Advertiser perceptions of fair compensation, confidentiality, and rapport.
Although macro-factors - such as the business environment, mergers, organization size, and product category - can have a dramatic effect on agency-client relationships (Michell, 1988; Michell and Sanders, 1995), these factors are usually beyond the control of advertising agency executives. In contrast, an agency's service behaviors can be directed by agency executives. The current study questions whether advertising agencies' poor standing with advertisers may be attributed to agency service behaviors as a step toward developing stronger and mutually rewarding agency-client relationships. Specifically, this paper reports on how client perceptions of fair agency compensation, rapport, and agency confidentiality in ongoing relationships are influenced by two key agency service behaviors: cooperativeness and diligence. The clients' perceptions of agency behaviors are then compared to their advertising agencies' perceptions to determine how well the agencies understand their clients' view of the agencies' services. Finally, the results are discussed to support recommendations for how advertising agencies can use client service behaviors to strengthen agency-client relationships.
Prior Studies of Advertising Agency-Client Relationships
The dissolution or weakening of an advertising agency-client relationship represents a serious and costly problem for both the agency and the client (Doyle, Corstjens, and Michell, 1980; Henke, 1995; Warner, 1992). To reduce the incidence of such relationship terminations, authors have proposed improving the agency selection process (Cagley and Roberts, 1984), charting the determinants of client loyalty (Michell and Sanders, 1995), and understanding the development and maintenance of these relationships (Henke, 1995; Verbeke, 1988; Wackman, Salmon, and Salmon, 1986). To this end, researchers have asked agencies and clients numerous questions to identify the key variables that influence agency selection and client satisfaction. Several studies factor-analyzed these variables to identify a smaller set of general factors that underlie productive agency-client relationships (Cagley, 1986; Verbeke, 1988). This approach helps identify the underlying dimensions but assumes the dimensions are related to client loyalty and not to each other.
The questions associated with the general factors identified in prior research range from accounts of agency behaviors (e.g., agency reacts quickly) to client beliefs (e.g., agency compensation is fair) and client confidence (e.g., we are willing to trust our agency). Unlike prior studies that have mixed client perceptions of agency capabilities, behaviors, and client sentiments by using factor analysis for data reduction, this paper examines whether these factors can be represented by a structural model. This model proposes interrelationships between factors in addition to examining the dimensionality of the underlying factors. Thus, a framework is offered for further categorizing these variables and investigating the relationships between the general factors the measures represent [ILLUSTRATION FOR FIGURE 1 OMITTED]. Emphasis is placed on the consequences of work pattern (i.e., agency behaviors) on relationship factors (Wackman, Salmon, and Salmon, 1986).
Agency Behaviors. This study investigates a variety of agency behaviors that represent two general factors: cooperativeness and diligence. Agency cooperativeness is defined as the degree to which an agency acts as though it is dedicated to helping the client achieve its objectives (Crosby, Evans, and Cowles, 1990). Illustrative behaviors include acting as a team member, reacting quickly to customer requests, and dedicating key people to the account (Wackman, Salmon, and Salmon, 1986). Agency diligence is defined as the degree to which an agency takes the initiative and responsibility for correcting problems that arise during interactions (Gundlach and Murphy, 1993; Hotz, Ryans, and Shanklin, 1982). Problems that may plague agency-client relationships include poor strategic decisions, mismanaged personnel assignments, production errors, and legal mishaps. Emphasis is placed on how well the agency responds when a problem is recognized because the quality of an agency's response is viewed by its client as an important symbol of its agency's capability and respect for the relationship. Effective responses to minor failures can prevent such problems from becoming critical incidents that lead to the breakup of the relationship (Michell and Sanders, 1995).
Client Beliefs. Three client beliefs of keen interest to advertising agency executives are: (1) agency confidentiality, (2) fair compensation, and (3) rapport. Agency confidentiality is defined as the client's perception that its agency will keep proprietary information from being disclosed (Moorman, Zaltman, and Deshpande, 1992). Fair compensation is defined as the client's perception that its agency charges fairly for its services (Wackman, Salmon, and Salmon, 1986). This factor is receiving increasing scrutiny by clients as they search for the best value for their marketing dollars (Michell, 1988). The client's perception of interpersonal relations has also been identified as a key dimension of agency selection criteria (Cagley, 1986; Cagley and Roberts, 1984) and successful relations (Michell and Sanders, 1995). For example, the level of rapport existing between an advertiser and its agency is often associated with successful agency-client relationships (Wackman, Salmon, and Salmon, 1986). Rapport is defined here as the client's perception that the personal relationships have the right "chemistry" and are enjoyable.
Client Confidence in Its Agency. Two relationship outcomes important to advertising agencies are: (1) building clients' trust, and (2) receiving current information from clients about their operations and future plans. Client trust is defined as an advertiser's willingness to rely on its advertising agency without feeling compelled to investigate the agency's motives or performance (see Moorman, Zaltman, and Deshpande, 1992; Zaltman and Moorman, 1988); their research has shown that, as agencies earn their clients' trust, interactions become more productive and clients become more committed to their agencies. A client's disclosure of information is also critical to its agency's development and implementation of effective advertising campaigns (Michell, 1986). Even so, Helgesen (1994) notes that agencies often struggle to obtain sufficient marketing information from clients.
Several propositions are offered to express the relationship between agency behaviors, client beliefs, and client confidence. Few of the relationships between the factors have been examined in prior research; even so the proposed relationships embody the logic often expressed in descriptive accounts of these relationships. The behaviors demonstrated by an advertising agency's personnel are expected to strongly influence its client's beliefs (Cagley, 1986; Wackman, Salmon, and Salmon, 1986).
It is postulated that greater agency diligence is likely to increase rapport in the relationship. When an agency adeptly handles problems it signals the client that the agency respects the client and values its business. These actions are also evidence that the agency is providing value to the client by assuming primary responsibility for successful advertising. In contrast, an agency's failure to resolve problems often requires the client to intervene. Such actions are time-consuming, annoying, and costly for the client. Consequently, the agency's value to the client diminishes as the client is forced to take greater responsibility for resolving problems. Moreover, the resulting conflict erodes the rapport and decreases the client's confidence that sensitive information will be handled properly.
P1: The more diligent an agency is, the stronger the client's belief that (a) its agency will keep information confidential, (b) its agency is fairly compensated, and (c) strong rapport exists in the relationship.
It is also posited that actions taken by an agency to express its dedication and commitment to a client increase the value and rapport perceived by the client. For example, an agency's assignment and retention of key personnel on a client's account are clear signals that the agency respects the client and desires to enhance the relationship (Henke, 1995; Michell, 1986), as are agency behaviors that are attuned to the client's needs and convey a sense of effort by the agency. Such demonstrations of a seller's effort and quality service inputs have been shown to increase a buyer's perception of value and fair compensation (Oliver and Swan, 1989). The steps an agency takes to enhance the interpersonal component of the business relationships should also build rapport. For example, agency efforts to become part of the client's team and to retain people on the account enable friendships to develop over time. Similarly, agency attempts to genuinely assist the client are often reciprocated by a client's greater willingness to permit a closeness to develop between the parties involved.
P2: The more cooperative an agency is, the stronger the client's belief that (a) its agency is fairly compensated, and (b) strong rapport exists in the relationship.
These client beliefs are likely to increase a client's willingness to trust its advertising agency and to disclose information (Michell, 1986; Verbeke, 1988; Zaltman and Moorman, 1988). Research on market research provider-user relationships has indicated that clients are more willing to disclose information when the research provider can be trusted and will keep the information confidential (Moorman, Zaltman, and Deshpande, 1992; Moorman, Deshpande, and Zaltman, 1993). Hence, it is proposed that clients behave according to their beliefs about their advertising agency.
P3: A client is more willing to disclose information when (a) the client believes its agency keeps client information confidential, and (b) the client is confident it can trust its agency.
The client's perception that its advertising agency is fairly compensated and providing a good value also supports the development of client trust and lasting relationships (Michell and Sanders, 1995; Oliver and Swan, 1989). Client feelings of friendship, mutual attraction, and congeniality are also expected to encourage the development of client trust (Moorman, Deshpande, and Zaltman, 1993).
P4: A client is more trusting when the client believes that (a) strong rapport exists in the relationship, and (b) its agency is fairly compensated.
Several steps were taken to develop suitable measures for the general factors examined in this study. First, in-depth interviews were conducted of advertising agency and client executives to establish the importance and nature of the factors considered in this study. Second, questions were drawn from published studies of advertising agencies or clients (Verbeke, 1988; Wackman, Salmon, and Salmon, 1986; Weilbacher, 1981). Third, a draft of the questionnaire was developed and then scrutinized by eight academic researchers and fourteen client executives. The questions arrived at for use in the final questionnaire are listed in Table 1 and the Appendix.
The focus of this study was on the relationship between manufacturers of consumer-packaged goods and heavily promoted industrial goods and the advertising agency who performed the largest share of their work. Two sources were used to provide a sample frame. First, a list of 865 names of senior marketing executives was obtained from Cahners Publishing in support of this study. Second, 300 names of senior marketing executives were randomly drawn from the 1994 AMA membership directory to broaden the sample frame. To increase the chances that each questionnaire reached a person intimate with an agency-client relationship and responsible for affecting changes in the relationship (e.g., switching agencies), each company on the list was contacted by telephone to ascertain the advertiser executive in charge of the agency-client relationships and responsible for affecting changes in the relationships. After up to three call-backs, 238 companies were deleted because they did not use an outside advertising agency or could not be reached. One hundred and five companies were duplicates or had incomplete contact information. In all, 814 key informants in client companies were verified.
A six-page questionnaire was then mailed with an offer of receiving an executive summary for participating. Three weeks later a reminder letter and a replacement questionnaire were mailed. This mailing was followed by telephone confirmation of the questionnaire's receipt. Two hundred and sixteen clients responded (26.5 percent). Given the emphasis on testing a conceptual framework rather than reporting descriptive statistics (Hunt, 1990), the response rate seems satisfactory. Of the 216 client respondents, 68 clients identified themselves. The other client respondents chose anonymity.
An agency questionnaire was then prepared and sent to these 68 clients. Each client participant was asked to forward the agency questionnaire to the top agency executive managing its account. Thirty-one agencies responded (45.6 percent). The organizational reference to "focus on the advertising agency that does the most work for your business unit" for advertisers changed to "the account team which services the advertiser that forwarded this questionnaire to [TABULAR DATA FOR TABLE 1 OMITTED] you" for advertising agencies. Instructions at the start of each section of the questionnaire were also adjusted. For example, the instruction to "please think about your agency's performance" for clients was changed to "please think about your agency's performance for this client" for agencies. The items in the agency questionnaire mirrored those in the client questionnaire. Generally, changes were limited to the replacement of "its" with "our," "them" with "us," or "their" with "our." For example, the statement "Our agency works hard to fix its mistakes, as soon as the problems are recognized" was changed to "Our agency works hard to fix our mistakes, as soon as the problems are recognized."
The data-collection effort succeeded in reaching a knowledge set of respondents in a broad array of companies. Respondents had many years of experience with advertising agencies in general (mean = 13.6 years; 84 percent [greater than] 4 years) and several years of experience with its current agency (mean = 4.52 years; 80 percent [greater than] 2 years) Respondents generally held senior management positions (vice presidents/directors, 61 percent; senior advertising/communication managers, 23 percent; marketing managers, 16 percent). Advertisers represented a range of company sizes ([less than]$50 million, 13.3 percent; $50-100 million, 17.1 percent; $101-300 million, 30.9 percent; $301-500 million, 11 percent; over $500 million, 27.6 percent), advertising and promotion budgets ([less than!$5 million, 67 percent; $5-15 million, 15.9 percent; $16-50 million, 8.8 percent; over $51 million, 8.2 percent) and served markets (consumer, 35.3 percent; business-to-business, 38.7 percent; both 26 percent). Their agencies were reasonably large in terms of total billings ([less than!$50 million, 55.6 percent; $50-100 million, 18.7 percent; $101-500 million, 12.9 percent; over $500 million, 12.9 percent).
To test for nonresponse bias, the responses were divided into early (73.6 percent) and late (26.4 percent) respondent groups based on the pattern of responses (Armstrong and Overton, 1977). No differences were found between the two groups in terms of agency of record (p [less than] .66), client sales (p [less than] .91), advertising and promotional budget (p [less than] .59), or served market (p [less than] .67). T-tests revealed no differences between early and late respondent groups for relationship age (p [less than] .97), informant experience with agencies in general (p [less than] .58), or its current agency (p [less than] .73), and for all of the focal constructs. These analyses suggest that nonresponse bias does not represent a problem.
Three separate analyses were conducted to achieve the objectives set forth earlier. First, the questions pertaining to agency cooperativeness and agency diligence were factor-analyzed to reveal the underlying factor pattern. The resulting factor pattern supports the contention that the set of measures represent two distinct factors (see Table 1). Second, the client data was tested using LISREL, a path analytic method for assessing the significance of the proposed relationships (see the Technical Appendix for details). Third, t-tests were conducted to examine how closely agencies' ratings of cooperativeness and diligence matched their clients' ratings.
Strong empirical support was found for the framework presented in this paper. Thus, the proposed model can be used confidently as a valid guide for understanding and diagnosing agency-client relationships. For example, agency executives who sense that a client is reluctant to disclose information should question whether they have done enough to earn the client's trust. Under such circumstances, the framework shows that agencies can affect improved client trust and beliefs by adjusting their agency's cooperativeness and diligence. These findings are based on a path analysis of data collected from 184 top advertiser executives which established that the proposed relationships are significant and explain much of the variation in the key factors [ILLUSTRATION FOR FIGURE 2 OMITTED]. Specifically, the model explains large portions of each key factor (confidentiality, 43 percent; fair compensation, 64 percent; rapport, 74 percent; trust, 29 percent; and information disclosure, 26 percent).
What if Client Disclosure and Client Trust Are Low? Given that client trust and information disclosure are key to effective long-term agency-client relationships (Michell, 1984; Zaltman and Moorman, 1988), agencies would be wise to ensure that the client's confidence is maintained. This study's results suggest that agencies who find their clients reluctant to share information may need to work to restore client perceptions of agency confidentiality ([Beta] = .28) and to rebuild client trust ([Beta] = .36). These two factors explain 26 percent of the variation in client disclosure of information. The results show that client trust can be strengthened by acting to strengthen client perceptions of fair agency compensation ([Beta] = .17) and rapport ([Beta] = .41) by explaining 29 percent of the variation in client trust. This study documents, for the first time, the consequences of rapport or "chemistry" felt by the client. These findings also suggest that agencies might productively explore their clients' beliefs about their performance. Such a performance audit will likely identify other areas upon which the agency could improve.
What if Client Perceptions of Rapport and Fair Agency Compensation Are Low? Since rapport or "chemistry" has'been frequently mentioned as essential in agency-client relationships (Michell, 1986), this study examines how agency behaviors influence such client beliefs. The results show that, if an agency senses rapport is waning, efforts to be more cooperative ([Beta] = .50) and diligent ([Beta] = .42) will likely reverse this trend given the positive influence of these factors on rapport observed in this study. Likewise, the results also suggest that during times when agency confidentiality is being questioned, agencies should make extra efforts to assume responsibility for any problems or mistakes that surface ([Beta] = .65). This factor explains almost half of the variation in agency confidentiality (43 percent). Similarly, if perceptions of fair compensation becomes a crucial issue, the findings suggest that improved [TABULAR DATA FOR TABLE 2 OMITTED] cooperativeness ([Beta] = .58) or better diligence ([Beta] = .27) can help restore the value its client believes its agency is providing. These two factors explain a sizable 64 percent of the variation in client perceptions of fair agency compensation. Agencies might also examine their performance on other service behaviors associated with cooperativeness and diligence to identify other areas where agency value can be enhanced and conveyed.
What if You Think Your Agency Is Providing Superb Service? There is a real possibility that you might be wrong and ignoring emerging problems in the relationship. The results show that agencies rate their cooperativeness and diligence significantly higher than their clients' evaluation (see Table 2). Several researchers have noted that advertising agencies have trouble understanding their client's perspective due to different backgrounds or objectives (Cagley, 1986; Doyle, Corstjens, and Michell, 1980; Zaltman and Moorman, 1988). A comparison of 30 matched agency-client pairs reveals a strong tendency for agencies to overestimate the level of service they are providing. The results of the comparison of agency and client perceptions of agency behaviors are presented in Table 2. Since these behaviors are shown to influence key client beliefs, such optimism seems dysfunctional. As a remedy, advertising agencies are advised to pay particular attention to the behaviors with the largest discrepancy between the agency and client views when seeking to build client perceptions of fair compensation, confidentiality, and rapport.
For 12 of the 15 agency behaviors examined, clients rated their agency significantly lower than their own advertising agency did. This finding should be somewhat alarming to senior advertising agency executives. The reason is that the results show that agency executives are overly optimistic when evaluating their cooperativeness and diligence, despite the fact that agencies tend to rate these behaviors more important than advertisers (Verbeke, 1988). Thus, agency executives may be assuming that knowing what behaviors are important translates into behaviors exhibited by their account team. The findings clearly challenge this assumption. This shortcoming is troubling because agency cooperativeness and diligence explain much of a client's beliefs about agency confidentiality, fair compensation, and rapport. Agencies seeking to strengthen these dimensions of their client relationships should therefore ask their clients for objective feedback rather than relying on their own perceptions. Failure to address the "gap" identified in this study seems likely to ultimately increase a relationship's vulnerability to dissolution. In fact, the poor comprehension of their client's perspective may be partially responsible for the persistence of known causes of client disaffection (Michell, Cataquet, and Hague, 1992).
This paper shows dramatically how agency service behaviors influence ongoing agency-client relationships. Thus, it can be argued that agencies should pay greater attention to servicing their clients relative to the efforts made to win creative awards, develop new services, and attract new clients. To this end, advertising agencies could use these results to develop improved client monitoring systems (Michell, Cataquet, and Hague, 1992). Such systems should involve, at a minimum, three steps. First, an advertising agency should assess their clients' perceptions of the current level of agency service performance. Since account executives close to the account may be conditioned against any self-criticism, it is often beneficial to de-personalize the problem by using survey instruments or conducting the assessment at a higher organizational level. The measures listed in Table 1 and the Appendix provide a good starting point. Second, the results from the assessment should be used to identify the largest and most important service gaps. Results similar to Table 2 could be produced. Third, advertising agencies should initiate programs to remedy these gaps by providing service training, rewarding the provision of superior service, and assessing progress toward stronger client relationships. Michell and Sanders (1995) report that such processes to identify and respond to early signals of failure are characteristics of successful long-term agency-client partnerships.
To summarize, the first purpose of this study was to present and test a framework for understanding the influence of agency cooperativeness and diligence on client beliefs. Since the model was supported empirically, the findings provide insights into what agencies can do to strengthen their client relationships. A logical extension of this research would be to examine additional agency behaviors, such as timeliness and creative problem-solving, that may also influence clients' beliefs about confidentiality, fair compensation, and rapport. There are other outcomes that might be included in further efforts to understand agency-client relationships, such as client intentions to continue the relationship or to expand the scope of its advertising agency's duties. Future researchers may also develop more specific measures of agency behaviors in order to help agencies identify more precise areas for improvement.
The second purpose of this study was to examine the gap between agency and client perceptions of agency behaviors pertaining to cooperativeness and diligence. For most of the agency behaviors examined, clients provided a significantly lower rating than their agency partner. Evidence of this perceptual "gap" reinforces the view that advertising agencies have trouble understanding their clients' perspective. Since such gaps may lead to termination of the relationship (Doyle, Corstjens, and Michell, 1980), future research should identify which "gaps" have the severest consequences and, thus, are most needy of immediate attention. Then, practical and cost-efficient means for reducing these discrepancies could be identified, tested, and implemented.
Advertising Age. "Role of Performance Gains in Agency Pay." November 27, 1995a.
Advertising Age. "Bayer's Global $225 Account Moves to BBDO." December 11, 1995b.
Armstrong, J. Scott, and Terry S. Overton. "Estimating Nonresponse Bias in Mail Surveys." Journal of Marketing Research 14, 3 (1977): 396-402.
Artzt, Edwin L. "Advertising's Grave New World." Speech presented at the annual meeting of the American Association of Advertising Agencies, May 1994.
Bentler, Peter M. "Comparative Fit Indexes in Structural Models." Psychological Bulletin 107, 2 (1990): 238-46.
Business Week. "Now Mad Ave Really Has To Sing For Its Supper." December 18, 1995.
Cagley, James W. "A Comparison of Advertising Agency Selection Factors: Advertiser and Agency Perceptions." Journal of Advertising Research 26, 3 (1986): 39-44.
-----, and C. Richard Roberts. "Criteria for Advertising Agency Selection: An Objective Appraisal." Journal of Advertising Research 24, 2 (1984): 27-31.
Crosby, Lawrence A., Kenneth R. Evans, and Deborah Cowles. "Relationship Quality in Services Selling: An Interpersonal Influence Perspective." Journal of Marketing 54, 3 (1990): 68-81.
Doyle, Peter, Marcel Corstjens, and Paul Michell. "Signals of Vulnerability in Agency-Client Relations." Journal of Marketing 44, 4 (1980): 18-23.
Gerbing, David W., and James C. Anderson. "An Updated Paradigm for Scale Development Incorporating Unidimensionality and Its Assessment." Journal of Marketing Research 25, 2 (1988): 186-92.
Gundlach, Gregory T., and Patrick E. Murphy. "Ethical and Legal Foundations of Relational Marketing Exchanges." Journal of Marketing 57, 4 (1993): 35-46.
Helgesen, Thorolf. "Advertising Awards and Advertising Agency Performance Criteria." Journal of Advertising Research 34, 4 (1994): 43-53.
Henke, Lucy L. "A Longitudinal Analysis of the Ad Agency-Client Relationship: Predictors of An Agency Switch." Journal of Advertising Research 35, 2 (1995): 24-30.
Hotz, Mary R., John K. Ryans, Jr., and William L. Shanklin. "Agency/Client Relationships as Seen by Influentials on Both Sides." Journal of Advertising 11, 1 (1982): 37-44.
Hunt, Shelby D. "A Commentary on an Empirical Investigation of a General Theory of Marketing Ethics." Journal of the Academy of Marketing Science 18, 2 (1990): 173-77.
Joreskog, Karl G., and Dag Sorbom. LISREL 8: User's Guide. Chicago, IL: Scientific Software, Inc., 1993.
Michell, Paul C. "Accord and Discord in Agency-Client Perceptions of Creativity." Journal of Advertising Research 24, 5 (1984): 9-23.
-----. "Auditing of Agency-Client Relations." Journal of Advertising Research 26, 6 (1986): 29-41.
-----. "The Influence of Organizational Compatibility on Account Switching." Journal of Advertising Research 28, 3 (1988): 33-38.
-----, Harold Cataquet, and Stephen Hague. "Establishing the Causes of Disaffection in Agency-Client Relations." Journal of Advertising Research 32, 2 (1992): 41-48.
-----, and Nicholas H. Sanders. "Loyalty in Agency-Client Relations: The Impact of the Organizational Context." Journal of Advertising Research 35, 2 (1995): 9-22.
Moorman, Christine, Rohit Deshpande, and Gerald Zaltman. "Factors Affecting Trust in Market Research Relationships." Journal of Marketing 57, 1 (1993): 81-101.
-----, Gerald Zaltman, and Rohit Deshpande. "Relationships Between Providers and Users of Market Research: The Dynamics of Trust Within and Between Organizations." Journal of Marketing Research 29, 3 (1992): 314-328.
Oliver, Richard L., and John E. Swan. "Consumer Perceptions of Interpersonal Equity and Satisfaction in Transactions: A Field Survey Approach." Journal of Marketing 53, 2 (1989): 21-35.
Verbeke, Willem. "Developing an Advertising Agency-Client Relationship in The Netherlands." Journal of Advertising Research 28, 6 (1988): 19-27.
Wackman, Daniel B., Charles T. Salmon, and Caryn C. Salmon. "Developing an Advertising Agency-Client Relationship." Journal of Advertising Research 26, 6 (1986):21-28.
Warner, Fara. "Should You Nuke Your Ad Agency?" Brand-week, October 12, 1992.
Weilbacher, William M. Auditing Productivity: Advertiser-Agency Relationships Can Be Improved. New York, NY: Association of National Advertisers, 1981.
Zaltman, Gerald, and Christine Moorman. "The Importance of Personal Trust in the Use of Research." Journal of Advertising Research 28, 5 (1988): 16-24.
[TABULAR DATA FOR APPENDIX OMITTED]
Procedure Used to Estimate the Path Coefficients
Step 1. Measurement Assessment. Following the procedure outlined by Gerbing and Anderson (1988), the set of measures for each scale was examined by looking at the item-to-total correlations and conducting exploratory factor analysis to assess dimensionality. The results of the exploratory factor analysis suggested that all but one item could be retained (see Appendix and Table 1). As a more rigorous test, the scales were submitted to confirmatory factor analysis using item covariance-variance matrices. The fit indices suggested by Bentler (1990) and Joreskog and Sorbom (1993) were used to assess model adequacy. The estimates generated by LISREL 8 provided evidence of adequate model fit, thereby supporting convergent validity and unidimensionality ([[Chi].sup.2[.sub.(413)]] = 821.50, CFI = .91, TLI = .90, NFI = .84, SRMR = .07). The resulting scales all demonstrated sufficient reliability (all exceeded .82 except fair compensation, .75 and client disclosure, .66).
Step 2. Hypothesis Tests. The next step was to test a model with only the proposed relationships between constructs permitted to be non-zero. This model was estimated by using the maximum likelihood fitting function of LISREL 8 to examine the fit of the proposed model to the data. The data used was the variance-covariance matrix of the individual measures. Listwise deletion reduced the 216 observations to 184. LISREL 8 produced estimates of the path coefficients and their standard errors [ILLUSTRATION FOR FIGURE 1 OMITTED!, estimates of the variance explained in each dependent construct, and estimates of the model fit. According to the criteria set forth by Bentler (1990), an acceptable model fit was obtained ([[Chi].sup.[s.sub.(424)]] = 841.90, CFI = .91, TLI = .90, NFI = .83, SRMR = .07).
Note. Structural equation modeling (e.g., LISREL) is preferable to regression analysis or two-stage least squares for several reasons. One benefit of LISREL is the ability to partition the variance of each measure into that arising from error and that arising from the underlying theoretical factor. Each item is restricted to load on only one factor, thereby preventing the overlap between two measures from contributing to the correlation between the underlying constructs. LISREL is also ideally suited for testing direct and indirect relationships between constructs.
Douglas Labahn gratefully acknowledges the research support provided by Cahners Publishing and the Institute for the Study of Business Markets at Pennsylvania State University. The author also thanks William Cook and Fred Zandpour (California State University Fullerton) for their helpful comments on drafts of this paper.
DOUGLAS W. LABAHN is associate professor in the department of marketing at the School of Business Administration and Economics, California State University, Fullerton. He has a Ph.D. in marketing from the University of Maryland. This paper describes one of a series of studies he has conducted on effectiveness in strategic business-to-business relationships. He has published papers in the Journal of Business Research, the Journal of Product Innovation Management, and Industrial Marketing Management. He was an account manager before entering academics and now consults on account development and retention.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||includes appendices|
|Author:||LaBahn, Douglas W.|
|Publication:||Journal of Advertising Research|
|Date:||Mar 1, 1996|
|Previous Article:||Using judgment profiles to compare advertising agencies' and clients' campaign values.|
|Next Article:||Ad executives grade new grads: the final exam that counts.|