a comparison of the performance and utilization of reading
TRANSCRIPT
A comparison of the performance and utilization of reading strategies of candidates in tests of
expeditious and careful reading
Author Name: Richard Spiby University of Reading
British Council ELT Master’s Dissertation Awards: Commendation
1
A comparison of the performance and utilization of reading strategies of candidates in tests of
expeditious and careful reading.
October, 2014
2
Contents
Page
i List of Figures 4
ii List of Tables 5
iii Abbreviations used in the study 6
iv Abstract 7
v Acknowledgements - removed from this unbranded version 8
1 Introduction 9
1.1 Background 9
1.2 Rationale for the research 9
1.3 Purpose of the study 10
1.4 Importance of the study 10
1.5 The dissertation 11
2 Literature review 12
2.1 Introduction 12
2.2 Reading – definition and purpose 12
2.3 Variety and flexibility in reading 12
2.4 Models of reading 13
2.5 Skills and strategies 15
2.6 Componential and unitary views of reading in English language
testing
16
2.7 Reading speed 19
2.8 University reading needs 20
2.9 Conclusions 21
2.10 Research questions 22
3 Methods 23
3.1 Introduction 23
3.2 Reading constructs and operations 23
3
3.3 Informants 23
3.3.1 Test-taker profile 23
3.3.2 Expert informants 24
3.4 Data collection instruments 24
3.4.1 Reading subtests 24
3.4.2 Student questionnaires 26
3.4.3 Student interviews 26
3.4.4 Expert interviews 27
3.5 Piloting 27
3.6 Main study procedure and administration 27
3.7 Data analysis 28
3.7.1 Test scoring 28
3.7.2 Interview data 28
3.7.3 Questionnaire data 28
3.8 Limitations of the study 29
4 Results 30
4.1 Introduction 30
4.2 A Posteriori test analysis 30
4.3 Correlation of test performances 31
4.4 Questionnaire and interview results 32
4.4.1 Previewing behaviour 32
4.4.2 Overall approaches to the text and task 34
4.4.3 Responding to test items 39
4.4.4 Further factors arising from the interview data 45
4.4.5 Difficulties experienced during the tests 47
4.4.6 Differences in test purpose 49
4.4.7 General view of reading skills 50
4.4.8 Text coverage 51
5 Discussion of results 52
5.1 Introduction 52
4
5.2 Interpretation of results 52
5.3 Comparison with existing research 55
6 Conclusions 59
6.1 Introduction 59
6.2 Evaluation of the study and methodology 59
6.3 Areas for future research 60
6.4 Recommendations for pedagogy and assessment 61
6.5 Summary of conclusions 61
Bibliography 63
Appendix A Expeditious reading test 67
Appendix B Careful reading test 73
Appendix C Scoring key for the reading tests 80
Appendix D Expeditious reading strategies questionnaire 81
Appendix E Careful reading strategies questionnaire 83
Appendix F Verbal protocol procedure 85
Appendix G Extract from proficiency exam information for candidates 86
Appendix H Pilot test data 89
Appendix I Sample page of coded interview 91
Appendix J Item analysis of the main administration of the test 92
Appendix K Test-taker previewing behaviour (data) 93
Appendix L Test-taker text and task approaches (data) 94
Appendix M Test-taker item response strategies (data) 97
Appendix N Additional interview data 100
Appendix O Test difficulty (data) 102
Appendix P Test aims (data) 103
Appendix Q Ethics documents - removed from this unbranded version 104
Appendix R Urquhart and Weir’s matrix of reading types (1998) 105
5
List of Figures Page
Figure 1 Score distribution for the expeditious reading test. 30
Figure 2 Score distribution for the careful reading test. 31
Figure 3 Scattergram of careful and expeditious reading scores. 32
Figure 4 Test-taker previewing behaviour. 33
Figure 5 Number of test items previewed before referring to the text. 33
Figure 6 Test-taker approaches to the text and task. Student questionnaire
data.
36
Figure 7 Test-taker approaches to the text and task. Student interview data. 37
Figure 8 Test-taker approaches to the text and task. Teacher interview data. 38
Figure 9 Frequency of different strategies while responding to items. Student
questionnaire data.
42
Figure 10 Frequency of different strategies while responding to items. Student
interview data.
43
Figure 11 Frequency of different strategies while responding to items. Teacher
interview data.
44
Figure 12 Additional coding categories from the student interview data. 46
Figure 13 Additional coding categories from the teacher interview data. 47
Figure 14 Greatest difficulties encountered with the reading tests. 48
Figure 15 Most difficult reading subtest. 48
Figure 16 Perceived aims of the expeditious reading test. 49
Figure 17 Perceived aims of the careful reading test. 50
6
List of Tables Page
Table 1 Importance of different skills at university. 50
Table 2 Difficulty of different skills at university. 50
Table 3 Estimated text coverage by interview informants. 51
Table 4 Item analysis of pilot ER test. 89
Table 5 Item analysis of pilot CR test. 90
Table 6 Item analysis of main ER test. 92
Table 7 Item analysis of main CR test. 92
Table 8 Task previewing. Questionnaire data. 93
Table 9 Number of test items read before referring to the text. 93
Table 10 Test-taker approaches to the text and task. Student questionnaire
data.
94
Table 11 Test-taker approaches to the text and task. Student interview data. 95
Table 12 Test-taker approaches to the text and task. Teacher interview data. 96
Table 13 Numbers of students using different strategies at least once during
the test. Questionnaire data.
97
Table 14 Numbers of students reporting different strategies while responding
to items. Interview data.
98
Table 15 Numbers of students reporting different strategies while responding
to items. Teacher interview data.
99
Table 16 Additional coding categories from the student interview data. 100
Table 17 Additional coding categories from the teacher interview data. 101
Table 18 Greatest difficulties encountered with the expeditious reading test. 102
Table 19 Greatest difficulties encountered with the careful reading test. 102
Table 20 Most difficult reading subtest. 102
Table 21 Perceived aims of the expeditious reading test. 103
Table 22 Perceived aims of the careful reading test. 103
7
Abbreviations used in this study:
AERT – Advanced English Reading Test
CEFR - Common European Framework of Reference for Languages
CR – Careful Reading
EAP – English for Academic Purposes
ELT – English Language Teaching
ER – Expeditious Reading
IELTS – International English Language Testing System
KR-20 – Kuder-Richardson Formula 20
L1 – First language
L2 – Second language
MARSI – Metacognitive Awareness of Reading Strategies Inventory
SAQ – Short Answer Question
SR – Student Respondent (interview)
TEEP – Test of English for Educational Purposes
TOEFL – Test of English as a Foreign Language
TR – Teacher Respondent (interview)
wpm – words per minute
Abstract
Despite an abundance of research into the nature of second language reading, there is little consensus
on the components or processes that it involves. In the field of language testing, debate also continues
over the existence of separate testable components. The aim of this study was to investigate the
behaviour of 88 students and 3 teachers at a Turkish university during the completion of two reading
subtests, one of expeditious reading and one of careful reading, developed with reference to the
framework of Urquhart and Weir (1998). After the tests, quantitative and qualitative data were
collected through questionnaires and interviews in order to compare test-taker behaviour and strategy
use on the two tests. The results of the study indicate that there were significant differences between
several strategies used on the tests. However, there was also considerable overlap in several areas,
particularly in terms of the incorporation of expeditious strategies into careful reading. The
implications of these results for componential versus unitary views of reading are discussed.
Keywords: second language reading, second language assessment, reading at university, reading
strategies, expeditious reading, careful reading, construct validity
9
Acknowledgements
Removed from this unbranded version
10
1. Introduction
1.1 Background
Proficiency in reading is viewed as a prerequisite to effectively functioning in society, enhancing
individuals’ earning potential and general quality of life (Grabe & Stoller, 2011). This is especially true
for reading in English as a second language (L2), as English currently retains its influence despite
perceived threats to its global hegemony (Graddol, 2006). In fact, the importance of L2 reading skills is
probably increasing due to the global reach of electronic media (Afflerbach & Cho, 2009). Reading
also plays an essential role in academic development (Anderson, 1999), with a prevalence of academic
texts produced in English (Bernhardt, 2011).
This importance is reflected in a vast body of theoretical and empirical literature on the subject.
Nevertheless, relatively little is known about the reading process, with an abundance of conflicting
theories and research findings. The reason for this is that reading is largely a private, cognitive process
(Alderson, 2000). It happens within the mind, so there is little directly observable behaviour occurring
during reading, rendering it difficult to model the processes involved.
In addition, reading has always had a very prominent position in language tests, and all of the major
tests include a dedicated reading component (Moore et al., 2012). However, it can be difficult to define
the psychological constructs involved, resulting in fundamental validity issues for L2 tests of reading
(Alderson, 2000). As a result, much reading research has been indirect, product-based and quantitative,
relying on inferences made about the nature of reading.
1.2 Rationale for the research
In 2005, I was part of a project to revise the assessment instruments used on a pre-sessional English
course at a Turkish university. I was involved in developing a task of ‘expeditious reading’, i.e. quick,
selective reading of a text, in which test-takers match descriptions of main ideas with paragraphs in the
text (see 3.4.1). In addition, I also helped develop new tasks for ‘careful reading’, in which candidates
read more slowly to understand main ideas and supporting details.
Since then, limited attempts to validate the tests, through item analysis, feedback, and monitoring of
student performance within the university, have suggested that the revised test batteries as a whole
have an acceptable level of reliability and predictive validity. However, there has been no serious
examination of the constructs tested in the two reading subtests. Indeed, claims for the validity and in
particular the divisibility of the constructs underlying the tests have been based on ‘expert intuition’
11
and anecdotal evidence. Thus, at an institutional level, further validation evidence needs to be collected
to justify inferences made about test-takers’ reading competence and the decisions taken on the basis of
those inferences.
1.3 Purpose of the study
The purpose of the present study, then, was to explore the actual reading behaviour of test-takers. A
mixed methods approach (Dörnyei, 2007) was used with the aim of discovering how test-takers engage
with reading texts and which strategies they use when under two distinct sets of conditions. I have
chosen this topic because I believe that reading is a very important area in English for academic
purposes (EAP), but research into this area has received relatively little attention. This study aims to
highlight the different reading needs of students at university and enhance our ablility to test them
effectively.
1.4 Importance of the study
In English language teaching (ELT), these concerns reflect a number of unresolved theoretical issues in
reading research and the testing of reading. Firstly, there has been little agreement over terms such as
skills, strategies and ‘types’ of reading, with a confusing mixture of terms used in the literature
(Afflerbach et al., 2008). Secondly, differing models of reading emphasise the unitary or componential
nature of reading ability, with disagreement on the precise number of components that may be involved
(Grabe & Stoller, 2011). Thirdly, the focus of reading tests has historically been ‘careful reading’, with
underrepresentation of expeditious reading styles (Weir, 2005). As a result, there is a clear need for
test-generated data to shed light on the cognitive processes that test-takers undergo when they read in
different ways for different purposes.
Such an investigation is also desirable on practical grounds. First, the basic question underpinning
validity in language testing is thought to be whether a test measures what it claims to measure (Brown,
1996). For example, in order to carry out a test of expeditious reading, do test takers really display the
kinds of abilities and undergo the kind of cognitive processes envisaged in the development of the test?
If so, it can be claimed that the test has construct validity. If not, it should be revised or redesigned. As
Coughlan and Duff (1994: in Moore et al., 2012) remind us, the ‘task assigned’ may not match the
‘task performed.’ Second, the capacity to distinguish between reading constructs is important since if
two tests are genuinely measuring two different sets of abilities, then there is a justification for
administering two different tests. However, if it is not possible to differentiate between the constructs
12
measured in the two tests, then they may test the same ability and one of the tests may be redundant
(Weir, 2005).
Thus, this study relates to a number of theoretical and practical issues in the testing of reading. It can
contribute to the development of language tests at institutions where tests of student’s reading abilities
are required for academic purposes. It can also contribute to the understanding of reading skills and
strategies in language teaching and assessment.
1.5 The dissertation
Chapter 1 has introduced the study as a whole. In Chapter 2, relevant literature is summarised,
concluding with the study’s research questions. Chapter 3 describes the research methodology and
rationale. In Chapter 4, the findings are presented. In Chapter 5, these findings are discussed in relation
to the research questions and other literature. Chapter 6 considers some limitations of the study, its
methodology and some pedagogical suggestions, before finishing with some final conclusions.
13
2. Literature review
2.1 Introduction
This chapter deals with some of the theoretical perspectives and empirical studies relating to different
types of reading. These are considered from a variety of standpoints – purpose, cognitive activity,
strategies, and reading speed, with their relevance to a university context. This is followed by statement
of the research questions.
2.2 Reading – definition and purpose
At its most basic, reading can be defined as the extraction of meaning from written text together with a
suitable interpretation of that meaning (Grabe & Stoller, 2011). This, however, in no way reflects the
multifaceted complexity of the reading process. By its very nature, reading is an individual act, and
both the processes and products of reading are determined by a number of different factors.
Firstly, both process and product are affected by the fact that reading is an interactive process
(Alderson, 2000). Thus, reading comprises interaction between the information in the text and the
reader’s prior background knowledge (or ‘schema’) and linguistic knowledge. As such, it is not only
the nature of the text that determines how the text is read and comprehended, but the role of the reader
is also important. The text merely has potential for meaning, from which actual extraction and
construction of the meaning by the reader takes place (Wallace, 1992; Bernhardt, 2011).
Secondly, reading is a sociocognitive process. As a real-world communicative act, it is clearly set
within a social context, and written texts appear against the backdrop of situational, institutional and
wider societal context (Wallace, 1992). The role of the reader and the social aspect of reading are
particularly important in that they provide the conditions for the reader’s purpose. As Wallace (1992)
notes, the purpose of the reader when they approach the text determines how the text will be read in
terms of the level of selectivity of reading and the selection of parts of the text to be read carefully.
2.3 Variety and flexibility in reading
The infinite variety of reading acts results from the interplay of different factors. These include
individual reader characteristics, for example, L1, background knowledge, L2 proficiency, and
personal preferences (Koda, 2005). Then there is the text itself – genre, linguistic difficulty, register
and so on (Weir, 2005). Reading also takes place within a specific time and place and is surrounded by
a wider social context (Wallace, 1992). Reading involves a variety of purposes, be they for learning,
14
pleasure, or to locate information. Thus, by necessity a reader engages in a number of different types of
reading.
Indeed, pedagogically speaking, flexibility of reading style is a desirable trait (Anderson, 1999;
Pressley, 2006) and a reflection of a high level of competence in balancing reading needs and available
resources. Grabe and Stoller (2011) speak of the ‘strategic reader’ (p.12), who is able to adjust their
reading to suit changing purposes and monitor their comprehension accordingly. Jordan (1997)
mentions variation in reading speed as a means of dealing with different types of material, while
Nuttall (1996) suggests that students should be taught to be flexible: firstly, in terms of level of
comprehension, in that a reading task may not always involve careful reading or complete
understanding; secondly, in terms of reading speed, since the actual text covered by a reader in a
certain time, measured in words per minute (wpm), should increase when strategies such as skimming
and scanning are introduced.
This notion of flexibility is closely associated with independence and metacognition. The reader needs
to know not only how to engage with text in different ways, but also under which conditions different
strategies might be appropriate. According to Bernhardt (2011), one of the challenges for the L2 reader
is to learn to cope with high-level texts without the help of a teacher.
2.4 Models of reading
One way of describing reading is through componential models, in which the basic elements of reading
ability are outlined. The ‘simple’ developmental view of reading consists of two components –
decoding and comprehension (Koda, 2005); in other words, word recognition and linguistic
comprehension. However, Urquhart & Weir (1998) argue that although the “evidence is impressive,
doubts must remain as to the simplicity of the ‘simple’ view” (p.48). This is due to the complexity of
the two components and the difficulty in defining either of them precisely.
A 3-component model of comprehension was proposed by Coady (1979), consisting of conceptual
abilities, background knowledge and process strategies. Conceptual abilities refer to the general
cognitive capacity of the reader, in other words their general intellect. Background knowledge is the
cultural and schematic framework the reader brings to the text, so that, for example, a text about the
family may be understood in different ways by people from different cultures. Meanwhile, process
strategies are the application of linguistic resources, meaning syntactic and lexical knowledge of a
language and the capacity to use them in practice. Similarly, Bernhardt (2011) describes the
15
development of a 3-component model comprising language, literacy and world knowledge
(background knowledge). The distinctive element here is literacy, meaning the reader’s knowledge of
how to approach a text.
Other models of reading emphasise the cognitive processes involved, the so-called ‘higher-level’ and
‘lower-level’ processes. The higher-level processes tend to focus on the semantic construction of a
propositional model of the text. Background knowledge and the meaning of the text are integrated so
that a wider situational model of the text is built up (Urquhart & Weir, 1998). The emphasis is on prior
knowledge, the reader’s own resources and application of those resources in executive control. On the
other hand, lower level processes deal with basic linguistic recognition and decoding in terms of lexical
items, syntax and basic semantic propositions (Grabe and Stoller, 2011).
The importance of these processes has been debated in the “top-down” and “bottom-up” models
exemplified by Goodman (1967) and Eskey (1988), respectively. Goodman (1967) claimed that readers
are reliant upon syntactic and semantic information to anticipate and construct meaning from the text.
In this model, graphic cues – the words on the page – are sampled just sufficiently to allow higher
processes to function. Eskey (1988), on the other hand, asserted the importance of word recognition,
arguing that reading comprehension for L2 readers is much more dependent on lower level processes.
These two opposing views were reconciled by the interactive-compensatory model of Stanovich (1980,
2000), who convincingly theorised that comprehension is based on complex interaction between higher
and lower level processes using a synthesis of information from different sources. Crucially, Stanovich
posits that the role of these sources will vary according to the reader’s individual profile. This means
that processing deficiencies at one level may be offset by information from others. Therefore,
prediction from context may indeed assist word recognition skills, as Goodman had claimed, but,
according to Stanovich (2000), such compensation is indicative of a poor reader, as stronger readers are
less likely to use context to aid text recognition. Similarly, Bernhardt (2011) outlined a compensatory
2-component model, based on English and Spanish language tests conducted on English native
speakers (Bernhardt & Kamil, 1995). According to this model, the components are 30% L2 knowledge
and 20% L1 reading ability, the other 50% remaining unaccounted for.
Stanovich’s work is important, since the interactivity of different processes is a plausible explanation of
the complexity of reading. Furthermore, the compensatory element can account for individual
differences in reading method and proficiency. Perhaps most importantly for the present study, it
16
demonstrates how the reader exploits all available cognitive resources to achieve comprehension. In
fact, McNeil (2012), drawing on the work of Bernhardt and Stanovich in particular, suggests a 4-
component compensatory model, comprising L2 ability, background knowledge, L1 literacy and,
significantly, reading strategies. The addition of reading strategies was largely based on the work of
Phakiti (2006), whose analysis of cognitive and metacognitive strategies demonstrated that they were
both strongly associated with each other and with L2 reading test performance.
The link between reader purpose (2.2 above) and type of reading is specified by Urquhart and Weir
(1998) and Khalifa and Weir (2009). Their models of reading involve a goalsetter. The goalsetter is the
metacognitive function through which the desired type of reading behaviour is chosen according to
purpose. When the purpose of reading is clear, then the goalsetter can select the optimum process
necessary to extract the required information. This is supported by a monitor, which evaluates the
success of strategies used, leading to their modification or the adoption of new ones. This central role
for metacognition is also supported by Alderson (2000), who remarks that poor readers can be
characterised by reduced knowledge of strategies and a lack of awareness over applying their
knowledge.
2.5 Skills and strategies
In terms of flexibility and independence, strategies are seen as central to reading competence (Grabe,
1991). However, the precise usage of the terms ‘skills’ and ‘strategies’ is often unclear. They are rarely
defined and, confusingly, may be used interchangeably or denote different levels of ability.
Fortunately, recent focus on this issue has led to clarification. Afflerbach et al. (2008) usefully outline
the different origins of the words to account for the confusion: the term ‘skill’ has long been used in
psychology and education mainly as a behavioural description of activity, whereas ‘strategy’ became
common with the emergence of information-processing models of reading to describe efforts to
develop cognition. These origins illustrate that the basic difference between skills and strategies is
intentionality (Manoli & Papadopoulou, 2012). This distinction is summarised by Grabe & Stoller
(2011), who define skills as processes not under the conscious control of the reader, while strategies are
potentially conscious processes, meaning that they can be deliberately modified where necessary.
Further to this, Weir et al. (2000) consider skills to be predominantly text-driven activities, while
strategies are reader-driven, being derived from the reader’s own specific purposes. These views are
also reflected in the Common European Framework of Reference for Languages (CEFR) in which
skills are described as procedural, while strategies are processes which are not routine and vary
17
according to the task and specific context at hand (CoE, 2001). Significantly, strategies are seen as a
link between metacognitive processes (planning, monitoring, allocating resources) and the application
of learner resources to a particular task. In addition, the CEFR states that strategies may be used as a
way of “compensating for deficiencies” elsewhere. This is important since it links compensatory
models of reading with strategy use, where strategies may be used as compensatory resources, as also
suggested by Bernhardt (2011).
Nevertheless, Grabe and Stoller (2011) concede that the dividing line between skills and strategies can
be ambiguous. Readers gain automaticity in certain learned strategies. This means that the same
behaviour could begin as a strategy as the object of deliberate attention, but then become a skill when it
eventually becomes an effortless process (Afflerbach et al., 2008).
2.6 Componential and unitary views of reading in English language testing
In L2 assessment, the debate over the nature of reading is central to construct validity (see 1.4). If the
reading construct is indivisible and all tests ultimately measure the same trait, then test development
can focus on other issues, such as reliability and practicality (Weir, 2005). On the other hand, if
divisibility can be demonstrated, then the test developer may need to produce different tests to measure
different reading components. Consequently, researchers in L2 reading assessment have attempted to
address the question of componentiality in reading ability through both quantitative and qualitative
analysis.
Alderson (2000) has argued that should such subskills exist, reading is simply too complex a process
for them to be described fully. In earlier work (1990a; 1990b) he had 18 experienced teachers attempt
to categorize TEEP (Test of English for Educational Purposes) and IELTS (International English
Language Testing System) reading test items according to ‘higher’ and ‘lower’ skills. Finding major
disagreement among the teachers over the identification of skills, Alderson called into question the
validity of the tests, saying that test items could not be matched “unambiguously” to any particular
skill. Weir et al. (1990) questioned Alderson’s use of untrained ‘expert’ judges and basic premise that
skills can be equated with level of difficulty. They concluded that by necessity items will involve a
combination of skills. Thus it is unsurprising that single discrete skills cannot be matched to specific
items.
Researchers have since continued to disagree about the existence of discrete testable skills. Lumley
(1993) tried to address some of the problematic issues raised by Alderson (1990a). Also employing 5
18
teachers as expert judges, Lumley found that with clarification of the skills through discussion it was
possible to reach broad agreement on the assignment of subskills to items. He also found some
statistical support for skill categories using Rasch analysis. Of particular significance to the present
study, Lumley claimed that skimming and scanning could not be isolated and linked to specific items,
but were pervasive and combined with other different skills across several items. Similarly, Weir and
Porter (1994) commented on the overlapping nature of skimming and detailed reading, and suggested
that componentiality in reading may be a feature of the reader’s language proficiency, as readers may
adopt a more unitary model as they progress.
Further support for the unitary view, however, came from Rost (1993), who investigated L1 reading
comprehension using factor analysis of several reading subtests. Rost concluded that most likely a
single factor, ‘general reading competence,’ accounted for the results, with the possibility that
vocabulary could also be a separate factor.
Weir et al. (2000) developed the Advanced English Reading Test (AERT) for Chinese undergraduates
reading scientific and technical papers. In the development of the test, they emphasized its ecological
validity – the issue of whether candidates follow similar processes as those intended by the test’s
developers. For the AERT, using Urquhart and Weir’s matrix of reading types (1998), the test
developers produced a test comprising careful reading (global and local), skimming, scanning and
search reading. They found that candidate performance differed significantly among the sections, while
a panel of teacher expert judges strongly agreed on the skills and strategies being tested on each of the
subtests. They also indicated expeditious reading skills as a strong secondary focus in some sections,
highlighting the multiplicity of reading skills involved in responding to each task type. Overall, the
writers saw the data and the validation procedure as a whole as providing support for the
componentiality of reading test scores.
In a quantitative study, Song (2008) analysed the results of a listening and reading test given to 110
test-takers at UCLA with items categorized according to subskills. For reading, Song found statistical
differences in the results as evidence for divisible subskills and determined with structural equation
modeling the best fit for the data to be a two-component model, i.e., a basic model of decoding and
comprehension. However, the precise nature of comprehension was unclear, with the possibility that
the reading component may be further divisible.
19
Weir et al. (2009a) investigated the academic reading component of the IELTS exam using
retrospective questionnaires with 352 students in the UK and Taiwan who had taken parts of the IELTS
reading test. Results revealed that higher scorers tended to focus on items before looking at the text
(goalsetting behaviour) and tended to employ the strategies of search reading and careful rereading of
the text to locate and comprehend item-specific information. Such strategies cut across sections, with
both expeditious and careful reading used together in different sections of the test. This suggested that
there is sometimes only a weak connection between the task type and a candidate’s response strategy.
In a companion study, Weir et al. (2009b) analysed the IELTS tasks themselves, using a team of
trained expert judges at the University of Bedfordshire to assess the constructs tested in the academic
reading module. The results indicated that there was too much dependence on local, careful reading,
and even in sections where the aim was to test expeditious reading types, such was the design of the
task that careful reading seemed a more appropriate strategy. For example, when matching headings to
paragraphs, test-takers could adopt a careful reading style to ‘eliminate’ irrelevant headings. Thus a
candidate with ‘test-wiseness’ could avoid the strategies intended by the test designers and use their
own methods of obtaining the answer.
The authors made two overall conclusions: firstly, that protocol-based research is required to shed light
on behaviours underlying performance on expeditious and careful reading items. Secondly, they
contended that expeditious skills and strategies were underrepresented on the IELTS test. This view has
also found support from other small-scale studies, for example, Devi Krishnan (2011), who conducted
interviews with 2 experienced test-takers after they had completed IELTS reading test components.
According to the findings, expeditious items constituted less than a quarter of all reading items on the
test.
Khalifa (2010) performed a validation study with 973 test-takers for the Graduate Proficiency Test at
Alexandria University, containing a range of both expeditious and careful reading subtests.
Quantitative and qualitative analysis showed some evidence to justify the multidimensionality of the
reading test. There were only low correlations found between expeditious and careful reading test
scores, and factor analysis revealed at least two factors affecting test performance. Test-takers also
reported using different strategies and approaches. Khalifa therefore concluded that it is possible to test
separable subskills in reading.
20
To conclude, the testing literature seems equivocal. There has been relatively little empirical research
on different reading constructs (Weir, 2005), and the analyses conducted have produced conflicting
results. Reflecting on their development of the AERT, Weir et al. (2000) outlined the advantages of the
product-oriented and componential view of reading which they advocate. They claimed that in ELT,
this approach to reading can provide applicable, clear objectives upon which to base test design,
research and teaching. This would also explain the prevalence of subskill and strategy approaches
promoted by authors such as Nuttall (1996) and Anderson (1999). These approaches are a convenient
way of organising teaching, and there is evidence to show that they are effective (Salataci & Akyel,
2002). However, doubts remain as to whether reading components can really be identified and tested.
2.7 Reading speed
Different ways of approaching texts which are determined by different purposes for reading will have
implications for reading rate and levels of comprehension. A lack of fluency can result in reduced
comprehension (Anderson, 1999). As Nathan and Stanovich (1991) state, fluency is a necessary
condition for effective comprehension. Text processing must be rapid so that the brain receives a
constant flow of information and can link propositions together. Grabe (1991) suggests that there is a
threshold level reading speed at which efficient cognitive processing can take place, somewhere
between 180 and 200 wpm (words per minute). Other writers put the figure slightly higher. Alderson
(2000) believes that 300 wpm is the optimum rate found in ‘good’ readers. Pressley (2006) suggests a
‘relaxed’ L1 reading rate of 250-300 wpm, and 200 wpm for adults who are reading to learn, whereas
Khalifa and Weir (2009) cite figures in the range 140-600 wpm depending on purpose and reader
characteristics, highlighting the variety in published findings.
Perhaps the most thorough treatment of reading speed is outlined by Carver (1992). He proposes five
loosely defined types of reading (in decreasing order of reading rate): scanning (600 wpm), skimming
(450 wpm), ‘rauding’ (300 wpm), learning (200 wpm) and memorising (138 wpm or slower). Carver
argues that each mode of reading adds a different level of processing, accounting for the corresponding
reduction in speed. He claims that the central reading process is ‘rauding’, a default, linear style of
reading for general comprehension. Rauding takes place at a natural and maximally efficient rate,
producing accuracy in comprehension and speed. Carver also claims that readers ‘shift gears’ as they
read for different goals. He claims that ‘process flexibility’ – switching between the five reading styles
– is a trait of good readers, especially in an academic context.
21
Corresponding reading rates in L2 can be much lower. Haynes and Carr (1990; in Khalifa & Weir,
2009) compared Chinese L2 and American L1 readers. They found that the Chinese readers averaged
86.5 wpm, while the American readers averaged 254 wpm and also displayed better comprehension.
Jensen (1986) found L2 reading could be under 100 wpm, even for advanced readers, while Nuttall
(1996) claimed that in some cases, university students may be reading in L2 as slowly as 50 wpm. Both
Jensen (1986) and Nuttall (1996) advocate a target of 300 wpm for the L2 reader in order to match L1
reading rates, partly through the use of skimming, while Heaney (2009) claims to have students process
text at a putative 800 wpm through the use of timed exercises on PowerPoint slides.
There are several important points to emerge from investigations into reading rate. The first is that
there is a large discrepancy between L1 and L2 readers. Although L1 reading rates are seen as desirable
for L2 readers, it may be optimistic to suppose L2 readers can bridge this gap. Second, there is an
acknowledged need for readers to be flexible with their reading style. Third, this flexibility entails the
adjustment of reading speed according to reading purpose and required level of comprehension.
Finally, expeditious reading styles (skimming, scanning) are considered ways of raising effective
reading rate, even though this means that large parts of the text will be skipped or processed in a
cursory manner.
2.8 University reading needs
In English-medium academic settings for L2 learners, reading is often recognised as the most important
of the language skills (Grabe, 1991). With written text the primary means of content delivery on
academic courses, students’ greatest needs in the target language situation are coping with large
volumes of set reading material (Urquhart & Weir, 1998) and locating information in the material
quickly (Evans & Morrison, 2011).
Jordan (1997) claims that the ability to read set textbooks effectively is the main requirement for
student success at university, with students often needing to read selectively for particular information
or ideas, finding the author’s opinion or searching for evidence to be used in an essay. However, he
claims that L2 readers almost invariably read slowly, which can then lead to difficulty in completing
reading assignments. In terms of expeditious reading, Jordan lists skimming and scanning among his
list of strategies and skills used with EAP books and materials, suggesting that their importance lies in
the fact that they can be used to gain an overview of a text and so provide an initial context for framing
the content.
22
As for more extensive empirical data, Moore et al. (2012) analysed tasks given to students and
interviewed academic staff at two Australian universities. They discovered that reading was required
for all disciplines, with basic understanding of textbooks in technical ‘hard disciplines’, and multiple
sources in ‘soft subjects’, such as the humanities. Effective participation in many other components of
courses, for example, lectures, seminars, and tutorials, was dependent on having read at least the core
course material. Lecturers felt that reading was generally problematic for students in terms of difficulty
and especially quantity, with students often failing to cover required texts.
Weir et al. (2009b) collected questionnaire data from 764 native-speaker and overseas students at the
University of Bedfordshire in order to identify important reading activities carried out and difficulties
encountered during their academic careers. According to the student responses, the most important
purpose for reading was to search through texts and find required information relevant for later use, for
instance, in assignments. As this activity suggests, the strategies used most often by students were
related to expeditious, global reading styles, i.e., skimming and search reading, with careful reading
strategies ranked much lower. Many of the difficulties students faced were associated with the time
available to complete assignments, the quantity of material involved, and the difficulty in finding
relevant information.
To summarise, university students are expected to sift through large amounts of material in limited
time, largely to collate information for assignments, all of which is key to their success at university.
These views are supported by needs analysis data from the university in the present study (School of
Languages, 2004) and anecdotal evidence from students overwhelmed by the sheer quantity of reading
that they are expected to do on university courses.
A vital method of coping with the demands of academic courses appears to be expeditious reading
strategies, in which the reader processes text quickly and selectively. Thus, as expeditious reading is an
important feature of the target language situation, there is a strong case for it to have prominent
inclusion within EAP reading courses and to be assessed in reading exams.
2.9 Conclusions
In summary, a review of the literature reveals several key issues pertaining to the present study.
1. One of the main difficulties faced by university students in English-medium courses is the sheer
quantity of reading material which they are expected to cover in order to complete assignments and
other tasks.
23
2. The reading speed of L2 learners is appreciably slower than their native-speaking counterparts.
This reduced fluency further compounds the problem of coping with the amount of assigned
reading, and may even reduce the level of comprehension attained.
3. Successful readers utilise a variety of skills and strategies which enable them to process text
efficiently. These depend to a large degree on the purpose for reading, and strong readers are
flexible in their use of strategies and reading speed.
4. Several useful models of reading have been produced, but there is still debate even over the main
components involved. However, there is evidence to suggest that reading involves complex
interaction between top-down and bottom-up processes, which reflects the interaction between the
reader and the text, and emphasises that reading is both a text-driven and reader-driven process.
5. In the realm of language testing, there has been a similarly inconclusive debate over whether
reading is a unitary or divisible construct, with disagreement about whether it is possible for
different subtests or individual items to test discrete reading elements.
6. A distinction has been made between expeditious and careful reading styles, both of which would
appear to be essential for university students studying in English. As a result, these constructs need
to be investigated to determine to what extent they exist in reality, and whether it is possible for
reading test items to assess them.
2.10 Research Questions
To shed light on these unresolved issues, four research questions were formulated. The first is the main
question related to the principal theoretical issues described above in sections 2.4 and 2.6. The
remaining sub-questions deal with the more tangible behavioural manifestations of these issues in EAP
tests. By answering the sub-questions, the main question can then be addressed.
To what extent do tests of expeditious and careful reading of the type used at the university provide
evidence to support the componentiality of reading as defined by Urquhart and Weir (1998) and the
divisibility of the reading construct?
a) To what extent does performance vary on tests of expeditious and careful reading?
b) According to candidates’ self-reports, what are the differences between the strategies they
employ on tests of expeditious and careful reading?
c) According to self-reports, how similar are the strategies and cognitive processes of
candidates during the test of expeditious reading to those identified in the literature and by
test developers?
Now that the research questions have been identified, the next chapter describes the design of
instruments to gather comparative data which could answer them.
24
3. Methods
3.1 Introduction
The purpose of this study was to investigate the performance of test-takers on two different types of
test. The study involved the collection of qualitative and quantitative data. This mixed methods
approach is advocated as a way of gaining a rich variety of data (Dörnyei, 2007). It was an appropriate
approach for this study considering the product-based nature of reading tests and the process-oriented
methodology required to investigate behavioural aspects of reading performance. Although drawing on
reading theory and previous research, the research project and data collection instruments were
designed and produced by myself exclusively for use in this study.
3.2 Reading constructs and operations
An expeditious reading test and careful reading test were used in this study. The operational
descriptions of the constructs are based on Urquhart and Weir’s (1998) matrix of reading styles along
two dimensions: expeditious-careful; and global-local (Appendix R).
3.3 Informants
Data was collected from two groups of informants: a group of 88 students studying on pre-sessional
university courses, and a group of 3 ‘expert’ informants working as instructors on the same course.
3.3.1 Test-taker profile
The 88 student informants were all studying on the pre-sessional English course at a small private
English-medium university in Turkey. Of these, 87 were Turkish, with one (Turkish-speaking) student
from South Korea. They were all young adults, 18-20 years old. The 16-week upper-intermediate
course, (16 hours instruction per week) is the final stage of the pre-sessional English programme and
has a specified exit level of approximately B2 (CoE, 2001). A passing grade (65%) is needed to begin
studying on faculty courses.
Thus the profile of these informants is relatively homogeneous in terms of linguistic level, educational
background, recent language learning experience, influence of L1 and target language situation (core
university courses). This particular cohort had all studied the topic of the environment, and had been
exposed to a similar approach to reading in the course materials (Bağ et al., 2003).
From the whole population, 2 classes were involved in piloting (n=18), while the other 10 classes were
the source of the subjects of the main study. This was done on the principle that when conducting pilot
25
studies (Weir, 2005; Dornyei, 2007), the sample should approximate the main population as closely as
possible. Signed consent was obtained from all informants prior to participation in the project
(Appendix Q).
A subgroup of these participants (n=10) attended a semi-structured retrospective interview about their
test-taking experience. The participants were selected in advance in collaboration with classroom
teachers according to the following criteria – that interviewees should be:
of mixed reading proficiency judged on the basis of
o previous scores on reading tests
o classroom performance
able to reflect upon their reading processes.
willing and available to attend.
3.3.2 Expert informants
The 3 ‘experts’, were instructors at the university with at least 10 years of teaching experience at
university level, and more than 5 years’ experience in test development. All of them had been involved
in constructing expeditious (ER) and careful reading (CR) tests of the same type used in this study.
3.4 Data collection instruments
During the study, qualitative and quantitative data were collected through administration of ER and CR
subtests, two semi-structured questionnaires and 13 interviews.
3.4.1 Reading subtests
Two texts were chosen from environmental sciences textbooks. The texts were mapped by myself and
a colleague to ensure coherence and suitability (Weir, 2005). They then underwent limited editing for
content and vocabulary using the Lextutor vocabulary profiler (Cobb, 2003).
The framework used to create the two tests is that of Urquhart and Weir (1998) shown in Appendix R.
Expeditious reading is characterised as reader-driven since it is fast, selective and involves sampling
the text often in a non-linear fashion. It is further subdivided into skimming, search reading and
scanning. Careful reading is text-driven in that it usually follows the order of the text in a linear style,
and is characterised by full comprehension A distinction is also made between local reading within the
bounds of a sentence and global reading at suprasentential or even intertextual level (Weir et al.,
26
2009a). Although both local and global items were used in the tests here, an investigation of this
dimension was outside the scope of this project.
The expeditious reading test is designed to cover all aspects of Urquhart & Weir’s (1998) expeditious
reading types, i.e., skimming, scanning and search reading. The test may be described predominantly
under the heading of ‘search reading’, as it involves locating information on known topics with
predetermined goals. Items consist of descriptions of the main ideas of paragraphs within the text. It is
also possible that occasional careful reading is sometimes required (Weir et al., 2009a). In addition, the
test contains some elements of skimming and scanning: first, skimming to ascertain the macrostructure
of the text and narrow down the potential search area; second, scanning to find specific words or
numbers.
The test used here (Appendix A) has a text of 3031 words and time limit of 16 minutes. The response
format is for candidates to write the number of the paragraph which matches each of the eight items.
Accordingly, the chances of test-takers choosing the correct answer by guessing are extremely low. In
addition, test-takers are not required to write anything in the target language, so scoring is objective
and reliable. The items are not in the same order as the information in the text, as advocated by Weir
(2005).
The careful reading test is designed to cover the different aspects of Urquhart & Weir’s careful reading
types, i.e., global and local. Following their classification, the test could most probably be termed
‘global careful reading’ since most items necessitate comprehension at suprasentential level. In terms
of response format, the items most closely resemble the Short Answer Questions (SAQs) advocated by
Weir (2005), where candidates write answers in 1-8 words, generally using the language of the text.
Although the response is less controlled than in the ER test, limiting the number of answers allowed
helps maintain a high level of reliability in scoring. The text was similar in length (3391 words) to that
used for ER, meaning that information in the text used to respond to items is distributed across the text.
Before piloting, two experienced teachers working on the upper-intermediate level provided detailed
feedback on both tests, and items were amended accordingly.
*Note on terminology. Institutionally, the terms ‘skimming’ (ER) and ‘detailed reading.’ (CR) are used
when referring to these tests. These have been converted to the terms ‘expeditious reading’ and ‘careful
reading’ in this paper.
27
3.4.2 Student questionnaires
The second test instrument was a set of two semi-structured questionnaires, one pertaining mainly to
the ER test, the other to the CR test. The questionnaires had two main aims: first, to record the reading
approaches and strategies used by students while doing the tests; second, to gain insight into students’
understanding of such reading tests and their place in the university context. They were both translated
into Turkish and checked for accuracy by a Turkish colleague. The items were formulated with
reference to expeditious reading and careful reading constructs in Urquhart and Weir (1998) and
Khalifa and Weir (2007), questionnaires in Weir et al. (2009a), the Metacognitive Awareness of
Reading Strategies Inventory (MARSI) (Mokhtari & Reichard, 2004) and the university advice sheet
for test-takers (Appendix G).
The ER questionnaire (Appendix D) comprised 30 questions in 5 sections. Part A was concerned with
previewing of the text/task. Part B related mainly to how students engaged with the text as a whole.
Part C dealt with specific questions, to discover the behaviours of students as they responded to the
items. Part D asked about alternative strategies used, while Part E aimed to elicit the perceived
importance and difficulty of reading. The CR questionnaire (Appendix E) consisted of 29 items in 4
sections. Parts A, B and C, and the first 2 items in Part D were identical to the ER questionnaire, while
the remaining 2 items asked about the purpose of the tests. The balance of open and closed item types
was determined in light of relevance to the purpose of the research and anticipated limits of informant
motivation.
3.4.3 Student interviews
The aim of the 10 semi-structured, open-ended retrospective interviews was to gain access to the
cognitive processes and mental states underlying the informants’ behaviour (Nunan, 1992). If strategies
are defined as conscious processes, then they should be (at least partly) open to this method.
Retrospective ‘self-observation’ was chosen here. ‘Think aloud’ introspection was rejected for several
reasons (Cohen, 2006):
In a task (ER) requiring fast, direct processing, the intrusivity of the ‘think aloud’ process
would subvert the very nature of the task.
Test conditions for the reading task would cease to be applicable.
An unfeasible amount of training may be needed to achieve meaningful results.
In the interviews, 2 informants used Turkish exclusively. The others used both English and Turkish.
The recall of the students was stimulated through recourse to their test papers and questionnaires
28
(Dörnyei, 2007), and by limited prompting from the interviewer as predetermined in the verbal
protocol procedure (Appendix F). The interviews were recorded and notes were taken during the
interview. All interviews were transcribed.
3.4.4 Expert interviews
Interviews with the three expert informants were conducted in the same way as the student interviews,
following completion of the test tasks.
3.5 Piloting
Pilot versions of the subtests and questionnaires were administered to 18 students. Then 2 students
were selected on a voluntary basis to participate in retrospective interviews. The data obtained from the
participants contributed to the revision of all the data collection tools. The tests were revised in light of
the item analysis in Appendix H. In addition, the questionnaires were shortened as they seemed to
contribute to student fatigue. Part C of both questionnaires was changed to elicit strategy use for
specific items on the test (4 in ER; 5 in CR). These items were selected because they were considered
representative of the test as a whole. The questionnaires were then revised again after feedback from
the ‘expert’ informants. The pilot interviews revealed that students were articulate in describing their
test-taking experiences. The ease with which the students dealt with the retrospective protocol implied
that specific training would not be required.
3.6 Main study procedure and administration
The research was presented to students one week in advance as an opportunity for exam practice and
reflection (it was scheduled for two days before the course final exam). This had two potential
advantages. First, it would encourage maximum participation and motivation. Second, student
performance would be representative of the exit point of the course. The tests were conducted under
strict exam conditions (Weir, 2005) in a single lecture hall. The tests, questionnaires and interviews
were conducted one immediately after the other according to the schedule below.
Expeditious Reading Test (16 minutes)
Expeditious Reading Questionnaire (15 minutes)
Careful Reading Test (45 minutes)
Careful Reading Questionnaire (15 minutes)
Student Interviews (20 minutes each)
29
3.7 Data analysis
Quantitative and qualitative data was analysed from the test results, questionnaire items and interview
transcripts, using the procedures described below.
3.7.1 Test scoring
A detailed scoring key was prepared prior to test administration (Appendix C). A reliable dichotomous
scoring model was used for both tests, with each answer awarded one point or zero. In the CR test, the
scope for subjectivity in marking was very limited, with little or no variation in response possible, as
recommended by Hughes (1994).
Both tests were also scored by a second rater. Inter-rater reliability was over 95%, with discrepancies
resolved through discussion and agreement. To ensure intra-rater reliability, some test papers were
remarked 2 weeks afterwards. Classical item analysis, descriptive statistics and correlation coefficients
were produced using MS Excel 2007 and SPSS version 20.
3.7.2 Interview data
The interviews were transcribed, with relevant parts translated into English and checked with a Turkish
colleague. The transcripts were then coded according to Dörnyei (2007). The coding system was
initially deductive, based on the reading strategies in Parts A, B and C of the questionnaires, and coded
according to the item number. Further codes were created inductively to represent recurring topics and
patterns emerging from the data. An example of a coded transcript is given in Appendix I. For
reliability purposes, two transcripts were coded by an experienced colleague. The same transcripts were
then recoded 2 weeks later. Both inter-coder and intra-coder reliability were high. Minor inter-coder
disagreements were resolved through discussion of the code parameters.
3.7.3 Questionnaire data
The quantitative data from the closed items of the questionnaires were recorded. The points were
categorized according to item and frequencies of response. Chi-squared tests were conducted to
determine significant differences. Qualitative data were first translated from Turkish to English and
checked. The responses were coded according to naturally emerging patterns and classifications. A
good level of intercoder and intra-coder reliability was established as described above (3.7.2).
30
3.8 Limitations of the study
Despite attempts to control variables and enhance relevance and applicability, this study has several
limitations:
The focus of this study is two specific formats of reading test. The test-takers are also largely
monocultural and come from a specific educational institution. This has an impact on external
validity and the generalisability of these findings in other academic settings.
There are issues of reliability with the reading tests themselves. Only one test of each type was
administered. With more time and resources, further parallel tests would be desirable,
increasing reliability by reducing the significance of variations in candidate performance,
increasing the number of items, reducing the effect of text topic bias, etc.
The sample was smaller than expected. A hundred or more respondents would have enabled
more complex and reliable analysis to be carried out. The test-taker sample was also self-
selecting as attendance was voluntary.
Student motivation is an issue. Anecdotal evidence seemed to indicate that the tests,
questionnaires and interviews were taken seriously. However, there were no stakes attached to
the test, so student behaviour may not completely have reflected the behaviour shown in a ‘real
exam.’
The use of verbal protocols can be problematic for expeditious reading. The retrospective
method used avoids issues of intrusivity, but at the cost of some accuracy. Respondents may
lack uniformity in their responses. Protocols may also miss unconscious skills employed while
reading (Cohen, 2006).
31
4. Results
4.1 Introduction
In order to investigate the behaviours of test-takers on an expeditious reading test and careful reading
test, data from the test results, questionnaires and interviews were subjected to quantitative and
qualitative analysis. The test results are presented first, followed by questionnaire and interview data,
organised according to the different sections of the questionnaires. Qualitative interview data is
inserted throughout where it has illustrative value.
4.2 A Posteriori Test Analysis
Figure 1. Score distribution for the expeditious reading test.
The score distribution for the expeditious reading test (Figure 1) shows that most test-takers were in the
upper range of scores, consistent with the status of the test as a practice achievement exam. The
distribution is to some extent skewed because of the cut-off at the maximum score. The careful reading
test revealed a normal distribution (Figure 2), with a greater distribution of scores.
Mean: 5.20 SD: 1.72
32
Figure 2. Score distribution for the careful reading test.
The item analysis results are given in full in Appendix J. All items were within an acceptable range of
facility of between 0.2 and 0.8 with the exception of one ER item at 0.81. All items showed a good
level of item discriminability. The KR-20 (Kuder-Richardson Formula 20) reliability measures were
not high for either test (ER=0.5, CR=0.62) and the Standard Error of Measurement for ER was
calculated as 1.21, while for CR it was 1.67. While these figures do not suggest a particularly high
level of reliability and accuracy, this is to some extent a consequence of the small number of items (ER,
n=8; CR, n=13).
4.3 Correlation of test performances
The Pearson Product Moment Correlation Coefficient was calculated to compare candidate
performance on the ER and CR tests. There was a statistically significant positive correlation between
scores on the 2 subtests, r=0.44, p<0.001. Thus the variance shared by the tests was found to be r2=0.19.
This means that approximately 19% of variance in ER performance can be explained by performance
on the CR test, i.e. it is a weak explanatory variable, with over 80% of variance explained by unknown
factors. The plot in Figure 3 illustrates the loose linear relationship between the 2 variables. The data
points are far apart from each other and are distributed loosely around the regression line, reflecting the
weak strength of association. Owing to the issues related to score distribution mentioned above, as a
check the non-parametric Spearman’s rho was calculated and was found to match the Pearson
coefficient (ρ=0.43, p<0.001).
Mean: 6.39 SD: 2.72
33
Figure 3. Scattergram of careful and expeditious reading scores.
4.4 Questionnaire and interview results
Questionnaire items specifically related to reader behaviour during the test (Parts A-D) were identical
for both tests and are presented together for ease of comparison.
4.4.1 Previewing behaviour
In Part A, students were asked about their initial approach to the exam task and whether they
previewed the text. The results for both subtests are given in Figure 4 (and in tabular form in Appendix
K).
None of the test-takers first read the ER text carefully and only 9% looked at the text quickly, the vast
majority (91%) preferring to read the items instead. For CR, the proportion reading the items first was
lower (66%), although it still constituted a firm majority. Only a quarter of the test-takers claimed that
they had read the text slowly. These overall differences in previewing behaviour were statistically
significant (chi squares; p<0.01). In contrast, teachers seemed more likely to preview the text quickly.
Two of the 3 teachers interviewed did so for ER, and 1 previewed the CR text before the questions.
34
Figure 4. Test-taker previewing behaviour.
Figure 5 gives a further breakdown for those students who looked at the questions first before looking
at the text according to the number of questions they read.
Figure 5. Number of test items previewed before referring to the text.
The vast majority read all or at least some of the ER items (over 90%). The unpredictability in the
order of information in the ER task led test-takers to read more items and try to retain them in working
memory. The approach adopted by all the student interviewees was summarised by a typical response
(in all quotations, ER/CR denotes the reading test being discussed, SR and TR refer to ‘student
respondent’ and ‘teacher respondent’, respectively):
35
ER – Firstly I just look at the questions. I just circle some keywords in the
questions. And then I start to read. SR6
In contrast, in the CR test, a much lower proportion of students read through all the items. Almost half
read only the first item. According to the interviewees, the sequential nature of the task meant they
could alternate between items and text in an ordered manner. However, several of the students and one
of the teachers used the items as a way of previewing topic areas in the text.
CR – I read all the questions here and tried to create a structure, for
understanding how the text can be ordered, like I did in the skimming. SR 3
4.4.2 Overall approaches to the text and task
In Part B, students were asked about their approaches to the text and task as a whole. The questionnaire
data are displayed in Figures 6 and 7 (and Appendix L). The data can be categorised according to
perceived importance and similarity between the tests.
Firstly, 4 strategies were used by only a relatively small number of students (approximately 20% or
fewer) and did not exhibit significant differences between the ER and CR tests. These were: reading the
last paragraph before other parts of the text, looking at connections between ideas in different parts of
the text, using background knowledge, and translation into L1.
The second category comprises 2 strategies which were commonly used (over 50% of respondents)
with practically equal frequency on the 2 tests. These were: attempting to work out the organisation of
the text and varying reading speed.
The third category involves statistically significant strategies more common in ER. Only one item fit
this description: looking at the title and subheadings before reading the text. Over half of students
(53.9%) did this in ER, while less than a third (29.9%) did so in CR.
Finally, 6 strategies were used significantly more (p<0.01) in the CR test. The greatest differences were
found in the linear processing of the task. In terms of the CR items, 85.1% of test-takers tried to deal
with them in the order given, and 55.2% read the text in a linear fashion, greatly outnumbering those
who did the same for ER (11.4% and 27.3% respectively), implying that linear processing in the ER
task was relatively unimportant. Surprisingly, 36% of respondents reported that in ER they tried to read
36
only text related to the test items, while over half (52.9%) did this for CR, raising the point that
selective reading may be a significant factor in CR. In terms of reading the whole of the first paragraph,
very few carried out this surveying strategy in ER (8%), while only a third did so in CR. Other
significant differences were found for rereading text and trying to understand all the ideas in the text.
Rereading was a popular strategy in both tests, but was high in CR at 66.7%. Trying to grasp all the
ideas in the text was less important, being highest in CR (31%).
In their accounts of the test-taking process (Figure 7, Appendix L), the students appeared to confirm
the relative lack of importance of certain strategies, such as paying particular attention to the first and
last paragraphs, translating into L1 and looking at relationships across the text. Varying reading speed
and rereading were mentioned frequently for both tests, while understanding all the ideas and dealing
with questions in order appeared important for CR.
A recurring theme in student comments was selective reading, focusing only on information relevant to
test items.
CR – For example, if there is ‘for instance’, I can pass over this because
this example is about the thing that we have just mentioned and it does not
interest me for the answers. SR7
Rereading of the text was carried out for two different reasons according to the student interviews. In
ER, it was mainly used as a check at the end of the test.
ER – And how did you check your answers?
I’m looking... I go back to the paragraph, and what I wrote. And that’s
the way I check. SR7
For CR, in all cases rereading was used as an aid to text comprehension.
CR – But I didn’t read the paragraph well. I couldn’t understand it,
actually. And I read it again and again. After reading a couple of times
I understood the meaning. SR3
Figure 6. Test-taker approaches to the text and task. Student questionnaire data.
38
Figure 7. Test-taker approaches to the text and task. Student interview data.
39
Figure 8. Test-taker approaches to the text and task. Teacher interview data.
40
The data from the teacher interviews (Figure 8, Appendix L) was in general agreement with the student
data. There was little or no mention of paying particular attention to the first and last paragraph, or
looking at relationships across the text. Teachers also emphasised answering items in order, rereading,
and working out text organisation. The only striking difference was that all three teachers mentioned
using their background knowledge in CR.
4.4.3 Responding to test items
In Part C, students were asked to indicate the strategies they used to answer a sample of individual
items. The data below show the number of students who used the strategies at least once during the
tests (Figure 9, Appendix M). Of the 11 strategies, 3 displayed no significant difference between ER
and CR. The most commonly used strategy was looking for keywords related to the topic of the test
items, used by over 80% of test takers in both tests. Another widely used method was matching words
in the item with identical words in the text (ER=54.5%, CR=62.8%). Finally, about half the students
made inferences in both ER and CR.
For all of the other strategies in this section significant statistical differences were found. In ER,
students had a greater tendency to search for specific names and numbers, match words in the item
with synonyms in the text, use the subheadings, and check the beginning and end of paragraphs. In CR,
students were relatively more likely to guess vocabulary from context, read the whole paragraph
containing an answer slowly, and look at the connections between sentences. While these might be
expected findings, it should also be noted that searching for names and numbers (73.3%) and matching
with synonyms from the test items (58.1%) were also frequent in CR. In fact, after matching keywords,
these were the second and third most frequently reported strategies for CR.
The questionnaire data were largely supported by the interview findings (Figure 10, Appendix M).
Again, expected trends for each type of reading were found, with keyword matching being the overall
most cited strategy. For ER, the strategies used were associated with skimming, search reading and
scanning the text, and students had a clear idea of how to locate relevant information.
ER – Because the main idea is generally at the beginning or the end of the
paragraph, I focus on those parts. SR4
42
In addition, since students had a strong tendency to read all the test items first (see 4.4.1), they also
tended to ‘attack’ the text with clear aims in terms of important language to look for.
ER – I found some of them with keywords, exact certain words, and one of
them with synonyms. SR1
ER – In ‘d’, there’s a specific situation, ‘the increase in the number of very
large cities in the world’. When I first read this question I thought I need a
specific number to answer this. SR2
However, as the figures indicate, the strategies that most students used in CR were remarkably similar.
CR – ‘Why did Soviet authorities not destroy...?’ There is ‘Soviet authorities’ and
‘Chernobyl’. I passed to the Chernobyl part and started to search for ‘food’. I found the
foods – meat, milk, et cetera. SR7
Selective reading strategies were often used to identify the approximate location of information, before
more careful reading took over.
CR – For example, in number 3, once I found what the question was asking
for directly in the paragraph, when I found ‘Bernard Cohen’, I read that
complete paragraph. SR9
However, this method was not always guaranteed to be successful. Students sometimes did not read
enough text in sufficient detail and responded incorrectly without realising it. Others were more able to
monitor themselves and adapt their approach accordingly.
CR – I couldn’t find this question easily, it confused me a bit because I
skimmed this paragraph so quickly and I couldn’t find the answer... Before
answering the first question, I found the second question’s answer here,
and I realised it had to be in this paragraph and I [went back and] found it.
SR3
43
Students corrected themselves in this way on several occasions. It is notable that in all cases of self-
correction, the result was a successful answer.
ER – I made a mistake while answering this question. Firstly, I said it was
the sixth paragraph, and then realised it is not about air pollution. It is
about general pollution. And then I realized, I erased my answer, and
looked for some air pollution paragraphs. SR3
The teachers reported a wide range of strategies, with all 3 teachers looking for matching words,
synonyms and keywords based on the test items, and also looking at subheadings and the beginning of
paragraphs in the text for both ER and CR, although they were used far more frequently in ER.
Searching for keywords was again the most frequently mentioned strategy (ER=23, CR=17).
In terms of CR, teachers seemed more conscious than students of the connections between sentences,
and referred to these 11 times during the interviews. Again the idea emerged that in CR, expeditious
strategies were useful, but needed to be supplemented with slower, detailed processing to complete the
task.
CR – I again looked for synonyms, similar phrases, I mean, paraphrases of
the questions. And I again looked at the first sentences of the paragraphs,
but then I had to read further and look at it more carefully and see the link
between the sentences and everything. I had to reread some parts to make
sure that it is really talking about the same thing or it is the answer. TR1
However, as with the student respondents, self correction was necessary, demonstrated by this example
where scanning proved insufficient.
CR – … scanning was initially helpful, but it wasn’t actually going to get
me the answer and I might probably, usually in these cases, have to go
back before in the text to get the answer. TR2
Figure 9. Frequency of different strategies while responding to items. Student questionnaire data.
45
Figure 10. Frequency of different strategies while responding to items. Student interview data.
46
Figure 11. Frequency of different strategies while responding to items. Teacher interview data.
47
4.4.4 Further factors arising from the interview data
Analysis of the student and interview transcripts also produced further coding categories which
emerged inductively from the data (Figures 12 & 13, Appendix N). Both teachers and students used
test-taking strategies to locate information. In ER, test-takers predicted an even spacing of answers
throughout the text.
ER – For example there are three paragraphs under one heading. If I can
identify one of them [as an answer], I can eliminate the other two and save
time. SR5
In CR, test-takers tried to benefit from the predictable sequencing of answers.
CR – I also used the question…I also know that [the answer] is after the
first question and before the third question, so I also read the third
question. SR7
The most frequent behaviour to emerge was the deliberate omission of whole sections of text from the
reading process. For ER, skipping stretches of text was an integral part of the process and was seldom
mentioned in an explicit way.
ER – I try to read a whole paragraph quickly, and if I decide it’s one of
[the answers], I quit reading and write the number, and then put a cross on
it not to waste time. SR1
For CR, this was a very common behavior, with 7 students and all 3 teachers mentioning it. Again, this
was used in conjunction with expeditious reading strategies to eliminate ‘unnecessary’ sections of text.
CR – I read a little here, then I decided to move onto another paragraph. I
didn’t read here. I think I read... No no I didn’t read here also because I
was looking for major accidents, about major accidents. SR2
Two other categories emerged from the interview data. For ER, 8 of the 10 students mentioned limited
careful reading (less than a whole paragraph) as a strategy for finding relevant information or
confirming main ideas found earlier.
48
ER – In this part there firstly I saw ‘attract’. These words also have the
same derivation as attractive. And that’s why I thought that it was [answer]
‘a’. And after that I did a little bit of detailed reading. ‘Beautiful cities’,
also, ‘beautiful cities’, is also a property of cities. SR7
The final category to emerge was ‘identifying relationships between ideas’. This did not necessarily
have to come from different parts of the text or be connected linguistically across adjacent sentences,
so carried a different emphasis from questionnaire items B8 and C11. It was significant because it was
mentioned frequently for CR by both students and especially teachers, but was not mentioned at all for
ER. All the teachers and 3 of the students mentioned it as a key element of careful reading.
CR – I think they try to measure your comprehension skills, how you can
understand the whole text – can you interpret or can you conceive of the
cause-and-effect relationships and the other relationships. SR2
Figure 12. Additional coding categories from the student interview data.
49
Figure 13. Additional coding categories from the teacher interview data.
In Part D, students were asked whether they had used any other strategies. Few students responded (ER,
n=31; CR n=14). The vast majority of responses simply reiterated points from other questionnaire
items.
4.4.5 Difficulties experienced during the tests
Part D asked about the most difficult aspect of the reading subtests (Figures 14 & 15, Appendix O).
The most frequent response overall was time, being the most frequent for ER and second most frequent
for CR, with a statistically significant difference between them (p<0.05). A related aspect was text
length, which was seen as a difficulty for 22 ER and 17 CR test-takers. Vocabulary and sentence
structure also emerged as factors. However, vocabulary was especially important, being the most
frequent CR difficulty. There was also a statistically significant difference between CR and ER in this
respect (p<0.01). Text topic was also an important factor in CR.
50
Figure 14. Greatest difficulties encountered with the reading tests.
Students were then asked which type of reading test they had most difficulty with in general (not the
specific tests they had done at that time).
Figure 15. Most difficult reading subtest.
51
The figures for each type of reading were very similar. This contrasts with teacher views that students
may find careful reading more challenging because of its potentially greater linguistic demands:
CR – They need more language to understand the text and answer the
questions correctly, so I think careful reading is more difficult. TR1
4.4.6 Differences in test purpose
Students were asked about the different aims of the tests (Figures 16 & 17, Appendix P). For ER,
almost 80% of responses fell into 3 categories: getting the main idea (skimming), reading quickly, and
searching for information (search reading). On the other hand, for CR, 80% of responses fell into the
categories of unspecified general comprehension, ‘full’ comprehension and extracting information.
In the interviews, however, some uncertainty emerged.
ER/CR – I’ve never understood the purpose of [expeditious reading]. I
think it comes before careful reading. SR8
ER/CR – They are the same. A little bit the same. I do not have to look at
all of the text in careful reading. SR10
Figure 16. Perceived aims of the expeditious reading test.
52
Figure 17. Perceived aims of the careful reading test.
4.4.7 General view of reading skills
In Part E, students were asked to rank English communication skills in order of importance and
difficulty at university.
Table 1. Importance of different skills at university.
Ranking Skill Average ranking (n=84)
1 Writing 2.06
2 Speaking 2.46
2 Listening 2.46
4 Reading 2.71
Table 2. Difficulty of different skills at university.
Ranking Skill Average ranking (n=83)
1 Speaking 2.45
2 Listening 2.45
3 Writing 2.48
4 Reading 2.60
53
Thus, as Tables 1 and 2 show, reading was deemed both the easiest and least important skill at
university.
4.4.8 Text coverage
Interviewees were asked to estimate the amount of text they read during each test. Although some
replies were vague or could not be quantified, some respondents were able to estimate the amount of
text they read. The responses are summarized in Table 3.
Table 3. Estimated text coverage by interview informants.
Interviewee ER text coverage (%) CR text coverage (%)
SR1 - 70-75
SR2 70 75
SR3 50 80
SR4 40 -
SR5 - -
SR6 70-80 90
SR7 - 30
SR8 - -
SR9 - -
SR10 only keywords 100
TR1 30-40 70
TR2 - 60-70
TR3 30 two-thirds
In the 6 cases where comparison is possible, respondents claimed to have read a higher proportion of
the CR text. However, only one respondent claimed to have read the whole CR text, with one student
claiming to have read only 30%. The responses of the teachers were consistent with each other, in the
range of 60-70% coverage.
54
5. Discussion of results
5.1 Introduction
The implications of the quantitave and qualitative data presented in Chapter 4 are discussed below. The
original research questions are presented in the text where related issues are dealt with. The three sub-
questions are addressed first in 5.2, while the main research question is discussed with reference to
other theoretical issues in the literature in 5.3.
5.2 Interpretation of results
Research sub-question a)
To what extent does performance vary on tests of expeditious and careful reading?
As may be expected, a positive correlation (r=0.44, p<0.001) was found between the scores on the
expeditious reading and careful reading tests. The two tests represent different aspects of the same
communicative skill, so a lack of any association would be surprising. However, the low to moderate
correlation means that 80% of variance in scores is unaccounted for, with many individuals exhibiting
large differences in their respective test scores. From this high unexplained variance we can conclude
that a score on careful reading does not appear to be an effective predictor of performance on
expeditious reading, and that success on each test requires proficiency in a different set of skills and/or
strategies.
Research sub-question b)
According to candidates’ self-reports, what are the differences between the strategies they employ on
tests of expeditious and careful reading?
Many of the behaviours associated with careful reading were displayed in statistically significant
quantities by students in the careful reading test:
Rereading particular sections of text.
Guessing the meaning of vocabulary.
Reading the whole paragraph related to a test item.
Looking for connections between sentences and logical relationships between ideas.
In addition, (compared to expeditious reading), the following was also observed.
High level of text coverage.
55
These behaviours imply greater engagement with the text at a deeper level of analysis. In other words,
they make a case for the construct validity of the test (Brown, 1996) according to the description of
careful reading provided in 3.4.1 and associated descriptions of ‘reading to learn’ (Carver, 1992;
Enright, 2000). However, a major unanticipated finding was the transfer of expeditious behaviours to
careful reading. Thus, test-takers frequently made use of the following strategies more usually
associated with quick, selective reading:
Searching for matching words, synonyms and keywords from the test items.
Searching for names and numbers.
Reading only text related to test items.
Not attempting to understand all the ideas in the text.
Not attempting to understand relationships between ideas in different parts of the text.
Selectivity in reading (less than 100% text coverage).
Therefore, we can conclude that the strategy of omitting sections of text, in other words, selective
reading, was the norm and not a rarity. This contradicts the notion that careful reading entails a full
understanding of the propositional content of the text (Alderson, 2000; Urquhart & Weir, 1998).
The reasons for this transfer are unclear. Certainly in the interviews (4.4.3), test-takers framed their
strategies in terms of necessity. If they felt that they were able to answer items with a minimum of text
coverage, then this is what they did (4.4.8). This strategy carried the inherent danger that the test-
taker’s perception of the minimum text coverage required to respond to the item would not match the
actual demands of the item. In many cases, this is what occurred. Strategies, such as scanning, were
only helpful to a certain extent, as a means of identifying relevant sections of text. Although all readers
used such strategies, more skilled readers recognized the limits of their usefulness. By monitoring their
strategy use, they were able to modify or compensate for the strategy’s lack of success, ‘widening the
net’ and engaging in detailed reading across a wider section of text. Weaker readers, on the other hand,
tended to scan to match a key word in the item and a keyword in the text, using a single sentence as the
basis of their answer.
Several factors may contribute to this adoption of expeditious strategies. The first is time pressure.
Time was perceived to be the main source of difficulty by a large proportion of test-takers (4.4.5).
However, teachers exhibited similar behaviour and, theoretically, the careful reading speed required
should be within the capability of the students in this study. Another factor is the orientation of the test-
takers. The sampled interviewees revealed that they were very goal-oriented, using test-taking
56
strategies whenever possible (4.4.1), for example, using the test items as a guide for which sections of
text they could omit. Certainly students seemed conscious of their methods as exam strategies to solve
the ‘problem’ of the test, rather than as reading strategies per se. Another factor is the length of the text.
The texts used here were approximately 3000 words in length, longer than typical academic reading
tests such as IELTS and TOEFL. It may be that when the text exceeds a certain length, then test-takers
will naturally begin to use expeditious reading techniques as the possibility of ‘redundant’ text
increases. Further investigation is needed to determine to what extent this strategy transfer is a feature
of the test itself or the test-taking population.
Overall, when the results for the two subtests are compared, there were significant differences in some
areas, such as the amount of slow and linear reading, attention to relationships between ideas and
working out vocabulary on the one hand; and selective use of certain text features (subheadings, names
and numbers, beginning and end of paragraphs) on the other. Nonetheless, behaviours were to some
extent shared, such as varying reading speed, background knowledge and matching words and
keywords. This suggests that there was no clear-cut division of strategies. Rather, the tests lie at
different points on a spectrum of different reading styles according to the level of text coverage and
engagement with the text.
Research sub-question c)
According to self-reports, how similar are the strategies and cognitive processes of candidates during
the test of expeditious reading to those identified in the literature and by test developers?
To a great extent, the behaviours displayed by students in the test of expeditious reading were those
described in the literature, especially those of skimming, scanning, and search reading given by
Urquhart & Weir (1998).
Previewing the task (goalsetting).
Using subheadings as a means of text orientation.
Non-linear processing of text.
Searching for matching words, synonyms and keywords from the test items.
Searching for names and numbers.
Paying particular attention to the beginning of paragraphs.
There was also a tendency for all interviewees to engage the text with limited stretches of careful
reading as a follow-up stage to the use of scanning and search reading techniques in order to confirm or
57
reject hypotheses about the relevance of information, also mentioned in Weir et al. (2009a). However,
the relative lack of careful reading highlights the importance of control over the timing of the test. The
test-takers saw time allocated and text length as the primary sources of difficulty on this test, so they
employed strategies to compensate for the lack of time allocated.
At this point it is worth considering the potential effect of approximate required reading speeds for
each subtest. Given that 5 (out of the total 16) minutes in the expeditious reading test could be assigned
to reading and responding to the items, students would have to process text at about 300 wpm or more
to read the whole ER text carefully, which is in the upper range of Carver’s (1992) figures for reading
speed, and would certainly be out of the reach of most L2 students (see 2.7). However, for careful
reading, the 45 minutes allowed for the task entails an important reduction in required speed. Given 20
minutes to respond to items, the test-taker would need to process the text at approximately 120 wpm,
within the capability of a student with a reasonable level of reading fluency.
In conclusion, the expeditious reading test here appears to measure the effectiveness of the reading
behaviours it is designed to measure. On the other hand, the careful reading test appears to involve
more behaviours than had been anticipated in the development of a careful reading test. However, it
could be argued that the careful reading test reflects genuine academic needs of mixed reading styles
and that the parametric qualities of the tests may be enough in and of themselves to justifiably make a
claim for differences in construct (Fulcher, 1999).
5.3 Comparison with existing research
One aim in the present study was to test the different types of expeditious reading in the same test in
the belief that it is very difficult to separate the different types from each other. As the results show,
operations associated with skimming, scanning and search reading were all represented in the test used
in this study (4.4.3). These behaviours were also present to a large degree in the careful reading test.
This is consistent with Carver’s view of reading (1992) outlined in 2.7 above, that types of reading can
be defined both by their speed and by the cumulative addition of extra levels of processing. So the
addition of careful reading may not eliminate the elements of expeditious reading it is built upon.
Furthermore, the behaviours displayed by students in this study would seem to fulfil the advice of
various authors (e.g. Grabe & Stoller, 2011; Nuttall, 1996) for readers to be flexible in combining
different reading styles.
58
Other studies have investigated the characteristics of expeditious reading – Weir et al. (2000) with the
AERT exam, and Weir et al. (2009a) with IELTS. However, both of these studies admitted that the
tests suffered from poor control over the condition of time. As implied at various points in this study,
time and task length are two defining characteristics of expeditious reading tests, so in this regard the
test in the present study may provide a more reliable indicator of expeditious reading behaviour. If this
is indeed the case, then it allows us to address the main research question:
Main research question
To what extent do tests of expeditious and careful reading of the type used at the university provide
evidence to support the componentiality of reading as defined by Urquhart and Weir (1998) and the
divisibility of the reading construct?
In terms of the componentiality of reading, this study did not produce conclusive results to support the
view that there is genuine divisibility in the reading construct. At this point, we may recall Alderson’s
(2000) view that it is impossible to measure the divisible components of reading. However, the present
study has shown that if reading conditions and purposes are manipulated carefully, it is to some extent
possible to elicit different behaviours in test-takers. Thus, in the expeditious reading test, the primary
behaviours reported by test-takers reflected those specified for skimming, scanning and search reading,
activities which, according to Urquhart and Weir (1998), are reader-driven and influenced by top-down
executive processes.
In terms of the component of language, it is difficult to see a direct connection between the present
study and componential models of reading. This may well be related to the particular aims and
methodology of this study and the fact that it deals more with strategies, under the conscious control of
the reader, than with unconscious skills. One of the most important elements of any model of L2
reading is the use of linguistic resources e.g. comprehension in the 2-component model, process
strategies in Coady’s model (1979) and language in Bernhardt’s model (2010). However, language use
is largely an automatic process at higher levels of proficiency, so its role in the present study was
negligible. The specific use of grammar and vocabulary was not a part of the questionnaire and did not
emerge as a significant factor during interviews.
Teachers in particular mentioned the greater dependence on linguistic factors in careful reading, and
students identified vocabulary as the primary difficulty, suggesting that language is a major component
of careful reading in L2. However, this could not be clearly identified elsewhere in the results here,
59
except for a minor role for cohesive elements (connections between sentences) and guessing
vocabulary, i.e. compensating for a lack of linguistic resources. By means of comparison, Weir et al.
(2009a) included grammar and vocabulary in a retrospective questionnaire to reading test-takers, but
received almost no response to them.
Similarly, background or world knowledge is an important feature of componential models described
in 2.4 above, (e.g. Coady, 1979). However, in the present study (see 4.4.2), fewer than a fifth of
respondents identified the use of background knowledge as a factor in their reading in the
questionnaire, while interview findings suggested that background knowledge played a role in careful
reading but was not a major factor.
It may be unrealistic to expect that discrete reading elements can be identified. Alderson (1990a; 1990b)
was unable to attribute discrete skills to test items, and the present study suggests that combinations of
overlapping skills and strategies are used, even on tests with purportedly different aims. Consequently,
the findings of this study would tend to support the conclusions of Lumley (1993) and Weir & Porter
(1994) – that it is extremely difficult to isolate expeditious reading types such as skimming and
scanning and that we should also acknowledge their role in careful reading, as determined by Weir et al.
(2000) for the AERT exam and Weir et al. (2009a) for IELTS.
Rather, test-takers will use whatever resources are available to deal with the task at hand. If time is
limited then, instead of reading the whole text, test-takers will use their own available cognitive
resources, i.e., adopt certain strategies, to compensate for the lack of time resources and the lack of
ability to read at a sufficiently high rate. Thus, expeditious reading strategies are compatible with
interactive theories of the reading process proposed by authors such as Stanovich (1990), Bernhardt
(2010), and especially McNeil (2012), who sees a specific role for strategies alongside the other major
components of reading.
One feature which emerged from the qualitative data in this study was the importance of metacognitive
factors, which are strongly associated with successful strategy use (Phakiti, 2003). In particular,
evidence emerged to confirm the importance of the ‘goalsetter’ and ‘monitor’ in the models of
Urquhart and Weir (1998) and Khalifa and Weir (2009). In both reading subtests examined here, the
vast majority of test-takers ensured they had specific goals before referring to the text and adapted their
approach accordingly. In addition, skilful readers were able to appraise the effectiveness of their
60
approach and adjust their strategies until there was a successful outcome, particularly when the
modified approach involved a slower and more careful reading of the text.
Unfortunately, according to the findings in 4.4.7, reading was considered the least important and
difficult skill at university by students. This may be due to the fact that the students in this study have
not yet experienced faculty courses and the accompanying heavy reading loads. It may also be that
students experience more immediate difficulty with the ‘real time’ processes of speaking and listening.
Nevertheless, students’ views of the lower importance of reading in academic life contrast with those
of surveyed university lecturers (Moore, 2012), students (Weir et al., 2009b) and various reading
experts, e.g. Grabe (1991) and Urquhart & Weir (1998). Such attitudes may make it more difficult for
students to adjust to reading demands on faculty courses.
61
6 Conclusions
6.1 Introduction
In this chapter, brief recommendations for further research, teaching and assessment are given together
with an evaluation of the study and some final conclusions.
6.2 Evaluation of the research and methodology
The overall methodology used in this study may be considered strong in certain areas. Test
development was principled and carefully staged. The overall administration of the data-gathering
instruments was highly controlled, from the planning of the test day to the efficiency of data collection
within a single morning. This condensed mixed methods approach of investigating student test
behaviour could also be usefully applied to a number of different contexts. On a theoretical level,
studies of this type can contribute to our understanding of the nature of reading and our ability to
differentiate between different reading constructs. On a more practical level, this kind of study can be
conducted as part of ongoing test validation. Although demanding of resources, it can be carried out on
a smaller scale, and should be feasible for multiple researchers. The results can then be compared with
course objectives and needs analysis data as a component of curriculum evaluation as a whole.
In terms of methodological limitations, there were several decisions which could have been made
differently with the benefit of hindsight. The data collected here is excessive for an MA dissertation.
This made the data difficult to manage when writing up the project. It also made it difficult to ‘do
justice’ to the data when carrying out the analysis, since many more perspectives on the current data
are possible. Partly, this was a natural consequence of adopting a mixed methods approach. However, it
was also related to the wide scope of the project, which was more exploratory than hypothesis-driven.
It is possible that a more specific focus would have rendered analysis more straightforward and
provided clearer conclusions.
One problem with data collection was encountered at the retrospective interview stage. This method
was chosen over a concurrent ‘think-aloud’ approach for sound reasons (3.4.3) and produced useful
insight into test-taker thought processes. Unfortunately, despite the effort put into the interview and
coding process, it is debatable how useful it was to turn the transcripts into quantitative data. The
student respondents varied in the detail of their responses (in Turkish and English) about the test-taking
experience, so measuring the frequency of strategy use had limited reliability. There may have been
ways of countering this, but they would have had different practical and methodological disadvantages.
For example, interviewees could have been selected and trained carefully. The interview guidelines
62
could also have been more detailed, with a more ‘active’ interviewer. Yet these procedures would also
have interfered with the validity of the data in some way. While teacher respondents performed better,
it is unreasonable to expect students to match their levels of interest and sophistication. This opens a
debate about the limits of introspective methods in expeditious reading in particular.
As for the tests themselves, they were created using long texts in excess of 3000 words. There were
separate time limits for each. This addresses some of the criticisms of major tests such as IELTS (Weir
et al., 2009a) and fulfils some of the conditions suggested by various writers (e.g. Weir, 2005; Nuttall,
1996; Hughes, 1994) for sufficiently long texts and time controls for each text. Observation during the
test suggested that the limits were realistic, but time stress could still be a factor in performance. It is
suggested that multiple trials be done and better ‘rules of thumb’ be formulated to address this.
Another concern is that topic may have had an effect, since 14 of the 88 students found it a major
source of difficulty in careful reading, but as Clapham (1996) observed, the importance of this can be
overstated. Additionally, in terms of response format, there may be concern that two different formats
may invalidate comparisons between the two tests. However, the different formats are instrumental to
the task and essentially help create the conditions and purpose for reading. It is argued here that since
the tests and tasks are designed to be different, the item types reflect this. Finally, at the end of the
teacher interviews, all three experts gave unsolicited positive feedback, volunteering the opinion that
these were ‘good tests’.
6.3 Areas for future research
This study has raised several questions which could be the basis of useful investigation.
To what extent were the responses in the current study culturally influenced? The population
investigated here had a very goal-oriented focus. How similar would strategy use be with
respondents from other countries and educational backgrounds?
What is the effect of the specific variables of text length and number of items on reading
behavior? The present study implies that reading style, in terms of such factors as speed, text
coverage, etc., will be different with a 500 versus a 3000-word text. Approaches may also differ
according to the density of items as a proportion of text length. The fact that there are 10 items
or 3 items per page of text may have a fundamental impact on the construct tested.
To what extent are the reading behaviours exhibited under test conditions indicative of those
observed in the classroom and in authentic academic reading situations? It may be that the
artificial conditions manufactured for test purposes may not reflect those in skills-focused
63
lessons or on faculty courses where students are dealing with L2 content. Further studies using
classroom observation, questionnaires and interviews may help to clarify this.
What is the effect of L2 proficiency on reading strategies? If there is variation in line with L2
language ability, this may have practical implications for the testing of reading, and may also
shed light on the componentiality of reading in terms of the impact of literacy, strategy use,
language ability and background knowledge.
6.4 Recommendations for pedagogy and assessment
In terms of the teaching of strategies, it should be noted that practical strategy use is a matter of
the reader’s individual choice. Since there are many sources of variation in L2 readers, students
need to be encouraged to practice a wide range of strategies so that they can develop strategy
routines which work for them.
It would be beneficial for students to learn to monitor their own strategy use and comprehension
through activities such as content prediction, self-questioning and post-reading discussion. By
becoming more conscious of strategy use, students will have a greater opportunity to show
flexibility in their approach and improve performance.
Reading lessons need to relate different purposes for reading to strategy use. Students can be
encouraged to discuss possible scenarios which may lead them to read in different ways
depending on such factors as the assigned task, text length, time allocated, etc. They could then
be encouraged to consider the kind of strategies they might use in each scenario. Students
currently studying on faculty courses could also be invited to the classroom to share their real-
life experiences in this regard and emphasise the importance of reading for academic study.
In terms of the validity of the tests in the present study, if we find that the careful reading
section is also testing expeditious reading, then important questions need to be answered.
Should we try to reduce the use of expeditious strategies by using different question types?
Alternatively, is the expeditious reading test redundant? My personal recommendation tends
towards the former, but this would be a matter for debate at an institutional as well as at a wider
theoretical level.
6.5 Summary of conclusions
This study investigated the reading behaviours of university students at a Turkish university as they
completed two reading subtests of expeditious and careful reading. It represents a useful contribution to
the area of reading and testing in ELT. First, it provides both quantitative and qualitative evidence of
test behaviour and a more holistic perspective on reading processes. Second, it deals with uncommon
64
test conditions, with long texts and controlled timing. As such, the findings constitute an important
addition to the existing literature. Third, there was no attempt to subdivide the expeditious reading test,
since it was felt to contain elements of skimming, scanning and search reading as described in the
literature.
The results here appear to support claims for the construct validity of an expeditious reading test. The
test-takers’ reported behaviour was consistent with quick and selective reading, with very limited
reliance upon careful reading elements. However, the careful reading test activated some expeditious
strategies. Unlike previous reports of this phenomenon, e.g. Weir et al. (2000), the present study has
briefly considered the implications of this and suggested subsequent areas for research.
The study suggests that the integration of skills and strategies is such that it may be futile to attempt to
separate them and claim to test them discretely. Instead, it may be more useful to think of them on a
scale with expeditious at one end and careful at the other, much as Carver (1992) describes levels of
processing from fast scanning at one extreme to slow memorization at the other. Therefore, if we want
to have meaningful tests which reflect real world applications, it is essential to recognise the different
conditions and different purposes for which reading takes place, and investigate the particular
combinations of strategies and skills that readers exhibit. Different tests may prompt overlapping skills
and strategies in the reader, but this is not sufficient grounds to reject a componential view of reading.
On the contrary, the various combinations and emphases of skills and strategies are important, and
provide the basis for designing multiple tests which will give a more complete picture of reading
competence.
65
Bibliography
Afflerbach, P., & Cho, B. (2009). Identifying and describing constructively responsive comprehension
strategies in new and traditional forms of reading. In S.E. Israel, & G.G. Duffy (Eds.),
Handbook of research on reading comprehension. New York: Routledge.
Afflerbach, P., Pearson, P., & Paris S.G. (2008). Clarifying differences between reading skills and
reading strategies. The Reading Teacher, 61(5), 364-373.
Alderson, J. C. (1990a). Testing reading comprehension skills. Reading in a foreign language, 6(2),
425-438.
Alderson, J.C. (1990b). Testing reading comprehension skills. Getting students to talk about a reading
test (a pilot study). Reading in a foreign language, 7(1), 465-503.
Alderson, J.C. (2000). Assessing Reading. Cambridge: Cambridge University Press.
Anderson, N. (1999). Exploring second language reading: issues and strategies. Boston, MA: Heinle &
Heinle.
Bağ, M.M. et al. (2003). Beyond the boundaries: English in an academic environment. Istanbul:
Sabanci University.
Bernhardt, E. B., & Kamil, M. L. (1995). Interpreting relationships between L1 and L2 reading:
Consolidating the linguistic threshold and the linguistic interdependence hypotheses. Applied
Linguistics, 16(1), 15-34.
Bernhardt, E.B. (2010). Understanding Advanced Second-Language Reading. New York: Routledge.
Brown, J.D. (1996). Testing in language programs. Upper Saddle River, NJ: Prentice Hall Regents.
Carver, R.P. (1992). Reading Rate: theory, research and practical implications. Journal of Reading,
36(2), 84–95.
Clapham, C. (1996). The development of IELTS: A study of the effect of background on reading
comprehension. Cambridge University Press.
Coady, J. (1979). A psycholinguistic model of the ESL reader. In R. Mackay, B. Barhman & R.R.
Jordan (Eds.), Reading in a second language. Rowley, MA: Newbury House.
Cobb, T. (2003). VocabProfiler, the complete lexical tutor. Retrieved May 5, 2014 from
http://www.lextutor.ca.
Cohen, A. D. (2006). The coming of age of research on test-taking strategies. Language
Assessment Quarterly, 3(4), 307–331
Council of Europe (2001). Common European Framework of Reference for Languages: Learning,
teaching and assessment. Cambridge: Cambridge University Press.
Devi Krishnan, K. S. (2011). Careful versus expeditious reading: the case of the IELTS reading
test. Academic Research International, 1(3), 25-35.
66
Dörnyei, Z. (2007). Research methods in applied linguistics: Quantitative, qualitative and mixed
methodologies. Oxford: Oxford University Press.
Enright, M., Grabe, W., Koda, K., Mosenthal, P., Mulcahy-Ernt, P., & Schedl, M. (2000). TOEFL 2000
reading framework. Princeton, NJ: Educational Testing Service.
Eskey, D. (1993). Holding in the bottom: an interactive approach to the language problems of second
language readers. In P. Carrell, J. Devine, & D. Eskey (Eds.), Interactive approaches to second
language reading. Cambridge: Cambridge University Press.
Evans, S., & Morrison, B. (2011). The first term at university: implications for EAP. ELT Journal,
65(4), 387-397.
Fulcher, G. (1999). Assessment in English for academic purposes: Putting content validity in its place.
Applied Linguistics, 20(4), 221—236.
Goodman, K.S. (1967). Reading: a psycholinguistic guessing game. Journal of the Reading Specialist,
6(4), 126-135
Grabe, W. (1991). Current developments in second language reading research. TESOL Quarterly, 25(3),
375-406.
Grabe, W. & Stoller, F.L. (2011). Teaching and researching reading. Harlow: Pearson.
Graddol, D. (2006). English Next. London: British Council.
Heaney, H. (2009). Eco-friendly reading practice at 800 wpm. Views 18(3), 166-168.
Henning, G. (1987). A guide to language testing: development, evaluation, research. Cambridge, MA:
Newbury House.
Hughes, A. (1994). Testing for language teachers. Cambridge: Cambridge University Press.
Jensen, L. (1986). Advanced reading skills in a comprehensive course. In F. Dubin, D.E. Eskey, & W.
Grabe (Eds.), Teaching second language reading for academic purposes. Reading, MA:
Addison-Wesley.
Jordan, R.R. (1997). English for Academic Purposes: a guide and resource book for teachers.
Cambridge: Cambridge University Press.
Khalifa, H. (2010). Construct validation of the reading module of an EAP proficiency test battery.
Cambridge ESOL: Research Notes 42, 8-14.
Khalifa, H. & Weir, C.J. (2009) Examining reading: Research and practice in assessing second
language reading. Cambridge: Cambridge University Press.
Koda, K. (2005) Insights into Second Language Reading. New York: Cambridge University Press.
Lumley, T. (1993). The notion of subskills in reading comprehension tests: An EAP example.
Language Testing, 10(3), 211-234.
Manoli, P. & Papadopoulou, M. (2012). Reading strategies versus reading skills: two faces of the same
67
coin. Procedia - Social and Behavioral Sciences, 46, 817-821
McNeil, L. (2012). Extending the compensatory model of second language reading. System, 40(1), 64–
76.
Mokhtari, K., & Reichard, C. (2004). Investigating the strategic reading processes of first and second
language readers in two different cultural contexts. System, 32(3), 379-394.
Moore, T., Morton, J., & Price, S. (2012). Construct validity in the IELTS Academic Reading test: A
comparison of reading requirements in IELTS test items and in university study. British
Council/IDP Australia IELTS Research Reports, 11, 120-211.
Nathan, R.G., & Stanovich, K.E. (1991). The causes and consequences of differences in reading
fluency. Theory Into Practice, 30(3), 176-184.
Nunan, D. (1992). Research methods in language learning. Cambridge: Cambridge University Press.
Nuttall, C.E. (1996). Teaching reading skills in a foreign language. Oxford: Heinemann.
Phakiti, A. (2003). A closer look at the relationship of cognitive and metacognitive strategy use to EFL
reading achievement test performance. Language Testing, 20(1), 26-56.
Pressley, M. (2006). Reading instruction that works. New York: Guildford Press.
Rost, D.H. (1993). Assessing the different components of reading comprehension: Fact or fiction.
Language Testing, 10(1), 79-92.
School of Languages (2004). Foundations Development Program needs analysis. Istanbul: Sabanci
University.
Salataci, R. & Akyel, A. (2002). Possible effects of strategy instruction on L1 and L2 reading. Reading
in a Foreign Language, 14(1), 1-17.
Song, M.Y. (2008). Do divisible subskills exist in second language (L2) comprehension? A structural
equation modeling approach. Language Testing, 25(4), 435-464.
Stanovich, K.E. (1980). Toward an interactive-compensatory model of individual differences in the
acquisition of literacy. Reading Research Quarterly, 16(1), 32-77.
Stanovich, K.E. (2000). Progress in understanding reading: Scientific foundations and new frontiers.
New York: Guilford Press.
Urquhart, A.H. & Weir, C.J. (1998). Reading in a second language: Process, product and practice.
London: Longman.
Wallace, C. (1992). Reading. Oxford: Oxford University Press.
Weir, C. (2005) Language testing and validation: An evidence-based approach. London: Palgrave.
Weir, C., Hawkey, R., Green, A., & Devi, S. (2009a). The cognitive processes underlying the academic
reading construct as measured by IELTS. British Council/IDP Australia IELTS Research Reports,
9, 157–189.
68
Weir, C. J., Hawkey, R., Green, A., Unaldi, A., & Devi, S. (2009b). The relationship between the
academic reading construct as measured by IELTS and the reading experiences of students in
their first year of study at a British university. British Council/IDP Australia IELTS Research
Reports, 9, 97–156.
Weir, C.J., Hughes, A. & Porter, D. (1990) Reading Skills: hierarchies, implicational relationships
and identifiability, Reading in a Foreign Language, 7, 505-10.
Weir, C.J. & Porter, D. (1994). The multi-divisibility or unitary nature of reading: The language tester
between Scylla and Charybdis. Reading in a foreign language, 10(2), 1-19.
Weir, C.J., Yang, H. & Jin, Y. (2000). An empirical investigation of the componentiality of L2 reading
in English for academic purposes. Cambridge: Cambridge University Press.
69
Appendix A – Expeditious reading test.
Name: _____________________________ ID No: ________
Part 1 Skimming (15%) 16 minutes
This part of the exam aims to test your ability to locate main ideas in a text. The text is about
major environmental issues related to cities.
Each of the following headings matches one of the paragraphs in the text. Write the paragraph
number next to the correct heading. The headings are not in the same order as the information in
the text. One of the answers is given as an example.
It may be useful to spend a few minutes previewing the text before you begin answering the
questions.
Each question is worth 1 point.
Paragraph
Number
Heading
12
e.g. The importance of waterways for the situation of a city.
a) Making cities attractive for people to live in.
b) Utilising the power of the sun as a source of energy.
c) A rare example of a city which has a very good position.
d) The increase in the number of very large cities in the world.
e) Optimistic and pessimistic views of the future development of cities.
f) The negative impact of change in the site of a city.
g) Different sources of air pollution in cities.
h) Growing awareness of the importance of the urban environment.
TOTAL:____ / 8
70
Cities – The Urban Environment
City Life
1. In the past, the emphasis of environmental action has most often been on wilderness, wildlife,
endangered species, and the impact of pollution on natural landscapes outside cities. Now it is time to
turn more of our attention to city environments. In the development of the modern environmental
movement in the 1960s and 1970s, it was fashionable to consider everything about cities bad and
everything about wilderness good. Cities were thought of as polluted, lacking in wildlife and native
plants, dirty, and artificial—and therefore bad. Wilderness was considered as unpolluted, clean, full of
wildlife and native plants, and natural—and therefore good. Although it was fashionable to disdain
cities, the majority of people live in urban environments and have suffered directly from their decline.
2. In the past, little public concern was focused on urban ecology; as a result, many urban people saw
environmental issues as outside their area. However, today, there is a rebirth of interest in urban
environments and in the development of urban ecology as city dwellers are at the center of some of the
most important environmental issues. People are now realizing that city and wilderness are inextricably
connected. We cannot fiddle in the wilderness while our Romes burn from sulfur dioxide and nitrogen
oxide pollution. Environmental scientists are realizing the importance of analyses of the urban
environment. The National Science Foundation has added two urban areas, Baltimore and Phoenix, to
its Long-Term Ecological Research Program, a program that supports long-term monitoring as well as
research on specific ecosystems and regions.
3. Worldwide, as we have seen, we are becoming an increasingly urbanized species. In the United States,
about 75% of the population live in urban areas, and about 25% live in rural areas. Today,
approximately 45% of the world’s population, 2.75 billion people, live in cities. It is projected that 62%
of the population, 6.5 billion people, will live in cities by the year 2025. Economic development leads
to urbanization; 75% of people in developed countries live in cities, but only 38% of the people in the
poorest developing countries are city dwellers.
4. Not only is human population increasingly urbanized, but there is rapid growth of huge metropolitan
areas with more than 8 million residents. In 1950, the world had only two such areas: New York City
with its nearby New Jersey areas (12.2 million residents) and greater London (12.4 million). By 1975,
Mexico City, Los Angeles, Tokyo, Shanghai, and Sao Paulo, Brazil, had joined this list. In 1995, there
were 23 such areas, 17 of them in the developing world. By 2015, there will be 36 megacities; and 23
of them will be located in Asia. Los Angeles and New York City are among the 10 largest megacities.
The City as a System
5. One of the ways in which we can improve the management of cities is to analyze the city as an
ecological system. Like any other life-supporting system, a city must maintain a flow of energy,
provide necessary material resources, and have ways of removing wastes. These ecosystem functions
are maintained in a city by transportation and communication with outlying areas. A city is not a self-
contained ecosystem; it depends on other cities and rural areas. A city takes in raw materials from the
surrounding countryside: food, water, wood, energy, mineral ores, everything that a human society
uses. In turn, the city produces and exports material goods and, if it is a truly great city, exports ideas,
innovations, inventions, arts, and the spirit of civilization. A city cannot exist without a countryside to
support it. As was said half a century ago, city and country, urban and rural, are one thing—one
connected system of energy and material flows—not two things.
71
6. Cities also export waste products to the countryside, including polluted water, air, and solids. It has
been estimated that the average city resident in an industrial nation annually uses (directly or indirectly)
208,000 kg (229 tons) of water, 660 kg (0.8 tons) of food, and 3,146 kg (3.5 tons) of fossil fuels and
produces 1,660,000 kg (1826 tons) of sewage, 660 kg (0.8 tons) of solid wastes, and 200 kg of air
pollutants. If these are exported without care, they pollute the countryside, reducing the countryside’s
ability to provide necessary resources for the city and making life in the surroundings less healthy and
less pleasant.
7. With such dependencies and interactions between city and surroundings, it is no wonder that
relationships between people in cities and in the countryside have often been strained. Why, country
dwellers want to know, should they have to deal with the wastes of those in the city? The answer is that
many of our serious environmental problems occur at the interface between urban and rural areas.
People who live outside but near a city have a vested interest in maintaining a good environment for
that city and maintaining a good system for managing the city’s resources.
8. There are many ways to make cities pleasing environments, including the development of parks,
connecting cities in an environmentally and aesthetically sound way to major landscape features such
as rivers and nearby mountains. As we will also see, there is a long tradition of city planning with the
goal of making cities a pleasing environment. By using both the long experience in city planning and
modern knowledge from environmental sciences, we can make cities of the future healthier and more
satisfying to people and better integrated within the environment. An argument has been made
frequently that beautiful cities are not only healthy but attract people, thereby relieving pressures on the
countryside.
9. With the growing human population, we can imagine two futures: In one, cities are pleasing and
livable; use resources from outside the city in such a way that those resources are sustainable; minimize
pollution of the surrounding country; and allow room for wilderness, agriculture, and forestry. In the
other future, cities continue to be seen as environmental negatives and allowed to decay from the
inside. People flee them to grander and more expansive suburbs that occupy much land; and the poor
who remain in the city live in an unhealthy and unpleasant environment; without care for the city, its
technological structure declines and it pollutes even more than in the past. Trends in both directions
appear to be occurring.
Site and Situation: The Location of Cities
10. There is one idea that our modern life hides with its rapid transportation and its many electronic tools.
This idea is that cities are not located at random but develop mainly because of local conditions and
regional benefits. In most cases, they grow up at crucial transportation locations (an aspect of what is
called the city’s situation) and can be readily defended with good building locations, water supplies,
and access to resources (qualities related to what is called site). The primary exceptions are cities that
have been located primarily for political reasons. For example, Washington, D.C. was located to be
near the geographic center of the area of the original 13 states; but the site was primarily swampland,
and nearby Baltimore provided the major harbor of the region.
11. The location of a city is influenced by the two factors just mentioned: site, which is the summation of
all the environmental features of that location; and situation, which is the placement of the city with
respect to other areas. A good site includes a good geologic substrate suitable for building construction,
such as a firm rock base and well-drained soils that are above the water table; good nearby supplies of
drinkable water; and good nearby lands suitable for agriculture, abundant timber, and other natural
resources. It is also easier to build a city where the climate is benign— meaning that it does not suffer
extremes of temperature and rainfall and is not subject to frequent storms. However, many important
72
cities have been built in difficult climates. For example, Minneapolis—St. Paul is a city with a cold
winter and hot summer; Houston, Texas, experiences hot, moist summers; and Miami, Florida, is
among many cities that lie in the path of hurricanes. In these cases, one negative aspect of site has been
overcome with modern engineering technology.
12. The environmental situation strongly affects the development and importance of a city, particularly
with regard to transportation and defense. Waterways are essential for transportation. Especially in
early times, before railroads, automobiles, and airplanes, cities depended on water for transportation.
Most early cities were located on or near waterways. In the ancient Roman Empire, for example, all
important cities were located near waterways. Waterways have continued to influence the locations of
cities; most major cities of the eastern United States are situated either at major ocean harbors or on
major rivers.
13. Cities are often founded at other kinds of crucial transportation points, growing up around a market, a
river crossing, or a fort. Newcastle, England, and Budapest, Hungary, are located at the lowest bridging
points on their rivers; other cities, such as Geneva, are located where a river enters or leaves a major
lake. Some well-known cities are located at the confluence of major rivers: Saint Louis lies at the
confluence of the Missouri and Mississippi rivers; Manaus, Brazil; Pittsburgh, Pennsylvania; Koblenz,
Germany; and Khartoum, Sudan are located at the confluence of several rivers. Many famous cities are
located at crucial defensive locations: such as on or adjacent to easily defended rock outcrops.
Examples include Edinburgh, Athens, and Salzburg, Austria. Other cities are situated on peninsulas—
for example, Monaco and Istanbul.
14. An ideal location for a city has both a good site and good situation, but such a place is difficult to find.
Paris is perhaps one of the best examples of a perfect location for a city—one with both a good site and
good situation. Paris began on an island more than 2,000 years ago, the situation providing a natural
moat for defense and waterways for transportation. Surrounding countryside, a fertile lowland called
the Paris Basin, affords good local agricultural land and other natural resources.
Site Modification
15. Site is provided by the environment, but technology and environmental change can alter a site for better
or worse. People can improve the site of a city and have done so when the situation of the city made it
important and when its citizens could afford large projects. An excellent situation can sometimes
compensate for a poor site. However, improvements are almost always required to the site so the city
can persist.
16. For example, New Orleans, at the mouth of the Mississippi, has a good situation but a poor site. An
important transportation center at the mouth of the Mississippi River, it lies on low mud flats of the
delta, which are unstable, are subject to frequent floods, and provide poor substrate for construction.
Backwaters and swamps offer little as a local resource for agriculture but provide breeding habitats for
mosquitoes. Fishing in the Gulf of Mexico is a plus, however. The situation of New Orleans was
especially important at the time of the founding of the United States and throughout the 19th century,
both for transportation and as a defense of the interior, as illustrated by the importance of the Battle of
New Orleans at the end of the War of 1812. Modern construction methods, including levees to prevent
flooding, have improved the site.
17. Transformation of a site over time can have adverse effects. For example, Bruges, Belgium, developed
as an important center for commerce in the 13th century because its harbor on the English Channel
permitted trade with England and other European nations. By the 15th century, however, the harbor
had seriously silted in and became blocked. The limited technology of the time did not make dredging
73
possible and the harbor could not be cleared. This problem, combined with political events led to a
decline in the importance of Bruges—a decline from which it never recovered. Today, Bruges still
lives, a beautiful city with many fine examples of medieval architecture. Ironically, the fact that these
buildings were never replaced with modern ones makes Bruges modern tourist destination. Ghent,
Belgium, and Ravenna, Italy are examples of other cities whose harbors silted. As human effects on the
environment extend to global change, there may be rapid, serious changes in the sites of many cities.
For example, if global warming occurs and sea levels rise, many coastal cities will be subject to
flooding.
The City as an Environment
18. A city changes the landscape; and because it does, it also changes the relationship between biological
and physical aspects of the environment. Many of these changes are related to the general
environmental aspects of pollution, water management, or climate. They are mentioned here with a
focus on how effective city planning can reduce the problems.
The Urban Atmosphere and Climate
19. Cities affect the local climate; as the city changes, so does its climate. Cities are generally less windy
than nonurban areas, because buildings and other structures obstruct the flow of air. But city buildings
also channel the wind, sometimes creating local wind tunnels with high wind speeds. The actual flow
of wind around one building is
influenced by nearby buildings.
The total wind flow through a city
is the result of the relationships
among all the buildings. Thus, in
planning a new building, its
location among other buildings as
well as its shape must be taken
into account. In some cases, when
this has not been done, dangerous
winds around tall buildings have
resulted in blown-out windows. A
famous example involved the
John Hancock building in Boston.
20. Recall that a city can receive less sunlight than the countryside because of the particles in the
atmosphere over cities. Often, urban areas have 10 or more times more particles than surrounding
areas. In spite of the reduced energy received from the sun, cities are warmer than surrounding areas (a
city is a heat island), for two reasons. One is increased heat production (the burning of fossil fuels and
other industrial and residential activities). The other is a decreased rate of heat loss, partly owing to the
abundance of building and paving materials, which act as solar collectors.
21. Until modern times, it was common to make use of solar power to heat city houses. Our century is a
major exception to this approach, because cheap and easily accessible fossil fuels have led us to forget
certain fundamental lessons. Cities in ancient Greece, Rome, and China were designed so that houses
and patios faced south and passive solar energy applications were accessible to each household. Today,
we are beginning to appreciate the importance of solar energy once again. Some cities have enacted
solar energy ordinances that make it illegal to shade another property owner’s building in such a way
that it loses solar heating capability.
74
Water in the Urban Environment
22. The construction of modern cities affects the water cycle greatly, in turn affecting soils and,
consequently, plants and animals in the city. Paved city streets and city buildings prevent water
infiltration. As a result, most rain runs off directly and is channeled into storm sewer systems. Hard city
surfaces also prevent water in the soil from evaporating to the atmosphere. In natural ecosystems,
evaporation is an important way of cooling the surface. City pavement increases the chances of local
flooding within the city, and the increased runoff from the city to the countryside can increase the
chances of flooding downstream. New methods of managing storm water can alleviate this problem by
controlling and reducing the speed and quality of water running off pavements and into streams. As an
example, a plan for the Alexandria, Virginia, central library parking lot includes wetland vegetation
and soils. These temporarily absorb runoff from the parking lot. The vegetation removes some of the
pollutants, and the plantings slow down the speed of water flow.
23. Because of reduced evaporations, midlatitude cities generally record a lower relative humidity (2%
lower in winter to 8% lower in summer) than the surrounding countryside. At the same time, cities can
have higher local rainfall than their surroundings, because dust above a city provides particles for
condensation of raindrops. Some urban areas have 5% to 10% more rain and considerably more cloud
cover and fog than do surrounding areas. Fog is particularly troublesome in the winter and may impede
ground and air traffic.
Pollution in the City
24. Everything is concentrated in a city, including pollutants. City dwellers are exposed to more kinds of
toxic chemicals in higher concentrations and to more human-produced noise, heat, and particles than
are their rural neighbors. This environment makes life riskier. Lives are shortened by an average of one
to two years in the most polluted cities in the United States. The city with the greatest number of early
deaths is Los Angeles, with an estimated 5,973 early deaths per year, followed by New York with
4,024, Chicago with 3,479 Philadelphia with 2,590, and Detroit with 2,123.
25. Some urban pollution comes from motor vehicles, which have contributed lead in gasoline (where it is
still used), nitrogen oxides, ozone, carbon monoxide, and other pollutants from exhaust. Stationary
power plants also produce harmful pollutants. Home heating is a third source, contributing particles,
sulfur oxides, nitrogen oxides, and other toxic gases. Industries are a fourth source, contributing a wide
variety of chemicals. The primary sources of particulate pollution, which consists of smoke and soot
and tiny particles formed from emissions of sulfur dioxide and volatile organic compounds, are older,
coal-burning power plants, industrial boilers, and gas- and diesel-powered vehicles.
26. Although it is impossible to eliminate exposure to pollutants in a city, it is possible to reduce the
exposure through careful design, planning, and development. For example, when lead was used in
gasoline, exposure to lead was greater near a road than away from it. Exposure to lead could be
reduced by placing houses and recreational areas away from roadways and by developing a buffer zone
that made use of trees that were resistant to the pollutant and that absorbed pollutants and slowed the
rate of spread.
75
Appendix B – Careful reading test.
Part 2 - Detailed Reading - (13 points)
You have 45 minutes for this part.
Read the text and write your answers in the spaces provided below. Where blanks are
provided, give short answers of one or more words.
The questions are in the order in which the information appears in the text (1 point each
question).
1. As nuclear plants became common, what unexpected disadvantage of nuclear power became
known?
________________________________________________________________________
2. Which aspect of coal-generated electricity is most harmful to human health?
________________________________________________________________________
3. What measurement is used by Dr. Bernard L. Cohen to show the danger of different methods of
generating electricity?
________________________________________________________________________
4. Why does the production of nuclear power produce greenhouse gases?
________________________________________________________________________
5. Which source of energy could possibly be replaced by nuclear power in electricity production?
________________________________________________________________________
6. Why are the young especially vulnerable to radiation?
________________________________________________________________________
76
7. Complete the following sentence.
The examples of Oak Ridge, Hiroshima, and Nagasaki show that it may be more dangerous to be
exposed to ____________________ levels of radiation over a ____________________ time.
8. Complete the following sentence.
Radionuclides in the environment are so harmful because they become ___________________ by
the time they enter the human body.
9. Why did the Soviet authorities not destroy the food products damaged by radiation from
Chernobyl?
________________________________________________________________________
10. Complete the following sentence.
At Chernobyl, operators carried out a test incorrectly. They then had to ___________________ the
reactor and this caused the reactor to explode.
11. At Three Mile Island, which component of the power plant prevented the accident being much
worse?
________________________________________________________________________
12. Why is it difficult to assess the harm caused by the Three Mile Island accident?
________________________________________________________________________
13. According to the writer, what is the main cause of nuclear accidents?
________________________________________________________________________
______/13
77
Nuclear Power and Safety
1. Nuclear power was first developed during World War II by the United States, but initially research
focused on military applications such as the use of nuclear technology to build bombs and power
submarines. The U.S. Atomic Energy Act of 1954 paved the way for private industry to develop
nuclear energy, and by 1983, 80 nuclear power plants were operating in the United States. As nuclear
plants multiplied in the 1960s and 1970s, however, the public naturally became increasingly concerned
about their safety. It was also discovered that nuclear power was not as inexpensive as had been
predicted. In fact, in the United States electricity generated by nuclear power plants is currently more
expensive than electricity generated by any other major technology. Nuclear-generated electricity can
cost two to four times as much as electricity generated at a plant powered by natural gas, coal, or wind.
Nuclear power plants are very expensive to build, mining and refining uranium ore is a considerable
undertaking, accidents and malfunctions result in significant monetary losses, and the ultimate
decommissioning of plants will require considerable sums of money (which could ultimately raise the
cost of nuclear power to even higher levels).
Advantages and Disadvantages of Nuclear Power
2. Advocates of nuclear power rightfully claim that in practice, nuclear power generation has thus far
proven to be the safest form of large-scale commercial power generation. Nuclear power plants, so far,
are much safer (in terms of human lives lost) and less damaging to the environment than fossil fuel—
burning power plants, and they are also safer than hydroelectric power plants. Nuclear plants have a
solid safety record despite several widely publicized accidents and the fact that minor accidents occur
routinely at nuclear facilities. Other sources of electricity generation also have many accidents, though
they do not seem to attract as much attention from the media and the public as accidents at nuclear
power plants. Furthermore, coal-fired power plants kill people on a routine basis. The air pollution they
give off literally kills innocent members of the public. Additionally, the acid rain and greenhouse gases
they release are causing untold property damage (estimated at billions of dollars a year in the United
States alone), disrupting ecosystems, and potentially changing the global climate. Hydroelectric power
plants, though apparently clean and natural, cause untold environmental havoc by flooding upstream
areas and reducing water flow downstream. The disruption they cause to natural water flow can
encourage the spread of disease-bearing organisms, and a large dam failure could conceivably kill
hundreds of thousands of people and cause billions of dollars worth of property damage.
3. Dr. Bernard L. Cohen, a professor of physics and radiation health at the University of Pittsburgh, has
studied the pros and cons of nuclear power for many years. He has attempted to quantify the risk to
human life that nuclear power plants pose as compared to other types of electric power generation. To
do this, he expresses risk as the loss of life expectancy. This is the average amount that a life will be
shortened by the risk under consideration. According to Cohen, nuclear power is extremely safe and the
hazards of high-level radioactive waste, low-level wastes, and routine emissions of radioactive gases
into the environment have been greatly exaggerated in the public’s mind. In fact, radioactive
substances that naturally occur in coal are emitted into the atmosphere when coal is burned. Cohen
believes even major accidents—the so-called ‘meltdowns’—in American-design reactors pose little
risk to the public (a Chernobyl-type accident would not occur in the United States). Averaged over time,
we could expect less than five deaths per year from nuclear power plant accidents. Cohen estimates that
the average meltdown (more accurately referred to as “core damage” to the reactor) would cause 400
fatalities, mostly from cancers caused by slightly increased exposures to radiation, and a few hundred
million dollars worth of off-site damage. From this he concludes that a meltdown would have to occur
in the United States every five days to make nuclear power as dangerous to the public as coal burning.
From the perspective of monetary damage, a meltdown would have to occur once every two months to
match the off-site property damage done by coal-fired power plants.
78
Drawbacks of Nuclear Power
4. Two major categories of negative environmental impacts are associated with the use of conventional
nuclear fission reactors. The first are the types of impacts that occur with any large power plant that
uses a bulky fuel. Enormous amounts of energy, land, and materials must be utilized to build the plant;
mining the enormous amounts of uranium ore needed to feed the plant involves substantial energy and
potential environmental degradation; and during routine operations, large quantities of water are used
for cooling, and disruptive amounts of waste heat are dumped into the environment (such thermal
pollution can be extremely damaging to the natural flora and fauna). Eventually, after 30 or 40 years or
less, an aging nuclear power plant must be decommissioned; this too requires enormous amounts of
energy and materials.
5. One of the chief advantages of nuclear power is that it does not appear to produce harmful, carbon-
based greenhouse gases such as carbon dioxide, nor does it emit particulates, sulfur dioxide, and
similar harmful substances into the environment. However, greenhouse gases are produced, as we can
see by looking at the entire process, not just at the actual operation of a nuclear power plant. Many
stages of the process involve energy derived from fossil fuels. These include the building and later
decommissioning of nuclear power plants, the mining and processing of uranium ore and the
transportation and storage of the uranium fuel and spent fuel. It is sometimes suggested that nuclear
power helps to decrease the U.S. dependence on foreign oil, but in fact many processes associated with
nuclear power are driven by oil. Furthermore, nuclear power is used almost exclusively to generate
electricity, and at most a mere 6% of the oil used in the United States goes toward generating electricity.
The real competitor with nuclear power for electricity generation is coal. Coal burning is extremely
dirty and is known to be causing severe environmental degradation. Nuclear advocates rightfully point
out that the large-scale substitution of nuclear power plants for coal-burning plants would avoid many
of the problems inherent in coal-burning technology. However, nuclear power generation contributes
its own set of wastes and attendant problems.
6. The second category of disadvantages of nuclear power is inherent and specific to this technology: the
dangers of radioactivity. The very basis of nuclear power production involves radioactivity. A typical
modern nuclear power plant contains radiation equivalent to that of a thousand Hiroshima bombs, and
this leads to fears that radioactive wastes may inadvertently leak into the environment, accidents may
occur (perhaps resulting in a reactor meltdown or explosion), or fissionable isotopes may fall into the
wrong hands. Compounding these concerns are a number of widely publicized accidents, and the fact
that there is still no long-term, satisfactory method of disposing of the radioactive wastes generated by
nuclear power plants. Many people are concerned that a single “worst-case” nuclear accident could
nullify all the potential benefits of nuclear power. A 1982 study by the Sandia National Laboratory in
New Mexico predicted that a major accident in the United States might cause 50,000 to 100,000
immediate deaths and up to another 40,000 subsequent deaths due to radiation-induced cancer;
monetary damages would amount to at least $100 billion.
Radiation
7. Radioactive atoms, known variously as radioisotopes or radionuclides, are unstable atoms that undergo
spontaneous disintegration (radioactive decay) and in the process give off radiation. At the
temperatures and pressures that humans live under, nothing can stop radioactive decay from occurring.
When a radionuclide decays, it gives off particles and/or rays, and this radiation will affect any atoms
or molecules it encounters. If the radionuclide happens to be inside a human body, it will damage
atoms, molecules, and cells as it disintegrates. Cells that grow and divide rapidly are usually most
affected by radiation. Thus, fetuses and children are generally more open to radiation damage than are
79
adults. In adults, cells in the bone marrow that make red blood cells, cells lining the digestive tract, and
cells of the thyroid, stomach, testes, ovaries, lungs, and breasts are particularly sensitive to radiation.
8. An important aspect of radiation, and also of many chemical carcinogens, is that no “threshold” of
exposure is necessary to potentially produce deleterious effects. In contrast, some poisons (for example,
carbon monoxide) will have no deleterious effects on the typical individual unless the exposure reaches
a certain “threshold” level. Since a single particle of radiation, or a single molecule of many chemical
carcinogens, can cause damage to a single molecule in a single cell, resulting in cancer or some other
disease, any level of radiation—no matter how small—is dangerous. Studies have also suggested that a
short, high-level dose of radiation may in some cases be less harmful in the long run than chronic
exposure to much lower levels of radiation over extended periods of time. A 1991 study of employees
exposed to low levels of radiation (supposedly acceptable levels) over long periods of time at the U.S.
Oak Ridge National Laboratory revealed that leukemia rates were 63% higher among those exposed
than among an equivalent nonexposed population. In contrast, the children of Japanese parents who
were exposed to large one-time blasts of radiation during the bombing of Hiroshima and Nagasaki do
not show such high incidences of leukemia.
9. Nuclear power plants, and the activities associated with them (such as fuel mining, processing, and
disposal), invariably release radionuclides into the environment. Even small releases of radionuclides
can be very dangerous for a number of reasons. In natural systems, these radioactive elements tend not
to disperse, but rather are concentrated by either physical or biological processes. If a radioactive
substance is spread evenly over an area, it will become approximately 10 times as dense in each step of
the food chain. Thus, for example, if radioactive strontium from a nuclear fallout covered a field, it
would first be taken up by the grass. Next cows eating the grass would concentrate the strontium
further in their milk and tissues, and finally very dense amounts of the material would be found in the
humans who drank the cow’s milk or ate their flesh. Furthermore, radionuclides are not dispersed
throughout the organism that consumes them, but are concentrated in certain tissues or organs. Thus,
strontium-90 replaces the calcium atoms in bones and there does its damage, increasing the risk of
leukemia and other blood-related diseases. Radionuclides, like other pollutants and toxins in general,
also tend to be more damaging to the young and otherwise vulnerable (for instance, the weak or very
old).
10. Of course, radiation from nuclear power plants is not a problem if it is not released into the
environment. Unfortunately, however, radiation is sometimes inadvertently released, as occurred with
the Chernobyl and Three Mile Island accidents.
The Chernobyl Disaster
11. In April 1986, a major nuclear reactor blew up, creating the single worst nuclear power plant disaster
so far. The Chernobyl nuclear reactor site, located on the Pripet River about 80 miles (130 km) from
Kiev in Ukraine, consisted of four reactors. The Number 4 reactor exploded, releasing an estimated 185
to 250 million curies of radioactivity into the environment (official Soviet estimates initially stated that
50 million curies were released, but it is now known that these numbers are too low). Within the first
few months, according to official reports, 31 people died as a direct result of the explosion and the
release of radiation (this includes a dozen firefighters who died of radiation poisoning), but the
unofficial reports of mortalities are much higher. Dr. Vladimir Chernousenko, a nuclear physicist who
was the scientific supervisor of the emergency damage control team sent into Chernobyl after the
accident, has suggested that as many as 5000 to 7000 people may have died as a result of being
exposed to radiation during the cleanup operations. The number who will ultimately die prematurely
because of the incident could number in the tens of thousands. Immediately following the accident
hundreds of people were diagnosed as having radiation sickness, and over 100,000 people living within
80
18.5 miles (30 km) of Chernobyl had to be evacuated. It has been estimated that as many as 40,000 to
70,000 additional cancer deaths (many of them occurring outside the former Soviet Union) may result
from the effects of Chernobyl over the next several decades. Chernousenko has suggested that perhaps
35 million people have been damaged by the high levels of radiation released Chernobyl.
12. In Ukraine people in the contaminated areas suffered from numerous sicknesses, their immune systems
weakened by the radiation. Trees and animals very close to the reactor simply died, but further away
they suffered mutations manifested as deformities and abnormalities. Despite the dangers, the Soviet
government did not ban or destroy large quantities of contaminated meat, milk products, and
vegetables grown in regions adversely affected by the accident. Instead, the food reportedly was
shipped to other parts of the country and mixed with uncontaminated produce. The former Soviet
Union continually suffered from food shortages so the government apparently decided that this would
be a better course of action.
13. The financial costs of the Chernobyl accident totaled about $13 billion four years after the incident, and
no one knows what the ultimate costs may be. The direct costs are expected to be greater than the
Soviet government’s total investment in nuclear power before the accident. Crude financial costs
include not only the loss of an expensive nuclear reactor and the “cleanup” and evacuation costs, but
also medical expenses, lost food production, loss of agricultural land, loss of villages and cities that had
to be abandoned, and lost business investments and potential. Damage, and therefore costs, due to the
accident will continue to manifest themselves for centuries.
14. The Chernobyl accident was initiated by operator error, but various design defects (such as inadequate
control systems and the lack of a containment vessel) made matters worse. The accident occurred
during a test of the backup electrical system. The operators were supposed to slow down the reactor so
that it would run at a low level during the test, but they found they had slowed it too much. To speed up
the reactor again, they pulled out too many control rods and reduced the amount of water cooling the
reactor. The reaction sped up, but quickly went out of control. Intense heat caused an explosion that
blew the roof, weighing about 1000 tons (900 metric tons), off the reactor and shot radioactive
elements into the atmosphere, and a major fire started.
15. Only the brave work of firefighters and other emergency workers kept the fire from spreading to an
adjacent nuclear reactor. Thousands of tons of boron, lead, and other radiation- absorbing materials had
to be dropped onto the reactor, but unfortunately much or all of this material may have missed the
actual core. Nitrogen was pumped under the reactor vessel to cool it, and finally the damaged reactor
was entombed in reinforced concrete to try to contain the remaining radiation. No one is sure how long
this tomb will hold; the radiation may begin to destroy it and leak through. Already there are reports
that this concrete tomb is riddled with holes, structurally unstable, and could collapse at any time.
Approximately 1000 square miles (2600 km2) of land around Chernobyl will remain contaminated with
high levels of radioactivity into the indefinite future.
16. Irrespective of the reasons for the Chernobyl accident, the fact is that a major disaster occurred.
Furthermore, over a dozen reactors of the same design type are still operating in Russia, Ukraine, and
Lithuania. Despite the early predictions of nuclear advocates that the probability of a major catastrophe
was negligible, we now know that catastrophes do occur. Chernobyl was not even the first major
nuclear accident.
Three Mile Island and Other Accidents
17. In March 1979, a nuclear meltdown nearly occurred in the United States. At the Three Mile Island
nuclear power plant near Harrisburg, Pennsylvania, a minor problem developed in the plumbing of
81
TMI Unit 2 with the end result that cooling water drained away from the reactor and the core began to
partially melt—the worst commercial nuclear power plant accident in the United States to date.
Operator errors, a stuck valve, faulty sensors, and design errors are all partially to blame for what
happened at Three Mile Island. Although some radioactive gas was released due to the accident (only
one thousandth as much as at Chernobyl), fortunately the containment structure around the reactor held
most of the radioactivity in—about 18 billion curies of radioactivity that could have been released had
the structure failed (compare this to the mere 185—250 million curies released at Chernobyl). Nuclear
advocates have hailed the Three Mile Island accident as “proof” that a more serious accident like
Chernobyl would not occur in the United States—for the containment structure did hold. Many experts,
however, have concluded we were lucky that Three Mile Island accident was not worse. Some
scientists have calculated that the core, which reached temperatures in the range of 4800 to 5000°F
(2650—2750°C) or higher, was just short of becoming hot enough to totally melt down. If emergency
measures had not been initiated when they were, given another 20 to 30 minutes, the core might have
completely melted through the steel reactor vessel and containment unit, releasing all 18 billion curies
of radioactivity.
18. Decades after the accident, the reactor core is still highly radioactive and dangerous despite a billion
dollars spent on cleaning up (the final cost is expected to be at least twice that amount). No one knows
for certain what the health effects of Three Mile Island are. In large part that is because no one is
certain how much radiation was actually released. A few days after the accident the then-governor of
Pennsylvania, Richard Thornburgh, evacuated all young children and pregnant women from within a 5-
mile (8 km) radius of Three Mile Island as a safety precaution. Over 150,000 other people within an
approximately 15-mile (24 km) radius of the plant voluntarily left the area for several days or more. At
the least there is evidence that due to radioactive- induced damage to immune systems, Three Mile
Island contributed to the premature deaths of some elderly people in the affected area. Dairy farmers
reported that many animals died shortly after the accident, local residents have come down with
leukemia and other cancers, and one study suggested that an increase in infant mortality and severe
thyroid disorders in babies born after the accident was due to the effects of radiation exposure.
19. Besides these major accidents, over the past decades there have been literally tens of thousands of
minor accidents at nuclear power plants, thousands of emergency shutdowns, and hundreds of
thousands of cases where nuclear power plant workers were exposed to higher than permissible levels
of radiation. Overall, one can argue that the safety record of the nuclear industry is abysmal. Between
1979 and 1990, the Nuclear Regulatory Commission (NRC) recorded 33,000 mishaps at nuclear power
plants in the United States; a thousand of these were considered to be “particularly significant.” In the
United States, in 1985 alone there were 430 emergency nuclear plant shutdowns, and in 18 of these
cases, the core of the reactor was damaged. Among 14 Western countries, 150 serious accidents
occurred at nuclear power plants from 1971 to 1984.
20. While equipment and systems failures are clearly reason for considerable concern, these incidents are
primarily attributable to human error. A classic case in point is the fire that broke out at the Brown’s
Ferry nuclear power plant near Decatur, Alabama, in 1975. It was initiated by a maintenance worker
using a candle to check for air leaks around electrical cables. A meltdown was just barely averted;
several core-cooling systems were destroyed during the fire, but a supplemental pump did go into
action and directed cooling water to the reactor core. It seems that given human imperfections, nuclear
accidents are unavoidable—indeed, at least minor accidents appear to be routine for the industry.
82
Appendix C – Scoring key for the reading tests. Skimming – 1 point each – 8pts
a) 8
b) 21
c) 14
d) 4
e) 9
f) 17
g) 25
h) 2
Detailed Reading – 1 point each – 13 pts
1. As nuclear plants became common, what unexpected disadvantage of nuclear power became known?
They were expensive/cost more (than predicted)
2. Which aspect of coal-generated electricity is most harmful to humans?
Air pollution
3. What measurement is used by Dr. Bernard L. Cohen to show the danger of different methods of
generating electricity?
The loss of life expectancy
4. Why does nuclear power produce greenhouse gases?
Because stages of the process involve energy derived from fossil fuels
5. Which source of energy could possibly be replaced by nuclear power in electricity production? coal
6. Why are the young especially vulnerable to radiation?
Cells that grow and divide rapidly are usually most affected by radiation
7. The examples of Oak Ridge, Hiroshima, and Nagasaki show that it may be more dangerous to be
exposed to low levels of radiation over a long/extended time.
8. Radionuclides in the environment are so harmful because they become (more) concentrated / dense by
the time they enter the human body.
9. Why did the Soviet authorities not destroy the food products damaged by radiation from Chernobyl?
Soviet Union suffered from food shortages
10. The Chernobyl accident was caused when operators took action to speed up the reactor and the reactor
exploded.
11. At Three Mile Island which component of the power plant prevented the accident being much worse?
The containment structure
12. Why is it difficult to assess the level of harm caused by the Three Mile Island accident?
Because no one is certain how much radiation was actually released
13. According to the writer, what is the main cause of nuclear accidents?
Human error / human imperfections
83
Appendix D – Expeditious reading strategies questionnaire.
Reading Strategies Questionnaire - Skimming
Name: _____________________ Section: ___________
Part A
What was the first thing you did at the start of the exam? Tick ONE box.
a. I first read the text slowly and carefully.
b. I first read some parts of the text quickly.
c. I first looked at the questions before looking at the text.
If you selected this option, please complete the sentence below.
I read _____________ question(s). (Write the number here, e.g. 1, 2, all)
Part B
Which of the following describe how you did the task? Tick the appropriate boxes. You can tick
more than one box.
1. I read the title and the subheadings before reading the text.
2. I only read the parts of the text which seemed related to specific questions.
3. I read all of the first paragraph.
4. I read the last paragraph before some other parts of the text.
5. I tried to answer the questions in the same order as they are written.
6. I read difficult or important parts of the text twice or more.
7. I tried to understand the organization of the text.
8. I looked for relationships between different ideas in different parts of the text.
9. I read different parts of the text at different speeds.
10. I read the text in order from beginning to end.
11. I thought about the background knowledge I have about this topic.
12. I tried to understand the ideas in every sentence very clearly.
13. I translated important words and ideas into Turkish/my own language.
84
Part C
Now look at questions a, d, e & g on your skimming paper. How did you try to answer each
question? Tick the appropriate boxes. You can tick more than one box.
Question Number
a d e g
1. I guessed the meanings of unknown words in the text using the context.
2. I searched for specific names or numbers.
3. I matched words in the question with the same words in the text.
4. I matched words in the question with synonyms in the text.
5. I searched for keywords in the text related to the general topic of the
question.
6. I read the whole paragraph slowly.
7. I made inferences about the information in the text.
8. I used the subheadings in the text.
9. I looked at the beginning of the paragraph.
10. I looked at the end of the paragraph.
11. I looked at the connections between sentences.
Part D
1. Did you use any other methods to answer the questions? Please explain below:
_____________________________________________________________________________
2. What did you find most difficult about this test? (e.g., time, topic, text length, sentence structure,
unknown words)
_____________________________________________________________________________
Part E - Please answer the questions about skills in English.
1. Rank the following English language skills from 1 to 4 in their order of importance for your
university studies in the future. (1 = most important, 2 = second most important, etc)
Listening Writing
Reading Speaking
2. Rank the following English language skills from 1 to 4 in their order of difficulty for you in your
university studies so far. (1 = most difficult, 2 = second most difficult, etc)
Listening Writing
Reading Speaking
3. Rank the following reading skills in order of difficulty for you. (1 = most difficult, 2 = second most
difficult)
Skimming
Detailed Reading
85
Appendix E – Careful reading strategies questionnaire.
Name: _____________________ Section: ___________
Part A
What was the first thing you did at the start of the exam? Tick ONE box.
a. I first read the text slowly and carefully.
b. I first read some parts of the text quickly.
c. I first looked at the questions before looking at the text.
If you selected this option, please complete the sentence below.
I read _____________ question(s). (Write the number here, e.g. 1, 2, all)
Part B
Which of the following describe how you did the task? Tick the appropriate boxes. You can tick
more than one box.
1. I read the title and the subheadings before reading the text.
2. I only read the parts of the text which seemed related to specific questions.
3. I read all of the first paragraph.
4. I read the last paragraph before some other parts of the text.
5. I tried to answer the questions in the same order as they are written.
6. I read difficult or important parts of the text twice or more.
7. I tried to understand the organization of the text.
8. I looked for relationships between different ideas in different parts of the text.
9. I read different parts of the text at different speeds.
10. I read the text in order from beginning to end.
11. I thought about the background knowledge I have about this topic.
12. I tried to understand the ideas in every sentence very clearly.
13. I translated important words and ideas into Turkish/my own language.
86
Part C
Now look at questions 3, 4, 6, 8 & 11 on your detailed reading paper How did you try to answer
each question? Tick the appropriate boxes. You can tick more than one box.
Question Number
3 4 6 8 11
1. I guessed the meanings of unknown words in the text using the context.
2. I searched for specific names or numbers.
3. I matched words in the question with the same words in the text.
4. I matched words in the question with synonyms in the text.
5. I searched for keywords related to the general topic of the question.
6. I read the whole paragraph slowly.
7. I made inferences about the information in the text.
8. I used the subheadings in the text.
9. I looked at the beginning of the paragraph.
10. I looked at the end of the paragraph.
11. I looked at the connections between sentences.
Part D
1. Did you use any other methods to answer the questions? Please explain below:
_____________________________________________________________________________
2. What did you find most difficult about this test? (e.g., time, topic, text length, sentence structure,
unknown words)
_____________________________________________________________________________
3. What do you think is the purpose of the detailed reading test?
_____________________________________________________________________________
4. What do you think is the purpose of the skimming test?
_____________________________________________________________________________
87
Appendix F – Verbal protocol procedure.
Student Interview Procedure
Inform students of the purpose and procedure of the interview:
Thank you for coming to this interview. It is very helpful. The interview will last about 15 minutes and
I will record it here. Everything you say will be anonymous and your name will not be used anywhere
in the research report. You can use English and/or Turkish when you speak.
The aim of the research is to find out about how students do exam readings. So in this interview, I’ll
give you your exam papers to look at again. I’d like you to just describe what you did, from when you
opened the exam paper, to when the time finished. Try to say as much as you can about what you did,
and be honest – we are interested in what you really did, not what you should do should have done. I
will just let you speak, but I may ask you one or two questions if necessary.
We’ll start with skimming…..
Let’s go on to the detailed reading…
Ask the following questions if students do not cover them as they talk.
What did you do first after you opened the paper?
Did you do anything to get a general idea about the whole text?
How did you find the answers? What did you do/look for with each question?
Did you read the whole text? If you didn’t, how much did you read? Which parts did you miss out?
Did you check your answers? How?
Did you finish before the end of the time allowed? By how much?
Ask these questions at the end of the interview.
What do you think is the aim of the skimming/detailed reading test?
Which part was more difficult for you – skimming or detailed reading? Why?
What do you think are the differences between the way you do skimming and the way you do
detailed reading?
Is there anything else you would like to add – about the exam or the questionnaire (or this
interview)?
Thank you for your help. Good luck with the final exam on Wednesday.
88
Appendix G – Extract from proficiency exam, information for candidates.
Useful Strategies for Taking the Stage 2 ELAE
Section 3 - Reading
Part 1 - Skimming (15%)
Exam strategies for Skimming
When you skim a text, you try to find out quickly what the main ideas of the text are, and where these
main ideas are. You do not have enough time to read the whole text so you need to concentrate on
some parts of the text which will help you to see where you can find different information.
Before Answering the Questions
Before you start trying to answer the questions, it’s usually useful to preview the text to find out the
general topic and how the text is organised. This should only take you a few minutes but it will make it
easier to find specific paragraphs later on.
Read the instructions – find the part of the instructions which tells you the topic of the whole text
(“The text is about....”). This will help you to understand the more specific topic areas in the text more
quickly.
Look at the title – this may give you a better idea of the general topic. It may give you a clue about
the author’s opinion or attitude to the topic so that you will know what to expect later.
Look for diagrams/charts/tables/illustrations – not all texts have these features. But if you see them,
they could give you more clues about the text. A diagram may show you that experiments are
explained. Or a table might show in which section the results of a study are analysed.
Read the subtitles in the text – there are usually subtitles (in bold) at different points which divide the
text into different sections. The subtitles often tell you what the topic of each section is. You can look
at certain sections first when looking for a specific paragraph.
Quickly read the introduction and conclusion – just like the subtitles, the introduction (almost
always the first paragraph) might give the different areas of the topic which are discussed in the text.
In some texts there are sentences in the introduction which give an outline of the text. In the same way,
the last paragraph is often the writer’s conclusion and could contain a summary of the main points in
the text.
89
While Answering the Questions
Read through the headings on the question paper and underline key words – this will help you to
focus on the most important ideas as you look through the text. Remember that the headings are not in
the same order as the paragraphs that match them in the text.
Underline key parts of the text – this helps you to identify which sections may be useful in matching
a paragraph with one of the headings. It also helps you to remember which parts of the text you have
already looked at. You can find these parts again later if necessary.
Use your time sensibly – if it is taking you a long time to find the paragraph that matches one of the
headings, leave it and concentrate on another heading. You can come back to the difficult one at the
end or you may find the answer while looking for something else.
Read the first two sentences of each paragraph – this is very important. Very often the first or
second sentence of the paragraph introduces the main idea which will be explained in the paragraph.
You will at least get an idea of whether the paragraph is relevant and whether you should continue
reading it.
Read the last sentence of each paragraph – this is also very helpful. The last sentence often contains
a summary of the paragraph as a whole or a comment from the author. Be careful, it might also look
forward and introduce the paragraph that follows it.
Search for key words – sometimes certain words and their synonyms are repeated in the paragraph. If
many of these words are related to each other, they will give you an idea of the main topic.
Search for visual clues – there may be other visual clues in a paragraph or section of text. For
example, a lot of dates might show that the paragraph contains a history or chronology; citations might
show that the writer is discussing the literature of a topic. Also look for specific names and words that
may match those in the heading.
Read some parts of a paragraph in detail – it may sometimes be necessary to read more of the
paragraph than the first two sentences if you are unable to work out the main idea using the strategies
above. Or, if you have time, you may need to check an answer to a question. However, remember that
because of the time limit, you will not be able to read large sections of text in such detail.
Part 2 - Detailed Reading (25%)
Exam Strategies for Detailed Reading Tasks 1 and 2 When you read a text for details, you are interested in establishing the relationship between the main
ideas of the text and the important supporting information. Unlike skimming, where a you don’t have
time to read the whole text, here you need to read most or all of the text, and you are given 60 minutes
to do this.
Again different from skimming, detailed reading is linear. In other words, to understand the text, you
read from the beginning until the end. For this reason, it is important to try and answer each of the
questions in the order in which they are presented.
90
Read the instructions carefully and remind yourself of the tasks - even though you have read all the
exam information you may not remember exactly what you have to do.
Very quickly look at the titles, the subheadings, the introduction and the conclusion of both texts - this will help you get the general idea of the topic of the texts.
Read all of the questions in the task before reading the text - this will guide you to some of the
main ideas in the texts before you read them, and give you greater understanding of the organisation of
the texts. Remember that the questions are in the same order as the answers appear in the text.
Use your time sensibly - organising your time properly will help you to answer all the questions. It’s
recommended that you spend no more than 25 minutes on either task.
If you have problems answering a question, come back to it later - it’s very easy to get stuck on a
question. If you spend too much time on one question you may not have enough time for the other
ones. Perhaps you can’t see the answer at that moment, but if you have a break from it, or you review
all your answers at the end, the answer may become much clearer.
Try to identify the relationship between different ideas/information within or across paragraphs
e.g. cause/effect, differences/similarities - in order to understand the text. It is very important to
establish the relationship between its different parts. The topic sentence, which contains the main idea
paragraph, is often the first sentence of the paragraph, but it can be in the middle or at the end of the
paragraph as well. It usually signals the type of relationship that exists between different parts of the
text, and it will help you to distinguish main ideas from supporting information and details.
Write clear and concise answers - most questions can be answered no more that 10- 15 words. With
some questions, only a few words or a key phrase is needed. You can use either your own words or
words from the text. However, if you are writing out long sentences from the text, you may be
including irrelevant information, and you may lose points. Most importantly, is the answer you give
meaningful to you?
91
Appendix H – Pilot test data.
Table 4. Item analysis of pilot ER test.
Item no. Item Facility Item Discriminability
a 0.67 0.75
b 0.94 0.55
c 0.78 0.75
d 0.83 0.60
e 0.89 0.60
f 0.83 0.55
g 1.00 0.50
h 0.5 1.00
Average Test Score (out of 8) 6.44
Standard Deviation 1.65
KR-20 0.67
Standard Error of Measurement 0.95
92
Table 5. Item analysis of pilot CR test.
Item no. Item Facility Item Discriminability
1 0.72 0.71
2 0.67 0.67
3 0.94 0.50
4 0.67 0.71
5 0.67 0.86
6 0.61 0.56
7 0.72 0.63
8 0.56 0.67
9 0.72 0.44
10 0.56 0.63
11 0.44 0.63
12 0.94 0.45
13 0.89 0.50
Average Test Score (out of 13) 9.11
Standard Deviation 1.71
KR-20 0.12
Standard Error of Measurement 1.60
93
Appendix I – Sample page of coded interview.
Coding
Sample
Here
94
Appendix J – Item Analysis of the main administration of the test.
Table 6. Item analysis of main ER test.
Item no. Item Facility Item Discriminability
a 0.65 0.72
b 0.80 0.64
c 0.63 0.71
d 0.77 0.63
e 0.73 0.70
f 0.47 0.76
g 0.81 0.60
h 0.36 0.83
Average Test Score (out of 8) 5.20
Standard Deviation 1.72
KR-20 0.50
Standard Error of Measurement 1.21
Table 7. Item analysis of main CR test.
Item no. Item Facility Item Discriminability
1 0.57 0.71
2 0.58 0.68
3 0.45 0.77
4 0.53 0.70
5 0.61 0.72
6 0.55 0.70
7 0.57 0.72
8 0.36 0.78
9 0.45 0.79
10 0.43 0.78
11 0.25 0.94
12 0.44 0.78
13 0.58 0.74
Average Test Score 6.39
Standard Deviation 2.72
KR-20 0.62
Standard Error of Measurement 1.67
95
Appendix K – Test-taker previewing behaviour (data).
Table 8. Task previewing. Questionnaire data.
Previewing behaviour ER CR
n % n %
a. I first read the text slowly and carefully. 0 0 21 23.9
b. I first read some parts of the text quickly. 8 9.2 9 10.2
c. I first looked at the questions before looking at the text. 79 90.8 58 65.9
Total 87 100 88 100
Table 9. Number of test items read before referring to the text.
Number of questions ER CR
n % n %
One 7 8.9 28 48.3
More than one but less than all 11 13.9 14 24.1
All 61 77.2 16 27.6
Total 79 100 58 100
96
Appendix L – Test-taker text and task approaches (data).
Table 10. Test-taker approaches to the text and task. Student questionnaire data.
ER
n=88
CR
n=87
n % n %
1. I read the title and the subheadings before reading the text. 47 53.4 26 29.9
2. I only read the parts of the text which seemed related to specific
questions. 32 36.4 46 52.9
3. I read all of the first paragraph. 7 8.0 29 33.3
4. I read the last paragraph before some other parts of the text. 4 4.5 4 4.6
5. I tried to answer the questions in the same order as they are
written. 10 11.4 74 85.1
6. I read difficult or important parts of the text twice or more. 37 42.0 58 66.7
7. I tried to understand the organization of the text. 46 52.3 48 55.2
8. I looked for relationships between different ideas in different
parts of the text.
19 21.6 18 20.7
9. I read different parts of the text at different speeds. 55 62.5 55 63.2
10. I read the text in order from beginning to end. 24 27.3 48 55.2
11. I thought about the background knowledge I have about this
topic. 17 19.3 13 14.9
12. I tried to understand the ideas in every sentence very clearly. 16 18.2 27 31.0
13. I translated important words and ideas into Turkish/my own
language. 17 19.3 20 23.0
97
Table 11. Test-taker approaches to the text and task. Student interview data.
ER
n=10
CR
n=10
No. stu
den
ts
Freq
uen
cy
% to
tal freq
uen
cy
No. stu
den
ts
Freq
uen
cy
% to
tal freq
uen
cy
1. I read the title and the subheadings before reading the text. - - - - - -
2. I only read the parts of the text which seemed related to
specific questions.
- - - 1 1 2.2
3. I read all of the first paragraph. - - - 2 2 4.4
4. I read the last paragraph before some other parts of the
text.
- - - - - -
5. I tried to answer the questions in the same order as they
are written.
2 2 7.7 6 7 15.6
6. I read difficult or important parts of the text twice or
more.
6 8 30.8 7 9 20.0
7. I tried to understand the organization of the text. 3 4 15.4 2 3 6.7
8. I looked for relationships between different ideas in
different parts of the text.
1 1 3.8 1 1 2.2
9. I read different parts of the text at different speeds. 3 5 19.2 6 10 22.2
10. I read the text in order from beginning to end. 4 4 15.4 3 3 6.7
11. I thought about the background knowledge I have about
this topic.
1 1 3.8 3 4 8.9
12. I tried to understand the ideas in every sentence very
clearly.
- - - 2 5 11.1
13. I translated important words and ideas into Turkish/my
own language.
1 1 3.8 - - -
Total 26 100 45 100
98
Table 12. Test-taker approaches to the text and task. Teacher interview data.
ER
n=10
CR
n=10
No
. tea
ch
ers
Freq
uen
cy
% to
tal fre
qu
en
cy
No
. tea
ch
ers
Freq
uen
cy
% to
tal fre
qu
en
cy
1. I read the title and the subheadings before reading the text. 1 2 16.7 1 1 4.2
2. I only read the parts of the text which seemed related to
specific questions.
- - - - - -
3. I read all of the first paragraph. - - - 1 1 4.2
4. I read the last paragraph before some other parts of the
text.
- - - - - -
5. I tried to answer the questions in the same order as they
are written.
3 3 25.0 3 5 21.0
6. I read difficult or important parts of the text twice or
more.
1 1 8.3 2 7 29.2
7. I tried to understand the organization of the text. 2 2 16.7 2 2 8.3
8. I looked for relationships between different ideas in
different parts of the text.
1 1 8.3 - - -
9. I read different parts of the text at different speeds. 1 2 16.7 1 3 12.4
10. I read the text in order from beginning to end. 1 1 8.3 1 2 8.3
11. I thought about the background knowledge I have about
this topic.
- - - 3 3 12.4
12. I tried to understand the ideas in every sentence very
clearly.
- - - - - -
13. I translated important words and ideas into Turkish/my
own language.
- - - - - -
Total 12 100 24 100
99
Appendix M – Test-taker item response strategies (data).
Table 13. Numbers of students using different strategies at least once during the test.
Questionnaire data.
ER
n=88
CR
n=86
n % n %
1. I guessed the meanings of unknown words in the text using the
context. 24 27.3 41 47.7
2. I searched for specific names or numbers. 78 88.6 63 73.3
3. I matched words in the question with the same words in the text. 48 54.5 54 62.8
4. I matched words in the question with synonyms in the text. 62 70.5 50 58.1
5. I searched for keywords in the text related to the general topic of
the question. 72 81.8 75 87.2
6. I read the whole paragraph slowly. 32 36.4 73 61.6
7. I made inferences about the information in the text. 40 45.5 46 53.5
8. I used the subheadings in the text. 56 63.6 34 39.5
9. I looked at the beginning of the paragraph. 61 69.3 35 40.7
10. I looked at the end of the paragraph. 45 51.1 32 37.2
11. I looked at the connections between sentences. 35 39.8 54 62.8
100
Table 14. Numbers of students reporting different strategies while responding to items. Interview
data.
ER
n=10
CR
n=10
N
o. stu
den
ts
Fre
qu
ency
% to
tal fre
qu
ency
No
. stud
ents
Fre
qu
ency
% to
tal fre
qu
ency
1. I guessed the meanings of unknown words in the text
using the context. 2 2 1.4 4 5 4.7
2. I searched for specific names or numbers. 7 12 8.5 6 9 8.4
3. I matched words in the question with the same words in
the text. 7 13 9.2 8 22 20.6
4. I matched words in the question with synonyms in the
text. 8 18 12.8 4 7 6.5
5. I searched for keywords in the text related to the
general topic of the question. 10 38 27.0 7 17 15.9
6. I read the whole paragraph slowly. 2 4 2.8 6 18 16.8
7. I made inferences about the information in the text. 1 1 0.7 3 9 8.4
8. I used the subheadings in the text. 9 23 16.3 8 13 12.1
9. I looked at the beginning of the paragraph. 9 17 12.1 2 3 2.8
10. I looked at the end of the paragraph. 7 13 9.2 1 1 0.9
11. I looked at the connections between sentences. - - - 3 3 2.8
Total 141 100 107 100
101
Table 15. Numbers of students reporting different strategies while responding to items. Teacher
interview data.
ER
n=3
CR
n=3
No
. teach
ers
Fre
qu
ency
% to
tal fre
qu
ency
No
. teach
ers
Fre
qu
ency
% to
tal fre
qu
ency
1. I guessed the meanings of unknown words in the text
using the context. - - - - - -
2. I searched for specific names or numbers. 1 3 4.1 2 6 7.6
3. I matched words in the question with the same words in
the text. 3 3 4.1 3 9 11.4
4. I matched words in the question with synonyms in the
text. 3 10 13.5 3 10 12.7
5. I searched for keywords in the text related to the
general topic of the question. 3 23 31.1 3 17 21.5
6. I read the whole paragraph slowly. - - - 2 11 13.9
7. I made inferences about the information in the text. - - - 1 2 2.5
8. I used the subheadings in the text. 3 16 21.6 3 9 11.4
9. I looked at the beginning of the paragraph. 3 14 18.9 3 4 5.1
10. I looked at the end of the paragraph. 1 4 5.4 - - -
11. I looked at the connections between sentences. 1 1 1.4 3 11 13.9
Total 74 100 79 100
102
Appendix N – Additional interview data.
Table 16. Additional coding categories from the student interview data.
ER
n=10
CR
n=10
N
o. stu
den
ts
Fre
qu
ency
% to
tal fre
qu
ency
No
. stud
ents
Fre
qu
ency
% to
tal fre
qu
ency
1. Test-strategy shortcuts to locate answers to items 2 5 12.5 5 6 14.6
2. Use of graphic clues - - - - - -
3. Making predictions of text organisation 3 10 25 2 4 9.8
4. Explicitly stated non-linear reading 2 2 5 - - -
5. Consciously skipping sections of text 4 4 10 7 15 36.6
6. Guessing answers 1 3 7.5 2 2 4.9
7. Identifying relationships between ideas - - - 3 9 22.0
8. Limited careful reading – less than one paragraph 8 16 40 3 3 7.3
9. Analysing the questions - - 2 2 4.9
Total 40 100 41 100
103
Table 17. Additional coding categories from the teacher interview data.
ER
n=3
CR
n=3
No
. teach
ers
Fre
qu
ency
% to
tal fre
qu
ency
No
. teach
ers
Fre
qu
ency
% to
tal fre
qu
ency
1. Test-strategy shortcuts to locate answers to items 1 1 6.2 3 7 11.4
2. Use of graphic clues 1 3 18.8 - - -
3. Making predictions of text organisation 2 4 25 3 9 14.8
4. Explicitly stated non-linear reading - - - - - -
5. Consciously skipping sections of text 2 4 25 3 18 29.5
6. Guessing answers - - - 1 1 1.6
7. Identifying relationships between ideas - - - 3 11 18.0
8. Limited careful reading – less than one paragraph 2 4 25 3 8 13.1
9. Analysing the questions - - - 3 7 11.4
Total 16 100 61 100
104
Appendix O – Test difficulty (data).
Table 18. Greatest difficulties encountered with the expeditious reading test.
Difficulty Frequency
Time 34
Text length 22
Unknown vocabulary 7
Sentence structure 5
Text organization 5
Distracting paragraphs 4
Topic 3
Concentration 1
Understanding/answering items 1
Finding keywords 1
Others 4
Table 19. Greatest difficulties encountered with the careful reading test.
Difficulty Frequency
Unknown vocabulary 25
Time 24
Text length 17
Topic 14
Sentence structure 8
Concentration 4
Understanding/answering items 3
Finding keywords 1
Table 20. Most difficult reading subtest.
Expeditious Reading Careful Reading
n % n %
45 54.2 38 45.8
105
Appendix P – Test aims (data).
Table 21. Perceived aims of the expeditious reading test.
Aim Frequency
Reading for overall meaning/main idea 31
Reading quickly 15
Searching for information 15
General comprehension 7
Understanding paragraph ideas 6
Extracting/focusing on certain details 2
Understanding test organisation 1
Table 22. Perceived aims of the careful reading test.
Aim Frequency
General reading comprehension 26
Full/detailed text comprehension 23
Extracting/focusing on certain details 14
Searching for information 10
Reading for overall meaning/main idea 3
Concentrating 2
106
Appendix Q – Ethics documents
Removed from this unbranded version – available on request
107
Appendix R – Urquhart and Weir’s matrix of reading types (1998)
Additional characteristics provided in italics.
Reading Types
Global level Local level
Expeditious
Reading
Potentially non-
linear
Reader-driven
Fast processing
Selective sampling
of text
Skimming quickly to
establish discourse topic
and main ideas, or
macrostructure of text, or
relevance to needs
Scanning to locate
specific points of
information
Search reading to locate
quickly and understand
information relevant to
predetermined needs
Careful Reading
Linear
Text-driven
Slow Processing
Full comprehension
Establishing accurate
comprehension of
explicitly stated main ideas
and supporting details
across sentences
Establishing accurate
comprehension of
explicitly stated main
idea or supporting
details within a
sentence
Making propositional
inferences
Identifying lexis
Establishing how ideas and
details relate to each other
in a whole text
Understanding syntax