993), Constructivist assessment. Sage, New York, NY
Neimeyer, G.J. (2002), Towards reflexive scrutiny in repertory grid methodology,
Journal of. Constructivist Psychology, 15, pp. 89–94
Neimeyer, R.A. (1993), An appraisal of constructivist psychotherapies, Journal of
Consulting and Clinical Psychology, 61, pp. 221-234
Neimeyer, R.A. and Mahoney, M.J. (1995), Constructivism in psychotherapy.
Washington, D.C.: American Psychological Association
Neimeyer, R.A. and Neimeyer, G.J. (2002), Advances in Personal Construct
Psychology. New York: Praeger
264
Novak, T.P. and MacEvoy, B. (1990), On Comparing Alternative Segmentation
Schemes: The List of Values (LOV) and Values and Life Styles (VALS), Journal
of Consumer Research, 17, (June), pp. 105-109.
o. V., 2007, Enquire Within software: Used in market research,
http://www.enquirewithin.co.nz/market.htm, accessed: June 2007
o. V., 2008, Wikipedia, http://en.wikipedia.org/wiki/Aspirational_brand, accessed: May
2008
o. V., 2009, http://dictionary.reference.com/browse/premium, accessed: January 2009
O'Guinn, T.C. and Wells, W.D. (1989), Subjective Discretionary Income, Marketing
Research: A Magazine of Application and Practice, 1, pp. 32-41
Olson, J.C. and Reynolds, T.J. (1983), Understanding consumers' cognitive structures:
implications for advertising strategy, in Percy, L., Woodside, A.G. (Eds),
Advertising and Consumer Psychology, Lexington Books, Lexington, MA, pp.
77-90
Oosterhuis, F., Rubik, F. and Scholl, G. (1996), Product policy in Europe: new
environmental perspectives. Springer
Orth, U. and Krska, P. (2002), Quality Signals in Wine Marketing, International Food
and Agribusiness Management Review, 4, pp. 385-397
Ozanne, J.L. and Anderson, L. (1989), Exploring diversity in consumer research, in
Hirschman, E. (Eds), Interpretive Consumer Research, ACR, Provo, UT, pp. 1-9
Paulssen, M. (1999), Individual Goal Hierarchies as Antecedents of Market Structure.
Wiesbaden: Gabler Edition Wissenschaft
Peter, J.P. and Olson, J.C. (1990), Consumer Behavior and Marketing Strategy, 2nd
ed., Irwin, Homewood, IL
Peter, J.P. and Olson, J.C. (1994), Understanding Consumer Behavior, Burr Ridge, IL:
Richard D. Irwin
Peter, J.P., Olson, J.C. (2004), Consumer Behavior and Marketing Strategy, 7nd Ed.,
Irwin Professional Pub
Petzold, M. (2006), Persönlichkeitsorientiertes Management von Automobilmarken -
Messung der Persönlichkeit von Automobilmarken und Steuerung durch
Markentransfers, Dissertation, Universität Mannheim
Pieters, R. (1993), A Control View on the Behavior of Consumers: Turning the
Triangle, in Van Raaij, W.F. and Bamossy, G.J. (Eds), European Advances in
Consumer Research, Vol. 1, pp. 507-512
265
Pieters, R., Baumgartner, H. and Allen, D. (1995), A means–end chain approach to
consumer goal structures, International Journal of Research in Marketing, 12,
pp. 227–244
Pincus, J.D. (2004), The Consequences of Unmet Needs: The Evolving Role of
Motivation in Consumer Research, Journal of Consumer Behaviour, June,
Volume 4, pp. 375-387
Puca, R.M. and Langens, T.A. (2008), Motivation, Volition and Emotion, in Müsseler,
J. (Hrsg.): Allgemeine Psychologie. 2 Aufl. Berlin, pp. 191-229
Reynolds, T.J. and Gutman, J. (1984), Laddering: Extending the repertory grid
methodology to construct attribute-consequence-value hierarchies, in Pitts R.E.
and Woodside A.G. (Eds.) Personal Values and Consumer Psychology,
Lexington Books, Toronto, pp. 155-167
Reynolds, T.J. and Gutman, J. (1988), Laddering theory, method, analysis and
interpretation, Journal of Advertising Research, 28(1), pp. 11-31
Reynolds, T.J. and Whitlark, D.B. (1995), Applying laddering data to communications
strategy and advertising practice, Journal of Advertising Research, Vol. 35,
No.4, pp. 9-17
Rice, C. (1993), Consumer Behavior: Behavioral Aspects of Marketing. Butterworth
Heineman, London
Rinsche, G. (1961), Der aufwendige Verbrauch, in Kreikebaum, H. and Rinsche, G.,
Das Prestigemotiv in Konsum und Investition, Berlin: Duncker & Humblot, pp.
105-221
Rokeach, M. (1973), The nature of human values. New York: Free Press
Rosengarten, P.G. and Stürmer, C.B. (2005), Premium power. Das Geheimnis des
Erfolgs von Mercedes-Benz, BMW, Porsche und Audi, 2. Aufl., Wiley-VCH,
Weinheim
Rossister, J.R. and Percy, L. (1991), Emotions and Motivation in Advertising,
Advances in Consumer Research, Vol.18, pp. 100-110
Rundell, M. and Fox, G. (2007), Macmillan English Dictionary: For Advanced Learners,
2nd Ed., Macmillan Publishers Limited, Oxford
Rychlak, J.F. (1981), Introduction to personality and psychotherapy. A theory-
construction approach, 2nd Ed., Boston: Houghton Mifflin
Ryle, A. (1976), Some clinical applications of grid technique, in Slater, P. (Ed.), The
Measurement of Intrapersonal Space by Grid Technique. 2 vols., London: John
Wiley
266
Saegert, J. and Fennell, G. (1991), Qualitative Research In The Textbooks: A Review,
Advances in Consumer Research, Vol. 18, The Association for Consumer
Research, pp. 262-270
Salmon, P. (1976), Grid Measures with child subjects, in Slater, P. (Ed.), The
Measurement of Intrapersonal Space by Grid Technique, Vol. 1, London: John
Wiley
Sampson, P. (1972), Using the repertory grid test, Journal of Marketing Research, Vol.
9, pp. 78-81
Sanjiv K. Bhatia and Qi Yao, (1993), A new approach to knowledge acquisition by
repertory grids, CIKM '93: Proceedings of the second international conference
on Information and knowledge management, Washington, D.C., pp. 738-740
Schulze, G. (1992), Die Erlebnisgesellschaft: Kultursoziologie des Gegenwart.
Frankfurt am Main
Schuster, H.-W. (1994), Prestigegeleitetes Konsumentenverhalten – Teil 1: Typologie
des Distinktkonsums, in Jahrbuch der Absatz- und Verbrauchsforschung, Nr. 2,
pp. 108-120
Schwartz, S. H. (1992), Universals in the Content and Structure of Values: Theory and
Empirical Tests in 20 Countries, in M. Zanna (Ed.), Advances in Experimental
Social Psychology, Vol. 25, New York: Academic Press, pp. 1-65
Senior, B. (1996), Team performance: using repertory grid technique to gain a view
from the inside, Journal of Managerial Psychology, Vol. 11, No. 3, pp. 26-32
Sharp, H., Hall, T., Baddoo, N. and Beecham, S. (2007), Exploring motivational
differences between software developers and project managers, ESEC-FSE
companion '07: The 6th Joint Meeting on European software engineering
conference and the ACM
Shaw P.J.A. (2003), Multivariate statistics for the Environmental Sciences, Hodder-
Arnold, London
Shaw, M.L.G. (1980), On Becoming a Personal Scientist: Interactive Computer
Elicitation of Personal Models of the World. Academic Press, London
Shaw, M.L.G. (1981), Recent Advances in Personal Construct Technology. London:
Academic Press
Shaw, M.L.G. (1994), Methodology for sharing personal construct systems, Journal of
Constructivist Psychology, 7, pp. 35-52
267
Shaw, M.L.G. and Gaines, B.R. (1987), KITTEN: Knowledge initiation and transfer
tools for experts and novices, International Journal of Man-Machine Studies, 27,
pp. 251-280
Shaw, M.L.G. and Gaines, B.R. (1992), Kelly's geometry of psychological space and
its significance for cognitive modeling, The New Psychologist, pp. 23-31
Shaw, M.L.G. and Gaines, B.R. (2005), Expertise and expert systems: emulating
psychological processes, in Fransella, F. (Ed.), The Essential Practitioner's
Handbook of Personal Construct Psychology. Wiley, Chichester, UK, pp. 87-94
Shaw, M.L.G. and McKnight, C. (1981), Think Again: Personal Problem-Solving and
Decision-Making. Prentice-Hall
Shrivastava, P. (1985), Rigor and practical usefulness of research in strategic
management, Strategic Management Journal, Vol. 8, pp.77-92
Simon, A.F. and Xenos, M. (2004), Dimensional reduction of word-frequency data as a
substitute for intersubjective content analysis, Political Analysis, 12, pp. 63-75
Slater, P. (1964), The Principal Components of a Repertory Grid. London; Vincent
Andrews
Slater, P. (1976), The measurement of intrapersonal space by grid technique, Vol. 1
Explorations of intrapersonal space. London: Wiley
Slater, P. (1977), The measurement of intrapersonal space by grid technique. Vol. 2
Dimensions of intrapersonal space. London: Wiley
Smith, H.J. (2000), The reliability and validity of structural measures derived from
Repertory Grids, Journal of Constructivist Psychology, 13, pp. 221-230
Smith, J.A. (1995), Semi-structured interviewing and qualitative analysis, in J. Smith,
R. Harre and L.V. Langenhove (Eds.) Rethinking Methods in Psychology.
London: Sage, pp. 9-26
Sohlberg, S. (2006), Concepts of the unconscious, in Steiner R., The Philosophy of
Freedom, Rudolf Steiner Press,
http://www.philosophyoffreedom.com/index.php?q=node/693 accessed on 22
September 2008
Solas, J. (1991), A constructive alternative to family assessment, Journal of Family
Therapy, 13, pp. 149-169
Solomon, M. (2004), Consumer Behavior, 6 Ed., Prentice Hall
Soukhanov, A.H., Jost, D.A., Ellie, K. and Severynse, M. (2000), The American
Heritage Dictionary of the English Language, 4 Ed.
268
Steed, A. and McDonnell, J. (2003), Experiences with Repertory Grid Analysis for
Investigating Effectiveness of Virtual Environments, Presence 2003, The 6th
Annual International Workshop on Presence, Aalborg, Denmark, October 6th-
8th, 2003.
Stewart, D.W. (1990), Cognition and personal structure: computer access and
analysis, Journal of Marketing Research, Vol. 27, pp. 112-121
Stewart, V., Stewart, A. and Fonda, N. (1981), Business Applications of Repertory
Grid. McGraw-Hill, London
Strauss, C. (1992), Models and motives, in Roy D'Andrade and Claudia Strauss
(Eds.), Human Motives and Cultural Models. Cambridge: Cambridge University
Press, pp. 1-20
Sullivan, M and Adcock, D (2002), Retail marketing, Thomson, London, UK
Thomas, L. and Shaw, M.L.G. (1976), Focus: A manual. Uxbridge, England: Centre for
the Study of Human Learning, Brunel University
Trommsdorff, V. (2009), Konsumentenverhalten, 7. Auflage, Kohlhammer, Stuttgart
Valette-Florence, P. and Rapacchi, B. (1991), Improvements in means-end chain
analysis: using graph theory and correspondence analysis, Journal of
Advertising Research, 31, pp. 30-45
Van Ooyen, A. (2001), Theoretical aspects of pattern analysis, in Dijkshoorn, L.,
Towner, K.J., and Struelens, M. (Eds.), New Approaches for the Generation and
Analysis of Microbial Typing Data, Amsterdam: Elsevier, pp. 31-45
Veblen, T. (1981 [1899]), Theorie der feinen Leute. Eine ökonomische Untersuchung
der Institutionen, München
Veludo-de-Oliveira, T.M., Ikeda, A.A. and Campomar, M.C. (2006), Discussing
laddering application by the means-end chain theory, The Qualitative Report,
11(4), pp. 626-642
Verlinden J.C. and Coenders M.J.J. (2000), Qualitative usability measurement of
websites by employing the repertory grid technique, CHI '00 extended abstracts
on Human factors in computing systems, The Hague, The Netherlands, pp. 143
- 144
Vershofen, W. (1940), Handbuch der Verbrauchsforschung. Bd. 1: Grundlegung.
Berlin.
Vershofen, W. (1959), Die Marktentnahme als Kernstück der Wirtschaftsforschung,
Berlin.
269
Vigneron, F. and Johnson, L. W. (1999), A review and a conceptual framework of
prestige-seeking consumer behavior, Academy of Marketing Science Review, 1,
pp. 1-15
Vrience, M. and Hofstede, F.T. (2000), Linking attributes, benefits, and consumer
values, Marketing Research, 12(3), pp. 4-10
Vygotsky, L.S. (1989), Thought and Language. Cambridge, Massachusetts: MIT Press
Walker, B., Celsi, R., Olson, J. (1987), Exploring the structural characteristics of
consumers' knowledge, Advances in Consumer Research, Vol. 14, pp. 17-21
Walker, B.A. and Olson, J.C. (1991), Means-end chains: connecting products with self,
Journal of Business Research, Vol. 22, No.2, pp.111-19
Wansink, B. (2000), New techniques to generate key marketing insights, Marketing
Research, Vol. 12(2), pp. 28-36
Wansink, B. (2003), Using Laddering to Understand and Leverage a Brand’s Equity,
Qualitative Market Research, Vol. 6(2), pp. 111-118
Wertsch, J.V. (1985), Culture, Communication and Cognition: Vygotskian
Perspectives. Cambridge: Cambridge University Press
Wind, Y. (1978), Issues and advances in segmentation research, Journal of Marketing
Research, Vol. 15, pp. 317-337
Wiswede, G. (1973), Motivation und Verbraucherverhalten, München
Wiswede, G. (1995), Einführung in die Wirtschaftspsychologie, 2. Aufl., München/
Basel
Wright, R.P. (2004), Mapping cognitions to better understand attitudinal and behavioral
responses in appraisal research, Journal of Organizational Behavior, 25, pp.
339-374
Wright, R.P. and Cheung, F.C.K.K. (2007), Articulating Appraisal System
Effectiveness Based On Managerial Cognitions, Personnel Review, Vol. 36,
Issue 1/2, pp. 206-230
Yadov, V.A. (1991), Strategies and methods of qualitative analysis, Sociology, pp.14-
31 (Russian periodical: Ядов, В.А. Стратегия и методы качественного
анализа данных// Социология 4М. 1991. №1. стр. 14-31)
Yang, S., Allenby G.M. and Fennell G. (2002), Modeling Variation in Brand
Preference: The Roles of Objective Environment and Motivating Conditions,
Marketing Science, 21(1), pp. 14-31
Yorke, D.M. (1985), Administration, analysis and assumption: Some aspects of
validity, in N. Beail (Ed.), Repertory grid technique and personal constructs:
270
Applications in clinical and educational settings. London: Croom Helm, pp. 383-
398
Yorke, D.M. (1989), The intolerable wrestle: words, numbers, and meanings,
International Journal of Personal Construct Psychology, 2, pp. 65-76
Young, S. and Feigin, B. (1975), Using the Benefit Chain for Improved Strategy
Formulation, Journal of Marketing, 39, pp. 72-74
Young, S., Ott, L. and Feigin, B. (1978), Some Practical Considerations in Market
Segmentation, Journal of Marketing Research, Vol. 15, No. 3, pp. 405-412
Zanoli, R. and Naspetti, S. (2002), Consumer motivations in the purchase of organic
food: A means-end approach, British Food Journal, Vol. 104(8), pp. 643-653
Zinkhan, G.M and Biswas, A. (1988), Using the repertory grid to assess the complexity
of consumers' cognitive structures, Advances in Consumer Research, Vol. 14,
pp. 493-497
271
Appendix
Appendix 1: Car classification
Car classification
American English
British English Segment Euro NCAP Examples
Microcar Microcar, Bubble car
- - BMW Isetta, Smart Fortwo
- City car A-segment Daewoo Matiz, Renault Twingo, Toyota Aygo, VW Lupo
Subcompact car Supermini B-segment
Supermini Hyundai Accent, Ford Fiesta, Opel Corsa, Suzuki Swift
Compact car Small family car C-segment Small family car
Ford Focus, Toyota Corolla, Opel Astra, VW Golf
Mid-size car Large family car Ford Mondeo, Opel Vectra, Toyota Avensis, VW Passat
Entry-level luxury car
Compact executive car
D-segment Large family car Audi A4, BMW 3 Series, Mercedes C-Class,
Volvo S60
Full-size car Ford Crown Victoria, Holden Commodore, Opel Omega, Chrysler 300C
Mid-size luxury car
Executive car E-segment Executive car
Audi A6, BMW 5 Series, Mercedes E-Class
Full-size luxury car
Luxury car F-segment - Audi A8, BMW 7 Series, Mercedes S-Class
Sports car Sports car - Chevrolet Corvette C6, Porsche 911
Grand tourer Grand tourer - Jaguar XK, Maserati GranTurismo
Supercar Supercar - Bugatti Veyron, Ferrari Enzo, Lamborghini Gallardo
Convertible Convertible - Mercedes CLK, Volvo C70, VW Eos
Roadster Roadster
S-segment
Roadster sports
Audi TT, BMW Z4, Porsche Boxster, Lotus Elise
- Leisure activity vehicle
Peugeot Partner, Škoda Roomster
- Mini MPV Opel Meriva, Fiat Idea, Ford Fusion
Compact minivan
Compact MPV, Midi MPV
Small MPV
Mazda5, Opel Zafira, Renault Scénic, VW Touran
Minivan Large MPV
M-segment
Large MPV Ford Galaxy, Toyota Previa, Renault Espace
Mini SUV Mini 4x4 Daihatsu Terios, Mitsubishi Pajero iO , Suzuki Jimny
Compact SUV Compact 4x4
Small Off-Road 4x4 BMW X3, Ford Escape, Honda CR-V, Subaru
Forester, Toyota RAV4
- Coupé SUV - Isuzu VehiCROSS, SsangYong Actyon, BMW X6
Mid-size SUV Ford Explorer, BMW X5, Jeep Grand Cherokee, VW Touareg
Full-size SUV
Large 4x4
J-segment
Large Off-Road 4x4 Cadillac Escalade, Chevrolet Suburban, Range
Rover, Toyota Land Cruiser
Pickup truck Pick-up - Pick-up Ford F-150, Mitsubishi Triton/L200, Nissan Navara
(Source: Wikipedia: http://en.wikipedia.org/wiki/Car_classification)
272
Appendix 2: Minkowski metric
k-Minkowski metric measures the distance between two points.
The Minkowski (power) metric from a point p to a point pi in Rm is defined by
dLk(p, pi) = [j=1∑m | xj - xij |
k]1/k Equation 1
where (x1, x2, . . . , xm) and (xi1, xi2, . . . , xim) are the Cartesian coordinates of p and
pi, respectively. Customarily the symbol Lk is used for the Minkowski metric, where k
refers to the degree of the power.
The parameter k varies in the range of 1 ≤ k < ∞.
i. If k = 1, equation 1 becomes
dL1(p, pi) = j=1∑m | xj - xij |
which is called the Manhattan metric, the city-block distance or the taxi-cab
distance. Here, the distance between two points measured along axes at right
angles.
ii. If k = 2, equation 1 becomes
dL2(p, pi) = [j=1∑m | xj - xij |
2]1/2
which is called the Euclidean distance. The straight line distance between two
points is measured.
iii. If k = ∞, the equation 1 becomes
dL∞(p, pi) = [maxj { | xj - xij | j Im }
which is called the Supermum metric or dominance metric.
(Source: Muhammed, R.B. (o. J.), Computational geometry,
http://www.personal.kent.edu/~rmuhamma/Compgeometry/MyCG/CG-
Applets/Minkowski/vorMinkow.htm, accessed: 15 November 2007)
273
Appendix 3: Repertory grid form and questionnaire developed for interviews
(translated from Russian)
Combination
Em
erg
en
t p
ole
(1
)
sdfg
sgd
sg
sdg
sg
sg
sdf
sdg
sdfg
Imp
licit
po
le (
5)
1 141312111098765431 2
274
Laddering down “What does it mean?”
Preferred pole Construct
№ Preferred pole
Construct №
275
Laddering up “Why is it important for you?”
Preferred pole Construct
№ Preferred pole
Construct №
276
Resistance-to-change
A -A
B -B
C -C
D -D
E -E
F -F
G -G
H -H
I -I
Number of revealed values: ______
Number of ValuesEither Or 5A -B B -AA -C C -AA -D D -AA -E E -AB -C C -BB -D D -BB -E E -BC -D D -CC -E E -CD -E E -D
Either Or 4A -B B -AA -C C -AA -D D -AB -C C -BB -D D -BC -D D -C
Either Or 3A -B B -AA -C C -AB -C C -B
Either Or 2A -B B -A
277
Questionnaire Which car do you currently own and use? (1) Brand _________________ Model/Type ____________Model year _________________ (2) Brand _________________ Model/Type ____________Model year _________________ (3) Brand _________________ Model/Type ____________Model year _________________ In case of more than one auto to answer further questions, please, either choose only one which is mostly used or indicate with (1), (2), (3) answers according to the named cars. Who is the owner of your car? □ You □ Your organization:
□ national □ private □ Other: __________________________________________________________________ Who is driving your car? □ You □ Your driver □ Your guard-driver □ You and your driver:
During working hours _________________ During leisure time _________________ □ Other: __________________________________________________________________ Who makes the final decision about an auto purchasing (if differs then considering the last purchased auto)? □ You □ Other: __________________________________________________________________ Who influences the choice of technical equipment and characteristics of your purchased auto? □ You □ Other: __________________________________________________________________ What does it mean for you “to be confident about your auto”? __________________________________________________________________________________________________________________________________________________________________ Is it important for you to be confident about your auto? □ very important □ important □ not that important □ not important
Aim of purchase of an auto (current or next)?
278
□ Need for transportation during leisure time □ Need for transportation at work □ To fit social status □ To enjoy driving (ride) □ Other: ________________________________________________________________________________ Do you use enhanced security devices and measures □ in auto? □ at home? □ in office? □ Other: ________________________________________________________________________________ Describe a situation related to the car usage or ownership which
1. you are afraid of or you would like to avoid (for example, in certain circumstances such as during the work, during family ride or trip, etc.) ______________________________________________________________________________________________________________________________________________________
2. affects your image that you are afraid of or you would like to avoid? ______________________________________________________________________________________________________________________________________________________
3. affects any property (e.g., car, laptop, briefcase, bag, etc.) that you are afraid of or you would like to avoid? ______________________________________________________________________________________________________________________________________________________
4. affects you personally (physically or psychologically) that you are afraid of or you would like to avoid? ______________________________________________________________________________________________________________________________________________________
Suppose you realize a serious danger regarding you or your family and you need special security measures. Please, evaluate your preference to either high-tech or physical security methods in terms of confidence (select one option):
1 2 3 4 5 Entire rigging with high-tech devices as far as possible.
Rigging with high-tech devices as far
as possible and physical
involvement only if still needed.
Does not matter which one until it
works and guards / Prefer having both
First of all physical protection and
high-tech devices if needed.
Nothing can replace or be better than
physical protection (man as a
bodyguard)
Estimate your agreement with the following statements:
absolutely
agree agree not always disagree
absolutely disagree
No matter how fast my income goes up, I never seem to get ahead.
I have more to spend on extras than most of my neighbors.
My income is high enough to satisfy nearly all of my important desires.
279
Your sex? □ Female □ Male Your marital status? □ Single □ Married How many kids do you have? __________ Your age? □<=18 □19-24 □25-29 □30-34 □35-39 □40-44
□45-49 □50-54 □55-59 □60-64 □>=65
Education (according to Russian educational institution classification)? □ Higher (university graduate) □ Incomplete higher □ Specialized school □ Secondary □ Incomplete secondary Your field of activity? □ Business (entrepreneurship) □ Politics □ Science □ Education □ Other: _______________________________________________________________________________ Your position? □ Owner (joint owner) □ Chief □ Top manager □ Manager □ Other: _______________________________________________________________________________ Pattern of ownership of your organization? □ National enterprise □ Own organization □ Private (not own) organization □ International company □ Other: _______________________________________________________________________________
280
Information about the interview
Date:
Time:
Duration:
Location:
281
Appendix 4: Categorized elicited constructs
Com
fort
1.2
com
fort
able
: d
river
's c
omfo
rt t
o p
rovi
de
less
em
otio
nal
an
d p
hys
ical
effo
rts
28
1.5
com
fort
able
2:
com
fort
ele
ctro
nic
s an
d t
ech
niq
ues
1
4.2
2.3
com
fort
able
: ap
pro
pria
te s
ize,
com
fort
able
pas
sen
ger
sea
ting
arr
ang
emen
t2
.11
com
fort
able
2:
com
fort
ele
ctro
nic
s an
d t
ech
niq
ues
3.5
com
fort
able
: d
on't
feel
hu
mm
ocks
, p
oth
oles
an
d b
um
py
road
s at
driv
ing
3
.7 c
omfo
rtab
le2
: fe
el c
omfo
rtab
le o
n t
he
pas
sen
ger
sea
t (d
on't
feel
bu
mp
y ro
ads
and
pot
hol
es)
3.8
com
fort
able
3:
enou
gh
sp
ace
to li
e (s
tret
ch le
gs)
so
I ca
n t
ake
a re
st d
urin
g a
rid
e 4
.6 c
omfo
rtab
le:
app
rop
riate
siz
e fo
r ta
ll p
eop
le5
.6 c
omfo
rtab
le:
com
fort
ele
ctro
nic
s an
d t
ech
niq
ues
5
.11
com
fort
able
2:
don
't fe
el h
um
moc
ks,
pot
hol
es a
nd
bu
mp
y ro
ads
at d
rivin
g
6.3
com
fort
able
: sm
ooth
ru
nn
ing
, fe
el c
omfo
rtab
le o
n t
he
pas
sen
ger
sea
t (d
on't
feel
bu
mp
y ro
ads
and
6
.6 c
omfo
rtab
le2
: g
ood
erg
onom
ics
of s
aloo
n6
.11
com
fort
able
3:
feel
com
fort
able
on
th
e p
asse
ng
er s
eat
8.5
com
fort
able
: co
mfo
rt e
lect
ron
ics
and
tec
hn
iqu
es,
larg
e -
feel
my
own
sp
ace
bei
ng
far
from
oth
er
peo
ple
9.7
com
fort
able
: d
on't
feel
hu
mm
ocks
, p
oth
oles
an
d b
um
py
road
s (s
oft
susp
ensi
on)
10
.1 c
omfo
rtab
le:
driv
er's
com
fort
, re
lativ
ely
larg
e1
2.4
com
fort
able
: la
rge
and
hig
h fo
r ta
ll p
eop
le1
2.8
larg
e/co
mfo
rtab
le2
: co
mfo
rtab
le fo
r ta
ll p
eop
le t
o g
et in
13
.1 c
omfo
rtab
le:
driv
er's
com
fort
at
driv
ing
13
.8 s
tylis
h2
/com
fort
able
2:
attr
activ
e, c
omfo
rtab
le a
nd
not
har
sh in
terio
r1
4.8
com
fort
able
: co
mfo
rt e
lect
ron
ics
and
tec
hn
iqu
es1
5.2
com
fort
able
: d
river
's c
omfo
rt in
car
han
dlin
g1
6.3
com
fort
able
--
un
com
fort
able
: n
o b
um
py
ride
at h
igh
sp
eed
17
.2 c
omfo
rtab
le:
com
fort
able
en
joyi
ng
ste
erin
g1
7.1
3 c
omfo
rtab
le2
--
un
com
fort
able
2:
don
't fe
el h
um
moc
ks,
pot
hol
es a
nd
bu
mp
y ro
ads
18
.1 c
omfo
rtab
le:
for
self-
driv
ing
at
hig
h s
pee
d:
stee
ring
, n
ot n
oisy
, g
ood
vie
w1
9.5
com
fort
able
: g
ood
noi
se in
sula
tion
19
.11
com
fort
able
2:
soft
car
susp
ensi
on,
not
sp
orty
car
su
spen
sion
com
fort
able
--
un
com
fort
able
Con
stru
ctC
ateg
ory
Bip
olar
cat
egor
y's
pol
esS
um
,
% in
gen
eral
sa
mp
le
282
Pre
stig
ep
rest
igio
us
-- o
rdin
ary
4.4
pre
stig
iou
s: o
ther
s w
ill n
otic
e m
e 2
54
.5 u
niv
ersa
l exp
loita
tion
/larg
e/p
rest
igio
us
-- n
arro
w e
xplo
itatio
n/c
omp
act/
ord
inar
y: fo
r b
usi
nes
s an
d
12
.75
.1 p
rest
igio
us/
rep
rese
nta
tive:
not
you
th5
.4 p
rest
igio
us2
: b
lack
col
or -
sp
ecta
cula
r, c
onvi
nci
ng
5.8
pre
stig
iou
s3 -
- or
din
ary3
: su
per
imag
e, fo
r m
en,
not
wom
en7
.4 p
rest
igio
us:
ref
lect
s m
y st
atu
s8
.2 p
rest
igio
us:
imag
e of
bra
nd
an
d p
rod
uce
r8
.4 p
rest
igio
us2
: st
ylis
h,
influ
ence
s p
erce
ptio
n o
f my
stat
us
8.1
2 p
rest
igio
us3
: re
spec
t of
oth
er d
river
s on
roa
ds
wh
o m
ake
(giv
e) m
e a
way
1
0.4
pre
stig
iou
s: in
fluen
ces
per
cep
tion
of m
y st
atu
s 1
0.6
pre
stig
iou
s2:
resp
ect
of o
ther
driv
ers
on r
oad
s w
ho
mak
e (g
ive)
me
a w
ay
11
.2 p
rest
igio
us
-- o
rdin
ary:
leg
end
ary
imag
e or
bra
nd
(e.
g.
Por
sch
e)
12
.7 p
rest
igio
us:
ref
lect
s m
y st
atu
s1
3.6
un
rep
rese
nta
tive:
for
you
th w
ho
are
mor
e ris
ky (
not
for
mat
ure
peo
ple
)1
5.1
pre
stig
iou
s: e
xpen
sive
, im
mod
est
imag
e1
5.6
pre
stig
iou
s2:
resp
ect
of o
ther
driv
ers
on r
oad
s w
ho
mak
e (g
ive)
me
a w
ay
15
.11
pre
stig
iou
s3:
for
exp
erie
nce
d u
sers
of p
rem
ium
car
s w
ho
are
goo
d ju
dg
es1
6.8
pre
stig
iou
s: s
up
er im
age,
pro
per
prid
e1
7.5
pre
stig
iou
s: e
xpen
sive
, im
mod
est
imag
e1
8.5
pre
stig
iou
s: e
xpen
sive
, im
mod
est
imag
e1
8.6
com
pac
t/p
rest
igio
us2
--
larg
e/or
din
ary2
: ot
her
s w
ill n
otic
e m
e, a
esth
etic
ally
acc
epta
ble
19
.3 u
niv
ersa
l exp
loita
tion
/ord
inar
y--n
arro
w e
xplo
itatio
n/p
rest
igio
us:
fo
r lif
e -
luxu
ry (
for
leis
ure
fun
)1
9.4
pre
stig
iou
s2/r
epre
sen
tativ
e2:
for
mat
ure
exp
erie
nce
d u
sers
wh
o ar
e g
ood
jud
ges
of a
uto
s, n
ot fo
r th
ose
wh
o sh
ow o
ff (y
outh
an
d g
irls)
20
.1 p
rest
igio
us:
su
per
imag
e2
0.2
of h
igh
tec
hn
ical
qu
ality
/pre
stig
iou
s2:
pro
du
cer's
pro
ved
hig
h q
ual
ity (
pro
du
cer
rep
uta
tion
)
283
Rep
rese
nta
tiven
ess
1.8
rep
rese
nta
tive
-- u
nre
pre
sen
tativ
e: p
rest
igio
us,
wh
ich
mea
ns
refle
cts
my
stat
us
23
2.8
rep
rese
nta
tive:
pre
stig
iou
s, w
hic
h m
ean
s re
flect
s m
y st
atu
s 1
1.7
3.4
rep
rese
nta
tive
-- u
nre
pre
sen
tativ
e: fo
r b
usi
nes
s b
ut
not
for
ente
rtai
nm
ent,
fam
ily o
r va
catio
n3
.6 r
epre
sen
tativ
e2:
not
for
self-
driv
ing
4.1
un
rep
rese
nta
tive
-- r
epre
sen
tativ
e: s
elf-
driv
ing
: g
o w
her
e an
d h
ow I
wan
t5
.1 p
rest
igio
us/
rep
rese
nta
tive:
not
you
th5
.5 r
epre
sen
tativ
e2 -
- u
nre
pre
sen
tativ
e2:
resp
ectfu
l bu
t al
so t
o m
ake
a sp
lash
6.1
rep
rese
nta
tive:
with
driv
er6
.7 r
epre
sen
tativ
e2:
for
bu
sin
ess,
not
for
fam
ily7
.11
rep
rese
nta
tive
-- u
nre
pre
sen
tativ
e: fo
r b
usi
nes
s an
d m
ysel
f 7
.12
un
iver
sal e
xplo
itatio
n2
/un
rep
rese
nta
tive2
: ca
n fi
t an
d s
eat
my
dog
, w
hen
car
is t
oo r
epre
sen
tativ
e, I
ca
n't
seat
in m
y d
og8
.7 r
epre
sen
tativ
e: fo
r st
atu
s b
ut
not
for
calm
life
10
.5 r
epre
sen
tativ
e: c
an b
e u
sed
with
a d
river
for
wor
k1
3.7
rep
rese
nta
tive
-- u
nre
pre
sen
tativ
e: r
efle
cts
my
stat
us
14
.4 r
epre
sen
tativ
e: fo
r b
usi
nes
s, n
ot fo
r en
tert
ain
men
t1
4.7
rep
rese
nta
tive2
: n
ot fo
r se
lf-d
rivin
g1
5.3
rep
rese
nta
tive:
with
driv
er1
5.7
rep
rese
nta
tive2
: re
flect
s h
igh
sta
tus
of it
s ow
ner
17
.1 u
nre
pre
sen
tativ
e --
rep
rese
nta
tive:
sel
f-d
rivin
g1
8.8
rep
rese
nta
tive:
to
show
wh
o's
got
hig
her
cre
dit
19
.1 r
epre
sen
tativ
e: u
pm
arke
t (f
or w
ork)
, n
ot fa
mily
car
19
.4 p
rest
igio
us2
/rep
rese
nta
tive2
: fo
r m
atu
re e
xper
ien
ced
use
rs w
ho
are
goo
d ju
dg
es o
f au
tos,
not
for
thos
e w
ho
show
off
(you
th a
nd
girl
s)2
0.3
un
rep
rese
nta
tive:
sp
eed
y se
lf-d
rivin
g o
n c
ool a
uto
rep
rese
nta
tive
--
un
rep
rese
nta
tive
284
Tec
hn
ical
qu
ality
1.1
rou
gh
-ter
rain
hig
h -
- lo
w -
per
form
ance
/of h
igh
--
low
tec
hn
ical
qu
ality
: sa
fe "
anyw
her
e" r
ide
in t
oug
h
clim
ate
con
diti
ons
(mu
d,
ice)
2
0
10
.22
.4 o
f hig
h t
ech
nic
al q
ual
ity:
time
and
mon
ey s
avin
gs
2.7
of h
igh
tec
hn
ical
qu
ality
2:
pro
du
cer's
pro
ved
hig
h q
ual
ity (
pro
du
cer
rep
uta
tion
)3
.2 o
f hig
h t
ech
nic
al q
ual
ity:
won
't b
reak
dow
n,
time
and
mon
ey s
avin
gs
4.7
of h
igh
tec
hn
ical
qu
ality
: sa
fe in
ou
r cl
imat
e (s
now
drif
t, g
lare
ice)
6.2
of h
igh
tec
hn
ical
qu
ality
: h
igh
est
leve
l of q
ual
ity t
o sa
ve t
ime
and
mon
ey7
.2 o
f hig
h t
ech
nic
al q
ual
ity:
won
't b
reak
dow
n,
time
and
mon
ey s
avin
gs
7.7
rou
gh
-ter
rain
hig
h-p
erfo
rman
ce2
/of h
igh
tec
hn
ical
qu
ality
: rig
id s
usp
ensi
on,
rug
ged
au
to w
ill s
urv
ive
9.1
of h
igh
tec
hn
ical
qu
ality
: to
ug
h b
ody
and
su
spen
sion
(in
cas
e of
ditc
h)
9.6
of h
igh
tec
hn
ical
qu
ality
: tim
e an
d m
oney
sav
ing
s1
1.6
of h
igh
tec
hn
ical
qu
ality
: tim
e an
d m
oney
sav
ing
12
.1 o
f hig
h t
ech
nic
al q
ual
ity:
rigid
fram
e an
d s
usp
ensi
on1
2.6
of h
igh
tec
hn
ical
qu
ality
2/s
afe/
secu
re:
adeq
uac
y of
mai
nte
nan
ce e
xpen
ses
= a
fter
run
nin
g g
ear
is
chan
ged
an
au
to c
an b
e st
ill s
old
at
it's
hig
h p
rice
12
.11
of h
igh
tec
hn
ical
qu
ality
2:
avai
lab
le g
ood
ser
vice
13
.4 o
f hig
h t
ech
nic
al q
ual
ity:
relia
ble
, n
eed
s le
ss s
ervi
cin
g1
3.5
of h
igh
tec
hn
ical
qu
ality
: p
rod
uce
r's p
rove
d q
ual
ity (
e.g
. G
erm
any,
Jap
an)
16
.2 o
f hig
h t
ech
nic
al q
ual
ity:
won
't b
reak
dow
n,
time
and
mon
ey s
avin
gs
17
.12
of h
igh
tec
hn
ical
: rig
id fr
ame
and
su
spen
sion
19
.2 o
f hig
h t
ech
nic
al q
ual
ity:
won
't b
reak
dow
n,
time
and
mon
ey s
avin
gs
20
.2 o
f hig
h t
ech
nic
al q
ual
ity/p
rest
igio
us2
: p
rod
uce
r's p
rove
d h
igh
qu
ality
(p
rod
uce
r re
pu
tatio
n)
of h
igh
tec
hn
ical
qu
ality
--
o
f lo
w t
ech
nic
al q
ual
ity
285
4.3
nar
row
exp
loita
tion
: sp
orty
for
ple
asu
re,
not
city
-un
iver
sal
17
4.5
un
iver
sal e
xplo
itatio
n/la
rge/
pre
stig
iou
s --
nar
row
exp
loita
tion
/com
pac
t/or
din
ary:
for
bu
sin
ess
and
8
.66
.5 u
niv
ersa
l exp
loita
tion
: fo
r ye
ar-r
oun
d
6.8
un
iver
sal e
xplo
itatio
n2
: h
igh
ly fu
nct
ion
al7
.5 u
niv
ersa
l exp
loita
tion
--
nar
row
: fo
r ev
eryd
ay li
fe n
ot o
nly
for
ente
rtai
nm
ent
7.1
2 u
niv
ersa
l exp
loita
tion
2/u
nre
pre
sen
tativ
e2:
can
fit
and
sea
t m
y d
og,
wh
en c
ar is
too
rep
rese
nta
tive,
I
can
't se
at in
my
dog
8.3
un
iver
sal e
xplo
itatio
n:
not
pom
pou
s8
.6 u
niv
ersa
l exp
loita
tion
2 -
- n
arro
w2
: ca
n b
e on
ly o
ne
in g
arag
e1
1.3
un
iver
sal e
xplo
itatio
n -
- n
arro
w e
xplo
itatio
n:
can
/can
't b
e on
ly o
ne
in g
arag
e1
1.4
nar
row
exp
loita
tion
2:
with
str
ong
ly m
arke
d p
ecu
liarit
ies
(e.g
. sp
orty
or
extr
a p
ower
ful)
bu
t n
ot
aver
aged
or
with
ave
rag
ed c
omfo
rt1
2.2
un
iver
sal e
xplo
itatio
n:
for
all-y
ear
rou
nd
, fo
r cl
imat
e co
nd
ition
s (n
o en
gin
e w
eep
ing
)1
3.3
un
iver
sal e
xplo
itatio
n:
not
sp
orty
16
.5 u
niv
ersa
l exp
loita
tion
/rou
gh
-ter
rain
hig
h-p
erfo
rman
ce:
for
ever
ywh
ere
(wor
k an
d le
isu
re)
17
.4 u
niv
ersa
l exp
loita
tion
--
nar
row
exp
loita
tion
: sp
orty
-for
mu
d1
8.7
un
iver
sal e
xplo
itatio
n -
- n
arro
w e
xplo
itatio
n:
doe
sn't
fit a
nyt
hin
g,
bu
mp
y rid
e1
9.3
un
iver
sal e
xplo
itatio
n/o
rdin
ary-
-nar
row
exp
loita
tion
/pre
stig
iou
s:
for
life
- lu
xury
(fo
r le
isu
re fu
n)
19
.6 u
niv
ersa
l exp
loita
tion
2:
all y
ear
rou
nd
, fa
mily
use
Siz
e la
rge
-- c
omp
act
1.6
larg
e: fo
r se
vera
l peo
ple
or
lug
gag
e1
71
.7 la
rge2
: co
mfo
rtab
le lu
gg
age
tran
spor
tatio
n (
don
't n
eed
to
pu
sh in
an
d s
tuff)
8.6
2.6
larg
e: b
ette
r ob
serv
atio
n (
pan
oram
ic v
iew
) an
d fe
el s
afe
4.2
larg
e: s
tead
y, n
ot s
hor
t -
not
to
go
off f
rom
ru
t 4
.5 u
niv
ersa
l exp
loita
tion
/larg
e/p
rest
igio
us
-- n
arro
w e
xplo
itatio
n/c
omp
act/
ord
inar
y: fo
r b
usi
nes
s an
d
4.1
1 la
rge2
: la
rge
and
hea
vy e
nou
gh
to
feel
saf
e b
ut
rath
er c
omp
act
in s
ize
6.4
larg
e: fo
r m
ore
than
2 p
eop
le6
.12
larg
e2:
not
sh
ort
8.8
larg
e --
com
pac
t: c
an g
o fa
r w
ith fr
ien
ds
and
fit
skis
9.4
com
pac
t --
larg
e: b
ette
r to
mov
e on
in c
ase
of n
arro
w r
oad
s or
tra
ffic
jam
10
.8 la
rge
-- c
omp
act:
can
fit
peo
ple
an
d lu
gg
age
11
.1 la
rge:
SU
V (
jeep
) b
ut
not
sed
an
12
.8 la
rge/
com
fort
able
2:
com
fort
able
for
tall
peo
ple
to
get
in1
7.6
larg
e --
com
pac
t: t
o si
t h
igh
er -
low
er
17
.7 c
omp
act2
: n
o n
eed
to
fit p
eop
le a
nd
lug
gag
e1
8.6
com
pac
t/p
rest
igio
us2
--
larg
e/or
din
ary2
: ot
her
s w
ill n
otic
e m
e, a
esth
etic
ally
acc
epta
ble
20
.4 c
omp
act:
bet
ter
to m
ove
on in
cas
e of
nar
row
roa
ds
or t
raffi
c ja
m
Inte
nd
ed a
pp
licat
ion
(U
niv
ersa
lity)
un
iver
sal e
xplo
itatio
n -
- n
arro
w e
xplo
itatio
n
286
Des
ign
styl
ish
--
un
attr
activ
e2
.2 s
tylis
h:
brig
ht
attr
activ
e ap
pea
ran
ce1
73
.3 s
tylis
h -
- u
nat
trac
tive:
sty
lish
ap
pea
ran
ce8
.64
.8 s
tylis
h:
attr
activ
e5
.2 s
tylis
h:
des
ign
7.3
sty
lish
: b
eau
tifu
l7
.6 s
tylis
h2
/new
mod
el:
eleg
ant
bec
ause
of n
ovel
ty,
un
iqu
enes
s1
0.2
sty
lish
: b
eau
tifu
l1
1.5
sty
lish
: b
righ
t at
trac
tive
des
ign
13
.2 s
tylis
h -
- u
nat
trac
tive:
att
ract
ive
des
ign
13
.8 s
tylis
h2
/com
fort
able
2:
attr
activ
e, c
omfo
rtab
le a
nd
not
har
sh in
terio
r1
4.2
sty
lish
: b
eau
tifu
l1
5.4
sty
lish
: d
emon
stra
tive,
brig
ht
des
ign
15
.8 s
tylis
h2
: ca
n b
e ex
pec
ted
un
com
mon
, n
ot b
orin
g c
olor
ing
1
6.7
sty
lish
/new
mod
el:
mod
ern
, n
ot d
ated
17
.11
sty
lish
: sh
arp
, ag
gre
ssiv
e fo
rms
(like
BM
W)
18
.2 s
tylis
h:
not
Am
eric
an2
0.8
sty
lish
: ot
her
s w
ill n
otic
e m
eS
pee
d/P
ower
1.3
hig
h-s
pee
d/p
ower
ful:
adeq
uac
y b
etw
een
pow
er a
nd
wei
gh
t1
34
.12
hig
h-s
pee
d/p
ower
ful -
- lim
ited
: sp
eed
y6
.69
.2 h
igh
-sp
eed
/pow
erfu
l: p
ower
ful
10
.3 h
igh
-sp
eed
/pow
erfu
l: sp
orty
= s
pee
dy
and
pow
erfu
l1
1.7
hig
h-s
pee
d/p
ower
ful:
com
bin
atio
n o
f pow
er a
nd
eco
nom
y (lo
w-c
onsu
mp
tion
en
gin
e)1
2.1
2 h
igh
-sp
eed
/pow
erfu
l: p
ower
ful
15
.5 h
igh
-sp
eed
/pow
erfu
l: sp
eed
y1
6.1
hig
h-s
pee
d/p
ower
ful/r
oug
h-t
erra
in h
igh
-per
form
ance
: co
mb
inat
ion
of a
llow
able
hig
h s
pee
d a
nd
ro
ug
h-t
erra
in h
igh
-per
form
ance
17
.3 h
igh
-sp
eed
/pow
erfu
l: p
ower
ful i
n e
very
thin
g:
spee
d,
pow
er,
des
ign
18
.3 h
igh
-sp
eed
/pow
erfu
l: n
ot g
irly
18
.4 h
igh
-sp
eed
/pow
erfu
l2:
allo
wab
le h
igh
sp
eed
19
.7 li
mite
d:
for
com
fort
able
, ca
lm d
rivin
g (
not
for
thos
e w
ho
driv
e lik
e a
clap
per
s (i.
e. e
xtre
mel
y fa
st)
20
.6 h
igh
-sp
eed
/pow
erfu
l: al
low
able
hig
h s
pee
d
hig
h-s
pee
d/p
ower
ful -
- lim
ited
287
Cro
ss-c
oun
try
abili
ty1
.1 r
oug
h-t
erra
in h
igh
--
low
-p
erfo
rman
ce/o
f hig
h -
- lo
w t
ech
nic
al q
ual
ity:
safe
"an
ywh
ere"
rid
e in
tou
gh
cl
imat
e co
nd
ition
s (m
ud
, ic
e)
12
6
.12
.1 r
oug
h-t
erra
in h
igh
per
form
ance
: ci
ty,
cou
ntr
ysid
e, m
oun
tain
s5
.7 r
oug
h-t
erra
in h
igh
-per
form
ance
: ci
ty,
cou
ntr
ysid
e, m
oun
tain
s7
.1 r
oug
h-t
erra
in h
igh
--
low
-p
erfo
rman
ce:
for
city
/eve
ryw
her
e7
.7 r
oug
h-t
erra
in h
igh
-per
form
ance
2/o
f hig
h t
ech
nic
al q
ual
ity:
rigid
su
spen
sion
, ru
gg
ed a
uto
will
su
rviv
e 9
.3 r
oug
h-t
erra
in h
igh
-per
form
ance
: fo
r "a
nyw
her
e"1
2.5
rou
gh
-ter
rain
hig
h-p
erfo
rman
ce:
for
ever
ywh
ere
= p
rese
nce
of f
ron
t ax
el1
6.1
hig
h-s
pee
d/p
ower
ful/r
oug
h-t
erra
in h
igh
-per
form
ance
: co
mb
inat
ion
of a
llow
able
hig
h s
pee
d a
nd
ro
ug
h-t
erra
in h
igh
-per
form
ance
16
.5 u
niv
ersa
l exp
loita
tion
/rou
gh
-ter
rain
hig
h-p
erfo
rman
ce:
for
ever
ywh
ere
(wor
k an
d le
isu
re)
16
.6 r
oug
h-t
erra
in h
igh
-per
form
ance
3:
no
wor
ry a
bou
t p
oth
oles
18
.11
rou
gh
-ter
rain
hig
h -
- lo
w -
per
form
ance
: ev
eryw
her
e-ci
ty2
0.5
rou
gh
-ter
rain
low
-per
form
ance
: p
refe
r ci
ty s
pee
dy
cars
Saf
ety/
Sec
urit
y1
.4 s
afe/
secu
re:
app
rop
riate
for
secu
rity
mea
sure
s (b
odyg
uar
d,
driv
er,
arm
orin
g)
10
8.1
saf
e/se
cure
--
un
safe
: p
ract
ical
= n
ot n
otab
le o
n t
he
road
s, le
ss r
isk
of c
rimin
al a
ttem
pt
5.1
8.1
1 s
afe/
secu
re:
no
risk
of c
rimin
al a
ttem
pt
bec
ause
of g
ood
car
bra
nd
9
.5 s
afe/
secu
re:
man
euve
rab
ility
9.8
un
safe
2 -
- sa
fe/s
ecu
re2
: ris
k of
crim
inal
att
emp
t b
ecau
se o
f goo
d c
ar b
ran
d (
not
pas
sen
ger
)9
.11
saf
e/se
cure
: ar
mor
eq
uip
pin
g w
ithou
t a
dam
age
of r
ide
per
form
ance
12
.6 o
f hig
h t
ech
nic
al q
ual
ity2
/saf
e/se
cure
: ad
equ
acy
of m
ain
ten
ance
exp
ense
s =
afte
r ru
nn
ing
gea
r is
ch
ang
ed a
n a
uto
can
be
still
sol
d a
t it'
s h
igh
pric
e1
4.1
saf
e/se
cure
: ap
pro
pria
ten
ess
for
arm
orin
g (
Au
di Q
7 is
not
ap
pro
pria
te a
s it
is a
n o
ff-ro
ad v
ehic
le (
too
big
, ea
sier
to
reac
h,
shoo
t d
own
)1
4.5
saf
e/se
cure
2:
initi
ally
bet
ter
tech
nol
ogic
al e
qu
ipm
ent
in s
ense
of s
afet
y1
7.1
4 s
afe/
secu
re:
no
risk
of c
rimin
al a
ttem
pt
bec
ause
of g
ood
car
bra
nd
rou
gh
-ter
rain
hig
h
per
form
ance
--
rou
gh
-te
rrai
n lo
w p
erfo
rman
ce
safe
/sec
ure
--
un
safe
288
Pric
e2
.5 e
xpen
sive
--
obta
inab
le p
rice:
not
ob
tain
able
by
anyo
ne
87
.8 e
xpen
sive
--
obta
inab
le p
rice:
rat
her
exp
ensi
ve b
ut
with
out
extr
a fr
ills
that
mak
e ca
r ov
erp
riced
4.1
10
.7 e
xpen
sive
--
obta
inab
le p
rice
11
.8 o
bta
inab
le p
rice
-- e
xpen
sive
: af
ford
able
for
maj
ority
- e
xpen
sive
16
.4 e
xpen
sive
--
obta
inab
le p
rice:
nee
d t
o w
ork
har
d t
o af
ford
th
is a
uto
1
7.8
exp
ensi
ve1
9.8
exp
ensi
ve -
- ob
tain
able
pric
e: v
alu
e fo
r m
oney
20
.7 e
xpen
sive
: ov
erp
riced
to
mak
e a
spla
shN
ovel
ty3
.1 n
ew m
odel
: n
ew m
ean
s te
chn
ical
ly in
nov
ativ
e, n
ot d
ated
75
.3 n
ew m
odel
--
old
mod
el:
un
iqu
e, u
nco
mm
on3
.67
.6 s
tylis
h2
/new
mod
el:
eleg
ant
bec
ause
of n
ovel
ty,
un
iqu
enes
s1
2.3
new
mod
el:
new
mea
ns
tech
nic
ally
inn
ovat
ive,
bet
ter
stee
ring
, ca
n't
be
bu
rnt
anym
ore
14
.3 n
ew m
odel
: n
ovel
ty1
4.6
new
mod
el2
: n
ew m
ean
s te
chn
ical
ly in
nov
ativ
e1
6.7
sty
lish
/new
mod
el:
mod
ern
, n
ot d
ated
Tot
als
19
71
00
.0
new
mod
el -
-
old
mod
el
exp
ensi
ve -
- ob
tain
able
p
rice
289
Appendix 5: Customers’ values and most preferred car attributes
Novelty
Price
Safety / Security
Cross-country ability
Speed/Power
Design
Size
Intended application (Universality)
Technical Quality
Representativeness
Prestige
Comfort
Self-actualization, achievement
Pleasure, enjoyment
Family happiness
Security
Freedom, Independence
Power / Social Recognition
Affiliation / Recognition
xx
xx
x2
12
xx
xx
31
xx
xx
61
2
xx
10
12
xx
14
21
xx
x1
91
2
xx
xx
xx
x1
22
31
xx
xx
x1
52
31
xx
xx
x1
1
xx
xx
81
2
xx
xx
x9
12
x1
11
xx
xx
xx
13
21
xx
xx
x1
62
31
xx
xx
x5
21
xx
xx
x7
31
2
xx
xx
xx
x1
72
13
xx
xx
xx
41
23
xx
xx
x1
81
2
xx
xx
20
12
Su
m in
g
rou
p1
12
55
66
11
12
11
11
21
21
34
33
25
56
21
11
13
11
32
31
33
11
11
31
3 20
Car
eeris
t-ex
trov
ert
Fre
qu
ency
of a
ttrib
ute
pre
sen
ce
Pre
senc
e of
att
ribute
in m
ost
impor
tant
dim
ensi
on
Respondent
Car
eeris
t-in
trov
ert
Prior
itize
d a
ssoc
iate
d p
erso
nal
val
ues
Hed
onis
t-ex
trov
ert:
"O
ther
s n
otic
e m
e"
Seg
men
t
Tot
als
Car
eeris
t-in
trov
ert:
"I
know
how
su
cces
sfu
l I a
m"
Car
eeris
t-ex
trov
ert:
"O
ther
s kn
ow
how
su
cces
sfu
l I a
m"
Hed
onis
t-in
trov
ert :
"I
enjo
y m
y lif
e"
Hed
onis
tic c
aree
rist-
extr
over
t:
"Oth
ers
shou
ld k
now
how
su
cces
sfu
l I a
m"
Hed
onis
t-ex
trov
ert
Hed
onis
tic c
aree
rist-
extr
over
tH
edon
ist-
intr
over
t
Am
oun
t of
res
pon
den
ts
290
Appendix 6: Brands of cars named for element categories
Brands mentioned for "Work" (1)
Lexus; 6
BMW; 2
Toyota; 2
Range Rover ; 1
Maserati ; 2
Mercedes ; 7Mercedes
Lexus
BMW
Toyota
Range Rover
Maserati
Brands mentioned for "Vacation" (2)
Mercedes ; 3
Lexus; 7
BMW; 3
Toyota; 1
Audi; 1
Porsche; 3
Chevrolet; 1
VW ; 1 Mercedes
Lexus
BMW
Toyota
Audi
Porsche
Chevrolet
VW
291
Brands mentioned for "Evening leisure time" (3)
Mercedes ; 3
BMW; 3
Toyota; 1
Maserati ; 1Porsche; 8
Jaguar; 1
Bentley ; 1
Cadillac ; 1
Infinity ; 1Mercedes
BMW
Toyota
Maserati
Porsche
Jaguar
Bentley
Cadillac
Infinity
Brands mentioned for "In no way" (4)
Range Rover; 1
Mitsubishi; 1
Ssangyong ; 1
Chrysler; 2
KIA; 1
Opel ; 1
Volga GAZ; 1
Skoda; 1
Volvo ; 1
Ford ; 1
Mercedes ; 1BMW; 2
Audi; 2
VW ; 3
Jaguar; 1
Mercedes
BMW
Range Rover
Audi
VW
Jaguar
Mitsubishi
Ssangyong
Chrysler
KIA
Opel
Volga GAZ
Skoda
Volvo
292
Brands of "Previously owned" (5) cars
Mercedes ; 4
Lexus; 3
BMW; 1Toyota; 8
Audi; 1
Porsche; 1
Mitsubishi; 1
Saab 93; 1
Mercedes
Lexus
BMW
Toyota
Audi
Porsche
Mitsubishi
Saab 93
Brands mentioned for
"Car that I like but it doesn't fit" (6)
Lexus; 1
BMW; 2
Toyota; 1
Maserati ; 1
Audi; 1
Skoda; 1
Hummer; 2
Mazda ; 1
Ferrari; 3
Lincoln Limo ; 1
Aston Martin ; 1
Porsche; 5
Lexus
BMW
Toyota
Maserati
Audi
Porsche
Skoda
Hummer
Mazda
Ferrari
Lincoln Limo
Aston Martin
293
Brands mentioned for
"Alternative choice during last purchase" (7)
Mercedes ; 4
Lexus; 4
BMW; 2
Toyota; 2
Range Rover; 1
Audi; 3
Porsche; 1
Infinity; 1
Honda ; 1
Land Rover; 1 Mercedes
Lexus
BMW
Toyota
Range Rover
Audi
Porsche
Infinity
Honda
Land Rover
Brands of "current" (8) cars
Lexus; 6
BMW; 4
Mercedes; 3
Audi; 1
Mitsubishi; 1
Bentley; 1
Toyota; 1
Range Rover; 1
Jaguar ; 1 Infinity ; 1Lexus
BMW
Mercedes
Audi
Mitsubishi
Bentley
Toyota
Range Rover
Jaguar
Infinity