Article Reviews

ATTITUDINAL BALANCEBASIL AND HERR

JOURNAL OF CONSUMER PSYCHOLOGY, 16(4), 391–403 Copyright © 2006, Lawrence Erlbaum Associates, Inc.

Attitudinal Balance and Cause-Related Marketing: An Empirical Application of Balance Theory

Debra Z. Basil University of Lethbridge Centre for Socially Responsible Marketing

Paul M. Herr University of Colorado at Boulder

We examine the effects of pre-existing organizational attitudes on consumer response to cause- related marketing (CRM) alliances, using a Balance Theory framework. Two experiments dem- onstrate that balanced attitudes (either both positive or both negative) resulted in perceptions of appropriateness, but did not necessarily lead to positive affect. The positive balance scenario led to a synergistic attitudinal boost when both pre-existing attitudes were positive. Attitudinal contamination was evident when either pre-existing attitude was negative. Fit operated within the balance scenario to enhance perceptions of the strength of the CRM alliance, leading to more positive responses.

Cause-related marketing (CRM) involves the pairing of a firm and a charity in a marketing effort. Alliances are often formed by well-known firms pairing with well-known orga- nizations, as in American Express’ CRM alliance with the Ronald McDonald House. American Express donates to the Ronald McDonald House for every transaction made on an American Express card. Presumably, both American Express and the Ronald McDonald House hope to benefit from ally- ing with an organization for which consumers hold positive pre-existing attitudes.

CRM alliances continue to grow in popularity (Cone/ Roper, 1999; PMA/Gable Group, 2000). The body of re- search addressing CRM issues is also growing, but many questions remain. Attitudes toward the CRM alliance and the alliance partners have been examined, but the exact nature of these attitudes has yet to be assessed. Does affect toward CRM alliances depend on the perceived appropriateness of the alliance, or are these issues independent? Pre-existing or- ganizational attitudes impact attitude toward the CRM alli- ance, but little if any research has examined the dynamic na- ture of this impact. For instance, how do these attitudes jointly influence CRM attitude? Fit between the allying orga- nizations influences attitude toward the alliance, but, again, the mechanism remains unknown. Our goal is to clarify the

Correspondence should be addressed to Debra Z. Basil, Associate Pro- fessor, University of Lethbridge Centre for Socially Responsible Marketing, Lethbridge, Alberta CANADA, T1K 3M4. E-mail: [email protected]

role of pre-existing organizational attitudes and fit in determining consumer response to CRM alliances, using a Balance Theory framework. Balance Theory addresses situa- tions where an individual evaluates the pairing of two sepa- rate elements—precisely the situation with CRM.

LITERATURE REVIEW AND THEORETIC FRAMEWORK

CRM enhances product choice (Barone, Miyazaki, & Taylor, 2000; Lichtenstein, Drumwright, & Braig, 2004; Yechiam, Barron, Erev, & Erez, 2003,). However, socially oriented messages are perceived differently, depending upon the sponsor (Szykman, Bloom, & Blazing, 2004), and percep- tions of a firm’s motive for forming the CRM partnership can impact resulting attitudes (Barone et al., 2000). Tying nega- tive information to the firm moderates response to CRM (Dean, 2003/2004; Deshpande & Hitchon, 2002), as do con- sumers’ elaboration levels (Menon & Kahn, 2003). More- over, CRM may negatively influence the charity (Basil & Herr, 2003). Hence, CRM alliances should be considered carefully, as alliances may not only influence immediate pur- chase decisions, but attitudes toward the partners as well.

Previous research has examined the impact of pre-existing firm and charity attitudes on attitude toward the CRM alli- ance, as well as attitude change toward the alliance partners (Lafferty, Goldsmith, & Hult, 2004). Pre-existing firm and

392 BASIL AND HERR

charity attitudes, and attitude toward the CRM alliance, are positively correlated. A similar relationship exists for atti- tude change toward the alliance partners. Although these re- sults offer insight into the impact of pre-existing attitudes, they do not address attitude dynamics. How do pre-existing organizational attitudes interact to impact CRM alliance atti- tude? We address this question.

Fit between the organizations also influences response to CRM alliances. “Fit” has been addressed in the branding lit- erature. Good fit between a brand extension and the firm’s current brand offerings (Aaker & Keller, 1990), and similar- ity, typicality, or relatedness between the extension and the core brand (Bottomley & Holden, 2001; Boush & Loken, 1991; Broniarczyk & Alba, 1994; Dacin & Smith, 1994; Herr, Farquhar, & Fazio, 1996) foster more favorable con- sumer attitudes toward a brand extension. Similarly, when two organizations’ brands and products are viewed as “fit- ting” together, consumer attitudes toward a cobranding effort are more favorable (Simonin & Ruth, 1998). Likewise, fit be- tween an event and its sponsor influences consumer response (Speed & Thompson, 2000).

Fit is important in cause-related marketing alliances, as well (Basil & Herr, 2003; Hamlin & Wilson, 2004; Lafferty et al., 2004; Menon & Kahn, 2003; Sen & Bhattacharya, 2001). Different types of fit have been proposed (e.g. Lafferty et al., 2004). Although fit matters, the means by which fit impacts CRM attitudes has received scant attention. Our goal is to clarify the process by which fit impacts attitude toward the CRM alliance.

CRM attitude measures vary dramatically in the literature. Some researchers report attitude toward the CRM alliance as a summary evaluation (Basil & Herr, 2002). Others report a blend of affective and cognitive measures, such as whether the alliance is good, positive, and favorable (Lafferty et al., 2004). Still others take a behavioral approach to measuring CRM alliance attitude (Barone et al., 2000; Strahilevitz & Myers, 1998). Although behavioral responses to CRM alli- ances have been extensively examined, no research has ad- dressed the distinction between cognitive and affective re- sponses to CRM alliances. Do these attitudinal responses differ? We also address this issue.

Three primary factors are common to CRM alliances. These include consumers’ pre- existing attitudes toward the firm, pre-existing attitudes toward the charity, and percep- tions of an alliance between these two. In some cases, con- sumers may be unfamiliar with one or the other organization (see Lafferty & Goldsmith, 2005), but such alliances are not the focus of this research. Rather, we examine CRM alliances involving organizations familiar to the individual, for which attitudes exist.

Different theoretical approaches have been used to exam- ine responses to CRM alliances, such as information integra- tion (Lafferty, Goldsmith, & Hult, 2004), cognitive elabora- tion (Menon & Kahn, 2003), and identification (Lichtenstein et al., 2004). Each approach has strengths and weaknesses.

We sought a parsimonious framework that effectively ad- dresses the three issues of interest: (1) the interactive effects of firm and charity pre-existing attitudes, (2) the mechanism by which fit impacts responses, and (3) differences between affective and cognitive attitudinal responses. None of the the- ories above addresses all of these issues. Balance Theory par- simoniously addresses each issue and guides hypothesis for- mation. Hence, we rely exclusively on Balance Theory to generate and test our specific research hypotheses.

Balance Theory

Balance Theory (Heider, 1946, 1958) examines relational tri- ads. Relationships between three individuals may be exam- ined, from the perspective of one of the individuals. For ex- ample, the relationship between Bob, Brad, and Bill’s attitude toward Bob and Brad’s relationship, as well as Bill’s attitude toward Bob and Brad individually, may be examined. Heider (1946, 1958) proposed that individuals seek balance among their interpersonal relationships and among attitudes toward these relationships. Balance may be ascertained by multiplying the signs in a triad of relationships (Cartwright & Harary, 1956). A positive result indicates balance (see Figure 1a).

Balance triads may contain relations between entities other than people. Relationships between people are referred to as sentiments, whereas relationships between entities are referred to as unit relationships (Heider, 1958). In a CRM scenario, the relationship between a firm and a charity is thus a unit relationship.

FIGURE 1 Balance Theory Triads.

ATTITUDINAL BALANCE 393

Jordan (1953) proposed that balance leads to a judgment of propriety or conforming to expectations, but for a situation to be pleasant both balance and a positive interpersonal rela- tionship are required. Hence, although two negative attitudes represent balance, the situation is not deemed pleasant. Judg- ments of unpleasantness require only the perception of im- balance or the perception of a negative relationship. Cacioppo and Petty’s (1981) findings that balance contrib- utes to a sense of propriety, whereas agreement and attraction between the individuals contribute to positive affect, further support this position.

These findings provide a foundation for hypothesizing consumer responses to CRM alliances. Cognitive and affec- tive responses to CRM alliances are expected to differ ac- cordingly. Previous research suggests that affect impacts judgments, but this area has not been thoroughly researched (Olsen & Pracejus, 2004; Pham, 2004). Consumers are ex- pected to judge as “appropriate” alliances for which their pre-existing attitudes yield balance. Judgments of propriety represent the cognitive element in a CRM alliance attitude. Either a positive–positive or a negative–negative alliance should be judged “appropriate.” Balance does not suggest positive affect toward the alliance itself. The negative–nega- tive balance situation may be deemed appropriate, but not be well liked, per Cacioppo and Petty (1981). This represents the affective element in a CRM alliance attitude. Thus,

H1: The positive impact of balance will be greater for judgments of propriety than for judgments of affect, as evidenced in a balance × judgment type interaction.

An individual’s attitude toward a CRM alliance should consist of some combination of his or her attitude toward the firm, the charity, and the pairing of these two, per Balance Theory. Hence, positive pre-existing attitudes toward the firm and the charity should contribute to a positive alliance attitude. Likewise, a positive view of the pair together should contribute to a positive alliance attitude (discussed later). However, Jordan’s (1953) findings regarding pleasantness perceptions demonstrated a benefit for the combination of balance plus a positive relationship. If, indeed, positive pre-existing attitudes toward the alliance partners are impor- tant for generating a positive attitude toward the alliance it- self, then the combination of balance plus positive pre-exist- ing attitudes should lead to a more positive response to the alliance, above and beyond the simple additive effects of each individual pre-existing attitude. This interactive affect of organizational attitudes has not been previously examined. Thus,

H2: Consumers’ pre-existing attitudes toward the firm and the charity will interact such that attitudes toward the alliance will be multiplicatively more positive when both pre-existing organizational attitudes are balanced and both are positive.

Fit

We propose that fit may be viewed as a measure of the strength of the relational tie between the two organizations. In a CRM alliance, the individual’s attitude toward the firm and the charity are two legs of a Balance triad. The presence (absence) of an organizational relationship may be viewed as the third leg of the triad. In Balance Theory’s original con- ceptualization, all relations were represented dichotomously. Extensions (e.g. Osgood & Tannenbaum, 1955) viewed the relations as continuous. This perspective allows for an exam- ination of the strength of the relationships in the triad. Fit in a CRM alliance can thus be viewed as strengthening the unit relationship between the firm and the charity, defining the na- ture of their association. A positive firm attitude, a positive charity attitude, and fit between the firm and charity reflect a strong, positive balance scenario (see Figure 1b).

We expect the relationship between organizations that fit to be judged stronger than the relationship between organiza- tions that do not fit. Moreover, if fit at least partly increases perceptions of relationship strength, then the impact of fit on attitude toward the CRM alliance should be at least partially mediated by perceptions of relationship strength. Hence,

H3a: The relationship between organizations that fit will be viewed as stronger than the relationship be- tween organizations that do not fit.

H3b: The effect of fit on CRM attitude will be partially mediated by perceptions of relationship strength.

If fit enhances perceptions of the strength of the relation- ship between a firm and a charity, then fit should similarly in- fluence judgments. Specifically, fit between the firm and charity should be viewed as more appropriate, regardless of attitudes toward the organizations. This is because fit should be seen as appropriate, based on individuals’ preference for balance. The same is not expected for affect, however. Two organizations may fit well, but an individual may not neces- sarily like the pairing, simply because they fit. Thus,

H4: Fit will more strongly influence judgments of ap- propriateness or propriety than judgments of positive affect, as reflected in a fit × judgment type interaction.

Balance is expected to influence target organizational atti- tude change in a manner similar to its anticipated effect on al- liance attitude. Consumers are expected to exhibit more posi- tive attitude change toward a target alliance member when they hold positive pre-existing attitudes toward the alliance partner, and are expected to prefer balanced attitudes. Hence, balanced attitudes with a liked pre-existing partner should enjoy additional positive response, beyond the benefit of simply owning positive pre-existing attitudes. The expecta- tion here is slightly different than that proposed in Hypothe-

394 BASIL AND HERR

sis 2: Only the partner’s pre-existing attitude valence must be positive, rather than requiring positive pre-existing attitudes toward both organizations. Practically speaking, however, this distinction is irrelevant, as pre-existing attitudes toward the target alliance member must in fact be positive for both balanced attitudes and a positive pre- existing attitude toward the alliance partner to exist. If pre-existing attitudes toward the partner are positive, then, by definition, pre-existing atti- tudes toward the target must be positive to attain balance. Hence, we predict a synergy for balanced attitudes and a pos- itive pre-existing partner attitude, not driven by pre-existing target organization attitudes, as follows:

H5: Consumers’ pre-existing attitudes toward the part- ner organization will interact with attitudinal balance. When pre-existing partner attitudes are positive and balanced, attitude change toward the target organiza- tion will be multiplicatively more positive.

EXPERIMENT 1

Experiment 1 tests these hypotheses. The experiment was ad- ministered via computer. The firm represents the target orga- nization and the charity represents the partner organization.

Pretests

All pretest and experiment participants were from the same major Western university. Pretests were conducted to select appropriate organizational profiles for fictitious firms and charities. The profiles were intended to create either positive or negative attitudes toward the fictitious organizations. Pre- test 1 involved 36 undergraduate business students, partici- pating for partial course credit. Participants were asked to list information about firms and charities that would lead them to hold either a positive or negative attitude toward the organi- zation. In Pretest 2, the most commonly cited information from Pretest 1 was further tested. Forty undergraduate busi- ness students participated in this Pretest for partial course credit. Pretest 3 combined these statements into firm and charity profiles to create reliably positive or negative atti- tudes toward the fictitious organization. Twenty-eight under- graduate business students participated for extra course credit. Pretest 4 determined product and charity categories that, when paired in a cause-related alliance, “fit” together, and pairs that did not “fit” together. “Fit” was defined for subjects in terms of whether the organizations’ purposes were complementary, and whether the organizations’ alli- ance “made sense.” Sixty-five undergraduates participated to partially fulfill a course requirement. The resulting firm and charity profiles created positive or negative attitudes, as well as CRM alliances that did or did not fit.

Participants and Design

One hundred sixty-eight undergraduate business students participated for extra course credit. A 2 (organizational fit) × 2 (firm attitude: positive or negative) × 2 (charity attitude: positive or negative) × 2 (judgment type: affect or propriety) mixed design was used. Fit and judgment type were within- subjects factors.

Independent Variables

Firm and charity attitudes were manipulated via the ficti- tious organizational profiles. Each organizational profile contained five statements about the organization, based on pretest results. These profiles were used to generate posi- tive or negative attitudes toward the firm and charity. (See Appendix A for sample profiles.) Profiles were randomly generated for each subject, from a pool of 15 possible state- ments, to assure that results were not due to excessive im- pact from any single statement. Both the statements dis- played and their order were randomized. Participants were randomly assigned to one of four between-subjects attitude conditions: positive firm attitude/positive charity attitude; negative firm attitude/negative charity attitude; positive firm attitude/negative charity attitude; or negative firm atti- tude/positive charity attitude.

Fit was manipulated through the pairing of firms and char- ities. Two fictitious firms and two fictitious charities were se- lected from Pretests. “Bakerman’s Bread” was paired with “Stop Starvation” in the fit condition, and with “Prevent Children’s Polio” in the no-fit condition. “Tikes Toys” was paired with “Prevent Children’s Polio” in the fit condition and “Stop Starvation” in the no-fit condition. Each partici- pant evaluated all four pairings.

Each subject was exposed to both propriety-based and af- fect-based adjectives. This exposure was a within-subjects factor for analysis purposes. Participants’ actual responses to these adjectives served as a dependent variable.

Dependent Variables

Attitude toward the CRM alliance was assessed by asking participants to agree or disagree with adjectives describing the alliance. These adjectives were: like, dislike, appealing, unappealing, good, bad, appropriate, and inappropriate. Atti- tude toward the CRM alliance was also assessed by asking participants to indicate their attitude toward the CRM on a 7-point scale, anchored by “very negative” at –3 and “very positive” at +3.

Attitude change toward the firm due to the CRM alliance was assessed. Participants were asked the extent to which the CRM alliance would make their attitude toward the firm more positive, and then asked the extent to which the alliance would make their attitude toward the firm more negative.

ATTITUDINAL BALANCE 395

Perceptions of the strength of the relationship between the firm and the charity were measured. Since a CRM alliance in- volves an overt donation from the firm to the charity, with no corresponding overt helping behavior from the charity to the firm, the items assessing relationship strength primarily fo- cused on perceptions of the firm’s assistance to the charity. Scale creation is discussed in the next section. Perceptions of the relationship were assessed through comments made in the thought-listing exercise also discussed in the next section.

Procedure

Participants were run in groups of 8 to 12. Stimuli were pre- sented and responses recorded via personal computer. Partic- ipants were first shown a firm profile containing five state- ments regarding the fictitious firm. Participants were told to examine the information until they felt comfortable with it. Participants then indicated their attitude toward the firm. They were presented with the charity profile, again contain- ing five statements, and indicated their attitude toward the charity. A definition of cause-related marketing was pro- vided next, along with an example of a CRM alliance. Partic- ipants completed a thought-listing task regarding their re- sponse to the CRM alliance between the fictitious firm and charity. They were given three minutes, and told to type ev- erything that came to mind when considering the alliance. Participants then were asked to consider a CRM alliance be- tween the two organizations, and to either agree or disagree with adjectives describing the alliance.

Immediately following the adjective-response activity, participants answered a series of attitudinal questions regard- ing the firm, charity, and alliance between them. Responses were collected on a 7-point scale, ranging from –3 to +3. This procedure was repeated for the remaining three CRM pair- ings for a total of four randomly ordered iterations per partic- ipant.

Analyses

First, manipulations were assessed and scales created. Thought-listing results were used to assess the fit manipula- tion. Two coders, naïve to hypotheses and study condition, coded subjects’ open-ended responses. Intercoder agreement was 91%. The primary researcher resolved disagreements. Participants made more comments about poor fit in the no-fit condition, compared to the fit condition (t = –5.4, p < .001), suggesting a successful manipulation.

Mean firm attitudes in the positive attitude conditions (M = 2.18) were significantly higher than in the negative attitude conditions (M = –1.89, p < .001). Similarly, mean charity at- titudes in the positive attitude conditions (M = 2.41) were sig- nificantly higher than in the negative attitude conditions (M = –1.13, p < .001). The firm and charity attitude manipulations were thus deemed successful.

Participants’ responses to the positive (appealing, appro- priate, good, like) and negative (unappealing, inappropriate, bad, dislike) adjectives were coded “1” for agree and “–1” for disagree. Adjective responses were combined into scales by averaging the positive adjectives and averaging the negative adjectives. The four positive adjectives were assessed for scale reliability. Scale reliability was good (α = .92). The four negative adjectives were assessed and the resulting scale was also reliable (α = .90).

Scales were also created to assess perceptions of propriety and affect separately. Scores for the negative adjectives were reverse coded. For the affect-based adjectives (like, appeal- ing, dislike reverse coded, unappealing reverse coded), reli- ability was good (α = .96). The affect scale was created by averaging these scores. Similarly, the propriety-based adjec- tives scale (good, appropriate, bad reverse coded, inappropri- ate reverse coded) reliability was good (α = .97). The propri- ety scale was created by averaging these scores.

A scale was created for the dependent variable firm atti- tude change. Responses to the negatively framed question were reverse coded, then combined with the positively framed question. Scale reliability was good (α = .83).

Results

Hypothesis 1 proposed a distinction between affect-based re- sponses and propriety-based responses. Specifically, balance was expected to increase perceptions of propriety more than feelings of positive affect. A repeated-measures ANOVA was conducted. The balance condition was recoded into two cate- gories, balance and imbalance. Two-level attitude balance (balanced/unbalanced) served as a between-subjects factor, and judgment type (affect-based adjectives, propriety-based adjectives) a within-subjects factor. Mean scores for adjec- tive responses served as the dependent variable. A main ef- fect for judgment type was evident, F(1, 158) = 48.4, p < .001, ε2 = .26. Responses to propriety-based adjectives (good/bad, appropriate/inappropriate) were significantly more positive than responses to affect- based adjectives (like/ dislike, appealing/unappealing). A significant interaction be- tween judgment type and balance was also evident, F(1, 158) = 5.3, p < .05, ε2 = .03, supporting Hypothesis 1. Attitudinal balance had a significantly larger impact on perceptions of propriety regarding the CRM alliance than on affect toward the alliance (see Figure 2). Specifically, balanced attitudes (positive–positive or negative–negative) generated a judg- ment of propriety (M = 1.26), although they did not necessar- ily generate positive affect (M = .21). Since the adjectives “good” and “bad” have been used in other research to indi- cate affect rather than propriety, an assessment was made us- ing only the adjectives “appropriate” and “appealing.” Con- sistent with the prior analysis, balanced attitudes were judged somewhat appropriate (M = 1.9) but not very appealing (M = .09), t (79) = 2.0, p = .05, per Hypothesis 1.

396 BASIL AND HERR

FIGURE 2 Balance × Judgment Type Interaction.

FIGURE 3 CRM Attitude.

Another repeated-measures ANOVA was run to test Hy- potheses 2 and 4. Fit and judgment type served as within-sub- jects factors, firm and charity attitudes as between-subjects factors. Responses to the valenced adjectives served as the dependent variable. Hypothesis 2 was supported, F(1, 161) = 4.6, p < .05, ε2 = .03. Firm and charity attitudes interacted such that responses were significantly more positive when both pre-existing attitudes were positive and balanced, above the simple additive effects expected from each of the pre-ex- isting attitudes individually (see Figure 3). Hypothesis 4 was also supported, F(1, 161) = 20.7, p < .001, ε2 = .09. Fit had a stronger impact on perceptions of propriety than on affect. Alliances that fit were judged appropriate even if not well liked.

Hypothesis 3 proposed a main effect for fit on perceptions of relationship strength (part a) and that perceptions of rela- tionship strength would mediate the impact of fit on CRM at- titude (part b). First, responses to the thought-listing exercise

(comments regarding the relationship between the firm and charity) were examined. A paired-samples t test was con- ducted. More negative comments regarding the relationship were made in the no-fit conditions, compared to the fit condi- tions, t(172) = 2.5, p < .05, supporting Hypothesis 3a.

To further assess Hypothesis 3, a mediation test was con- ducted (Baron & Kenny, 1986). First, a repeated measures ANOVA was run with fit serving as the repeated measure, and perceptions of relationship strength as the dependent variable. Fit significantly increased perceptions of relation- ship strength, F(1, 166) = 56, p < .001, ε2 = .25, again sup- porting Hypothesis 3a. Moreover, firm attitude, F(1,155) = 67.4, p < .001, ε2 = .29, and charity attitude, F(1, 166) = 9.9, p < .005, ε2 = .06, both predicted relationship strength. A re- peated-measures ANOVA was then conducted, with fit serv- ing as the repeated measure and CRM attitude as the depend- ent variable. Fit significantly predicted CRM attitude, F(1, 166) = 77, p < .001, ε2 = .32. Finally, a third repeated-mea-

ATTITUDINAL BALANCE 397

FIGURE 4 Firm Attitude Change.

sures ANOVA was conducted, with fit serving as the repeated measure, perceived strength of the relationship a covariate, and CRM attitude the dependent variable. Both fit, F(1, 164) = 23, p < .001, ε2 = .13, and perceived relationship strength, fit F(1, 164) = 47, p < .001, ε2 = .22; no fit F(1, 164) = 15, p < .001, ε2 = .08, significantly predicted CRM attitude. Since the impact of fit was reduced with the inclusion of relation- ship strength (ε2 reduced from .32 to .13), fit was partially mediated by perceived relationship strength, supporting Hy- pothesis 3b.

To test Hypothesis 5, change in attitude toward the firm was assessed using an ANOVA. Firm and charity pre-exist- ing attitudes served as between-subjects factors. The firm at- titude change scale served as the dependent variable. Pre-ex- isting firm and charity attitudes significantly interacted, F(1, 166) = 7.0, p < .01, ε2 = .04, supporting Hypothesis 5. A syn- ergy of organizational attitudes was evident. When both pre-existing firm and charity attitudes were positive, change in attitude toward the firm was significantly larger than the simple additive effects would suggest (see Figure 4).

Discussion

The results of Experiment 1 support the hypotheses regard- ing attitude toward the CRM alliance. These results enhance understanding of the impact of fit on attitudes toward the CRM alliance. Prior research has demonstrated an impact for fit, but the nature of this impact has not been thoroughly ex- amined. This research demonstrates that fit enhances the sense of relationship strength between the two organizations. Organizations that fit are seen to have a stronger relationship. Moreover, more negative thoughts regarding the alliance come to mind when the organizations do not fit. Thus, fit strengthens the relationship between the two organizations, creating a stronger unit relationship. The effect of fit on CRM attitude is partially mediated by perceptions of strength of the CRM alliance. Hence, when organizations fit, the CRM alli-

ance is seen as more appropriate, although this alone may not generate much positive affect. This finding is an important step toward explicating the nature of the impact of fit.

A similar effect is found for balance. The significant bal- ance by judgment-type interaction suggests that balanced at- titudes lead to perceptions of appropriateness, but not neces- sarily to positive affect. Collectively, these results suggest that individuals have a sense of propriety for CRM alliances, with views on whether the pairing is appropriate or not, inde- pendent of their liking for the CRM alliance.

The results also support hypotheses regarding attitude change toward the firm. Pre-existing charity attitude is a strong determinant of attitude change toward the firm. Con- sistent with Balance Theory, firm attitudes changed to be- come more consistent with attitudes toward the alliance part- ner. A synergistic interaction of pre-existing attitudes was also evident, yielding benefits to the firm attitude given bal- ance and a positive pre-existing charity attitude.

The interaction between firm and charity attitude consis- tently attained significance. Both when predicting attitude to- ward the CRM alliance and when predicting firm attitude change, an interaction between firm and charity attitude was evident. This may be due to a synergistic effect when both balance and positive attitudes are present. When everything is perfect (pre-existing attitudes are positive and balance ex- ists), responses are much more positive. Alternatively, this pattern of results may be due to a “contamination effect.” Specifically, any one negative element may lead to a more negative response. In order to assess which of these is the case, it is necessary to compare responses for positive, neu- tral, and negative pre-existing organizational attitudes. If positive firm and charity attitudes lead to more positive eval- uations than when one or both of these is neutral, a synergy effect exists. Alternatively, if the difference stems from dif- ferences between negative and neutral attitudes, rather than neutral and positive attitudes, a contamination effect is sup- ported. Thus, the competing hypotheses:

398 BASIL AND HERR

H6a: The interaction between pre-existing firm and charity attitudes is due to a synergy effect such that CRM alliance attitudes will experience an enhance- ment only when both pre-existing attitudes are posi- tive, compared to neutral attitudes.

-OR-

H6b: The interaction between pre-existing firm and charity attitudes is due to a contamination effect such that CRM alliance attitudes will experience a decre- ment whether one or both pre- existing attitudes are negative, compared to neutral attitudes.

EXPERIMENT 2

Experiment 1 tested the proposed hypotheses. However, the stimuli used were of fictitious firms and charities, and organi- zational attitudes were created as part of the study. Although germane to newly formed attitudes, it is impossible to deter- mine whether the results generalize to situations involving longstanding pre-existing attitudes. In Experiment 2, a con- ceptual replication is performed. Real firms and charities are used, allowing a test of the proposed hypotheses with pre-ex- isting attitudes. The use of real organization names also en- hances external validity by allowing for a test of the hypothe- ses in a situation where many complex elements have likely contributed to attitude formation, as opposed to the simplistic firm profiles used in Experiment 1. Moreover, a larger num- ber of pairings is used, to improve the generalizability of re- sults. Experiment 2 addresses Hypotheses 6a and 6b to fur- ther clarify the nature of the firm by charity attitude interaction. Finally, need for cognition (NFC) was added to this experiment as a covariate, as previous research indicates that extent of cognitive elaboration affects consumer re- sponse to advertising efforts (Priester, Godek, Nayakankuppum, & Park, 2004) and may impact response to CRM alliances (Menon & Kahn, 2003). NFC was used as a proxy to control for possible elaboration differences.

Participants and Design

Sixty undergraduate business students participated in Exper- iment 2 for extra course credit. Males comprised 58% of the

sample. No participant took part in any Pretest or Experiment 1. Experiment 2 used a 2 (fit) × 6 (pairings) within-subjects design, with firm and charity attitudes serving as measured (rather than manipulated) variables. Need for Cognition (Cacioppo & Petty, 1982) served as a covariate. Six firms were paired with 2 charities each, 1 fit and 1 no-fit charity, for a total of 12 CRM alliances (see Table 1).

Each participant evaluated all 12 of the CRM alliances in a random order. Participants were run in four groups, ranging in size from 12 to 19. Paper and pencil stimuli were used, which participants completed at their own pace.

Independent Variables

“Fit” was manipulated by the selection of specific firms and charities, based on pretesting. Each firm was paired with one charity deemed to fit, and one deemed not to fit. “Pairing” was a nontheoretical replication variable included to increase generalizability. Testing multiple CRM alliances helps as- sure that results are not idiosyncratic.

Pre-existing firm and charity attitudes were measured in- dependent variables. Participants indicated their pre-existing attitudes toward the firm and the charity on an 11-point scale ranging from –5 to +5, with anchors “very negative” and “very positive.” A larger response scale was used in Experi- ment 2 than in Experiment 1 (10 point vs. 7 point), in case measures of pre-existing attitudes toward charities were con- stricted to positive or neutral scale responses. Efforts were made to select charities toward which some people held neg- ative attitudes.

Participants responded to 22 questions regarding each al- liance. Each question was posed on an 11-point scale, an- chored by “not at all” at –5 and “very much” at +5. The de- pendent variables of interest were attitude toward the CRM alliance, perception that this was a good CRM alliance, atti- tude change toward the firm, and perceptions of the strength of the alliance.

Since pre-existing firm and charity attitudes were mea- sured, rather than manipulated, a standard within-subjects ANOVA was deemed unsuitable. Within-subjects regression (Judd, Kenny, & McClelland, 2001) was used instead. This procedure provides a test of the moderating effects of continu- ous variables in within-subjects designs. To conduct a within-subjects regression, the dependent variable of interest

TABLE 1 Firm and Charity Pairings

Fit No Fit

Nike Athletic Shoes/The American Heart Association Nike Athletic Shoes/Feed the Children Velveeta Cheese/Feed the Children Velveeta Cheese/The American Heart Association Nintendo Video Games/Youth at Risk Nintendo Video Games/National Rifle Association Smith & Wesson Guns/National Rifle Association Smith & Wesson Guns/Youth at Risk Gerber Baby Food/The Pro-Life Action League Gerber Baby Food/Greenpeace Environmental Conservation Charity Big 5 Sporting Goods/Greenpeace Environmental Conservation Charity Big 5 Sporting Goods/The Pro-Life Action League

ATTITUDINAL BALANCE 399

is regressed on the within-subjects factors for each participant separately. The resulting betas for the within-subjects factors represent the relationship between each within-subjects factor and the dependent variable for each participant. These betas are then regressed on the between-subjects factors and indi- vidual difference variables using the data set as a whole. For example, regressions were performed for each individual par- ticipant, regressing firm attitude change (one of the dependent variables) on pre-existing firm attitude, pre-existing charity attitude, pre-existing firm attitude × pre-existing charity atti- tude interaction, and fit. The resulting betas for pre-existing firm attitude represented the relationship between pre- exist- ing firm attitude and firm attitude change due to the CRM alli- ance for each individual participant when controlling for other variables in the equation. Similarly, the resulting betas for fit represented the relationship between fit and firm attitude change for each participant, and so forth. The purpose of these individual regressions, then, is to ascertain the relational pat- tern between the within-subjects independent variables and the dependent variables for each participant, not to determine any form of statistical significance. After this step, the result- ing within-subjects beta weights were used in a series of be- tween-subjects regressions that regressed the beta weight on NFC(Cacioppo&Petty,1982) inorder toassessstatistical sig- nificance.Thiswasrepeatedforeachdependentvariable.Fora more detailed description of this procedure, see Judd et al. (2001).

Pre-existing firm and charity attitudes were tested as con- tinuous variables, but trichotomized to simplify reporting. Responses between 1 and 5 were grouped as positive pre-ex- isting attitudes (N = 477 for firms, N = 521 for charities). Re- sponses between –1 and –5 were grouped together as nega- tive pre-existing attitudes (N = 107 for firms, N = 109 for charities). Responses of zero were categorized “neutral” (N = 136 for firms, N = 88 for charities).

Results

Throughout Experiment 2, NFC served as a control variable. This variable did not attain significance in any of the analyses (p > .05), so it is not discussed further.

Hypothesis 2 was tested using a within-subjects regres- sion. Pre-existing firm and charity attitudes significantly in- teracted to predict attitude toward the CRM alliance (t = 2.5, p < .05, ε2 = .09), supporting Hypothesis 2. The synergistic effect of positive pre- existing attitudes and balance was evi- dent, demonstrating more positive attitudes toward the CRM alliance in the positive–positive condition.

Hypothesis 3 was tested to examine the effect of fit. B-weights calculated from responses to the statement “This will be a long-lasting alliance,” answered on an 11-point scale, served as the dependent variable. Baron and Kenny’s (1986) approach to testing mediation was used, adapted to a within-subjects design. First, a within-subjects regression was conducted to assess whether fit and pre-existing firm and

charity attitudes served to predict relationship strength. Fit significantly predicted perception of relationship strength (t = 11.6, p < .001), supporting Hypothesis 3a. Additionally, firm attitude and charity attitude significantly interacted to predict perceptions of strength (t = 3.0, p < .005). In the sec- ond step of the mediation test, fit significantly predicted atti- tude toward the CRM alliance (t = 6.2, B = 1.3, SE = .21, p < .001). Finally, with perception of relationship strength and fit included in the calculation of the B-weights, strength signifi- cantly predicted CRM attitude (t = 15, B = .71, SE = .05, p < .001), as did fit (t = 2, B = .22, SE = .11, p < .05). To assess the difference in the predictive power of fit with the inclusion of perceived strength, thus assessing mediation, the change in unstandardized betas was compared in terms of standard er- rors. Specifically, the unstandardized beta with strength in- cluded (.22) was subtracted from the unstandardized beta without strength included (1.3), and this amount was divided by the standard error for the unstandardized beta without strength included (.21). The unstandardized betas differed by 5.1 standard errors, suggesting a significant difference (dif- ferences exceeding 2 standard errors are significant). The ef- fect of fit was reduced when perceived strength was included, suggesting the effect of fit is partially mediated by percep- tions of relationship strength; this supports Hypothesis 3b.

To test Hypothesis 5, the effect of the CRM alliance on at- titudes toward the firm was assessed next. Pre-existing firm and charity attitudes interacted marginally to predict firm at- titude change (t = 1.9, p = .06, ε2 = .06). The positive–positive condition elicited more attitude change than the other condi- tions, marginally supporting Hypothesis 5.

Hypothesis 6 queried whether the interactive effect of pre-existing organizational attitudes was due to a synergy or a contamination effect. In order to assess this, trichotomized firm and charity attitudes were examined using t tests. The dependent variable attitude toward the CRM alliance was used. When both firm and charity attitudes were positive, CRM attitude was significantly more positive than when ei- ther attitude was neutral, t (418) = 4.3, p < .001, suggesting a synergy, supporting Hypothesis 6a (see Figure 5).

However, when either firm or charity attitude was nega- tive, CRM attitude was less favorable than when both atti- tudes were at least neutral (all p < .001). A single negative or- ganizational attitude served to contaminate CRM attitude, supporting Hypothesis 6b (see Figure 6).

Discussion

These results are consistent with and clarify the findings of Experiment 1. Pre-existing firm and charity attitudes interact to determine attitude toward the CRM alliance. Two distinct effects occur. First, positive pre-existing attitudes synergisti- cally enhance attitude toward the CRM alliance. This effect is magnified when both attitudes are positive, suggesting a “balance boost.” Attitude toward the CRM alliance becomes multiplicatively more positive when both pre-existing atti-

400 BASIL AND HERR

FIGURE 5 Synergy Effect.

FIGURE 6 Contamination Effect.

tudes are positive. A synergy is obtained by “having every- thing right.” A contamination effect also occurs, as a result of “having anything wrong.” If either the firm or the charity atti- tude is negative, attitude toward the CRM alliance deterio- rates.

These findings do not simply reflect ordinal results whereby neutral attitudes fall between negative and positive attitudes. Rather, one negative pre-existing attitude is equiva- lent to both attitudes being negative, suggesting contamina- tion. Similarly, only when both pre-existing attitudes are pos- itive is a multiplicative enhancement to CRM attitude

evident. When either one or both of the pre-existing attitudes are neutral, CRM attitudes are depressed. This again demon- strates a deviation from ordinal results. Everything must be “right” to obtain multiplicative attitudinal benefit; if anything is “wrong,” an attitudinal penalty occurs.

Balance Theory helps to explicate the effect of fit in a CRM alliance, as well. Fit consistently influenced attitudes. Fit positively impacted both CRM attitude and attitude change toward the firm. Moreover, fit led to perceptions of a stronger unit relationship between the firm and the charity, partially mediating the impact of fit.

ATTITUDINAL BALANCE 401

GENERAL DISCUSSION

Prior work demonstrated that fit impacts consumer response to CRM alliances, but the nature of this impact was uncertain. Fit’s role was clarified in the present research. Fit increased perceptions of the relationship between the firm and the char- ity. This effect may be understood in terms of Balance The- ory (Heider, 1946, 1958). The beneficial impact of fit in a CRM alliance appears partly to stem from a perception that the relationship between the firm and the charity is stronger when fit exists. In Balance Theory terms, the unit relation- ship between the CRM (firm and the charity) becomes stron- ger when fit exists. Fit’s impact on alliance attitudes was par- tially mediated by perceptions of relationship strength. In part, fit influences CRM attitude by strengthening percep- tions of the unit relationship between the firm and the charity. These results suggest that when contemplating an alliance, fit should be a primary consideration.

Previous research has also demonstrated that pre-existing firm and charity attitudes impact attitude toward the CRM al- liance. This research advances our understanding of the role of pre- existing attitudes by demonstrating that this impact is interactive: The effect of pre-existing firm attitude depends upon the valence of pre-existing charity attitude, and vice versa. Specifically, a synergistic benefit was evident when pre-existing attitudes toward both of the organizations were positive. If pre-existing attitudes toward the target organiza- tion are negative, pre-existing attitudes toward the partner or- ganization become less important. This suggests that if con- sumer attitudes toward a firm are negative, adding a CRM partner will have reduced impact. Partner attitudes do have a significant impact on attitude toward the target organization, so if a firm with negative pre-existing attitudes partners with a charity with positive pre-existing attitudes, the firm will benefit from the alliance, but the benefit is far less than what would be enjoyed by a more positively viewed firm. If pre-existing firm attitudes are positive, the firm stands to gain a good deal through a CRM alliance, but only if the alliance partner enjoys positive consumer attitudes as well. These re- sults represent a synergistic effect for having both positive firm and charity attitudes, and a contamination effect for hav- ing either a negative firm or a negative charity attitude.

Attitudinal balance has differential impact depending upon the type of judgment being made. Balanced attitudes are seen as appropriate. It is appropriate for an organization to partner with another organization toward which pre-exist- ing attitudes are comparable. Balance, however, has far less influence on affect toward the alliance. Positive attitudes are necessary to generate positive affect, whereas only balance is necessary to generate a sense of propriety. This effect was mirrored with fit. Specifically, fit between two organizations has a stronger impact on perceptions of propriety than on positive affect. Collectively, these results suggest that when two organizations appear to “go together,” either because they share a common attitude valence (balance) or they share

a common purpose (fit), their alliance is seen to be appropri- ate. This does not indicate that the alliance will be well liked, however. Positive organizational attitudes are necessary to generate positive affect toward the alliance.

Practical Implications

The interaction between firm and charity pre-existing atti- tudes suggests that the organizations most likely to benefit from a CRM alliance may in fact be those that need it least. The greatest benefit is attained (a multiplicative enhance- ment to attitudes) when pre-existing attitudes toward both firm and charity are positive. Firms already enjoying positive consumer attitudes may be particularly good candidates for CRM campaigns, to further solidify their attitudinal advan- tage.

These results also suggest that an organization may not be able to overcome negative consumer attitudes by simply forming a CRM alliance. If attitudes toward the firm are neg- ative, it makes little difference whether the charity has posi- tive or negative consumer attitudes. Attitudes toward the firm change very little either way as a result of a CRM alliance.

Finally, fit is important for CRM alliances. CRM alliances are seen as more appropriate when they fit. At least in part, fit operates by strengthening perceptions of the firm and charity relationship. Given that consumers are often skeptical about firms’ motives for helping charities, enhancing perceptions of the strength of the firm/charity relationship may help to re- duce skepticism and thus should benefit the firm.

Limitations

This research faces the limitations common to many labora- tory experiments in which student participants are used, in- cluding questions of generalization. The homogenous sam- ple, however, is acceptable for theory testing (Calder, Phillips, & Tybout, 1981). Replications with “real-world” samples are clearly warranted before extending our results very far.

The benefit of a within-subjects design is that more CRM alliance pairings could be tested without increasing sample size. This also helps to assure that our results are not due to an idiosyncratic CRM pairing. The within-subjects design may, however, increase experimental awareness of the study par- ticipants. Future research should seek to replicate this work in a between-subjects design.

ACKNOWLEDGMENTS

We gratefully acknowledge funding from the University of Colorado at Boulder.

402 BASIL AND HERR

REFERENCES

Aaker, D. A., & Keller, K. L. (1990). Consumer evaluations of brand exten- sions. Journal of Marketing, 54(1), 27–41.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable dis- tinction in social psychological research: Conceptual, strategic, and statisti- cal considerations. Journal of Personality and Social Psychology, 51(6), 1173.

Barone, M. J., Miyazaki, A. D., & Taylor, K. A. (2000). The influence of cause-related marketing on consumer choice: Does one good turn de- serve another? Journal of the Academy of Marketing Science, 28(2), 248–262.

Basil, D. Z., & Herr, P. M. (2003). Dangerous donations? The effects of cause-related marketing on charity attitude. Journal of Nonprofit and Public Sector Marketing, 11(1), 59–76.

Bottomley, P. A., & Holden, S. J. S. (2001). Do we really know how consum- ers evaluate brand extensions? Empirical generalizations based on sec- ondary analysis of eight studies. Journal of Marketing Research, 38(4), 494–500.

Boush, D. M., & Loken, B. (1991). A process-tracing study of brand exten- sion evaluation. Journal of Marketing Research, 28(1), 16–28.

Broniarczyk, S. M., & Alba, J. W. (1994). The importance of the brand in brand extension. Journal of Consumer Research, 31(2), 214–228.

Cacioppo, J. T., & Petty, R. E. (1981). Effects of extent of thought on the pleasantness ratings of P-O-X triads: Evidence for three judgmental ten- dencies in evaluating social situations. Journal of Personality and Social Psychology, 40, 1000–1009.

Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42(1), 116–131.

Calder, B. J., Phillips, L. W., & Tybout, A. M. (1981). Designing research for application. Journal of Consumer Research, 8(2), 197–208.

Cartwright, D., & Harary, F. (1956). Structural balance: A generalization of Heider’s theory. Psychological Review, 63, 277–293.

Cone/Roper. (1999). Cause-Related Trends Report. Boston: Cone, Inc. Roper Starch Worldwide, Inc.

Dacin, P. A., & Smith, D. C. (1994). The effect of brand portfolio character- istics on consumer evaluations of brand extensions. Journal of Marketing Research, 31(2), 229–242.

Dean, D. H. (2004). Consumer perception of corporate donations. Journal of Advertising, 32(4), 91–102.

Deshpande, S., & Hitchon, J. (2002). Ethical implications of use of cause-re- lated marketing ads in the light of negative news. Journalism and Mass Communication Quarterly, 79(4), 905–926.

Hamlin, R. P., & Wilson, T. (2004). The impact of cause branding on con- sumer reactions to products: Does product/cause “fit” really matter? Jour- nal of Marketing Management 20(7&8), 663.

Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychol- ogy, 21, 107–112.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Herr, P. M., Farquhar, P. H., & Fazio, R. H. (1996). Impact of dominance and relatedness on brand extensions. Journal of Consumer Psychology, 5(2), 135–159.

Jordan, N. (1953). Behavioral forces that are a function of attitudes and of cognitive organization. Human Relations, 6, 273–287.

Judd, C. M., Kenny, D. A., & McClelland, G. H. (2001). Estimating and test- ing mediation and moderation in within-subject designs. Psychological Methods, 6, 115–134.

Lafferty, B. A., & Goldsmith, R. E. (2005). Cause-brand alliances: Does the cause help the brand or does the brand help the cause? Journal of Business Research, 58(4), 423.

Lafferty, B. A., Goldsmith, R. E., & Hult, T. M. (2004). The impact of the al- liance on the partners: A look at cause-brand alliances. Psychology & Marketing, 21(7), 509–531.

Lichtenstein, D. R., Drumwright, M. E., & Braig, B. M. (2004). The effect of corporate social responsibility on customer donations to corporate-sup- ported nonprofits. Journal of Marketing, 68(4), 16.

Menon, S., & Kahn, B. E. (2003). Corporate sponsorships of philanthropic activities: When do they impact perception of sponsor brand? Journal of Consumer Psychology, 13(3), 316–327.

Olsen, G. D., & Pracejus, J. (2004). Integration of positive and negative af- fective stimuli. Journal of Consumer Psychology, 14(4), 374–384.

Osgood, C. E., & Tannenbaum, P. H. (1955). The principle of congruity in the prediction of attitude change. Psychological Review, 61, 42–55.

Pham, M. T. (2002). The logic of feeling. Journal of Consumer Psychology, 14(4), 360–369.

PMA/Gable Group. (2000). Survey of cause marketing. Retrieved October 19, 2001, from http://www.pmalink.org

Priester, J. R., Godek, J., Nayakankuppum, D. J., & Kiwan, P. (2004). Brand congruity and comparative advertising: When and why comparative ad- vertisements lead to greater elaboration. Journal of Consumer Psychol- ogy, 14(1&2), 115–123.

Sankar, S., & Bhattacharya, C. B. (2001). Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. Jour- nal of Marketing Research, 38(2), 225–243.

Simonin, B. L., & Ruth, J. A. (1998). Is a company known by the company it keeps? Assessing the spillover effects of brand alliances on consumer brand attitudes. Journal of Marketing Research, 35(1), 30–42.

Speed, R., & Thompson, P. (2000). Determinants of sports sponsorship re- sponse. Academy of Marketing Science Journal, 28(2), 226–238.

Strahilevitz, M., & Myers, J. G. (1998). Donations to charity as purchase in- centives: How well they work may depend on what you are trying to sell. Journal of Consumer Psychology, 24(4), 434–446.

Szykman, L. R., Bloom, P. N., & Blazing, J. (2004). Does corporate sponsor- ship of a socially-oriented message make a difference? An investigation of the effects of sponsorship identity on responses to an anti-drinking and driving message. Journal of Consumer Psychology, 14(1&2), 13–20.

Yechiam, E., Barron, G., Erev, I., & Erez, M. (2003). On the robustness and the direction of the effect of cause-related marketing. Journal of Con- sumer Behaviour, 2(4), 320.

APPENDIX

Subjects rated their attitudes toward the following atti- tude-formation statements on 11-point Likert scales. The av- erage rating for each statement was then compared against the scale midpoint (5), using a one-sample t test (or z test). Mean scores significantly below five indicate statements that generated negative attitudes. Mean scores significantly above 5 indicate statements that generated positive attitudes. Statements not significantly different from five represent neutral attitudes. Each organizational profile contained three valenced statements (positive or negative) and two neutral statements.

Charity Statements

Average volunteer tenure is over 5 years, M = 7.0, t = 6.5, p < .0001

Average volunteer tenure is 3 months, M = 4.2, t = 4.3, p < .0001

In an organization-wide survey, volunteers ranked their job satisfaction as “very high”, on average, M = 8.3, t = 16.6, p < .0001

In an organization-wide survey, volunteers ranked their job satisfaction as “somewhat low”, on average, M = 2.7, t = 12.2, p < .0001

An independently conducted survey of those charity X has worked to help demonstrated very high levels of satisfaction, M = 8.5, t = 15.6, p < .0001

An independently conducted survey of those Charity X has worked to help demonstrated low to moderate levels of satis- faction, M = 3.3, t = 6.8, p < .0001

An industry consortium of charities voted to honor the presi- dent of Charity X for his exceptional management, M = 7.4, t = 8.6, p < .0001

The president of Charity X is under investigation for misuse of organization funds, M = 1.2, t = 18.2, p < .0001

Charity X has been operating for over 50 years, M = 7.9, t = 12.1, p < .0001

Charity X has been operating for almost 1 year, M = 4.5, t = 2.5, p = .016

Charity X was voted best overall charity by a consortium of charities in their field, M = 8.6, t = 16.1, p < .0001

Charity X has not received any honors from the consortium of charities, M = 4.3, t = 3.2, p = .003

Charity X has never been reported to the Better Business Bu- reau for inappropriate fund-raising methods, M = 7.2, t = 6.6, p < .0001

The Better Business Bureau recently received several com- plaints regarding fund-raising methods of Charity X, M = 1.9, t = 13.4, p < .0001

Less than 10% of all funds donated go toward overhead, M = 7.4, t = 8.6, p < .0001

Approximately 50% of all funds donated go toward over- head, M = 3.6, t = 4.9, p < .0001

Charity X operates in eight states, M = 5.4, t = 1.8, p = .078

Charity X has been in operation for 6 years, M = 5.6, t = 2.6, p = .012

Charity X is headquartered in Tucson, Arizona, M = 5.1, t = .62, p = .54

Firm Statements

ATTITUDINAL BALANCE 403

Average employee tenure is over 5 years, M = 6.3, t = 5.3, p < .0001

In a firm-wide survey, employees ranked their job satisfac- tion as “very high”, on average, M = 8.5, t = 21.4, p < .0001

Average employee tenure is 3 months, M = 2.9, t = 8.5, p < .0001

In an organization-wide survey, employees ranked their job satisfaction as “somewhat low”, on average, M = 2.5, t = 14.9, p < .0001

An independently conducted survey of customers demon- strated very high levels of customer satisfaction, M = 8.1, t = 13.5, p < .0001

An industry consortium of businesses voted to honor the president of Firm Y for his exceptional management, M = 7.5, t = 13.5, p < .0001

The president of Firm Y is under investigation for misuse of corporate funds, M = 1.7, t = 17.6, p < .0001

Firm Y has been operating for over 50 years, M = 7.7, t = 12.3, p < .0001

Firm Y has been operating for almost 1 year, M = 4.6, t = 3.6, p < .005

Industry analysts have ranked Firm Y as the best overall in- vestment in its industry, M = 8.3, t = 14.6, p < .0001

Firm Y has never been ranked by industry analysts, M = 5.3, t = .9, p < .4

Firm Y has never been reported to the Better Business Bu- reau for inappropriate marketing methods, M = 6.6, t = 5.2, p < .0001

The Better Business Bureau recently received several com- plaints regarding marketing methods of Firm Y, M = 2.5, t = 13.8, p < .0001

Firm Y uses only the highest-quality materials/ingredients in all aspects of production, M = 7.7, t = 10.2, p < .0001

The ingredients and materials used by Firm Y minimally meet legal requirements, M = 3.8, t = 4.5, p < .0001

Firm Y operates in eight states, M = 5.7, t = 4.3, p < .0001

Firm Y has been in business for 6 years, M = 5.6, t = 4.8, p < .0001

Firm Y is headquartered in Tucson, Arizona, M = 4.9, t = .5, p < .6