an empirical examination of relationship magnitude

AN EMPIRICAL EXAMINATION OF RELATIONSHIP MAGNITUDE

by

Susan L. GolicicUniversity of Oregon

and

John T. MentzerUniversity of Tennessee

INTRODUCTION

The business press and interorganizational relationship literature have advocated the need forfirms to build and manage closer, longer-term relationships with suppliers and customers. However,in their study on buyer-seller relationships, Cannon and Perreault (1999) found that some buyer firmsdo not want or need close ties with all of their suppliers. Their results show that different types ofinterorganizational relationships dominate in different situations, and each relationship requiresdifferent types and degrees of investment and produces different outcomes. Other authors agree thatthere is no one relationship that is appropriate or necessary for all situations (Day 2000; Lambert,Emmelhainz, and Gardner 1996; Mentzer, Min, and Zacharia 2000). One reason for this is that it isnot possible to pursue certain types of relationships, such as partnerships, with all suppliers or cus-tomers because the implementation costs are too great in terms of capital, time, and effort (Lambert,Emmelhainz, and Gardner 1996; Mentzer, Min, and Zacharia 2000). Therefore, analogous to main-taining a portfolio of different investments, a firm is involved in a wide range of different relation-ship structures with suppliers and customers.

Coalescing previous literature with the results of their qualitative study, Golicic, Foggin, andMentzer (2003) break relationship structure down into two distinct components, magnitude and type,to better explain the many possible interorganizational relationships. Relationship type, similar tothe extant literature, was conceptualized as the group or class of relationships that share commongovernance characteristics. The authors defined the little-researched relationship magnitude as thedegree or extent of closeness or strength of the relationship between organizations. The distinctionis important to both interorganizational relationship theory and practice as it provides a betterunderstanding of the variety of different relationships firms need to manage.

Thus, it is critical to study the nature of interorganizational relationship structure (i.e., magnitudeand type) to fully explain and understand a growing phenomenon – the existence of various forms

JOURNAL OF BUSINESS LOGISTICS, Vol. 27, No. 1, 2006 81

of relationships in the supply chain. It is equally critical for practitioners to have this understand-ing to better manage interorganizational relationships. Although the concept of relationship type hasbeen examined, the nature of relationship magnitude, and its role with relationship type in the struc-ture of an interorganizational relationship, has received limited empirical investigation.

Therefore, the principal objective and contribution of this research is to examine the constructof relationship magnitude in more detail, including operationalizing the construct, testing its com-ponents, and differentiating it from relationship type to understand its role in contributing to the per-ception of the value of a relationship. This will not only add new knowledge to existing literature,but it will also coalesce research findings in this area into a more parsimonious conceptualizationof interorganizational relationship structures. The study described was specifically designed toanswer the question, “What is the effect of the level of relationship magnitude on relationship typeand on the perception of value from the relationship?” The following section briefly describes thetheoretical foundation for the study. The quantitative methodology to test this foundation is then pre-sented, followed by a discussion of the results. Finally, implications and opportunities for futureresearch are provided.

CONCEPTUAL FOUNDATION

The theoretical framework addressing the research question is presented in Figure 1. Theframework was deduced from a combination of a review of the literature and observations of the rela-tionship magnitude phenomenon in practice through depth interviews. The purpose of the depth inter-views was to supplement knowledge obtained from the literature and to help guide the creation andadaptation of measures for the quantitative study. Therefore the interviews followed typical proto-cols recommended by Churchill (1979) and McCracken (1988). The sample was purposefullydrawn from employees involved in managing relationships with suppliers and/or customers. A totalof 14 depth interviews representing 3 different supply chains (automotive, pharmaceutical, andplastics) were conducted. Those interviewed held various positions within their companies, rang-ing from Materials Supervisor to Senior Vice President. All respondents were assured of confi-dentiality, and all interviews were audio taped for subsequent transcription to minimize researcherbias and support measure quality and reliability.

82 GOLICIC AND MENTZER

FIGURE 1

THEORETICAL FRAMEWORK OF RELATIONSHIP STRUCTURE

Relationship Magnitude

Relationship magnitude has always existed; it just has not been recognized as a distinct com-ponent of relationship structure until recently. This construct is defined as the degree or extent ofcloseness or strength of the relationship among organizations (Golicic, Foggin, and Mentzer 2003).Bove and Johnson (2001) deduced that relationship magnitude is a higher order construct com-prised of existing relationship constructs depending upon the context of the relationship (e.g., trust,commitment, and/or dependence). These authors stated that other relationship variables (e.g., sat-isfaction, cooperation, information) are antecedents of these dimensions. Donaldson and O’Toole(2000) operationalized a similar construct, relationship strength, through first order measures com-monly used for trust, commitment, and dependence.

Mentzer et al. (2001a and 2001b) state that relationships vary on their levels of trust, commitment,dependence, organizational compatibility, vision, leadership, and top management support. Thehigher the levels of these, the closer the firms are to an integrated relationship. Rinehart et al. (2004)identified trust, commitment, and the frequency of interaction. Trust, commitment, and depen-dence have each been considered a component of intimacy in personal relationships and linked tosocial and structural bonds in buyer-seller relationships (Mentzer, Min, and Zacharia 2000; Thibautand Kelley 1959; Wilson 1995). They have also been considered antecedents of interorganizationalrelationships in both conceptual and empirical studies (for example, see Ganesan 1994; Geyskenset al. 1996; Gundlach, Achrol, and Mentzer 1995; Gundlach and Cadotte 1994; Morgan and Hunt1994). Both trust and commitment are related to a market orientation (Baker, Simpson, and Siguaw1999) and, along with dependence, are also components of social exchange theory (Lambe, Wittman,and Spekman 2001) and relationship marketing theory (Lewin and Johnston 1997). Monczka et al.

Relationship Type

Relationship Value

Commitment Dependence Trust


Relationship Structure

H2

H1c H1b H1a

H3


(1998) conceptualize all three as attributes of a relationship. Trust, commitment, and dependencehave also been empirically related to each other (Geyskens et al. 1996; Kumar, Scheer, and Steenkamp1995; Wetzels, de Ruyter, and van Birgelen 1998). Therefore, trust, commitment, and dependenceappear to be the dimensions of the higher order construct, relationship magnitude.

Consistent with this literature, the interviews supported trust, commitment, and dependence asthe three most often mentioned components of relationship magnitude. One interview participantcommented on what was present in a close relationship that was not in others, “Trust is the numberone thing. And it’s built on both of us saying I’m going to trust you. That we’re going to deal witheach other as if we’re one company, and we’re not going to [hurt] you and you’re not going to [hurt]us. You know you said you were going to do this and this is what you did.” Another respondent said,“I don’t think you could have a close relationship without being honest.” The opposite situation wasalso addressed by respondents. If trust is lacking, a high level of magnitude is not possible, “There’sa deep seated lack of trust I think on both sides – at this point that would have to be overcome if wewere ever going to change the relationship.”

Anderson and Weitz (1989) said that you cannot get benefits from a relationship unless youbelieve it will last. This relationship continuity is a function of trust, which they define as the beliefthat needs will be fulfilled in the future by the other party’s actions. Ganesan (1994) states it is a will-ingness to rely on a party in whom one has confidence. Another definition of trust is confidence inthe reliability and integrity of the other party (Moorman, Deshpande, and Zaltman 1993; Morganand Hunt 1994). Although definitions vary slightly, most authors operationalize trust through hon-esty and benevolence (Andaleeb 1995; Doney and Cannon 1997; Ganesan 1994; Wetzels, de Ruyter,and van Birgelen 1998). For this study, trust follows prior research and is defined as the willingnessto rely on an exchange partner in whom there is confidence in their honesty and benevolence. Trustis hypothesized to be one dimension of relationship magnitude.

H1a: Trust is a dimension of relationship magnitude.

Besides trust, respondents spoke of an understanding of expectations between firms and beingcommitted to the relationship when the level of magnitude was high. An interviewee described a sit-uation with a customer, “Because they know we’ll do [what is asked of us] – it’s the nature of therelationship. That’s their expectation with me and that’s my expectation.” Another said to have anintense relationship, “it gets down to commitment and doing what you say you’re going to do.” Whendiscussing how strong one particular relationship was, a participant described the level of commitment,“And when you run into the inevitable bumps in the road where on a given day we don’t have as manytrucks as they want, the alliance is strong enough to endure those issues that periodically crop up.Because it’s in our mutual interest to work them out.”

Creating a close relationship requires each party to dedicate resources to the relationship andto assume risk. The willingness to make these sacrifices allows companies to realize long termbenefits (Anderson and Weitz 1992). Morgan and Hunt (1994) state that commitment is a belief thatthe relationship is so important it warrants maximum efforts to maintain it. It is the intention to con-


tinue the relationship (Geyskens et al. 1996). Many authors agree there are two components tocommitment – 1) attitudinal or affective commitment, which is an enduring positive regard for theother party, and 2) instrumental or calculative commitment, which is actions or investments takenthat demonstrate a party’s intention for the future of the relationship (Geyskens et al. 1996; Gundlach,Achrol, and Mentzer 1995; Sollner 1999; Wetzels, de Ruyter, and van Birgelen 1998). Commitmentis thus defined as the willingness to exert effort to continue the relationship and is the seconddimension of relationship magnitude.

H1b: Commitment is a dimension of relationship magnitude.

When discussing closeness in relationships, respondents said they were careful not to takeadvantage of the power they might have. They believed this would not only drive the magnitude down,but possibly end the relationship, “We have the ability to compel price in the market and to the extentthat we are ever perceived to abuse that power, I can see where one of them would go out and tryto find an alternative to us.” Rather, companies strive to develop mutual dependence by followingthrough on what is expected, “I think that’s how you create those relationships that they can dependon you.” Dependence should not be one-way; respondents frequently spoke of both working togetherand mutually needing each other, “If the two parties don’t work together then I’m telling you we aren’tgoing to be successful.”

In any dyad, both members are dependent upon the relationship to some degree. Dependenceexists when one party does not entirely control all of the conditions necessary to achieve a desiredoutcome performed by the other party (Emerson 1962). When one party is dependent upon another,that party wants to continue the relationship. However, when one party is not dependent upon theother, there is little motivation to develop a strong cooperative relationship (Ganesan 1994). Theseideas are often measured through importance, the number and attractiveness of alternatives, and switch-ing (Andaleeb 1995; Ganesan 1994; Gundlach and Cadotte 1994; Heide and John 1988; Wetzles,de Ruyter, and van Birgelen 1998). For this study, dependence is the perception of the need for oneparty to maintain the relationship to achieve desired goals (Frazier 1983) and is a dimension ofrelationship magnitude.

H1c: Dependence is a dimension of relationship magnitude.

Relationship Type

Interorganizational relationships have historically been characterized by where they fall on agovernance spectrum. The channels literature was the first to propose a range of relationships fromarms length transactions (market governance) to vertical integration (hierarchical governance) withcooperative relationships (hybrid governance) in between (Contractor and Lorange 1988; Heide 1994;Webster 1992). These studies attempted to characterize relationships under a particular type basedon how the transactions between parties were organized. Relationship type can be defined as the groupor class of relationships that share common governance characteristics and are operationalized


through variations under the primary accepted categories of arms length, cooperative, and inte-grated relationships (Bove and Johnson 2001; Golicic, Foggin, and Mentzer 2003).

However, the type of relationship only explains part of the structure of a relationship. Thedepth interviews conducted for this study revealed that executives discussed levels of relationshipmagnitude (i.e., from distant to very close) within different types of supplier and/or customer rela-tionships. When asked about their firm’s alliances, one interview participant responded, “I’d liketo think we have a very close relationship with [other firm], an open relationship.” Another gavean example of close relationships with two of their suppliers, “They attended all the family functionsthat we had. They ate lunch with us every day. I mean they were pretty much members of the fam-ily.” The interview participants also described distance, or a lack of closeness, in what appear to betransactional relationships, “We have thousands of other customers who come and go. There’s noopportunity to provide collaboration there.” Another described how they managed relationships thatwere not very close, “We could kind of lose that relationship with them so I would probably wantto replace them if that should happen. So I kind of keep those relationships viable so that if we doneed to change we can.”

Some authors argue a single type, such as alliances or partnerships, can span a variety of struc-tures along a continuum with varying levels of collaboration throughout its evolution (Iyer 2002;Lambert, Emmelhainz, and Gardner 1996). Therefore, another dimension of structure, relation-ship magnitude, is necessary to fully explain the wide variety of relationship structures that exist.Prior research has distinguished the two dimensions of relationship structure and proposed them tobe positively related, with magnitude antecedent to type (Bove and Johnson 2001; Golicic, Foggin,and Mentzer 2003). While this is intuitive, this specific relationship has not yet been empirically tested.Similar to many of the characteristics of relationships (Boyle et al. 1992), magnitude is generallyhighest in integrated, high in cooperative, and lower in arms length relationships (Golicic andMentzer 2005).

H2: An increase in the level of relationship magnitude increases the level of relationship type.

Relationship Value

The perception of value is the first step to quantifying or measuring the outcome of a relationship,one area that to date has eluded most firms due to the complexity of isolating the costs and bene-fits specific to a relationship. There is agreement in the literature that value is an outcome of the struc-ture or type of relationship (Barringer and Harrison 2000; Nevin 1995; Nooteboom 1999; Stern,El-Ansary, and Coughlan 1996). Doz and Hamel (1998) state that interorganizational relationshipshelp firms create value by sharing resources, sharing knowledge, and gaining access to markets. There-fore, value was believed to be a high priority outcome to test as part of the relationship magnitudetheoretical framework.

Value is defined by consumers as: whatever I want in a product, what I get for the price I pay,and what I get for what I give (Zeithaml 1988, p. 13). The author summarizes these into an overall


assessment of the utility of a product based on perceptions of what is received and what is given.Monroe’s (1990) definition of value is an evaluation of the benefits received versus the costs thatwere paid to obtain the benefits. Likewise, in the business-to-business literature, Novack, Langley,and Rinehart (1995) define value as a trade-off between a customer’s evaluation of the benefits andcosts.

When asked about value of relationships, interview respondents relied on examples of bene-fits and costs considered to explain their perceptions. Interviewees described perceived relationshipbenefits such as increased business, higher efficiencies, better visibility, decreased inventory, higherresponsiveness, allocation priority, and the sharing of knowledge. When discussing one relationship,an interviewee stated, “I think that naturally you have to spend less time and effort policing themand save more dollars.”

Not all agreed that you necessarily spend less time and effort on closer relationships. Some citedthis as a cost along with the commitment of additional resources, the decrease in leverage over theother party, information security, and other risks. One respondent admitted, “The truth is I can’t affordin terms of my time and the time of my team to engage all of those customers and try to grow theminto a close relationship. So it’s not in our interest to do that.” Many of those interviewed also saidthat they believed they get value from relationships that are not cooperative or close in that arms lengthrelationships cover their costs. One interviewee mentioned that they do not get the most value fromtheir closest alliance but from what he considered to be a lower level of relationship, “Well actuallythe customer that we get the most value from is one of the two partnerships that I referred to.” Sowhile practitioners consider the trade-off between benefits and costs when evaluating their differ-ent relationships, there is not necessarily a positive correlation between value and the structure ofthe relationship.

The results of the interviews showed that practitioners in business-to-business relationships per-ceive their overall value similar to the literature (Novack, Langley, and Rinehart 1995; Zeithaml 1988).Therefore, relationship value is defined for this study as the perception of benefits received versuscosts sacrificed from the relationship. This perceived value is not only a consequence of relation-ships, but is proposed to be a direct outcome of the type of relationship. Intuitively, this relationshipshould be a positive one; however, some of the interview findings contradict this. Because firms donot always purposefully structure their relationships and rarely measure the value of their rela-tionships, they do not always know if they are getting value from their relationships (Cannon andPerreault 1999; Cox 2001). It is therefore unknown if the relationship will be consistent and is leftfor testing to determine the direction.

H3: A change in the level of relationship type changes the level of relationship value.


SURVEY METHODOLOGY

Survey Design

Due to the covariate nature of the relationship structure model, structural equation modeling(SEM) was considered an appropriate technique to evaluate the research hypotheses (Loehlin 1998).A nonexperimental mail survey methodology was used to gather the data necessary to test themodel and its hypotheses. The unit of analysis was the respondent’s perceptions of the interorga-nizational relationship between his/her firm and a transportation provider. The sample for the sur-vey was taken from customers (or shippers) of the transportation industry (or carriers). The natureof the industry is that shippers have access to a wide variety of providers, and some may be consideredcore carriers while others are only used on an as-needed basis. Shippers therefore manage a port-folio of various relationships with their carriers. The constructs of interest were thus all expected tobe present in varying degrees.

Relationship magnitude is conceptualized here as a second order construct, and was measuredthrough its dimensions of trust, commitment, and dependence. Based on the interview information,existing measures from Morgan and Hunt (1994) were deemed appropriate for trust and commit-ment. The reported reliabilities for the two measures were 0.949 and 0.895 respectively. Dependenceitems were adapted from Ganesan (1994) and Rinehart et al. (2004). Ganesan achieved a reliabil-ity of 0.94 for his retailer dependence items. Existing measures for relationalism from Boyle et al.(1992) were appropriate for the operationalization of relationship type and were reported to achieveadequate levels of reliability and validity in their study. New items were created to supplementthese in order to more fully tap the construct definition. Since relationship value did not have exist-ing scales that were applicable to the interview perceptions, new scales were developed. Themethodology for new scale development and scale purification followed the procedures recommendedby Anderson and Gerbing (1991), Churchill (1979), and Mentzer and Flint (1997).

Survey Pretest

A pretest was conducted in order to validate the 31 measures created or adapted for thisresearch. Experts in interorganizational relationship research and survey design reviewed the draftsurvey instrument for readability and item clarity. This process provided support for the face valid-ity of the measures. A transportation provider supplied 96 total contacts that met the sample require-ments. The pretest survey implementation followed the five-step process recommended by Dillman(2000).1 To ensure variation, one of two versions of the survey (one referencing a relationship the


1 The first step involved an initial contact via telephone for prenotification. The second step was sending the sur-vey and a letter explaining the purpose of the research. A follow-up postcard was the third step. The fourth stepconsisted of another letter and survey, and the final step was a follow-up phone call to determine the status ofresponse.

respondent considered to be good and one referencing a relationship the respondent considered tobe not very good) were randomly sent in an even split to the sample. The participants contacted werethose individuals responsible for managing the relationships with carriers, and all were assuredthat their responses were confidential. Two incentives were offered to each survey recipient fortheir participation: 1) an executive summary of the results if requested, and 2) a $100 donation tothe charity of choice for five randomly selected survey respondents. Eighty-one completed sur-veys were returned for a pretest response rate of 86%.

Scale purification included tests for unidimensionality, reliability, and construct validity, followingthe procedures described by Anderson and Gerbing (1988), Garver and Mentzer (1999), Gerbing andAnderson (1988), and Steenkamp and van Trijp (1991). Unidimensionality tests run on the individualconstructs uncovered nine items that loaded poorly. These items were removed since they wereredundant and seemed to cause confusion for the respondents. Exploratory factor analysis of a ran-dom subset of the data uncovered five factors with appropriate item loadings explaining 73% of thevariance. Confirmatory factor analysis was then run on a different subset to validate the analysis,and 67% of the variance was explained by the resulting five factors. Reliability of the scales was testedusing coefficient alpha, the SEM scale reliability and variance extracted (Garver and Mentzer1999). The results for each construct were acceptable.

The magnitude, direction, and statistical significance of the estimated parameter loadingswere used to assess convergent validity. All items were significant, in the appropriate direction, andgreater than 0.60. Paired construct comparisons were conducted to support discriminant validity. Therewere no discrimination issues. The resultant final survey items are provided in the Appendix.

Final Survey Design

Once the survey instrument was deemed acceptable through the pretest, the survey was mailedto the final sample and followed the same process as described for the pretest. Four major trans-portation providers provided a total of 92 customer contacts. In addition, a sample of 544 shipperswas drawn from the membership list of the National Industrial Transportation League (one contactper firm), an organization that supports shippers and carriers, for a total sample size of 636. Respon-dents were individuals responsible for managing relationships and/or contracts with carriers attheir company.

The structural equation model and research hypotheses were evaluated using Anderson and Gerbing’s(1988) two-step approach supported by LISREL modeling software. Analysis of the measurementmodel provides a confirmatory assessment of reliability and construct validity. The test of the struc-tural model then constitutes a confirmatory assessment of nomological validity. A number of good-ness-of-fit measures were used in combination to assess the overall fit, comparative fit to the nullmodel, and model parsimony (Hair et al. 1998). These included the significance and magnitude ofthe hypothesized paths, the root mean square error of approximation (RMSEA), the comparative fitindex (CFI), and the normed chi-square adjusted for the degrees of freedom (CMIN). All of thesemeasures were chosen due to their appropriateness for use with larger samples (considered to be greater


than 150 data points like this one) and with higher order models (Marsh 1994). In addition, the tar-get coefficient, which is the ratio of the chi-square of the first order model (Mf) to the chi-square ofthe theoretical model (Mt), helped determine if the second order model was a more parsimoniousexplanation of the theory (Marsh and Hocevar 1985).

Survey Response

Of the 636 surveys sent, 22 were returned due to incorrect addresses. Another 26 indicated thatthey currently did not ship products with trucking or intermodal providers, reducing the total sam-ple to 588. Complete responses were returned from 326 shippers; however four surveys wereremoved from the data set due to the number of missing responses, for a final response rate of over54%. Each of the three waves was compared against the other two, and no significant differencesexisted (Armstrong and Overton 1977). In addition, a sample of nonrespondents were called and askedto complete five (one for each construct) of the survey items (Lambert and Harrington 1990;Mentzer and Flint 1997). The responses from 31 non-respondents were compared to the data fromrespondents, and again, no significant differences were found. To explore the possibility of false report-ing bias, the respondents were asked how long they personally and their firm have used the providerin question to transport their freight. Most of the firms (70%) and the respondents (60%) had beendealing with the carrier for over three years with the remainder having at least one year in the rela-tionship. The conclusion was made that the respondents are knowledgeable of the relationshipabout which they answered questions.

The customer respondents represented various levels of management (from supervisors toexecutives), mainly in the consumer packaged goods (14%), chemicals/plastics (14%), automotive(9%), and industrial products (19%) industries, with a wide range of sales revenue ($1 million to greaterthan $1 billion). To ensure the respondent dealt with a variety of relationships, they were asked howmany providers their firm used. Nearly all used at least six different providers, with more than halfusing 21 or greater. This is a good indication that most manage a large number of carrier relation-ships. No significant differences were found in the survey results for the above demographics.

The descriptive statistics for the survey are provided in Table 1. The means for the itemsranged from 3.02 to 5.10. Every item obtained the full range of possible answers (from 1 to 7). Stan-dard deviations for the items ranged from 1.556 to 1.900. Prior to analysis of the structural equa-tion model, the measurement model was first examined to assess the construct validity of the scales(Anderson and Gerbing 1988).



TABLE 1

FINAL SAMPLE DESCRIPTIVE STATISTICS

Correlation Matrix

Std KurtosisItem Mean Dev Skew s Trust Commit Depend Type Value

1 5.10 1.631 -0.797 -0.3902 4.97 1.657 -0.720 -0.5023 4.94 1.696 -0.787 -0.3625 5.01 1.637 -0.748 -0.4296 4.36 1.875 -0.423 -1.041

Trust 4.88 1.699 1.00

8 4.98 1.671 -0.704 -0.4779 4.30 1.848 -0.304 -1.03310 4.62 1.777 -0.571 -0.74111 3.40 1.850 0.284 -1.09112 4.49 1.900 -0.458 -0.981

Commit 4.36 1.809 0.76 1.00

15 4.32 1.652 -0.453 -0.73016 3.99 1.809 -0.207 -1.06818 4.31 1.816 -0.354 -1.026

Depend 4.21 1.751 0.60 0.63 1.00

21# 4.89 1.613 -0.735 -0.33322 3.96 1.661 -0.111 -0.97224 3.79 1.773 -0.055 -1.18825 4.71 1.621 -0.594 -0.47026 4.32 1.761 -0.291 -0.958

Type 4.19 1.686 0.77 0.81 0.64 1.00

27# 4.48 1.667 -0.506 -0.70628* 3.02 1.664 0.710 -0.38330 4.40 1.556 -0.378 -0.60831 4.64 1.616 -0.467 -0.560

Value 4.02 1.626 0.63 0.67 0.53 0.82 1.00

Rel Mag 4.48 1.753 0.85 0.89 0.71 0.91 0.75

*Negatively phrased items.#Subsequently removed items.

Measurement Model

The measures and scales were analyzed in both SPSS and LISREL. A confirmatory measure-ment model, allowing all latent variables to correlate with each other and with individual manifestvariables loading on their appropriate latent variable, was run. Even though the descriptive statis-tics for the data support a normal distribution, most data sets are not multivariate normal, which isrequired for estimation using the maximum likelihood method. Thus the data were analyzed usingthe asymptotic covariance matrix and a Satorra-Bentler scaled statistic, which provide results thatare not based on an assumed distribution (West, Finch, and Curran 1995). Various components ofthe SEM output – standardized regression weights, squared multiple correlations, modificationindices, and goodness of fit indicators – were used to confirm the unidimensionality, reliability, andconstruct validity of the scales.

Unidimensionality is demonstrated through the overall goodness of fit of the confirmatoryfactor model, and the convergence and discriminance of items (Anderson and Gerbing 1988; Gerb-ing and Anderson 1988). The fit of the measurement model was good with an RMSEA of 0.072, CFIof 0.98, and CMIN of 2.67. The regression weights of the items on the latent variables were all inthe appropriate direction and statistically significant to 0.01 (see Table 2). These results demonstratesupport for both convergent validity and unidimensionality. Four of the items were less than the rec-ommended value of 0.70 (Garver and Mentzer 1999). Modification indices show potential improve-ments if items were allowed to load on any latent variables. From these results, there were apparentissues with the potential for items 21 and 27 to cross-load on other constructs. Items 21 and 26 seemedto be phrased similarly. Item 27 was different from the other value items in that it only addressedthe benefit portion of value while the others addressed both benefits and costs. Therefore removalof these items had little impact on the meaning of the scales. The fit of the revised measurement modelimproved to an RMSEA of 0.058, CFI of 0.99, and CMIN of 2.07. The regression weights slightlyimproved as well with all above 0.60 and only three below 0.70 (these are included in Figure 2). Modification indices and squared multiple correlations were all acceptable, providing support for construct validity.


TABLE 2

CONFIRMATORY FACTOR ANALYSIS LOADINGS

Item TRU COM DEP TYPE VAL

1 0.932 0.943 0.915 0.936 0.93

8 0.899 0.8810 0.8411 0.8112 0.91

15 0.8316 0.8318 0.66

21 0.6922 0.6824 0.7825 0.8226 0.81

27 0.8728* -0.7030 0.6931 0.95

* Negatively phrased items.


FIGURE 2

THEORETICAL SECOND ORDER MODEL RESULTS

Path Estimate Critical Ratio (> 1.96) First Order Paths

RMag-Tru (�11) 0.85 18.27 Tru- > Type 0.41

RMag-Com (�21) 0.89 17.05 Com- > Type 0.33

RMag-Dep (�31) 0.71 11.03 Dep- > Type 0.22

RMag-Type (�41) 0.91 16.95

Type-Val (�54) 0.82 14.84 Type- > Val 0.83

Trust-Item 1 (�l1) 0.93 n/a*

Trust-Item 2 (�l2) 0.94 35.62

Trust-Item 3 (�l3) 0.91 25.61

Trust-Item 5 (�l5) 0.93 28.36

Trust-Item 6 (�l6) 0.93 31.44

Com-Item 8 (�l8) 0.89 n/a*

Com-Item 9 (�l9) 0.88 25.79

Com-Item 10 (�l10) 0.84 25.01

Com-Item 11 (�l11) 0.81 19.83

Com-Item 12 (�l12) 0.91 29.52

Dep-Item 15 (�l15) 0.76 n/a*

Dep-Item 16 (�l16) 0.92 16.11

Dep-Item 18 (�l18) 0.60 10.50


Relationship Type

�� γ41 Value

η5 η4 ξ1

Commitment

η2

Dependence

η3

Trust

η1

γ31 γ11 γ21

2 3 5 1 6

8 9 10 11

15 16 18

12

31 30 28 24 25 26 22

λ15-18 λ8-12

λ1-6

λ22-26 λ28-31


FIGURE 2 (CONT.)

THEORETICAL SECOND ORDER MODEL RESULTS

Path Estimate Critical Ratio (> 1.96) First Order Paths

Type-Item 22 (�l22) 0.67 12.90

Type-Item 24 (�l24) 0.77 16.40

Type-Item 25 (�l25) 0.80 16.17

Type-Item 26 (�l26) 0.80 n/a*

Value-Item 28 (�l28) -0.69 13.15

Value-Item 30 (�l30) 0.72 15.20

Value-Item 31 (�l31) 0.96 n/a*

*The t-values for these paths were not calculated as these items were fixed for scaling purposes.

The items for trust, commitment, and dependence were combined into a composite score foreach data record so the scale for relationship magnitude could be examined along with the firstorder constructs. All of the scales had coefficient alpha and construct reliability values greater thanthe acceptable minimum value of 0.70. The variance extracted for each construct was greater thanthe acceptable minimum value of 0.50 (see Table 3 for all reliability results).

TABLE 3

SUMMARY OF FINAL SCALE RELIABILITY

Reliability

VarianceConstruct Unidimensionality � Scale Extracted

Magnitude 3 items > 0.70 0.805 0.859 0.673

Trust 5 items > 0.70 0.968 0.969 0.861

Commitment 5 items > 0.70 0.937 0.939 0.755

Dependence 2 items > 0.70 (1 = 0.60) 0.809 0.811 0.596

Type (w/ 21 removed) 3 items > 0.70 (1 = 0.67) 0.861 0.848 0.584

Value (w/ 27 removed) 3 items > 0.70 0.920 0.841 0.643


RESEARCH FINDINGS

With the measurement model now purified, the hypotheses could be tested using the full struc-tural model. The first order model, which links trust, commitment, and dependence directly to rela-tionship type with no second order magnitude construct present (alternative #1) was run followedby the second order theoretical model. Both models achieved good fit (see Table 4), but were sig-nificantly different (change in chi-square of 14 with only 2 degrees of freedom). The path weightsin the second order model (Figure 2) and high target coefficient statistic (Table 4) provide supportfor the contention that relationship magnitude is a second order construct comprised of the three firstorder constructs of trust, commitment, and dependence. All path weights are large and significant;this therefore better explains the covariation among the first order dimensions and provides a moreparsimonious representation of relationship structure through magnitude and type.

TABLE 4

MODEL FIT COMPARISONS

1st Order Structural 2nd Order Structural(Alternative #1) (Theoretical Model)

Chi-square 385.31 (163 df) 399.21 (165 df)

RMSEA (< 0.08)1 0.065 0.066

CFI (> 0.90)1 0.980 0.980

CMIN (< 3.0)1 2.364 2.419

Target Coeff (approach 1.0)2 n/a 0.965

1Hair et al. (1998)2Marsh and Hocevar (1985)

Hypothesis 1

Empirical studies in the past have conceptualized trust, commitment, and dependence asantecedents to relationships or to each other (e.g., Andaleeb 1995; Ganesan 1994; Mentzer, Min, and Zacharia 2000; Monczka et al. 1998; Morgan and Hunt 1994; Wetzels, de Ruyter, and vanBirgelen 1998). However, the goodness of fit of the second order theoretical model strongly sup-ports these three constructs as dimensions. Not only are the loadings significant (0.85, 0.89, and 0.71respectively), supporting H1a, H1b and H1c, but the reliability using these three as measures of rela-tionship magnitude is high (a = 0.805, scale = 0.859, variance = 0.673). This finding is interestingin that while it helps tie some of the literature together, it may also challenge findings from prior studies.


Hypothesis 2

Hypothesis 2 states that an increase in the level of relationship magnitude increases the levelof relationship type. This hypothesis was supported as a strong positive relationship (path weight0.91) and was statistically significant to � = 0.01. Bove and Johnson (2001) state that magnitude isa function of the type of relationship, but this research conceptualizes type as a function (outcome)of magnitude. When empirically tested, this is supported. However, relationship magnitude accountsfor nearly all of the variance in relationship type. An additional exploratory factor analysis wastherefore done using the measures for type and composites of trust, commitment, and dependencefor three measures of magnitude. Two factors resulted, which were significant, unidimensional,and discriminant, adding more evidence that the two constructs are distinct. The distinction was alsosupported by those interviewed; when describing their relationships, the interviewees would men-tion different levels of magnitude (e.g., both close and distant) within one type of relationship (e.g.,a partnership). Because they are so closely related, the results support the position that magnitudeand type are components of relationship structure (i.e., they may be highly correlated rather than oneantecedent to the other). This finding is significant for supply chain management as higher levelsof one goes hand-in-hand with high levels of the other, and deserves additional consideration in futureresearch.

Hypothesis 3

Hypothesis 3 stated that a change in the level of relationship type changes the perceived levelof relationship value. Although intuitively this relationship should be positive, the hypothesis wasexploratory because prior research (the literature and depth interviews) is mixed on the perceptionof value resulting from relationships types. The survey findings support this hypothesis, however,with a strong and significant positive relationship (path weight 0.82, � = 0.01). The findings wereconsistent when the data set was analyzed by the various groups of relationship type (arms length,cooperative, and integrative, per a demographic question on the survey) – all type-value paths weresignificant and positive with path weights greater than 0.70. The survey data demonstrated thathigher levels of relationship type result in the perception of higher value and lower levels of rela-tionship type result in lower value. Overall this finding is not a surprise; however closer examina-tion of this is needed in future research to determine if this positive relationship holds true under allrelationship conditions.

DISCUSSION

The primary objective of this study was to examine the concept of magnitude in an interorga-nizational relationship and how this fit into the existing literature. Variations of magnitude in busi-ness-to-business relationships, such as embeddedness, intensity, and strength (Beekun and Glick 2001;Donaldson and O’Toole 2000; Granovetter 1973; Rindfleisch and Moorman 2001) have been dis-cussed in the literature, but none of these studies empirically linked the concept to other aspects ofrelationship structure (such as governance or type). Two studies have focused specifically on rela-


tionship magnitude. Bove and Johnson (2001) conceptualized magnitude as the strength, close-ness, or quality in business-to-business relationships and proposed it as a second order construct.However, they did not test this, but called for empirical research to do so. Expanding on this, Goli-cic, Foggin, and Mentzer (2003) conceptualized relationship magnitude as distinct from a secondcomponent of relationship structure, relationship type. These authors also proposed magnitude asa higher order construct comprised of existing relationship constructs. The current study takes thenext step of testing magnitude as a construct separate from relationship type and comprised of thethree first order constructs of trust, commitment, and dependence.

The loadings of the three dimensions of trust, commitment, and dependence on magnitude werelarge and significant, supporting H1 and the operationalization of magnitude through these corre-lated first order dimensions. Other research has implied potential causal paths between these con-structs (for example trust and commitment in Morgan and Hunt 1994). The current study does notcontradict this research, but adds more information to the data relating these constructs. All of theempirical evidence supports a statistical relationship among them. Because it is so difficult to showcausation, particularly with methods such as structural equation modeling, any new conceptualizationsof how these constructs are related enhances the knowledge and understanding of their role in supply chain relationships.

The results for the second hypothesis and the comparison of the first order model to the theo-retical model supports relationship magnitude as a construct distinct from, but related to, type. Thisfinding supports Stern and Reve’s (1980) internal political economy structure. Two distinct com-ponents are present in the structure of a relationship – the magnitude, which is associated with thepolity structure, and the type, which is associated with the economic structure.

The primary research question was the heart of the theoretical model, “What is the effect of thelevel of magnitude on relationship type and on the perception of value from the relationship?” It washypothesized that relationship magnitude would have a significant, positive effect on relationshiptype (H2) and that a change in relationship type would change the perception of value from the rela-tionship (H3). The results strongly support both of these hypotheses (with positive path weights of0.91 and 0.82 respectively). According to the model and the survey results, the effect of relationshipmagnitude on relationship value is mediated by relationship type. A high modification index of41.12 shows a potential for a direct path (with an expected value of 0.80) between relationshipmagnitude and relationship value (alternative #2). Paths to value from both magnitude and type wouldsuggest these constructs to be two dimensions of an even higher order construct, such as relation-ship structure. There was originally no theoretical basis to test this alternative model; however thispresents an interesting avenue to explore in future research.

The results showed a strong, positive relationship between type and value (H3). While this heldfor the carrier-shipper relationship when tested under different conditions (type of relationship,age of relationship), we feel that this prediction should not be automatic. The literature and quali-tative study provide evidence that there may be occasions when higher levels of relationship typedo not provide higher value. Closer relationships may sometimes cost more in terms of time to


develop and maintain it, as mentioned in the interviews. The association between the relationshipand value from it may therefore be contingent upon the specific relationship and conditions sur-rounding it. Future research should examine this with different populations under different antecedentconditions to determine if the association is consistent or varies with certain circumstances.

CONCLUSIONS AND FUTURE OPPORTUNITIES

This research is distinct in that it tests two components of relationship structure in the business-to-business context – relationship magnitude and relationship type. As stated earlier, relationship mag-nitude has always existed; it just has not been recognized as a distinct component of relationshipstructure. Relationship magnitude was measured as a second order construct comprised of the exist-ing, heavily researched relational constructs of trust, commitment, and dependence. This differs fromthe extant literature in that trust, commitment, dependence, and other constructs have been histor-ically treated as antecedent to relationship levels. Empirical results support the new conceptualization,which is an important finding.

A great deal of research has been conducted on interorganizational relationships in the chan-nels, logistics, and supply chain management literature. No theory, however, exists in the reviewedliterature that fully explains and organizes the components of relationship structure. This studyattempted to do this by clarifying how a relationship is structured. Bove and Johnson (2001) statedthat, “there is a clear need for empirical studies to validate the suggested relationship strength con-struct” which they conceptualized as a second order construct (p. 194). This empirical study doesjust what Bove and Johnson called for, and also found that relationship magnitude was distinctfrom, yet highly related to, relationship type.

A combination of existing and new measures were used and tested for relationship magni-tude and type. The dimensions of magnitude – trust, commitment, and dependence – successfullyutilized existing measures from Ganesan (1994) and Morgan and Hunt (1994), which substantiatesthese measures in the context of shipper-carrier relations. Interestingly, one existing item (21) forrelationship type from Boyle et al. (1992) did not produce good results. This stresses how impor-tant it is to retest measures, even those with no changes, in each study.

Lastly, this research substantiates the views of value from transaction cost economics andsocial exchange theory. In the interviews, perceived value was viewed as a judgment of benefits andcosts. Transaction cost economics favors exchange relationships that minimize transaction costs (i.e.,that provide economic value). Social exchange theory argues for exchanges that provide socialvalue. Retention of the relationship then occurs if the benefits provided in the relationship out-weigh the costs of the relationship. Transaction cost economics and social exchange theory there-fore favor closer relationships since, according to the results of this research, closer relationshipsresult in the perception of higher levels of value.

The research also provides implications for supply chain practitioners. The results offer evi-dence for trust, commitment, and dependence as components of relationship magnitude. To develop


a closer relationship, the parties must develop high levels of trust, commitment, and dependence.Much research has been done on these constructs, and the development of high levels of thesetakes time. Knowing this, suppliers can work on building higher levels of these components withthe customers with which they desire a close relationship.

A second managerial implication pertains to the value of interorganizational relationships.Practitioners may be apprehensive about investing time and resources into closer relationshipsbecause they seldom see a quick financial return on their investment. This research shows thathigher value is a perceived outcome of closer relationships. Companies must therefore realize thatpositive outcomes will result if they are patient with the relationship development process.

A very important implication of this research is that firms must understand how relationshipsare structured in order to manage a portfolio of many different relationships. The study resultsreveal a great deal of variation in the survey answers. This supports the argument that respondentsare dealing with many different relationship structures. When analyzed collectively in the theoret-ical model, the findings provide a foundation of knowledge on what is needed to help firms buildcloser relationships in certain business situations. Value can result from every relationship, but aseach relationship is different, how each of the components of magnitude is handled also differs. Oncea firm can identify relationship costs and benefits, they can develop strategies to obtain maximumvalue from their different relationships.

The study limitations provide opportunities for an ongoing stream of future research. In the inter-view phase of this research both sides of the relationship were represented. The survey, however,was sent to shippers only. Prior studies on interorganizational relationships have shown that researchcan be successfully conducted with one firm in the relationship; but results on interorganizationalrelationships that are obtained from dyadic samples are more informative. The quantitative studyshould be extended to the other side of the relationship and also to other dyadic pairs within the sup-ply chain for greater insight and generalizability. It would also be interesting to test this model onrelationships that have been terminated to determine if the paths among the constructs hold in thisextreme.

The theoretical model was deduced from the extant literature, coupled with the results ofdepth interviews. It is appropriate to acknowledge that other equivalent structural equation modelsmight provide acceptable fit as well. Modification indices in LISREL did not provide any evidencefor other paths in the model with the exception of one from relationship magnitude directly tovalue. However, there was no theoretical basis to test alternatives other than those already included.Additional research, such as experiments or longitudinal research, is needed to determine if otherconstruct relationships exist within the current framework (MacCallum et al. 1993).

Finally, more research is needed to help firms measure the value of their relationships, beyondrelying solely on perceptual evaluations. Longitudinal research may be needed to isolate theresources that are put into relationships and the benefits that result. From this, methods to attach coststo these resources could be developed. Metrics could then be formulated to quantify the evaluationof benefits and costs from the relationship to determine the overall value.


The primary contribution of this research is that it provides an empirically tested theoreticalfoundation from which to conduct future research on interorganizational relationship structure. Itis hoped this research provides a better understanding of the outcomes (i.e., higher value from thecustomer perspective) of closer relationships and some of the requirements (higher levels of trust,commitment, and dependence in the relationship) to achieve higher levels of relationship magnitude.

ACKNOWLEDGMENT

This research was made possible through the sponsorship of Standard Corporation and The Uni-versity of Tennessee College of Business Administration Scholarly Research Grant. We are partic-ularly indebted to Mr. Bill Gates of Standard Corporation and all of those organizations thatparticipated in the research.

APPENDIX

FINAL SURVEY ITEMS

Respondents were asked to choose their level of agreement or disagreement with all constructquestions according to the following scale.

NeitherStrongly Somewhat Agree Or Somewhat Strongly

Agree Agree Agree Disagree Disagree Disagree Disagree

7 6 5 4 3 2 1

Construct Questions

Trust 1. In our relationship, the provider…the willingness to rely on a. has high integrity. [1]an exchange partner in b. can be counted on to do what is right. [2]whom there is confidence c. is sincere in their promises. [3]of their honesty and d. treats my firm fairly and justly. [5]benevolence e. is a firm my firm trusts completely. [6]

Items from Morgan and Hunt (1994)

Commitment 2. The relationship my firm has with the provider…the willingness to exert a. is something my firm is very committed to. [8]effort to continue the b. is something my firm intends to maintain indefinitely. [9]relationship c. deserves my firm’s maximum effort to maintain. [10]

Items from Morgan and d. is something my firm would do almost anything to keep. [11]

Hunt (1994) e. is something my firm cares a great deal about long-term. [12]

Dependence 3. My firm…the perception of the need a. is dependent upon the provider. [15]for one party to maintain b. believes the provider is crucial to our success. [16]the relationship to achieve c. needs the provider to accomplish our goals. [18]desired goals

Items from Ganesan (1994)


APPENDIX (CONT.)

FINAL SURVEY ITEMS

Construct Questions

Relationship type 4. The business relationship my firm has with the provider could better be the group or class of described as “cooperative” rather than an “arms length.” [21]relationships that share 5. The business relationship my firm has with the provider could better be common governance described as an “integrated” rather than a “cooperative.” [22]characteristics 6. My firm and the provider coordinate some of our business functions as if

First two items from we were one company. [24]

Boyle el al (1992) 7. My firm’s relationship with the provider is more than just repeat transactions. [25]

8. My firm’s relationship with the provider could better be described as “strategic” than “transactional.” [26]

Relationship value 9. My firm receives a great deal of benefits from the relationship with the the perception of benefits provider. [27]received versus costs 10. The costs to my firm for the relationship with the provider do not justify sacrificed from the the benefits we receive. [28]relationship 11. My firm receives more benefits from the relationship with the provider

than costs put into maintaining it. [30]12. My firm gets a lot of value from the relationship with the provider. [31]

NOTES

Andaleeb, Syed Saad (1995), “Dependence Relations and the Moderating Role of Trust: Implica-tions for Behavioral Intentions in Marketing Channels,” International Journal of Research in Marketing, Vol. 12, No. 2, pp. 157-172.

Anderson, Erin and Barton Weitz (1992), “The Use of Pledges to Build and Sustain Commitmentin Distribution Channels,” Journal of Marketing Research, Vol. 29, No. 1, pp. 18-34.

Anderson, Erin and Barton Weitz (1989), “Determinants of Continuity in Conventional IndustrialChannel Dyads,” Marketing Science, Vol. 8, No. 3, pp. 310-323.

Anderson, James C. and David W. Gerbing (1991), “Predicting the Performance of Measures in aConfirmatory Factor Analysis With a Pretest Assessment of Their Substantive Validities,” Journalof Applied Psychology, Vol. 76, No. 5, pp. 732-740.

Anderson, James C. and David W. Gerbing (1988), “Structural Equation Modeling in Practice: AReview and Recommended Two-Step Approach,” Psychological Bulletin, Vol. 103, No. 3, pp. 411-423.

Armstrong, J. Scott and Terry S. Overton (1977), “Estimating Nonresponse Bias in Mail Surveys,”Journal of Marketing Research, Vol. 14, No. 3, pp. 396-402.


Baker, Thomas L., Penny M. Simpson, and Judy A. Siguaw (1999),“The Impact of Suppliers’Perceptions of Reseller Market Orientation on Key Relationship Constructs,” Journal of the Academy of Marketing Science, Vol. 27, No. 1, pp. 50-57.

Barringer, Bruce R. and Jeffrey S. Harrison (2000), “Walking a Tightrope: Creating Value ThroughInterorganizational Relationships,” Journal of Management, Vol. 26, No. 3, pp. 367-403.

Beekun, Rafik I. and William H. Glick (2001), “Organizing Structure from a Loose Coupling Perspective: A Multidimensional Approach,” Decision Sciences, Vol. 32, No. 2, pp. 227-250.

Bove, Liliana L. and Lester W. Johnson (2001), “Customer Relationships with Service Personnel:Do We Measure Closeness, Quality or Strength?” Journal of Business Research, Vol. 54, No. 2, pp. 189-197.

Boyle, Brett F., Robert Dwyer, Robert A. Robicheaux, and James T. Simpson (1992), “Influence Strate-gies in Marketing Channels: Measures and Use in Different Relationship Structures,” Journal of Marketing Research, Vol. 29, No. 4, pp. 462-473.

Cannon, Joseph P. and William D. Perreault, Jr. (1999), “Buyer-Seller Relationships in Business Markets,” Journal of Marketing Research, Vol. 36, No. 4, pp. 439-460.

Churchill, Gilbert A. (1979), “A Paradigm for Developing Better Measures of Marketing Con-structs,” Journal of Marketing Research, Vol. 16, No. 1, pp. 64-73.

Contractor, F. J. and P. Lorange (1988), “Competition vs. Cooperation: A Benefit/Cost Frameworkfor Choosing Between Fully-Owned Investments and Cooperative Relationships,” Management Inter-national Review, Vol. 28, Special Issue, pp. 5-18.

Cox, Andrew (2001), “Managing with Power: Strategies for Improving Value Appropriation fromSupply Relationships,” Journal of Supply Chain Management, Vol. 37, No. 2, pp. 42-47.

Day, George S. (2000), “Managing Market Relationships,” Journal of the Academy of MarketingScience, Vol. 28, No. 1, pp. 24-30.

Dillman, Don A. (2000), Mail and Internet Surveys: The Tailored Design Method, 2nd Ed. New York,NY: John Wiley & Sons, Inc.

Donaldson, Bill and Tom O’Toole (2000), “Classifying Relationship Structures: RelationshipStrength in Industrial Markets,” Journal of Business and Industrial Marketing,Vol. 15, No. 7, pp. 491-506.


Doney, Patricia M. and Joseph P. Cannon (1997), “An Examination of the Nature of Trust in Buyer-Seller Relationships,” Journal of Marketing, Vol. 61, No. 2, pp. 35-51.

Doz, Y. L. and G. Hamel (1998), Alliance Advantage, Boston, MA: Harvard Business School Press.

Emerson, Richard M. (1962), “Power Dependence Relations,” American Sociological Review, Vol. 27, No. 1, pp. 32-33.

Frazier, Gary L. (1983), “One the Measurement of Interfirm Power in Channels of Distribution,”Journal of Marketing Research, Vol. 20, No. 2, pp. 158-166.

Ganesan, Shankar (1994), “Determinants of Long-Term Orientation in Buyer-Seller Relationships,”Journal of Marketing, Vol. 58, No. 2, pp. 1-19.

Garver, Michael S. and John T. Mentzer (1999), “Logistics Research Methods: Employing Struc-tural Equation Modeling to Test for Construct Validity,” Journal of Business Logistics, Vol. 20,No. 1, pp. 33-57.

Gerbing, David W. and James C. Anderson (1988), “An Updated Paradigm for Scale DevelopmentIncorporating Unidimensionality and Its Assessment,” Journal of Marketing Research, Vol. 25,No. 2, pp. 186-192.

Geyskens, Inge, Jan-Benedict E. M. Steenkamp, Lisa K. Scheer, and Nirmalya Kumar (1996),“The Effects of Trust and Interdependence on Relationship Commitment: ATrans-Atlantic Study,”International Journal of Research in Marketing, Vol. 13, No. 4, pp. 303-317.

Golicic, Susan L., James H. Foggin, and John T. Mentzer (2003), “Relationship Magnitude and ItsRole in Interorganizational Relationship Structure,” Journal of Business Logistics, Vol. 24, No. 1,pp. 57-76.

Golicic, Susan L. and John T. Mentzer (2005), “Exploring the Drivers of Interorganizational Rela-tionship Magnitude,” Journal of Business Logistics, Vol. 26, No. 2, pp. 47-71.

Granovetter, Mark (1973), “The Strength of Weak Ties,” American Journal of Sociology, Vol. 78,No. 6, pp. 1360-1380.

Gundlach, Gregory T., Ravi S. Achrol, and John T. Mentzer (1995), “The Structure of Commitmentin Exchange,” Journal of Marketing, Vol. 59, No. 1, pp. 78-92.


Gundlach, Gregory T. and Ernest R. Cadotte (1994), “Exchange Interdependence and InterfirmInteraction: Research in a Simulated Channel Setting,” Journal of Marketing Research, Vol. 31, No. 4, pp. 516-532.

Hair, Joseph C., Rolph E. Anderson, Ronald L. Tatham, and William C. Black (1998), Multivari-ate Data Analysis, 5th Ed. Upper Saddle River, NJ: Prentice Hall, pp. 577-664.

Heide, Jan B. (1994), “Interorganizational Governance in Marketing Channels,” Journal of Mar-keting, Vol. 58, No. 1, pp. 71-85.

Heide, Jan B. and George John (1988), “The Role of Dependence Balancing in Safeguarding Trans-action-Specific Assets in Conventional Channels,” Journal of Marketing, Vol. 52, No. 1, pp. 20-35.

Iyer, Karthik N. S. (2002), “Learning in Strategic Alliances: An Evolutionary Perspective,” Acad-emy of Marketing Science Review [online], No. 10.

Kumar, Nirmalya, Lisa K. Scheer, and Jan-Benedict E.M. Steenkamp (1995), “The Effects of Perceived Interdependence on Dealer Attitudes,” Journal of Marketing Research, Vol. 32, No. 3, pp. 348-356.

Lambe, C. Jay, C. Michael Wittman, and Robert E. Spekman (2001), “Social Exchange Theoryand Research on Business-to-Business Relational Exchange,” Journal of Business-to-BusinessMarketing, Vol. 8, No. 3, pp. 1-36.

Lambert, Douglas M., Margaret A. Emmelhainz, and John T. Gardner (1996), “Developing and Imple-menting Supply Chain Partnerships,” The International Journal of Logistics Management, Vol. 7, No. 2, pp. 1-17.

Lambert, Douglas M. and Thomas C. Harrington (1990), “Measuring Nonresponse Bias in CustomerService Mail Surveys,” Journal of Business Logistics, Vol. 11, No. 2, pp. 5-25.

Lewin, Jeffrey E. and Wesley J. Johnson (1997), “Relationship Marketing Theory in Practice: A CaseStudy,” Journal of Business Research, Vol. 39, No. 1, pp. 23-31.

Loehlin, John C. (1998), Latent Variable Models: An Introduction to Factor, Path, and StructuralAnalysis, 3rd Ed. Mahwah, NJ: Lawrence Erlbaum Associates.

MacCallum, Robert C., Duane T. Wegener, Bert N. Uchino, and Leandre R. Fabrigar (1993), “TheProblem of Equivalent Models in Applications of Covariance Structure Analysis,” PsychologicalBulletin, Vol. 114, No. 1, pp. 185-199.


Marsh, Herbert W. (1994), “Confirmatory Factor Analysis Models of Factorial Invariance: A Mul-tifaceted Approach,” Structural Equation Modeling, Vol. 1, No. 1, pp. 5-34.

Marsh, Herbert W. and Dennis Hocevar (1985), “Application of Confirmatory Factor Analysis tothe Study of Self-Concept: First and Higher Order Factor Models and Their Invariance AcrossGroups,” Psychological Bulletin, Vol. 97, No. 3, pp. 562-582.

McCracken, Grant (1988), The Long Interview, Beverly Hills, CA: Sage Publications, Inc.

Mentzer, John T., William DeWitt, James S. Keebler, Soonhong Min, Nancy W. Nix, Carlo D.Smith, and Zach G. Zacharia (2001a), “What is Supply Chain Management?” in John T. Mentzered., Supply Chain Management, Thousand Oaks, CA: Sage Publications, Inc., pp. 1-25.

Mentzer, John T., William DeWitt, James S. Keebler, Soonhong Min, Nancy W. Nix, Carlo D.Smith, and Zach G. Zacharia (2001b), “Defining Supply Chain Management,” Journal of BusinessLogistics, Vol. 22, No. 2, pp. 1-26.

Mentzer, John T. and Daniel J. Flint (1997), “Validity in Logistics Research,” Journal of BusinessLogistics, Vol. 18, No. 1, pp. 199-216.

Mentzer, John T., Soonhong Min, and Zach G. Zacharia (2000), “The Nature of Interfirm Partner-ing in Supply Chain Management,” Journal of Retailing, Vol. 76, No. 4, pp. 1-21.

Monczka, Robert M., Kenneth J. Peterson, Robert B. Handfield, and Gary L. Ragatz (1998), “SuccessFactors in Strategic Supplier Alliances: The Buying Company Perspective,” Decision Sciences,Vol. 29, No. 3, pp. 553-577.

Monroe, Kent B. (1990), Pricing: Making Profitable Decisions, New York, NY: McGraw-HillPublishing Company.

Moorman, Christine, Rohit Deshpande, and Gerald Zaltman (1993), “Factors Affecting Trust in Market Research Relationships,” Journal of Marketing, Vol. 57, No. 1, pp. 81-101.

Morgan, Robert M. and Shelby D. Hunt (1994), “The Commitment-Trust Theory of RelationshipMarketing,” Journal of Marketing, Vol. 58, No. 3, pp. 20-38.

Nevin, John R. (1995), “Relationship Marketing and Distribution Channels: Exploring Funda-mental Issues,” Journal of the Academy of Marketing Science, Vol. 23, No. 4, pp. 327-334.


Nooteboom, Bart (1999), Inter-Firm Alliances: Analysis and Design, New York, NY: Routeledge.

Novack, Robert A., C. John Langley, Jr., and Lloyd M. Rinehart (1995), Creating Logistics Value:Themes for the Future, Oak Brook, IL: Council of Logistics Management.

Rindfleisch, Aric and Christine Moorman (2001), “The Acquisition and Utilization of Informationin New Product Alliances: A Strength of Ties Perspective,” Journal of Marketing, Vol. 65, No. 2,pp. 1-18.

Rinehart, Lloyd M., James A. Eckert, Robert B. Handfield, Thomas J. Page, and Thomas Atkin (2004),“Structuring Supplier-Customer Relationships,” Journal of Business Logistics, Vol. 25, No. 1, pp. 25-62.

Sollner, Albrecht (1999), “Asymmetrical Commitment in Business Relationships,” Journal of Busi-ness Research, Vol. 46, No. 4, pp. 219-233.

Steenkamp, J. B. and H. van Trijp (1991), “The Use of LISREL in Validating Marketing Con-structs,” International Journal of Research in Marketing, Vol. 8, No. 4, pp. 283-299.

Stern, Louis W., Adel I. El-Ansary, and Anne T. Coughlan (1996), Marketing Channels, 5th Ed. UpperSaddle River, NJ: Prentice Hall, pp. 25-29.

Stern, Louis W. and Torger Reve (1980), “Distribution Channels as Political Economies: A Frame-work for Comparative Analysis,” Journal of Marketing, Vol. 44, No. 3, pp. 52-64.

Thibaut, John W. and Harold H. Kelley (1959), The Social Psychology of Groups, New York, NY:John Wiley.

Webster, Frederick E., Jr. (1992), “The Changing Role of Marketing in the Corporation,” Journalof Marketing, Vol. 56, No. 4, pp. 1-17.

West, Stephen G., John F. Finch, and Patrick J. Curran (1995), “Structural Equation Models withNonnormal Variables: Problems and Remedies,” in Rick H. Hoyle ed., Structural Equation Mod-eling: Concepts, Issues, and Applications, Ed. Rick H. Hoyle, Thousand Oaks, CA: Sage Publica-tions, pp. 56-75.

Wetzels, Martin, Ko de Ruyter, and Marcel van Birgelen, (1998), “Marketing Service Relationships: TheRole of Commitment,” Journal of Business and Industrial Marketing, Vol. 13, No. 4/5, pp. 406-423.


Wilson, David T. (1995), “An Integrated Model of Buyer-Seller Relationships,” Journal of theAcademy of Marketing Science, Vol. 23, No. 4, pp. 335-345.

Zeithaml, Valerie A. (1988), “Consumer Perceptions of Price, Quality, and Value: AMeans-End Modeland Synthesis of Evidence,” Journal of Marketing, Vol. 52, No. 3, pp. 2-22.

ABOUT THE AUTHORS

Susan Golicic is an Assistant Professor in the Department of Marketing at the University ofOregon. She received her Ph.D. in Logistics at the University of Tennessee, Knoxville. Her researchinterests include supply chain management, interorganizational relationships, forecasting, andlogistics. She has experience in logistics at DaimlerChrysler and more than 5 years in project man-agement and environmental engineering. She has presented at numerous academic and practitionerconferences and has published in Journal of Business Logistics, International Journal of PhysicalDistribution and Logistics Management, Supply Chain Management Review, and Journal of Forecasting.

John T. (Tom) Mentzer (Ph.D. Michigan State University) is the Harry J. and Vivienne R. BruceChair of Excellence in Business in the Department of Marketing and Logistics at the University ofTennessee. He has published more than 180 articles and papers in the Journal of Business Logistics,Journal of Marketing, Journal of Business Research, International Journal of Physical Distributionand Logistics Management, Journal of the Academy of Marketing Science, Industrial Marketing Management, and other journals. He has authored seven books on the topics of supply chain man-agement, sales forecasting and marketing. Tom served as President of the Council of LogisticsManagement during 2000 - 2001, and was the 2004 recipient of the CLM Distinguished ServiceAward.


an empirical examination of relationship magnitude

Documents