Identity Congruency Effects on Donations

기부에 대한 정체성 일치 효과

Journal of Marketing Research 351 Vol. XLV (June 2008), 351–361 © 2008, American Marketing Association ISSN: 0022-2437 (print), 1547-7193 (electronic) *Jen Shang is a postdoctoral fellow, Center on Philanthropy, Indiana University (e-mail: yshang@iupui.edu). Americus Reed II is Associate Professor of Marketing, the Wharton School, University of Pennsylvania (e-mail: amreed@wharton.upenn.edu). Rachel Croson is Professor of Economics and Management, School of Management and EPPS, and Director of the Negotiations Center, University of Texas at Dallas (e-mail: crosonr@utdallas.edu). The authors appreciate the financial support of dissertation awards from the National Science Foundation and the Aspen Institute and a summer research grant from Risk Management and Decision Processes Center at the Wharton School. They also graciously thank Daniela Lejtneker for invaluable assistance with data collection. Chris Janiszewski served as associate editor for this article. JEN SHANG, AMERICUS REED II, and RACHEL CROSON* This article describes several field and laboratory experiments that investigate an identity congruency effect on donations. Experiment 1 is a field experiment showing that consumers give more money to a public radio station if they are told that a previous donor who shares their identity also made a large contribution. This effect is more likely to occur when consumers have high collective-identity esteem (measured in Experiment 2a) and when attention is focused on others (manipulated in Experiment 2b). The authors measure these two moderators simultaneously and observe and replicate a three-way interaction. Again, the identity congruency effect is the strongest when consumers have high collective-identity esteem and when attention is focused on others (Experiment 3a and Experiment 3b). These results provide a novel understanding of the causes of the identity congruency effect on donations. The authors conclude with a discussion of the theoretical and substantive implications of these findings. Keywords: social identity, philanthropy, charitable giving, collective selfesteem, self focus, other focus, fundraising, nonprofit marketing Identity Congruency Effects on Donations Consumers’ behavior is often affected by what others have done (Amaldoss and Jain 2005; Reingen 1982). In this research, we examine a particular kind of others’ behavior—another person’s contribution to a nonprofit entity— and how it affects actual donation behavior and donation intentions. Prior research indicates that consumers may be affected by the behavior of others when they share some basis for social identification (see, e.g., Forehand and Deshpandé 2001). We examine the case when a target learns of a previous donor’s contribution, and we ask whether the match of the target and the previous donor’s identity increases the target donor’s actual donation or donation intentions. We call this effect the “identity congruency effect” and explore its dispositional and transitory moderators. Experiment 1 is a field experiment demonstrating the identity congruency effect. We find that actual call-in donors give more to a public radio station if they are told that a person with the same identity made a high contribution. The next four experiments investigate the moderating mechanisms behind the observed effect. We find that the identity congruency effect is more likely to occur when consumers have high identity esteem (measured in Experiment 2a) or when attention is focused on others (manipulated in Experiment 2b). The last two studies (Experiment 3a and Experiment 3b) explore the joint impact of these two moderators and identify and replicate a three-way interaction: When consumers have low identity esteem, the identity congruency effect does not emerge, regardless of whether their attention focuses on the self or on others, but when consumers have high identity esteem, the identity congruency effect is more likely to occur when attention focuses on others than when attention focuses on the self. The five experiments make several contributions to the field of marketing and to the knowledge of psychological processes in consumption and donation decision making. This is the first field experiment to demonstrate that identity congruency influences the level of actual charitable contributions. There has been a call for consumer research to focus more on “messier” environments to buttress the mundane realism of the body of knowledge in the field (Cohen 2005). Our first experiment answers this call. 352 JOURNAL OF MARKETING RESEARCH, JUNE 2008 We also contribute to the growing body of work in consumer behavior that illuminates the importance of identity processes on consumption decisions in general (Forehand, Deshpandé, and Reed 2002; Reed 2004) and how identities influence consumption decisions as a function of contextual and situational cues in particular (see DeMarree, Wheeler, and Petty 2005; Mandel 2003). We demonstrate the importance of identity processes and further build on this research in the domain of donation behavior. More specifically, this is the first study to investigate identity esteem and self/other focus as moderators of identity congruency effects and the first to demonstrate their joint impact on decisions (in this case, donations). Finally, this study is important because it provides an examination of key identity-based constructs in the substantive domain of nonprofit marketing and thus should be informative for nonprofit theoreticians and consumption behavior researchers, as well as for practitioners who aim to increase charitable contributions. THEORETICAL BACKGROUND OF IDENTITY CONGRUENCY EFFECTS IN THE DONATION CONTEXT In the consumer behavior field, there is extensive literature on social influence. In situations in which information is ambiguous or absent, consumers are often affected by what other consumers have done (Burnkrant and Cousineau 1975; Cohen and Golden 1972). However, it is not only the question of “what” that matters but also the question of “who.” For example, a consumer’s identity often serves a social adjustment function for those who are motivated to reduce the psychological distance to similar real or imagined others (Smith, Bruner, and White 1956, p. 42). Social identity facilitates this adjustment by directing the consumer to specific beliefs and behaviors that signal identification externally to society and internally to the self. In this sense, under certain conditions, the behavior of a consumer with whom a person shares some basis for social identification may be particularly influential on his or her own decision making. In this article, we identify charitable donations as a behavioral context in which others’ actions may influence target donors’ decisions, and identity congruency is of special theoretical and practical significance. Social Identities Can Be Diagnostic When social identities are diagnostic of the decisions or judgments at hand, other consumers’ behavior can become relatively more relevant to a judgment (Feldman and Lynch 1988) when aspects of these consumers connect to the target consumer’s identity (Reed 2004). Thus, when an identity dimension of a target consumer and the other consumers is congruent, it may prompt the target consumer to categorize him- or herself along that dimension (Forehand, Deshpandé, and Reed 2002; Reed 2004). Although consumers can consciously assess their relative similarity or dissimilarity with other people, a great deal of this selfcategorization occurs without conscious processing (see Eiser and Sabine 2001; Stapel and Koomen 2000). Nonetheless, this congruency may be a reason another consumer’s behavior becomes more diagnostic and possibly affects a donation decision. In the experiments presented in this article, we manipulate the social identity congruency of the target donor and a previous donor. Gender Is a Dimension of Social Identity Applicable in the Field of Charitable Fundraising The gender match and mismatch between the target donor and a previous donor is one of the many potential dimensions in which identity congruency effects may appear and affect the target donor’s behavior. Theoretically, gender is a well-established dimension of social identity in the psychological literature (Bem 1981; Deaux et al. 1995). For example, people for whom gender is important and accessible have “male” or “female” as part of their selfschemata (Bem 1981). Consequently, these consumers tend to attend and react to themselves and others in terms of gender rather than some other social dimension (Frable 1989). Males and females have also been shown to behave in ways that are consistent with the social targets of the same gender in other settings (Mussweiler, Ruter, and Epstude 2004). Therefore, in the domain of charitable contributions, social identity along the gender dimension can serve as a diagnostic cue in the decision of how much to give. From an implementation point of view, it is desirable to select a relatively “clean” dimension of social identity that is applicable and easily identifiable to all potential donors (Cialdini 2001). Previous research has used incidental similarities, such as first names, birthdays, or fingerprints; however, this information is not available to us in critical field settings. Conversely, gender is applicable to all potential donors, and callers’ gender is identifiable from their voice, with only a few exceptions. Thus, for theoretical and practical reasons, we selected gender as the dimension of social identity to study the effect of identity congruency on charitable donations. We believe that the target donors will be relatively more influenced by another donor’s previous behavior when their own gender matches the gender of the other donor. H1: Donations are higher when identity is congruent than when it is incongruent. EXPERIMENT 1: THE EFFECT OF IDENTITY CONGRUENCY ON REAL DONATIONS Method Design. We used a subtle between-subjects design with two experimental conditions. Before making their own contribution, callers were told the contribution of another donor. In the identity-congruent (gender was matched) condition, male (female) donors were told: “We had another member; he [she] contributed $240.” In the identityincongruent condition, gender was mismatched. Participants. Seventy-six new donors received the experimental treatments. The station limited our data collection efforts to new donors who had never previously contributed. Procedure. We conducted this field experiment in an anonymous public radio station on the East Coast in June and September 2004 during the station’s on-air fund drive. During the on-air drive, the station disc jockeys interspersed music with appeals for donations. Listeners responded to the on-air appeals during the drive and called the station to make a pledge. Experimenters answered the telephone as volunteers for the station, asked the routine Identity Congruency Effects on Donations 353 questions for the station, and implemented the manipulation in the appropriate place in the conversation. After answering the telephone with the station’s identifier, “Hello, [STATION_NAME] member line,” the experimenters asked, “Are you a new member or a renewing member of [STATION_NAME]?” For new members, experimenters read the following sentence (depending on treatment): “We had another member; he [she] contributed $240.” The dependent measure, the pledge amount, was collected in the following sentence: “How much would you like to pledge today?” The two experimental conditions were randomized with each experimenter and within each hour. We ensured that another member of each gender had indeed contributed the suggested amount ($240) previously in the fund drive, so that our statements would not constitute deception (for more information on the experimental procedure, see the Web Appendix, Part A1, at http:// www.marketingpower.com/jmrjune08). Results Our analysis of existing station data suggests that contributions can be dramatically different depending on the fundraising theme used in each drive, the thank-you gifts offered each day and hour, and the donors’ gender. This pattern is replicated in our current analysis. Although not all these factors significantly explained variance in our experimental data, we included them as control variables in our regression analysis. The coefficient of identity congruency was significant (t = 2.05, p < .05; N = 76), but the gender interaction with the identity congruency effect was not. This means that the degree of the identity congruency effect on actual donations did not differ between males and females. However, consistent with H1, the matched-gender condition (M = $141.88, SE = 14.09; N = 35) produced significantly higher contributions than the mismatched-gender condition (M = $105.70, SE = 8.84; N = 41) (for the regression results, see the Web Appendix, Part A2, at http:// www.marketingpower.com/jmrjune08). Discussion of Experiment 1 The influence of another consumer’s behavior on the target’s contribution is sensitive to the congruency between the identity of the other and the target. This result is consistent with the claim that behavior is affected by the match of the target consumer’s identity and the other consumer’s identity. The identity congruency effect we found in this experiment was shown in a close interpersonal context in which donors were engaged in a telephone conversation with the experimenter. To generalize this result to other fundraising situations, such as direct marketing, which is the most widely used fundraising method, we manipulate the identity congruency effect using printed manipulations in the next set of studies. In addition, we explore the mechanisms behind the identity congruency effect and investigate internal and dispositional identity-based trait and situational state-of-mind factors that may moderate this effect. EXPERIMENT 2A: THE EFFECT OF IDENTITY CONGRUENCY AND GENDER ESTEEM ON DONATIONS Experiment 1 provided evidence of an identity congruency effect on actual donation behavior. The literature on identity activation effects in judgment provides some insight into the mechanisms that moderate this effect. For example, differential sensitivity to identity-congruent information should be strongest when the relevant identity is an activated component of the consumer’s social self-schema (Reed 2004). This activation can be influenced by the degree to which that basis for social identity is personally significant, meaningful, and highly important to how a consumer views him- or herself (Bem 1981; see also Schmitt, Leclerc, and Dube-Rioux 1988) or how much value a person places on such social identity (Luhtanen and Crocker 1992). A consumer may be drawn to a particular social identity for many reasons, including instances in which that social identity is central or valuable to his or her selfconcept. For such people, this social identity serves as a “phenomenological lens” that deeply engulfs the consumer as a powerful basis for self-definition (Reed 2004) and results in a significant amount of positivity related to the most individualist (versus collectivist) aspect of a person’s identity esteem (i.e., membership esteem). Identity membership esteem is defined as “individuals’ judgment of how good or worthy they are as members of their social groups” (Luhtanen and Crocker 1992, p. 305). As such, social identities are often “core” aspects of self-construal and therefore are chronically accessible and likely to stimulate identity-based processing of the environment, even in the absence of social or contextual cues (Markus and Nurius 1986; Reed and Forehand 2007). H2: The effect of identity congruency on contribution is moderated by the target donor’s identity esteem; in particular, identity congruency is stronger when the target donor’s identity esteem is high than when it is low. Method Participants. Forty-three undergraduate students in a northeastern university completed a series of tasks as part of a one-hour experimental session. All participants were told that they would be compensated $10 for participation in the laboratory session. They were then actually given $14 along with the opportunity to donate some of their payment for future research. We assessed participants’ gender esteem, provided them with incidental information of either male or female previous donations, and measured their actual level of donations. Design. The study is a 2 (identity congruency: match versus mismatch) × 2 (identity esteem: low versus high) between-subjects design that manipulates the first factor and measures the second. Procedure. When participants entered the laboratory, they were told that they were going to participate in several unrelated experiments by different researchers and then were given written instructions. Buried in these tasks were the key experimental materials of interest to this study. For the purpose of our experiment, participants received an identity-esteem scale and two envelopes, together with a letter. This letter was developed to create a situation in 354 JOURNAL OF MARKETING RESEARCH, JUNE 2008 Figure 1 EXPERIMENT 2A: IDENTITY CONGRUENCY EFFECT ON DONATIONS EMERGES IN PEOPLE WITH HIGH BUT NOT LOW IDENTITY ESTEEM which participants could make charitable contributions anonymously in the lab. Participants were instructed to open Envelope 2, which contained $14 (one $5 bill and nine $1 bills). It was $4 more than their expected compensation for participation in the experimental session. First, we measured identity esteem by modifying Luhtanen and Crocker’s (1992) collective self-esteem scale to account for gender. We presented this scale to participants as a separate and unrelated study before they opened the envelopes and read the letter described. We measured the four items on a nine-point scale ranging from “strongly disagree” (1) to “strongly agree” (9): “I am a worthy member of my gender group,” “I feel I don’t have much to offer to my gender group,” “I am a cooperative participant in the activities of my gender group,” and “I often feel I’m a useless member of my gender group.” We combined these four items into an aggregate measure (α = .70) that formed an index of identity esteem. Second, we manipulated identity congruency in a letter that participants received with their $14 payment for the experimental session (see the Web Appendix, Part B, at http://www.marketingpower.com/jmrjune08). The letter thanks respondents for their participation and requests a contribution toward future research activities. The letter includes an example of another participant who contributed the entire $14. Similar to Experiment 1, we held the other person’s (high) donation constant ($14) and manipulated whether the previous contributor was male or female. In the identity-congruent (identity-incongruent) condition, participants’ gender matched (mismatched) the donors’ gender in the letter. Dependent variable. The amount of money participants donated was the dependent variable. Results We tested H2 by conducting a hierarchical regression on the dependent variable. In Step 1, we entered gender and age into the model. In Step 2, we entered the two main effects of the dummy-coded identity congruency variable (mismatch = 0, and match = 1) and the identity-esteem measure. In Step 3, we entered a two-way interaction as our direct test of H2. To minimize multicollinearity, we meancentered the covariate (gender identity esteem) before creating the cross-product terms (Aiken and West 1991). The main effects in Step 2 did not produce a significant change in the R-square for the omnibus test (F(2, 38) = .05, not significant [n.s.]), but the two-way interaction term did (F(1, 37) = 7.87, p < .05). The omnibus analysis showed a marginally significant main effect of identity esteem (t = 1.80, p < .1) that was qualified by a significant identity congruency × identity esteem two-way interaction (t = 2.81, p < .05). To understand the nature of these results, we graphed the identity congruency × identity esteem interaction using the procedures that Aiken and West (1991) recommend. In Figure 1, the identity congruency variable is on the x-axis, and the amount of the donation is on the y-axis; each slope represents this relationship at +1 and –1 standard deviations above and below the mean for gender identity esteem. This pattern supports H2. It shows that the effect of identity congruency on donations depends on dispositional identity esteem. This two-way interaction is driven by two patterns of the main effect of identity congruency. More specifically, the positive slope of identity congruency is significant in the high-identity-esteem condition (value of slope = 1.97, SE of slope = .95; t = 2.06, p < .05), whereas the negative slope is not significant in the low-identityesteem condition (value of slope = –.92, SE of slope = .82; t = 1.12, n.s.). Consistent with the theoretical underpinnings suggested in previous identity research, when identity esteem is high, the relevant identity cue of the other giver’s gender affects giving more. This relationship does not occur when identity esteem is low. Discussion of Experiment 2a In this experiment, we find that the identity congruency effect on donations is more likely to emerge when the target donor has high identity esteem. We found this pattern in actual contributing behavior in the laboratory setting and in a context in which participants’ giving was made privately. An advantage of this experiment is that the identityesteem variable was buried in another survey and was measured before the identity congruency manipulation, so the identity-esteem variable could not have been affected by the experimental manipulation of identity congruency. However, this study has limitations. Because we measured rather than manipulated identity esteem, it may correlate with some other unobserved factors that produce the effect. The manipulation of an identity-based factor of interest would provide additional insights into the identity congruency effect. Moreover, there may be something idiosyncratic to undergraduate students’ giving behavior that may not generalize to an adult population. Experiment 2b attempts to address these internal and external validity limi- Identity Congruency Effects on Donations 355 tations by manipulating an additional factor identified in the literature (self/other focus) and by exploring its moderating impact on the identity congruency effect in an adult population. EXPERIMENT 2B: IDENTITY CONGRUENCY AND SELF/OTHER FOCUS IN ADULTS The social identity function (Shavitt 1990) argues that identity-based attitudes and judgments not only help consumers classify themselves but also may become the embodiment of a unique social classification or reference group (Nelson et al. 1997; Shavitt 1989, 1990; Shavitt and Nelson 2000). Over time, these attitudes not only result from shared group beliefs but also come to embody group membership itself. This process is likely to result in a collectively anchored preference or attitude that is formed through identification processes (Kelman 1961) and is held, expressed, or used as a guide for behavior to establish, maintain, or even communicate that basis of social identification to others (Reed 2004). Identity-based judgments, attitudes, and beliefs are unsurprisingly resistant to counterpersuasion (Bolton and Reed 2004), and research indicates that this may be at least partially due to the notion that such beliefs are perceived as having social reality, in that they exist within a social network of others with similar views (Visser and Mirabile 2004). If such conjecture is valid, it leads to the prediction that the kind of identity congruency effects we present herein may be heightened if a person is thinking about others at the time he or she is making a decision to donate: H3: The identity congruency effect is moderated by self/other focus and is more likely to emerge when target donors focus on others than when they focus on themselves. Method Participants. Two hundred twelve professionals in a market in the downtown of a Midwest city completed a survey voluntarily while having lunch. Design. The study is a 2 (identity congruency: match versus mismatch) × 2 (focus: self versus other) betweensubjects design in which both factors are manipulated. Procedure. First, we manipulated self versus other focus by asking participants to read a short passage as part of a reading comprehension task (Aaker and Lee 2001; see the Web Appendix, Part C, at http://www.marketingpower.com/ jmrjune08). This task was presented to participants as a separate and unrelated study before they received the scenario explained in the next section. Participants then rated four manipulation-check items on a scale from 1 (“not at all”) to 7 (“a lot”): “Your thoughts were focused on just you,” “Your thoughts about the paragraph were focused on just yourself,” “Your thoughts were focused on other listeners,” and “Your thoughts about the paragraph were focused on other public radio listeners.” Participants also assessed two items on a scale from 1 (“you”) to 7 (“other public radio listeners”): “Your thoughts were focused on you or other public radio listeners,” and “Your thoughts about the paragraph were focused on you or other public radio listeners.” Second, we manipulated identity congruency in a scenario that was carefully developed to mimic the situation in public radio fundraising (see the Web Appendix, Part C, at http://www.marketingpower.com/jmrjune08). Respondents were told that they were prepared to contribute to a public radio station and that another donor had given $200. They were then asked for their contribution. We manipulated identity congruency similarly to Experiments 1 and 2a; we held the previous donor’s contribution constant ($200) and manipulated whether the previous donor was male or female. In the identity-congruent condition, participants’ gender matched the donors’ gender in the scenario and was switched in the identity-incongruent condition. Dependent variable. After this scenario, participants made the following decision: Now, this volunteer asks how much you want to contribute to this radio station: How much would you likely contribute to this radio station (hypothetically in this scenario)? $_______. Results Manipulation checks. The scenario included two manipulation checks (“In this scenario, how much did the other member contribute to the radio station?” and “In this scenario, what is the name of the other member?”). One hundred eighty-two participants correctly responded to the manipulation-check questions and answered all questions. We included only the data from these respondents. The six manipulation-check items for self/other focus were reliable (α = .81). The participants in the self-focus condition had a significantly lower score (M = 3.69, SD = 1.29; N = 88) than those in the other-focus condition (M = 4.53, SD = 1.25; N = 94), according to a two-way t-test (t = 4.48, p < .05). Thus, the reading-comprehension passage successfully shifted participants’ attention to either self or others. Hypothesis testing. We tested H3 by conducting a hierarchical regression on the dependent variable. In Step 1, we entered gender and age into the model. In Step 2, we entered in the dummy-coded identity congruency variable (mismatch = 0, and match = 1) and the dummy-coded self-/ other-focus variable (self = 0, and other = 1) to test two independent main effects. In Step 3, we entered a two-way interaction as our direct test of H3. The main effects in Step 2 did not produce a significant change in the R-square for the omnibus test (F(2, 177) = .44, n.s.), but the two-way interaction term did (F(1, 176) = 5.78, p < .05). The omnibus analysis also showed a significant main effect of self/other focus (t = 2.26, p < .05) that was qualified by a significant identity congruency × self/ other focus two-way interaction (t = 2.41, p < .05). People’s self-reported donation intentions were higher in the self condition than in the other condition. To understand the nature of these results, we graphed the identity congruency × self/other focus interaction. In Figure 2, the identity congruency variable is on the x-axis, and the amount of the donation is on the y-axis; each slope represents this relationship at either self or other focus. This pattern supports H3. It shows that the effect of identity congruency on donations depends on the self/other focus at the time participants made their decisions. This two-way interaction is driven by two patterns of the main effect of identity congruency. More specifically, the positive slope of identity congruency is significant in the otherfocus condition (value of slope = 47.04, SE of slope = 17.84; t = 2.64, p < .05), whereas the negative slope of 356 JOURNAL OF MARKETING RESEARCH, JUNE 2008 Figure 2 EXPERIMENT 2B: IDENTITY CONGRUENCY EFFECT ON DONATIONS EMERGES WHEN PEOPLE FOCUS ON OTHERS BUT NOT SELF identity congruency is not significant in the self-focus condition (value of slope = –38.87, SE of slope = 40.00; t = .97, n.s.). As participants’ attention shifts from themselves to others, the identity congruency effect emerges. Discussion of Experiment 2b In this experiment, we find that the identity congruency effect is stronger when the target donors focus their attention on others than when they focus their attention on themselves. This result indicates that the social reality surrounding consumers may be heightened in consumers’ minds when cognitive resources are directed outward. Experiment 2b also addresses a key limitation of Experiment 2a by manipulating (instead of measuring) a short-term situational moderator in a general (instead of an undergraduate student) population. This result is consistent with previous research showing that when consumers think collectively, their behavior is more affected by norms of the group (Bontempo and Rivero 1992; Miller 1994). In addition, this study shows that when consumers direct their mental resources outward, even just one other previous donor’s behavior is enough to trigger the social identification processes and thus affect consumers’ behavioral intentions. EXPERIMENTS 3A AND 3B: THE EFFECT OF IDENTITY CONGRUENCY, IDENTITY ESTEEM, AND SELF/OTHER FOCUS ON DONATION INTENTIONS IN TWO POPULATIONS Thus far, we have investigated a dispositional and transitory moderator of the identity congruency effects on donation. At the theoretical level, identity-based processes may be triggered by both stable individual difference variables (e.g., identity esteem) and transitory informational processing factors (e.g., self/other focus). To bolster this interpretation, in Experiments 2a and 2b, we examined the independent impact of these two moderators on the identity congruency effect and found evidence consistent with the previously described theoretical arguments. We now explore synergistic interaction effects of these two moderators on identity congruency effects on donation intentions in both an undergraduate student sample (Experiment 3a) and a general population (Experiment 3b). We test the following more stringent and complex prediction: H4: The identity congruency effect is the strongest when target donors’ identity esteem is high and the target donors exhibit an other focus (a three-way interaction). Suppose that a target donor has a particular level of dispositional gender identity esteem (i.e., either high or low). If the identity of a previous donor is congruent with the target donor along the gender dimension (i.e., their genders match with each other), the esteem of the identity is high (i.e., the target donor feels good about having a certain gender identity), and the target donor’s cognitive thoughts are currently focused on others who also share that identity (i.e., on other previous donors rather than the target donor him- or herself), the behavior of the previous donor will likely affect the target donor’s decision of how much to give. Therefore, the activation potential associated with a high level of identity esteem coupled with the focus of thoughts on others who share that identity may jointly heighten the effect of identity congruency on donation intentions. In other words, neither moderator can maximize the identity congruency effect. They jointly create the most potential to influence consumers’ decision making. The experimental procedure is identical in the two experiments; only the sample populations differ. Specifically, Experiment 3a draws from an undergraduate student sample, and Experiment 3b draws from a general population sample. Method Design. The study is a 2 (identity congruency: match versus mismatch) × 2 (identity esteem: low versus high) × 2 (self/other focus: self versus others) between-subjects design with the latter two factors as measured covariates. Procedure. We developed a scenario that was similar to the previous experiments (see the Web Appendix, Part C, at http://www.marketingpower.com/jmrjune08). The only difference between these experiments and Experiment 2b is that we added a planned contribution amount of $25 for the hypothetical donor when he or she called the radio station to make a contribution, and we changed the other person’s contribution from $200 to $70. This modification enables us to adjust for the small available budget to the undergraduate student population in Experiment 3a and to increase the generalizability of findings from Experiment 2b by testing a different dollar amount in a similar adult population in Experiment 3b. We manipulated identity congruency similarly to all previous experiments; we held the amount of contribution constant ($70) and manipulated whether the previous contributor was male or female. We measured identity esteem using the same scale items as in Experiment 2a. We tested two new questions to measure (instead of manipulate, as we did in Experiment 2b) the self/other focus of the participants. We asked participants, “How Identity Congruency Effects on Donations 357 would you describe your decision to donate the amount you wrote down?” They responded on a nine-point Likert scale on the following two dimensions: from “based on thoughts about me” (1) to “based on thoughts about others” (9) and from “internal considerations” (1) to “external considerations” (9). We combined these two items to form an index (α = .82) that was a measure of the extent to which the participants were more internally focused (self focused) or externally focused (other focused) at the time they decided their own contribution. In a pretest we conducted on 176 participants, these two items were internally consistent with our adapted Aaker and Lee (2001) manipulation check questions with an alpha of .76. Dependent variable. Immediately following the scenario described previously, participants responded to the following dependent variable: Now, this volunteer asks how much you want to contribute to this radio station: How much would you likely contribute to this radio station? $_______. Results Experiment 3a Participants. One hundred ninety-seven undergraduate students in a northeastern university completed a survey voluntarily. We provided participants with identitycongruent or -incongruent information (either male or female) and assessed their gender esteem identity, focus of attention, and donation intentions. Manipulation checks. The scenario included three manipulation checks (“How much did you contribute?” “How much does the other station member contribute?” and “In this scenario, what is the name of the other member?”). One hundred eighty-four undergraduate student participants correctly responded to the manipulation check questions and answered all questions on the survey. We include only the data from those respondents here. Identity self-esteem (t = .76, n.s.) and self-/other-focus (t = .44, n.s.) scores are not influenced by the match versus mismatch identity manipulation according to two-tailed t-tests. Hypothesis testing. We tested H4 by conducting a hierarchical regression on the dependent variable. In Step 1, we entered gender and age into the model. In Step 2, we entered in the dummy-coded identity congruency variable (mismatch = 0, and match = 1), the measure of the extent to which identity esteem was high, and the measure of the extent to which participants were thinking of others (externally) as three independent main effects. In Step 3, we entered in all three possible higher-order, two-way interactions. In Step 4, we entered the three-way interaction as our direct test of H4. As in the previous experiments, we mean-centered the covariates before creating the crossproduct terms to minimize multicollinearity. The main effects in Step 2 produced a significant change in the R-square for the omnibus test (F(3, 178) = 8.23, p < .05), and so did the three-way interaction term (F(1, 174) = 6.27, p < .05). The omnibus analysis showed a significant main effect of self/other focus (t = 2.60, p < .05) that was qualified by a significant gender congruency × identity esteem × self/other focus three-way interaction (t = 2.50, p < .05). We conducted two additional follow-up analyses. Again, the dependent variable was the intended donation amount. In Step 1, we added the gender of the participants and their age to the model. In Step 2, we added identity congruency and self/other focus. In Step 3, we added the identity congruency × self/other focus two-way interaction. We conducted this regression analysis twice, at two levels of gender identity esteem split at the median. To understand the nature of these results, we graphed the identity congruency × self/other focus interaction at both levels of gender identity esteem. In Figure 3, the identity congruency variable is on the x-axis, and the amount the participants wrote in as their donation intentions is on the y-axis. Each slope represents this relationship at +1 and –1 standard deviations above and below the mean for the self-/ other-focus measure. This pattern supports H4. This three-way interaction is driven by two patterns of two-way interactions at each level of identity esteem. This analysis shows that for participants with low identity esteem, there is a main effect of self/other focus (t = 2.78, p < .05), so that for every unit increase in their other focus, there is a corresponding increase in the amount donated, regardless of the other variables. In addition, at the level of low identity esteem, the congruency effect does not exist, either when participants focus on themselves (value of slope = –.62, SE of slope = 2.70; t = .23, n.s.) or when they focus on others (value of slope = –2.07, SE of slope = 2.29; t = .90, n.s.). In contrast, among those with high identity esteem, there is a significant identity congruency × self/other focus twoway interaction (t = 2.57, p < .05). When participants have high identity esteem, congruency effects emerge only when donors focus on others (value of slope = 7.07, SE of slope = 2.60; t = 2.72, p < .05) but not when they focus on themselves (value of slope = –1.51, SE of slope = 2.15; t = –.70, n.s.). This result is consistent with H4. Results Experiment 3b Participants. One hundred fourteen people were recruited when they attended a free concert in the downtown of a Midwest city. They completed the survey voluntarily. Manipulation checks. The survey included three manipulation checks (“How much did you contribute?” “How much does the other station member contribute?” and “In this scenario, what is the name of the other member?”). Ninety-one participants correctly responded to the manipulation check questions and answered all questions on the survey. We include only the data from those respondents here. Identity self-esteem (t = 1.13, n.s.) and self-/otherfocus (t = .40, n.s.) scores are not influenced by the match versus mismatch identity manipulation according to twotailed t-tests. Hypothesis testing. We tested H4 by conducting the same four-step hierarchical regression on the dependent variable as in the previous experiment. As in Experiment 3a, the main effects in Step 2 produced a significant change in the R-square for the omnibus test (F(3, 85) = 4.92, p < .05), as did the three-way interaction term (F(1, 81) = 4.28, p < .05). The omnibus analysis showed a marginally significant main effect of self/other focus (t = 1.83, p < .1) that was 358 JOURNAL OF MARKETING RESEARCH, JUNE 2008 qualified by a significant three-way interaction (t = 2.07, p < .05). To explore the nature of this effect, we conducted the same two additional follow-up analyses as in Experiment 3a and graphed them in Figure 4. This pattern is consistent with H4, replicating our results from Experiment 3a in the general population. This three-way interaction is again driven by two patterns of the two-way interactions at each level of identity esteem. When people have low identity esteem, they give more when their thoughts focus on others than when their thoughts focus on themselves (t = 2.75, p < .05). However, for this population, the congruency effect does not emerge when people focus on themselves (value of slope = –.85, SE of slope = 2.19; t = .39, n.s.) or when they focus on others (value of slope = –3.11, SE of slope = 2.12; t = 1.47, n.s.). Among those with high gender identity esteem, however, there is a marginally significant gender congruency × self/ other focus two-way interaction (t = 1.69, p < .1). When people have high identity esteem, the congruency effect emerges only when they focus on others (value of slope = 5.09, SE of slope = 2.57; t = 1.97, p < .05) but not when they focus on themselves (value of slope = –.98, SE of slope = 2.51; t = .39, n.s.). This result replicates our findings from Experiment 3a and further supports H4. Discussion of Experiments 3a and 3b In Experiments 3a and 3b, we provide further support for our identity-based explanation of the identity congruency effect by showing a three-way interaction in the undergraduate student and general population, respectively. More specifically, we find that the identity congruency effect is the strongest (the positive slope is significant) when identity esteem is high and the focus is on others. This is an additional demonstration that both factors—stable individual difference variables and temporary information processing variables—jointly moderate the identity congruency effect. Experiments 3a and 3b are the first studies to demonstrate this three-way interaction in the identity and marketing literature. OVERALL DISCUSSION: LIMITATIONS AND FURTHER RESEARCH In this research, we study the influence of identity-based congruency along the gender dimension in the context of nonprofit marketing. Experiment 1 demonstrates the identity congruency effect on actual donations to a public radio station; an identity-congruent previous donor’s donation has a significantly higher impact than an identity-incongruent previous donor’s donation on the target donor’s donation. To explore the identity-based processes of this effect, we conducted two laboratory and two field experiments in which we identified two moderators: identity esteem and self/other focus. We found that high identity esteem (measured in Experiment 2a in the laboratory) and other focus (manipulated in Experiment 2b in the field) increased the impact of the congruency effect on donation decisions. Finally, we identified a three-way interaction among identity congruency, identity esteem, and self/other focus in the lab using undergraduate students (Experiment 3a) and in the field with an adult population (Experiment 3b). These Figure 3 EXPERIMENT 3A: IDENTITY CONGRUENCY EFFECT ON AMOUNT OF HYPOTHETICAL CONTRIBUTION EMERGES ONLY IN HIGH IDENTITY ESTEEM WITH OTHER-FOCUSED THOUGHTS (UNDERGRADUATE STUDENT SAMPLE) A: Low Identity Esteem B: High Identity Esteem Identity Congruency Effects on Donations 359 Figure 4 EXPERIMENT 3B: IDENTITY CONGRUENCY EFFECT ON AMOUNT OF HYPOTHETICAL CONTRIBUTION EMERGES ONLY IN HIGH IDENTITY ESTEEM WITH OTHER-FOCUSED THOUGHTS (ADULT SAMPLE) A: Low Identity Esteem B: High Identity Esteem latter two studies found that when people have low identity esteem, the congruency effect does not emerge, regardless of self/other focus. Conversely, when people have high identity esteem, the congruency effect is significant only when people focus on others rather than themselves. These findings supported H4 that the identity congruency effect is the strongest when gender esteem is high (an internal and dispositional identity-based trait) and attention is focused on others (a situational state switching between the self and the external social reality). As with any study, this research has limitations. Although the examination of moderators is often an equal, if not sometimes better, way to explore the psychological mechanisms (Spencer, Zanna, and Fong 2005) of the identity congruency effect, an in-depth exploration of mediators is a fruitful avenue for further research. Potential mediators include the inclusion of the other scale (Aron, Aron, and Smollan 1992) or the inclusion of the in-group in the self scale (Tropp and Wright 2001). Another way to explore the mechanisms behind identity congruency effects is to include a condition in which no identity is activated and observe how that condition compares with the congruent or incongruent conditions. In addition, the generalizability of the moderating effects we found is potentially limited by type of identity we tested, the specific instance of public radio, and the domain of nonprofit marketing we chose. Next, we discuss each of these limitations and some further research projects that they suggest. An important limitation involves the identity we tested. Gender is a predetermined stable trait identity that consumers do not choose (like liberal versus conservative identity) or marketers do not intentionally create (like music lovers versus news lovers). In particular, self-chosen identities or marketing-context-specific identities may carry stronger diagnostic cues than predetermined identities. For example, being a liberal might be more diagnostic than being a female when it comes to the decision about whether to donate to public radio. Similarly, being a music lover might be more diagnostic than being a liberal when it comes to the decision about whether to donate to a news public radio station or a contemporary music public radio station. Further research is necessary to investigate how different identities might trigger different identity-based processes to influence consumer behavior. Another limitation involves the specific type of nonprofit organizations (i.e., public radio) that we studied in the field. The influence of the identity-based process on donation behavior that we identified in public radio should be generalizable to most nonprofit organizations. However, differences may exist between different types of nonprofits. For example, donors to international humanitarian relief and development nonprofit organizations, such as CARE and Oxfam, are typically not also beneficiaries of the organizations’ work, unlike the public radio situation in which donors listen to the programming. We expect an identity type × organization type interaction effect in such situations. More specifically, we might expect to find a relatively small congruency effect of stable and general identities, such as gender and race, in all nonprofit organizations, whereas we might expect to find a relatively large congruency effect of less stable but more specific (i.e., more diagnostic) identities, such as environmental protector, public 360 JOURNAL OF MARKETING RESEARCH, JUNE 2008 radio listener, or AIDS fighter, only in a few directly “identity-diagnosable” organizations. A final limitation pertains to the nonprofit domain we chose and the generalizability of this research to for-profit marketing. We were directly interested in nonprofit marketing because it is an unexplored domain that has been increasing in importance in recent years. In 2004, approximately $250 billion was raised by nonprofit organizations (American Association of Fundraising Counsel 2005). Over the past century, nonprofit organizations have become the third-largest economic sector, after government and forprofit firms (Salamon 2002). A significant portion of individual contributions are from small donors who contribute less than $500 per year (American Association of Fundraising Counsel 2005). Despite this steady growth, many nonprofit organizations face challenges in obtaining funds. With cuts in government funding and the general weakening of the economy over the past five years, marketing and fundraising have become among the biggest challenges for these organizations. Thus, consumer researchers need to understand the nonprofit marketing environment and to generate actionable knowledge grounded in rigorous marketing research (e.g., the identity-based processes in consumption). In addition, our findings should be useful and novel for for-profit practitioners and academics. We suggest that further research is necessary to pinpoint the generalizability of these findings in for-profit marketing because how the identity-based processes work in all consumer products might differ. In particular, products that people use to signal their uniqueness are likely to be less valued when others consume them. In such situations, the identity congruency effects may be reversed. IMPLICATIONS AND CONCLUSIONS These results have important implications for both practitioners and marketing academics. For practitioners, our results suggest techniques to raise revenues without additional costs—namely, providing information about others’ contributions with identity congruency and inducing an other focus. The effect size is large; in Experiment 1, donors give $105.70 in the identity-incongruent condition but $141.88 in the identity-congruent condition—an increase of 34%. The effect also appears to be lasting. Shang and Croson (2005) find that contributions one year later from donors who received information about a previous donor’s contribution are approximately $20 higher than contributions from donors in a control condition without such information. This result suggests that identity-based congruency effects can have a lasting influence on behavior. Furthermore, our identification of the mechanisms through which the effect operates can be helpful for practitioners. For example, practitioners can use fundraising letters to remind people of their high esteem in terms of belonging to a certain social category and to focus their attention on groups of other donors who share that social identity. Then, on the pledge form, they can suggest appropriate levels of identity-congruent information to increase donation revenue. This research is also important for academics. This is the first article to demonstrate the influence of identity-based congruency on decisions in the nonprofit domain. Furthermore, we are the first to examine identity esteem and self/ other focus as joint moderators of the identity-based congruency effect. Subject to the limitations discussed previously, our results can generalize to both other nonprofit settings and for-profit marketing domains. Overall, our results offer advice for practitioners and identify new directions for academics in the area of identity-based marketing. REFERENCES Aaker, Jennifer L. and Angela Y. Lee (2001), “‘I’ Seek Pleasures and ‘We’ Avoid Pains: The Role of Self-Regulatory Goals in Information Processing and Persuasion,” Journal of Consumer Research, 28 (1), 33–49. Aiken, Leona S. and Stephen G. West (1991), Multiple Regression: Testing and Interpreting Interactions. Thousand Oaks, CA: Sage Publications. Amaldoss, Wilfred and Sanjay Jain (2005), “Pricing of Conspicuous Goods: A Competitive Analysis of Social Effects,” Journal of Marketing Research, 41 (February), 30–42. American Association of Fundraising Counsel (2005), Giving USA. Glenview, IL: American Association of Fundraising Counsel. Aron, Arthur, Elaine N. Aron, and Danny Smollan (1992), “Inclusion of Other in the Self Scale and the Structure of Interpersonal Closeness,” Journal of Personality and Social Psychology, 63 (4), 596–612. Bem, Sandra L. (1981), “Gender Schema Theory: A Cognitive Account of Sex Typing,” Psychological Review, 88 (4), 354–64. Bolton, Lisa E. and Americus Reed II (2004), “Sticky Priors and Identification-Based Judgments,” Journal of Marketing Research, 41 (November), 397–410. Bontempo, Robert and J.C. Rivero (1992), “Cultural Variation in Cognition: The Role of Self-Concept in the Attitude Behavior Link,” paper presented at the meeting of the American Academy of Management, Las Vegas (August 8–13). Burnkrant, Robert and Alain Cousineau (1975), “Informational and Normative Social Influence in Buyer Behavior,” Journal of Consumer Research, 2 (3), 206–215. Cialdini, Robert B. (2001), Influence: Science and Practice, 4th ed. New York: HarperCollins. Cohen, Joel (2005), “Seeing If,” Journal of Public Policy & Marketing, 24 (Fall), 181–82. ——— and Ellen Golden (1972), “Informational Social Influence and Product Evaluation,” Journal of Applied Psychology, 56 (1), 54–59. Deaux, Kay, Anne Reid, Kim Mizrahi, and Kathleen A. Ethier (1995), “Parameters of Social Identity,” Journal of Personality and Social Psychology, 68 (2), 280–91. Demarree, Kenneth G., Christian S. Wheeler, and Richard E. Petty (2005), “Priming a New Identity: Self-Monitoring Moderates the Effects of Nonself Primes on Self-Judgments and Behavior,” Journal of Personality and Social Psychology, 89 (5), 657–71. Eiser, Richard J. and P. Sabine (2001), “Optimism, Pessimism, and the Direction of Self-Other Comparisons,” Journal of Experimental Social Psychology, 37 (1), 77–84. Feldman, Jack M. and John G. Lynch (1988), “Self-Generated Validity and Other Effects of Measurement on Belief, Attitude, Intention, and Behavior,” Journal of Applied Psychology, 73 (3), 421–35. Forehand, Mark R. and Rohit Deshpandé (2001), “What We See Makes Us Who We Are: Priming Ethnic Self-Awareness and Advertising Response,” Journal of Marketing Research, 38 (August), 335–48. ———, ———, and Americus Reed II (2002), “Identity Salience and the Influence of Differential Activation of the Social Self- Identity Congruency Effects on Donations 361 Schema on Advertising Response,” Journal of Applied Psychology, 87 (6), 1086–1099. Frable, Deborrah E. (1989), “Sex Typing and Gender Ideology: Two Facets of the Individual’s Gender Psychology That Go Together,” Journal of Personality and Social Psychology, 56 (1), 95–108. Kelman, Herbert C. (1961), “Processes of Opinion Change,” Public Opinion Quarterly, 25 (1), 57–78. Luhtanen, Riia K. and Jennifer Crocker (1992), “A Collective Self-Esteem Scale: Self-Evaluation of One’s Social Identity,” Personality and Social Psychology Bulletin, 18 (3), 302–318. Mandel, Naomi (2003), “Shifting Selves and Decision Making: The Effects of Self-Construal Priming on Consumer RiskTaking,” Journal of Consumer Research, 30 (1), 30–40. Markus, Hazel and Paula Nurius (1986), “Possible Selves,” American Psychologist, 41 (9), 954–69. Miller, Joan G. (1994), “Cultural Diversity in the Morality of Caring: Individual Oriented Versus Duty-Based Interpersonal Moral Codes,” Cross Cultural Research, 28 (1), 3–39. Mussweiler, Thomas, Katja Ruter, and Kai Epstude (2004), “The Ups and Downs of Social Comparison: Mechanisms of Assimilation and Contrast,” Journal of Personality and Social Psychology, 87 (6), 832–44. Nelson, Michelle R., Sharon Shavitt, Angela Schunnum, and Jacqueline Barkmeier (1997), “Prediction of Long-Term Advertising Effectiveness: New Cognitive Response Approaches,” in Measuring Advertising Effectiveness: Advertising and Consumer Psychology, William D. Wells, ed. Mahwah, NJ: Lawrence Erlbaum Associates, 133–55. Reed, Americus, II (2004), “Activating the Self-Importance of Consumer Selves: Exploring Identity Salience Effects on Judgments,” Journal of Consumer Research, 31 (2), 286–95. ——— and Mark Forehand (2007), “Managing Social Identity: Strategies for Creating Brand Identification and Community,” working paper, the Wharton School, University of Pennsylvania. Reingen, Peter H. (1982), “Test of a List Procedure for Inducing Compliance with a Request to Donate Money,” Journal of Applied Psychology, 67 (1), 110–18. Salamon, Lester M. (2002), The State of Nonprofit America. Washington, DC: Brookings Institution Press. Schmitt, Bernd H., France Leclerc, and Laurette Dube-Rioux (1988), “Sex Typing and Consumer Behavior: A Test of Gender Schema Theory,” Journal of Consumer Research, 15 (1), 122–28. Shang, Jen and Rachel Croson (2005), “Field Experiments in Charitable Contribution: The Impact of Social Influence on the Voluntary Provision of Public Goods,” working paper, the Wharton School, University of Pennsylvania. Shavitt, Sharon (1989), “Operationalizing Functional Theories of Attitude,” in Attitude Structure and Function, Anthony R. Pratkanis, Steven J. Breckler, and Anthony G. Greenwald, eds. Hillsdale, NJ: Lawrence Erlbaum Associates, 331–37. ——— (1990), “The Role of Attitude Objects in Attitude Functions,” Journal of Experimental Social Psychology, 26 (2), 124–48. ——— and Michelle R. Nelson (2000), “The Social-Identity Function in Person Perception: Communicated Meanings of Product Preferences,” in Why We Evaluate: Functions of Attitudes, Gregory R. Maio and James M. Olson, eds. Mahwah, NJ: Lawrence Erlbaum Associates, 37–57. Smith, M. Brewster, Jerome S. Bruner, and Robert W. White (1956), Opinions and Personality. New York: John Wiley & Sons. Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), “Establishing a Causal Chain: Why Experiments Are Often More Effective Than Mediational Analyses in Examining Psychological Processes,” Journal of Personality and Social Psychology, 89 (6), 845–51. Stapel, Diederik A. and Willem Koomen (2000), “How Far Do We Go Beyond the Information Given? The Impact of Knowledge Activation on Interpretation and Inference,” Journal of Personality and Social Psychology, 78 (1), 19–37. Tropp, Linda R. and Stephen C. Wright (2001), “Ingroup Identification as the Inclusion of Ingroup in the Self,” Personality and Social Psychology Bulletin, 27 (5), 585–600. Visser, Penny S. and Robert R. Mirabile (2004), “Attitudes in the Social Context: The Impact of Social Network Composition on Individual-Level Attitude Strength,” Journal of Personality and Social Psychology, 87 (6), 779–95. 

week12_identity congruency effects on donation.pdf

0.28MB

https://smartstore.naver.com/smallme/products/5605946432

컬러링 스몰미빌딩 (20개 도시 세트) 미술교육. 창의력 발달. 해외여행. 창의력교육. 색칠공부, :