When I’s Meet

Abstract

Believing one shares a subjective experience with another (i.e., I-sharing) fosters connections among strangers and alters perceptions of the ingroup and outgroup. In this article, the authors ask whether I-sharing also fosters liking for members of a salient outgroup. Study 1 establishes that I-sharing promotes liking for the other sex, even among people with salient social identities. Study 2 shows that I-sharing promotes liking for a member of the sexual orientation outgroup, whether it occurs before or after group memberships get revealed. Study 3 focuses on salient race categories and looks at the effects of I-sharing versus value-sharing as a function of shared group membership. For those high in existential isolation, I-sharing trumps value-sharing, regardless of the I-sharer’s social identity. I-sharing may offer a way of improving attitudes toward outgroup members that still enables people to embrace their differing social identities.

Keywords

I-sharing self existential isolation intergroup processes interpersonal processes similarity

I can tell we have an awful lot in common,

even though we look as different as can be.

We don’t even have to try, to see things eye-to-eye;

it just comes to us, naturally.

—Tiger, from An American Tail

People have a deep-seated preference for those who look like them. The assortative mating practices studied by comparative psychologists (Thiessen, Young, & Delgado, 1997), the matching hypothesis studied by social psychologists (Feingold, 1988), and the ingroup preferences that pervade cultures worldwide (Greenberg, Landau, Kosloff, & Solomon, 2009) all point to this robust tendency. Under certain conditions, however, people can and will look beyond differences in appearance, or, more generally, beyond group differences, in their dealings with others. When members of two different groups can see themselves as members of a superordinate category (Gaertner, Dovidio, Anastasio, Bachman, & Rust, 1993), or when they learn that they hold the same values or beliefs (Insko & Robinson, 1967; Rokeach, 1973), or when they work together toward a common goal (Aronson, Blaney, Stephan, Sikes, & Snapp, 1978; Sherif, Harvey, White, Hood, & Sherif, 1961), their liking for one another increases. In short, when similarity on some other dimension gets highlighted to people who otherwise view themselves differently, they tend to like each other. Looking at the matter from the opposite perspective, people tend to shy away from outgroup members because they anticipate more differences between them than may actually exist (Mallett, Wilson, & Gilbert, 2008).

Of course, exceptions accompany every rule, and even similarity can backfire. When similarities become salient between ourselves and people on whom we frown, we wind up rejecting or denigrating the person more and not less than we would have had those similarities not gotten called to our attention (Silvia, Graham, & Hawley, 2005; Taylor & Mettee, 1971). At the intergroup level, researchers have observed a black sheep effect—more extreme negative reactions to ingroup members with negative characteristics than to outgroup members with those same negative characteristics (Eidelman & Biernat, 2003; Marques & Yzerbyt, 1988). In a related vein, belief dissimilarity among members of the same group leads to greater repulsion than does dissimilarity between oneself and the outgroup (Chen & Kenrick, 2002). Finally, attention drawn to similarities between different groups can threaten the need for intergroup distinctiveness and can result in greater expressions of ingroup favoritism (Hornsey & Hogg, 2000), particularly for those people with high ingroup identity (Crisp, Walsh, & Hewstone, 2006; Jetten, Spears, & Postmes, 2004; see also Brewer, 2007).

In short, similarity may not always lead to attraction and may not always make discernible strides toward improving liking for members of the outgroup. In this article, we ask whether a variety of similarity that has just recently received explicit attention—I-sharing (Pinel, Long, Landau, Alexander, & Pyszczynski, 2006; Pinel, Long, Landau, & Pyszczynski, 2004)—contributes to greater liking for outgroup members.

I-Sharing

The construct of I-sharing derives from James’s (1890/1918) distinction between the self-as-object (the Me) and the self-as-subject (the I). The Me consists of everything that we would say if someone asked us to describe ourselves without leaving out one single piece of information. It includes everything about the self on which we can reflect. Key aspects of the Me include the following: our personal identities, our social identities, the people in our lives, our likes and our dislikes, memorable events, how we spend our days, our beliefs and our values, how we feel about ourselves, and our goals for the future. To use a mirror as an analogy, the Me represents the image we see in a mirror.

The I, on the other hand, represents the subjective self. It refers to the experiencer, that which does the thinking, feeling, doing, looking, dreaming, and eating: in short, the experiencing. The I changes from moment to moment, as one’s experience changes, and creates what James dubbed a “stream of consciousness.” The I, then, refers to passing states of consciousness. To refer back to our mirror analogy, the I is the aspect of self that gazes at the image, reflects on it, and experiences affect in response to it.

Keeping this distinction in mind, we define I-sharing as a moment during which people believe that they and at least one other person have had the same subjective experience in response to a given stimulus. I-sharing differs from the sharing of objective features of the self, or “Me-sharing.” When two people share any aspect of the self that they have reflected upon or upon which they are currently reflecting—social identity, values, appearance, performance, likability, you name it—they Me-share. In contrast, I-sharing has nothing to do with the sharing of objective features of the self. The I lacks content—it refers to a passing state of consciousness—and so I-sharing happens when people believe they have shared a state of consciousness.

Given this distinction between Me-sharing and I-sharing, two people can I-share without sharing any objective features of the self. A Zen Buddhist, a Jew, and an atheist can all appreciate the setting sun while sitting on the same sand dune, despite worlds of differences in their core belief systems.

The Inferential Nature of I-Sharing

Mirror neurons notwithstanding, people cannot ever know for certain whether their state of consciousness overlaps with that of another person. As such, people must use cues to infer I-sharing. The most foolproof evidence of I-sharing comes when two or more people respond simultaneously and identically to the same stimulus. Such cues lead people to believe that whatever they have experienced in response to a stimulus, the I-sharer experienced as well.

People may also infer I-sharing based on the extent to which they “Me-share” with someone else. When two people Me-share, they may conclude that this objective similarity reveals something about the extent to which they might I-share with one another. Certain objective features of the self—such as one’s self-reported values, future selves, beliefs—may get treated as a window into a person’s current subjective experience or may be considered predictive of a person’s past or future subjective experience. People who share the same values pertaining to meat eating, for example, may expect that they will have the same subjective experience in response to a pig roast. People who share the same vision for their future selves may expect that they will experience the same feelings of pride or disappointment in response to a personal success or failure. Because Me-sharing could imply I-sharing, and vice versa, we have developed a methodological paradigm that disentangles these two forms of similarity (described below).

I-Sharing and Liking

Perhaps ironically, the inferential quality of I-sharing ties into its allure. We can never know whether people truly I-share with us because of our unbridgeable experiential separateness from others (Yalom, 1980). Yet I-sharing draws us to others precisely because of that separateness—that existential isolation (Yalom, 1980). Whereas existential isolation threatens people’s fundamental needs for belief validation (e.g., Festinger, 1954; Solomon, Greenberg, & Pyszczynski, 1991; Swann, 1996) and for interpersonal connectedness (Baumeister & Leary, 1995; Bowlby, 1969), I-sharing satisfies those needs. What could be more validating and unifying than sharing the exact same phenomenology as another person? (For a more detailed discussion of I-sharing, see Pinel et al., 2006.)

To test whether I-sharing—distinct from Me-sharing—promotes liking, we have developed a methodology that provides participants with both I-sharing and Me-sharing information about interaction partners for whom participants then report their level of liking. The studies presented here provide this information in a within-participants design (see also Pinel, Long, & Crimin, 2008), so that participants interact with two partners, with one of whom they Me-share (i.e., the ingroup member) and with one of whom they do not Me-share (i.e., the outgroup member). We manipulate whether participants I-share with the ingroup member or with the outgroup member, and we measure participants’ liking for both partners. Our methodological approach allows us to disentangle the effects of I-sharing from the effects of Me-sharing. If we did not provide participants with both Me and I information about their partners, we would not know whether the effects of I-sharing stem from what I-sharing might say about Me-sharing, or vice versa. Note that the critical cells for disentangling I-sharing and Me-sharing involve those instances where participants I-share with a non-Me-sharer (in this case, an outgroup member) and when they do not I-share with a Me-sharer (in this case, an ingroup member). If, in our current set of studies, people prefer the former individual to the latter, it would suggest that I-sharing—disentangled from Me-sharing—overrides shared social identity when it comes to liking.

Previous studies have already established the prepotency of I-sharing when it comes to sharing certain aspects of the Me. Pinel and colleagues (2006, Studies 1-3) have observed repeated instances of people preferring those with whom they I-share (e.g., someone who giggles at the same time as them) but do not Me-share (e.g., someone from a foreign country) over those with whom they Me-share (e.g., someone from one’s hometown) but do not I-share (e.g., someone who does not giggle at the same time as them). However, the past work did not focus on social identities that people hold dear and that commonly get investigated in intergroup research (i.e., social identities based on gender, race, and sexual orientation). Moreover, the past work utilized a scenario-based methodology that may not have approximated the true construct of I-sharing as well as do the studies reported here.

The Current Studies

As an extension of the previous work, here we ask whether people prefer an outgroup member with whom they I-share over an ingroup member with whom they do not I-share. Furthermore, we ask whether I-sharing can exert this effect in the face of salient social identities. If so, I-sharing would get distinguished from other perspectives that aim to improve liking for outgroup members. Consider both the self-categorization (Turner, 2010) and common ingroup identity (Gaertner et al., 1993) perspectives, both of which highlight the intergroup benefits that result when people recategorize themselves and members of different groups into the same social category.

Although we fully agree that allowing previously articulated group differences to dissolve can improve liking for the outgroup, we believe that the group divide gets bridged in an altogether different form when subjective similarities emerge. Objective group differences remain salient, but people like one another nonetheless, thanks to the feelings of closeness that subjective similarity promotes.

Applying the above-mentioned methodology to the current set of studies, we provided participants in Studies 1 through 3 with Me-sharing and I-sharing information about their interaction partners. Specifically, one partner shared participants’ social identity and thus was a Me-sharer on this dimension. The other partner did not share participants’ social identity and thus was not a Me-sharer.

We also manipulated I-sharing through the use of a computerized version of the game “Imaginiff” (see Pinel et al., 2008). This game—described more thoroughly in Study 1—requires participants to provide gut-level reactions to nonsensical hypothetical questions about celebrities and their likenesses to random objects (e.g., if Jennifer Aniston were a tool, what would she be?). Participants first provide their gut-level reaction to the question and they immediately learn that of their interaction partners. The I-sharing partner responds identically to our participants on 75% of the Imaginiff items; the non-I-sharing partner never gives the same response as our participants.

In addition to manipulating whether participants I-shared with an ingroup member or an outgroup member, the first two studies also examined variables suspected to affect the salience of social identity. Study 1 looks at the importance people place on their social identities; Study 2 looks at the order in which participants received social identity information about the I-sharer and non-I-sharer. If I-sharing exerts an effect even in the face of salient differences with regard to social identity, neither one of these variables should moderate the effects of I-sharing on liking.

Study 3 puts I-sharing to the most challenging of tests, by pitting it against a form of similarity documented as trumping shared social identity: values (Rokeach, 1973). We provided participants with information about the degree to which they value-share and I-share with their interaction partners. If participants prefer an I-sharer who does not share their values to someone who does not I-share with them but who does share their values, these results will serve as a testament to the potency of I-sharing. We now turn our attention to the specific studies reported here.

Study 1

Previous research documents a general tendency for women to prefer other women in favor of men (Rudman & Goodwin, 2004). In Study 1, we examined the effect of I-sharing on women’s liking for a fellow woman (and thus a Me-sharer) and a man (and thus a non-Me-sharer). We looked at this question by measuring both liking and the desire to interact with a male versus a female interaction partner.

To address the question of whether I-sharing exerts its effect despite salient objective differences, we included importance of social identity as an individual difference measure. The more important a person’s social identity, the more important it is to maintain his or her group’s positive distinctiveness (Mummendey, Klink, & Brown, 2001; Simon, Kulla, & Zobel, 1995). If I-sharing fosters liking for an outgroup member, even among those who cling to their group’s distinctiveness, it would suggest that I-sharing fosters positive attitudes despite salient objective differences in social identity.

Female participants believed they were playing Imaginiff with two partners, one male and one female. We manipulated whether the same-sex or other-sex partner I-shared with the participant. We measured participants’ liking for each partner and their choice of with whom to interact in a subsequent task. Given previous work demonstrating a female preference for females over males (Rudman & Goodwin, 2004), we could expect this preference to manifest under conditions when female participants I-shared with fellow females and did not I-share with males. Thus, we predicted that participants would prefer (and choose) the same-sex partner when she I-shared with them. Most important, we expected that I-sharing with an outgroup member would undermine and perhaps even completely counteract any preference observed when the ingroup member also was the I-sharer. In keeping with the proposition that I-sharing fosters liking for outgroup members even when ingroup–outgroup distinctions remain salient, we did not expect the importance participants placed on their social identity to moderate our results.

Method

Participants and Design

Fifty-five female introductory psychology students participated in exchange for course credit (49 self-identified as White, 2 as Black, 3 as Asian, and 1 Hispanic). Participants came to the lab individually, where we randomly assigned them to condition in a 2 (I-sharer: ingroup member, outgroup member) × 2 (Liking: for ingroup member, for outgroup member) mixed factorial design, with repeated measures on the second factor.

Procedure and Materials

When participants arrived at the lab, an experimenter told them that the study examined online communication and involved interacting with two partners over the computer. The experimenter explained that the computer would present the same questions to the participant and the partners, and they would see some of one another’s responses. In actuality, the participants saw computer-generated responses.

After the experimenter left the room, the computer provided all instructions. Participants began by choosing a screen name and learning that their ostensible partners had selected the names Jamie and Alex. From this point on, the computer instructions addressed participants by their chosen name and referred to the ostensible partners as Jamie and Alex.

Next, participants completed several demographic questions (e.g., gender, race, age). After responding to each item, participants learned the ostensible responses of their partners (and believed their partners saw how they personally had responded to the items). These responses appeared on the screen almost immediately after participants provided their own response and remained there for 7 s. Jamie and Alex appeared similar to each other in every regard (both were White, 18, psychology majors, and currently involved in a relationship) except for their gender. For all participants, one partner was male and the other was female. Because all of our participants were female, we operationally define the female partner as the ingroup member (and thus the Me-sharer) and the male partner as the outgroup member (and thus the non-Me-sharer). To reinforce the gender of the two partners, the computer asked the participants to form a mental picture of their partners before moving on.

Next, the computer introduced the I-sharing manipulation: our computerized version of Imaginiff. In this game, participants imagine a celebrity (e.g., Oprah Winfrey) as some other category (e.g., a tool) and choose which instantiation of that category characterizes the celebrity (e.g., cocktail mixer, screwdriver, sledge hammer, toenail clippers). Because participants had never considered such questions, they could not use their extant knowledge or previous thoughts (i.e., their Me) to answer them; instead, they had to rely on their in-the-moment subjective experience (i.e., their I). Thus, we reasoned that when people respond the same way to such questions, they I-share.

Specifically, the computer presented 12 trials of the Imaginiff game; each trial included a different celebrity (half male, half female), a different category, and a different set of four multiple-choice options. After each trial, participants learned their ostensible partners’ responses and believed their partners learned their responses. Both partners’ responses appeared on the screen immediately after participants had provided their own response and remained there for 7 s. One of the partners provided the exact same response on 8 of the 12 trials and thus I-shared with the participants. The other partner never provided the same response. Participants were randomly assigned to I-share with either the same-sex partner or the different-sex partner.

We should note that, in the Imaginiff game, we deliberately chose targets that represented the two social identities of interest (in this study, male and female). By doing so, we ensured that participants would get exposed to an equal number of representatives from both social identities of interest and not a predominance of one group over the other. Had participants in the current study only gotten exposed to females, for example, this could have affirmed their social identity and could have had an unintended effect on our manipulations (for related effects, see work on self-affirmation theory; Steele, 1988). In contrast, had participants only gotten exposed to males, participants in all conditions could have had an excessive need to bolster their social identity, and this could have undermined any potential effect of I-sharing.

Several other aspects of this manipulation warrant underscoring. First, the information we glean from participants in the Imaginiff game is not about them but rather about their experience of a stimulus. As such, it has nothing to do with participants’ Me. Second, participants’ responses have no meaning in the world at large and thus are significant only insofar as they reveal participants’ immediate reaction to a thought question. Third, our manipulation differs from similarity with respect to values, which have been previously shown to improve liking for outgroup members (Rokeach, 1973). Whereas values are aspects of the objective self—not to mention ones that people hold especially dear—responses to an Imaginiff question have nothing to do with the objective self and have no special significance to participants.

In short, our use of Imaginiff as a way of manipulating I-sharing differs in critical ways from prior manipulations of similarity. Whereas researchers such as Rokeach (and others like Cadinu & Rothbart, 1996; Haslam, Turner, Oakes, Reynolds, & Doosje, 2002) have included similarity manipulations that involve highly meaningful components of people’s objective selves (e.g., their values), we used a similarity manipulation that involves a trivial response to a game of no importance aside from its ability to reflect the I.

Dependent measures

After the Imaginiff game, participants learned that their answers to the remaining questions would be confidential. On a scale ranging from 0 (very low) to 9 (very high), participants reported the extent to which they felt close to each partner, could imagine becoming friends with each partner, would feel comfortable meeting each partner, would look forward to meeting each partner, and liked each partner. We averaged these five items into composites of liking for the ingroup member (Cronbach’s α = .85) and liking for the outgroup member (Cronbach’s α = .90). Participants then selected one of the two partners as the person with whom they would like to work on an upcoming face-to-face task.

After the liking and partner choice measures, participants completed a version of the four-item importance subscale of Luhtanen and Crocker’s (1992) Collective Self-Esteem Scale, modified so that the items refer to social identity with regard to gender (e.g., being a man/woman is an important part of my self-image). We combined these items to create a composite of gender identification (Cronbach’s α = .75).

Note that we administered these items after the main manipulation because previous work demonstrates that completing measures pertaining to important features of the self—including social identities—can alter the impact of a manipulation (Steele, 1988; Steele & Aronson, 1995). To ensure that we did not run into the opposite problem, that is, our manipulations affecting our measure of gender identification, we compared gender identification scores in our two between-subject conditions and found no difference, F < 1. We thus feel comfortable using these scores to test for moderation.

Finally, the experimenter probed participants for suspicion, explained the purpose of the study, and excused them.

Results

Did participants who I-shared with an outgroup member and not with an ingroup member prefer that outgroup member to the ingroup member? Did the importance participants place on their social identities moderate this effect?

To answer these questions, we first dichotomized scores on our measure of importance of social identity.¹ Next, we submitted the liking composites for the ingroup member and outgroup member to a 2 (Social identity: high in importance, low in importance) × 2 (I-sharer: ingroup member, outgroup member) × 2 (Liking: for ingroup member, for outgroup member) ANOVA, with repeated measures on the last factor. Only the interaction between I-sharer and Liking reached conventional levels of statistical significance, F(1, 51) = 123.40, p < .001, η² = .70 (see details in Table 1). When participants I-shared with the ingroup member, they liked the ingroup member significantly more than the outgroup member, F(1, 51) = 89.62, p < .001; when participants I-shared with the outgroup member, they liked the outgroup member significantly more than the ingroup member, F(1, 51) = 40.97, p < .001. No other effects emerged as statistically significant; all other ps > .11.

Table 1.

Study 1: Mean Liking Ratings for a Gender Ingroup Member and a Gender Outgroup Member as a Function of Whether the Gender Ingroup Member I-Shared or Whether the Gender Outgroup Member I-Shared

	Liking
	For ingroup member			For outgroup member
I-sharer	M	SD	n	M	SD	n
Ingroup member	6.61	1.30	30	4.01**	1.30	30
Outgroup member	4.60	0.89	25	6.52**	1.18	25

Note: Asterisks denote that the means on the same row are statistically significantly different from one another.

p < .05.

Did these results extend to participants’ choice of with whom to work in a later experimental task? We calculated the chi-square likelihood ratio of choosing the male or the female interaction partner as a function of whether the male or the female I-shared with participants. A statistically significant chi-square emerged, χ²(1, N = 55) = 61.53, p < .001.

Ninety-three percent of participants who I-shared with the ingroup member chose her as their interaction partner; 100% of participants who I-shared with the outgroup member chose him.

Study 2

Consistent with our theorizing, Study 1 revealed that participants favored the ingroup member only when they I-shared with that ingroup member and that this effect emerged regardless of the importance participants placed on their social identities. Although consistent with our theorizing, Study 1 involved two groups that have a unique relationship. Although people of different races, religions, or sexual orientations can continue to exist without ever interacting, males and females generally rely on one another to perpetuate the species (Fiske & Stevens, 1993). It therefore seems possible that I-sharing with a member of the other sex has different effects on liking for an outgroup member than I-sharing with someone of a different race, religion, or sexual orientation. For this reason, Study 2 used sexual orientation as the intergroup dimension of interest. Research and news events alike attest to the challenges that confront gay men and lesbians (D’Augelli et al., 2005; D’Augelli, Grossman, & Starks, 2006; D’Augelli, Pilkington, & Hershberger, 2002; Gettleman, 2011). At the very university where we collected the data for Study 2, gay men and lesbians reported experiencing a weekly average of 2.03 experiences with discrimination based on their sexual orientation (Swim, Johnston, & Pearson, 2009; Swim, Pearson, & Johnston, 2007). Given the prevalence of heterosexism at the very university where we collected the data for Study 2, we could only expect that a preference for heterosexuals would manifest under conditions where a fellow heterosexual also happened to I-share with participants. Given the hypothesized power of I-sharing, however, we expected any preference for heterosexuals that we might observe in the ingroup–I-sharer condition to get undermined when the outgroup member (in this case a gay man or lesbian), and not the ingroup member, happened to I-share with participants.

Also new in Study 2, we asked whether I-sharing exerts its effect even if we made group differences particularly salient. This time, we broached this issue by manipulating the point in time at which participants learned their interaction partners’ group identities. Participants learned about the sexual orientation of their partners either before the I-sharing manipulation or after the I-sharing manipulation and immediately before completing the liking measure. We reasoned that the I-sharing experience could lead people who learned their partners’ sexual orientation before undergoing the I-sharing manipulation—but not those who learned their partners’ sexual orientation after the I-sharing manipulation—to push any differences in sexual orientation to the back of their minds. As such, the sexual orientation of their partner should have been especially salient for participants in the after group. If, as we have proposed, the salience of group differences does not moderate I-sharing’s effect on liking, people should like the outgroup I-sharer equally, regardless of when they learned the I-sharer was in the sexual orientation outgroup.

Method

Participants and Design

One hundred fifteen introductory psychology students (53 males, 62 females) participated in the experiment in exchange for course credit. All participants identified themselves as heterosexual, except one. We excluded this participant’s data, leaving us with data from 114 participants. Of these 114 participants, 97 self-identified as White, 8 self-identified as Black/African American, 8 self-identified as Asian, and 1 self-identified as a Native Hawaiian/Pacific Islander.

Participants came to the lab individually, where we randomly assigned them to condition in a 2 (Participant’s sex: male, female) × 2 (I-sharer: ingroup member, outgroup member) × 2 (Order of sexual orientation information: before I-sharing, after I-sharing) × 2 (Liking: for ingroup member, for outgroup member) mixed factorial design with repeated measures on the last factor. We included participant sex as a factor because research points to possible gender differences in attitudes toward gay men and lesbians (Herek, 2002).

Procedure and Materials

We employed the same procedure as in Study 1, with four key exceptions. First, participants believed that they and their two interaction partners shared the same gender. Second, we made sexual orientation the group dimension of relevance. One ostensible partner Me-shared with participants by sharing their sexual orientation (i.e., heterosexuality); one did not. Third, the targets of the Imaginiff game consisted of well-known celebrities who varied in their highly publicized sexual orientations (e.g., Ellen DeGeneres, Elton John, David Letterman, Mick Jagger).

Finally, and most importantly, we varied the point in time at which participants learned their partners’ sexual orientation. To this end, we divided the demographic questions into two sets, and administered one set before the Imaginiff game and one set after. Participants in the before condition encountered the set containing the sexual orientation question before the Imaginiff game; participants in the after condition encountered this set after the Imaginiff game and immediately before the liking items. Thus, for participants in the after condition, their partners’ sexual orientation was made salient immediately prior to the liking items.

Results

We expected people’s preference for the ingroup member over the outgroup member to depend on whether they I-shared with the ingroup member or the outgroup member. To test this prediction, we submitted the liking composites for the ingroup member and the outgroup member to a 2 (Participant’s sex: male, female) × 2 (I-sharer: ingroup member, outgroup member) × 2 (Order of sexual orientation information: before I-sharing, after I-sharing) × 2 (Liking: for ingroup member, for outgroup member) ANOVA with repeated measures on the last factor. This analysis yielded a main effect of Liking, F(1, 106) = 10.89, p = .001, η² = .07, such that participants liked the ingroup member more than the outgroup member. This main effect was qualified, however, by the predicted I-sharer × Liking interaction, F(1, 106) = 48.82, p < .001, η² = .29. As can be seen in Table 2, when participants I-shared with the ingroup member, they preferred the ingroup member to the outgroup member, F(1, 106) = 54.66, p < .001. However, when participants I-shared with the outgroup member, they preferred the outgroup member to the ingroup member, F(1, 106) = 6.59, p = .012. Once again, the ingroup preference that emerged when the ingroup member was also the I-sharer reversed itself when participants I-shared with an outgroup member. Importantly, order did not moderate these effects, p = .826, nor did any other effect reach significance (all other ps > .17).

Table 2.

Study 2: Mean Liking Ratings for a Sexual Orientation Ingroup Member and a Sexual Orientation Outgroup Member as a Function of Whether the Sexual Orientation Ingroup Member I-Shared or Whether the Sexual Orientation Outgroup Member I-Shared

	Liking
	For ingroup member			For outgroup member
I-sharer	M	SD	n	M	SD	n
Ingroup member	5.72	1.52	59	4.00**	1.43	59
Outgroup member	4.49	1.52	59	5.11**	1.79	55

Note: Asterisks denotes that the means on the same row are statistically significantly different from one another.

p < .05.

Did I-sharing moderate participants’ preference for the ingroup member when it came to their choice of interaction partner? To answer this question, we ran a chi-square test of independence, with I-sharer (sexual orientation ingroup member or sexual orientation outgroup member) as our column variable and partner choice (sexual orientation ingroup member or sexual orientation outgroup member) as our row variable. Results revealed a statistically significant likelihood ratio, χ²(1, N = 114) = 20.72, p < .001.² When participants I-shared with the ingroup member, they selected the ingroup member 75% of the time and the outgroup member 25% of the time. When participants I-shared with the outgroup member, their preferences almost completely reversed, such that they selected the outgroup member 67% of the time and the ingroup member 33% of the time.

Taken together, the results so far indicate that I-sharing can foster liking for an outgroup member despite salient group differences. This effect appears with regard to both feelings and behaviors. After I-sharing with a gay man or a lesbian, heterosexual participants chose to work with that person rather than with someone who shared their sexual orientation and did not I-share with them. Importantly, participant sex did not qualify this, indicating that even heterosexual men (who tend to have more negative attitudes toward homosexuals than do women; Herek, 2002) prefer a gay man with whom they I-share to a heterosexual man with whom they do not. That participant sex did not moderate our results further attests to I-sharing’s ability to forge intergroup connection despite salient group differences. Given the strong homo-negativity that characterizes heterosexual men in the United States (Herek, 2002) in general, as well as in Central Pennsylvania where we collected these data (D’Augelli & Rose, 1990), it seems safe to argue that our heterosexual male participants noted and kept salient their interaction partners’ sexual orientation. Nonetheless, this did not make a difference in our findings.

Of course, the strongest evidence for I-sharing’s ability to increase liking for an outgroup member relative to an ingroup member, despite salient group differences, comes from our results pertaining to our order manipulation. Recall that the order in which participants learned of their partners’ group memberships made no difference to the results. Even when participants learned their partners’ sexual orientation immediately before completing the liking measures, they preferred the I-sharing partner, regardless of that person’s sexual orientation.

Study 3

Studies 1 and 2 established, with two different ingroup–outgroup pairings and with two different means of keeping ingroup–outgroup distinctions salient, the impact that I-sharing with an outgroup member can have on feelings of liking and choice of interaction partner. In Study 3, we pushed the edge of the I-sharing envelope even further, by asking whether I-sharing can trump value-sharing, even when it comes to liking for outgroup members. To this end, we rigged it so that an all White sample interacted with both an I-sharer and a value-sharer. For half the participants, the I-sharer was Black and the value-sharer was White. For the remaining half, it was the reverse: the I-sharer was White and the value-sharer was Black.

Also in Study 3, we examined the role that existential isolation plays in liking for an I-sharer. I-sharing theorizing posits that the allure of I-sharing stems largely from its ability to quell feelings of existential isolation. As such, it stands to reason that people with high levels of existential isolation should respond particularly favorably to an I-sharer. Previous published work attests to the role that existential isolation plays in I-sharing (see Pinel et al., 2004; Pinel, Long, & Crimin, 2010); here we look at how it plays out in relative liking for I-sharers and value-sharers as a function of shared group membership.

Finally, to make social identities especially salient, we chose a stereotypically African American name—Latisha for women and Jamal for men—to represent the outgroup member, and we chose a stereotypically European American name—Allison for women and Andrew for men—to represent the ingroup member. This name got repeated with each presentation of the partners’ responses to the Imaginiff questions. Thus, with each Imaginiff question, participants received a reminder of the race of their interaction partners. Despite this continued salience of race similarity and difference, we expected participants high in existential isolation to prefer the I-sharer to the value-sharer, regardless of race. (Given the established strength of value-sharing, we did not know whether participants low in existential isolation would follow suit.)

Method

Participants and Design

Thirty-nine European American participants received credit in their introductory psychology class for their involvement in this study. Of these participants, 23 self-identified as female and 16 self-identified as male. Participants came to the lab individually, where they interacted over the computer with two ostensible partners, one of whom identified as European American (EA partner) and one of whom identified as African American (AA partner). We randomly assigned participants to condition in a 2 (I-share/value-share condition: EA I-shares, AA value-shares; EA value-shares, AA I-shares) × 2 (Liking: for I-sharer, for value-sharer) mixed factorial design with repeated measures on the last factor.

Procedure and Materials

We employed the same procedure as in Studies 1 and 2, with a couple of exceptions. First, after providing consent, participants completed Pinel, Long, Johnson, Murdoch, and Huneke’s (2012) measure of existential isolation. This measure consists of six items (Cronbach’s α = .83), four of which get reverse-scored, that tap into the extent to which participants feel as though they are alone in their experiences. Sample items include the following: “I often have the same reactions to things that other people around me do” (reverse-scored) and “Other people usually do not understand my experiences.” Participants indicated their level of agreement with each item on a scale ranging from 1 (strongly disagree) to 7 (strongly agree), and all six items then get averaged to form an index of existential isolation.

We also incorporated a values manipulation into our design. We asked participants to provide their opinion on 10 social issues, such as gun control, gay marriage, health care, and taxes. After answering these questions, participants saw a summary of their partners’ attitudes on these topics, specifically whether they were “in favor” or “against” each issue. Participants learned that one partner (the values-sharer) agreed with them on eight of these social issues, and that the other partner (the I-sharer) never agreed with them.

Finally, we took steps to ensure that the partners’ racial categories would remain salient to participants throughout the experiment. To do so, we identified the White target as “Allison/Andrew” and the African American target as “Latisha/Jamal.” As the experiment progressed, the partners’ names appeared along with their responses to each of the value and Imaginiff questions (e.g., “Allison/Andrew said _____. Latisha/Jamal said _____.”). Moreover, as in Studies 1 and 2, we selected a mix of celebrity targets for the Imaginiff game. In this study, half the celebrity targets were White and half were Black.

Results

Does I-sharing trump value-sharing, regardless of shared group membership? Is this especially true for those most sensitive to I-sharing considerations, for people high in existential isolation? To answer this question, we conducted a median split on existential isolation scores and submitted liking composites for the I-sharer and value-sharer to a 2 (Existential isolation: low, high) × 2 (I-share/value-share condition: White I-sharer/Black value-sharer; Black I-sharer, White value-sharer) × 2 (Liking: for I-sharer, for value-sharer) ANOVA, with repeated measures on the last factor. This analysis yielded a two-way interaction between existential isolation and liking, F(1, 35) = 7.14, p = .01, η² = .17. This interaction stemmed from two near-significant, opposing mean differences. As expected, participants high in existential isolation liked the I-sharer more than the value-sharer, regardless of the I-sharer’s group membership, F(1, 35) = 3.31, p < .08. In contrast, participants low in existential isolation liked the value-sharer more than the I-sharer, F(1, 35) = 3.84, p < .06. Table 3 presents the descriptive statistics relevant to this interaction.³

Table 3.

Study 3: Mean Liking Ratings for the I-Sharer and the Value-Sharer as a Function of Existential Isolation

	Liking
	For I-sharer			For value-sharer
Existential isolation	M	SD	n	M	SD	n
Low	5.60	1.36	19	6.38*	0.97	19
High	6.00	0.87	20	5.28*	1.78	20

Note: Asterisk denotes that the means on the same row are marginally significantly different from one another.

p < .10.

General Discussion

Here we extended the research on I-sharing to the intergroup realm. In Studies 1 and 2, people who I-shared with an outgroup member came to favor the outgroup member over the ingroup member. In contrast, when participants I-shared with an ingroup member, the classic ingroup preference that has been observed time and again emerged (Dovidio & Gaertner, 2010). Across Studies 1 and 2, neither the importance of social identity nor the time at which social identity similarity got revealed moderated the effects of I-sharing.

Study 3 pushed the edge of the I-sharing envelope even further by asking whether people would prefer an I-sharer to a value-sharer. Consistent with the theoretical rationale for the power of I-sharing, among people high in existential isolation, we observed a preference for an I-sharer over a value-sharer, regardless of the race of these individuals. Importantly, the last study distinguishes I-sharing from other seemingly related forms of similarity that have proven effective in bridging the intergroup divide (Rokeach, 1973). One wonders whether the power of value-sharing stems in part from what it may imply about the likelihood of future I-sharing. This will be an interesting topic for future research.

Taken together, these findings suggest a way of improving liking for outgroup members that stands apart from previous approaches that emphasize rearranging group memberships so that people once perceived as outgroup members can get perceived in a new light, as ingroup members (Dovidio, Kawakami, & Gaertner, 2000; Oakes, Haslam, & Reynolds, 1999; Turner, 2010). Rather than shifting people’s perceptions of objective similarity in this way, I-sharing seems to allow people to feel similar on an entirely different, subjective level. Consistent with this conclusion, when we asked whether I-sharing could trump the most potent form of objective similarity—value-sharing—we found that it did so for those people who theoretically would respond most strongly to I-sharers, that is, people high in existential isolation. People high in existential isolation rarely encounter people whose experiences mirror their own. High levels of existential isolation go hand in hand with low self-esteem, negative affect, and feelings of alienation (Pinel et al., 2012). Given the theoretical importance of I-sharing for reducing feelings of existential isolation and its concomitant mental health repercussions, it makes sense that people high in existential isolation would emerge as particularly enamored of I-sharers.

Because it involves subjective, and not objective, similarity, the I-sharing approach to improving liking for outgroup members might provide a solution to what to many seems like a tug-of-war between multiculturalism on one side and color-blindness on the other (Richeson & Nussbaum, 2004). With I-sharing, people can continue to celebrate their objective differences, but find commonalities with regard to a fundamental feature of being human—commonality in states of consciousness.

Of course, all findings come with caveats, and we have three of our own. For one, we must acknowledge that we never did include a pair of control conditions to ascertain whether we would observe preference for the ingroup in the absence of any I-sharing information. Given the resource intensive nature of our paradigm, we opted to turn to robust accounts of ingroup preference as evidence for our assumption that we would indeed observe ingroup favoritism in such a pair of control conditions. For the record, however, we do not believe it would undermine the interest value of our findings even if we observed flat lines in a pair of control conditions. To us, the most striking aspects of our findings concern the fact that I-sharing with an outgroup member makes people like that person significantly more than the ingroup member, not to mention that it makes people with high levels of existential isolation like that person significantly more than the value-sharer, regardless of the I-sharer’s social identity.

We also acknowledge that we did confound the value-sharing and I-sharing manipulations with order of presentation, such that the value-sharing information always came first and the I-sharing information always came second. Given this confound, one may wonder whether the results would have looked different if we had administered the value-sharing information after the I-sharing information. Because we did not witness order effects in Study 2, and because we recently replicated the null effect of order using face-to-face contact between ingroup members and outgroup members (Pinel & Brenna, 2012), we suspect that our findings from Study 3 would hold, regardless of the order of the value-sharing and I-sharing information.

Last, we should also add to our list of caveats a limitation with regard to our dependent measure. We assessed liking for individuals only, and yet we intimate that our results have implications for intergroup attitudes more generally. We devote the next section to this point.

Seeing I to I as a Means of Reducing Prejudice

Those wanting to improve intergroup relations at a faster pace than one intergroup contact at a time may wonder whether the effects observed in the current studies translate into attitudes toward the group at large. We have both theoretical and empirical reason to believe that I-sharing with just one outgroup member can alter more general attitudes toward the outgroup. From a theoretical perspective, we suspect that I-sharing with even just one outgroup member can shatter some of the strong assumptions that underlie ingroup favoritism to begin with, namely, that members of our ingroup will I-share with us but members of our outgroup will not (see also Mallett et al., 2008). This analysis would be consistent with the few studies that do report general attitude change as a function of individualized contact with an outgroup member, such as work on intergroup friendships (e.g., Pettigrew, 1997). We suspect that I-sharing moments contribute heavily to the formation of these friendships, and thus that I-sharing can help account for the effect they have on attitudes more generally.

Consistent with the theoretical claim that I-sharing’s effects on liking for individual outgroup members may generalize to attitudes to the outgroup more broadly, consider the work of Pinel et al. (2008), which showed that I-sharing with one outgroup member impacted perceptions of the warmth and competence of the group at large. Specifically, when it came to warmth ratings for African Americans and European Americans as a group, African Americans who I-shared with a fellow African American showed ingroup favoritism, but African Americans who I-shared with a European American did not. When it came to competence ratings, African Americans who I-shared with a fellow African American showed no signs of favoritism; however, African Americans who I-shared with a European American showed signs of outgroup favoritism! To the extent that perceptions of warmth coincide with feelings of trust, these findings fit nicely with Brewer’s (2007) commentary on the difficulties associated with increasing trust across the group divide. Researchers may thus want to explore whether I-sharing with outgroup members can indeed promote intergroup trust, or if its effects remain limited to more competency-based assessments of the outgroup. Either way, the findings from Pinel et al. 2008 illustrate the potential impact that I-sharing with just one outgroup member has on perceptions of the group at large.

So what are we waiting for? Realistically speaking, researchers interested in improving intergroup attitudes via I-sharing may encounter some challenges, as creating I-sharing moments that cross group boundaries might be easier said than done. Recent work coming out of our lab, as well as research on ingroup projection (Cadinu & Rothbart, 1996; Mummendey & Wenzel, 1999; Sechrist & Stangor, 2001), suggests that people might resist I-sharing with those whom they dislike (Long & Pinel, 2011). In one study, participants indicated that they would rather listen to different music than to the same music that a disliked target was listening to, even when the target’s music was objectively preferable. Thus, although the data we present here show that if we can somehow get members of different groups to I-share, we can improve their liking for one another, Long and Pinel’s work (2012) speaks to the challenges associated with attempts to create intergroup I-sharing moments. In the same way that forecasts about intergroup encounters suffer from inaccuracy (Mallett et al., 2008), so too may this aversion to I-sharing prove ineffective at predicting the actual effect of I-sharing once it does happen. Lingering questions notwithstanding, the data presented here attest to the intergroup possibilities that result when I’s meet.

Footnotes

Acknowledgements

The authors thank Sarah Gervais as well as several anonymous reviewers for their helpful comments on an earlier draft of this article. They also thank Lindsay Allan, Raymond Alvarez, Erin Barden, Rachel Bires, Mary Beth Coggshall, Justine Figner, Ashley Grad, Michelle Lindsay, Matilda Navarro, Mallory Podolsky, Kristine Rivera, Tamara Salomon, and Michael Smith for their invaluable assistance with data collection.

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: National Institute of Mental Health Grant R01MH067823-02 funded portions of this research.

Notes

References

Aronson

Blaney

Stephan

Sikes

Snapp

(1978). The jigsaw classroom. Beverly Hills, CA: SAGE.

Baumeister

R. F.

Leary

M. R.

(1995). The need to belong: Desire for interpersonal attachment as a fundamental human motivation. Psychological Bulletin, 117, 497-529.

Bowlby

(1969). Attachment and loss (Vol. 1). New York, NY: Basic Books.

Brewer

M. B.

(2007). The importance of being we: Human nature and intergroup relations. American Psychologist, 62, 728-738.

Cadinu

M. R.

Rothbart

(1996). Self-anchoring and differentiation processes in the minimal group setting. Journal of Personality and Social Psychology, 70, 661-677.

Chen

F. F.

Kenrick

D. T.

(2002). Repulsion or attraction? Group membership and assumed attitude similarity. Journal of Personality and Social Psychology, 83, 111-125.

Crisp

R. J.

Walsh

Hewstone

(2006). Crossed categorization in common ingroup contexts. Personality and Social Psychology Bulletin, 32, 1204-1218.

D’Augelli

A. R.

Grossman

A. H.

Salter

N. P.

Vasey

J. J.

Starks

M. T.

Sinclair

K. O.

(2005). Predicting the suicide attempts of lesbian, gay, and bisexual youth. Suicide and Life-Threatening Behavior, 35, 646-660.

D’Augelli

A. R.

Grossman

A. H.

Starks

M. T.

(2006). Childhood gender atypicality, victimization, and PTSD among lesbian, gay, and bisexual youth. Journal of Interpersonal Violence, 21, 1462-1482.

10.

D’Augelli

A. R.

Pilkington

N. W.

Hershberger

S. L.

(2002). Incidence and mental health impact of sexual orientation victimization of lesbian, gay, and bisexual youths in high school. Social Psychology Quarterly, 17, 148-167.

11.

D’Augelli

A. R.

Rose

M. L.

(1990). Homophobia in a university community: Attitudes and experiences of heterosexual freshmen. Journal of College Student Development, 31, 484-491.

12.

Dovidio

J. F.

Gaertner

S. L.

(2010). Intergroup bias. In Fiske

S. T.

Gilbert

D. T.

Lindzey

(Eds.), Handbook of social psychology (5th ed., Vol. 2, pp. 1084-1121). Hoboken, NJ: John Wiley.

13.

Dovidio

J. F.

Kawakami

Gaertner

S. L.

(2000). Reducing contemporary prejudice: Combating explicit and implicit bias at the individual and intergroup level. In Oskamp

(Ed.), Reducing prejudice and discrimination (pp. 137-163). Mahwah, NJ: Lawrence Erlbaum.

14.

Eidelman

Biernat

(2003). Derogating black sheep: Individual or group protection? Journal of Experimental Social Psychology, 39, 602-609.

15.

Feingold

(1988). Matching for attractiveness in romantic partners and same-sex friends: A meta-analysis and theoretical critique. Psychological Bulletin, 104, 226-235.

16.

Festinger

(1954). A theory of social comparison processes. Human Relations, 7, 117-140.

17.

Fiske

S. T.

Stevens

L. E.

(1993). What’s so special about sex? Gender stereotyping and discrimination. In Oskamp

Costanzo

(Eds.), Gender issues in contemporary society (pp. 173-196). Newbury Park, CA: SAGE.

18.

Gaertner

S. L.

Dovidio

J. F.

Anastasio

P. A.

Bachman

B. A.

Rust

M. C.

(1993). The common ingroup identity model: Recategorization and the reduction of intergroup bias. In Stroebe

Hewstone

(Eds.), European review of social psychology (Vol. 4, pp. 1-26). London, England: Wiley.

19.

Gettleman

(2011, January 28). Ugandan who spoke up for gays is beaten to death. New York Times, p. A4.

20.

Greenberg

Landau

Kosloff

Solomon

(2009). How our dreams of death transcendence breed prejudice, stereotyping, and conflict: Terror management theory. In Nelson

T. T.

(Ed.), Handbook of prejudice, stereotyping, and discrimination (pp. 309-332). New York, NY: Psychology Press.

21.

Haslam

S. A.

Turner

J. C.

Oakes

P. J.

Reynolds

K. J.

Doosje

(2002). From personal pictures in the head to collective tools in the world: How shared stereotypes allow groups to represent and change social reality. In McGarty

Yzerbyt

Spears

(Eds.), Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 157-185). Cambridge, UK: Cambridge University Press.

22.

Herek

G. M.

(2002). Gender gaps in public opinion about lesbians and gay men. Public Opinion Quarterly, 66, 40-66.

23.

Hornsey

M. J.

Hogg

M. A.

(2000). Subgroup relations: A comparison of mutual intergroup differentiation and common ingroup identity models of prejudice reduction. Personality and Social Psychology Bulletin, 26, 242-256.

24.

Insko

C. A.

Robinson

J. E.

(1967). Belief similarity versus race as determinants of reactions to Negroes by southern White adolescents: A further test of Rokeach’s theory. Journal of Personality and Social Psychology, 7, 216-221.

25.

James

(1918). Principles of psychology. Toronto, Ontario, Canada: General Publishing Company. (Original work published 1890)

26.

Jetten

Spears

Postmes

(2004). Intergroup distinctiveness and differentiation: A meta-analytic integration. Journal of Personality and Social Psychology, 86, 862-879.

27.

Long

A. E.

Pinel

E. C.

(2012). Seeking and avoiding I contact: When subjective overlap appeals and when it repels. Manuscript in preparation.

28.

Luhtanen

Crocker

(1992). A collective self-esteem scale: Self-evaluation of one’s social identity. Personality and Social Psychology Bulletin, 18, 302-318.

29.

Mallett

R. K.

Wilson

T. D.

Gilbert

D. T.

(2008). Expect the unexpected: Failure to anticipate similarities leads to an intergroup forecasting error. Journal of Personality and Social Psychology, 94, 265-277.

30.

Marques

J. M.

Yzerbyt

V.Y.

(1988). The black sheep effect: Judgmental extremity towards ingroup members in inter- and intra-group situations. European Journal of Social Psychology, 18, 287-292.

31.

Mummendey

Klink

Brown

(2001). Nationalism and patriotism: National identification and outgroup rejection. British Journal of Social Psychology, 40, 159-171.

32.

Mummendey

Wenzel

(1999). Social discrimination and tolerance in intergroup relations: Reactions to intergroup difference. Personality and Social Psychology Review, 3, 158-174.

33.

Oakes

P. J.

Haslam

S. A.

Reynolds

K. J.

(1999). Social categorization and social context: Is stereotype change a matter of information or of meaning? In Abrams

Hogg

M. A.

(Eds.), Social identity and social cognition (pp. 55-79). Cambridge, MA: Blackwell.

34.

Pettigrew

T. F.

(1997). Generalized intergroup contact effects on prejudice. Personality and Social Psychology Bulletin, 23, 173-185.

35.

Pinel

E. C.

Brenna

R. L

III . (2012). I-sharing and face-to-face encounters with objectively similar and dissimilar others. Manuscript in preparation.

36.

Pinel

E. C.

Long

A. E.

Johnson

L. A.

Murdoch

Huneke

(2012). A prisoner of one’s own mind: Measuring individual differences in existential isolation. Manuscript in preparation.

37.

Pinel

E. C.

Long

A. E.

Crimin

L. A.

(2008). We’re warmer (they’re more competent): I-sharing and African Americans’ evaluations of the ingroup and outgroup. European Journal of Social Psychology, 38, 1184-1192.

38.

Pinel

E. C.

Long

A. E.

Crimin

L. A.

(2010). I-sharing and a classic conformity paradigm. Social Cognition, 28, 277-289.

39.

Pinel

E. C.

Long

A. E.

Landau

Alexander

Pyszczynski

(2006). Seeing I to I: A pathway to interpersonal connectedness. Journal of Personality and Social Psychology, 90, 243-257.

40.

Pinel

E. C.

Long

A. E.

Landau

Pyszczynski

(2004). I-sharing and the problem of existential isolation. In Greenberg

Koole

Pyszczynski

(Eds.), Handbook of experimental existential psychology (pp. 352-368). New York, NY: Guilford.

41.

Richeson

J. A.

Nussbaum

R. J.

(2004). The impact of multiculturalism versus color-blindness on racial bias. Journal of Experimental Social Psychology, 40, 417-423.

42.

Rokeach

(1973). The nature of human values. New York, NY: Free Press.

43.

Rudman

L. A.

Goodwin

S. A.

(2004). Gender differences in automatic ingroup bias: Why do women like women more than men like men? Journal of Personality and Social Psychology, 87, 494-509.

44.

Sechrist

G. B.

Stangor

(2001). Perceived consensus influences intergroup behavior and stereotype accessibility. Journal of Personality and Social Psychology, 80, 645-654.

45.

Sherif

Harvey

L. J.

White

B. J.

Hood

W. R.

Sherif

C. W.

(1961). The Robbers Cave experiment: Intergroup conflict and cooperation. Middletown, CT: Wesleyan University Press.

46.

Silvia

P. J.

Graham

J. S.

Hawley

C. N.

(2005). Changing attitudes toward prison reform: Effects of similarity to prisoners on attraction and rejection. Journal of Applied Social Psychology, 35, 248-258.

47.

Simon

Kulla

Zobel

(1995). On being more than just a part of the whole: Regional identity and social distinctiveness. European Journal of Social Psychology, 25, 325-340.

48.

Solomon

Greenberg

Pyszczynski

(1991). Terror management theory of self-esteem. In Snyder

C. R.

Forsyth

D. R.

(Eds.), Handbook of social and clinical psychology: The health perspective (Vol. 162, pp. 21-40). New York, NY: Pergamon Press.

49.

Steele

C. M.

(1988). The psychology of self-affirmation: Sustaining the integrity of the self. In Berkowitz

(Ed.), Advances in experimental social psychology (Vol. 21, pp. 261-302). San Diego, CA: Academic Press.

50.

Steele

C. M.

Aronson

(1995). Stereotype threat and the intellectual test performance of African-Americans. Journal of Personality and Social Psychology, 69, 797-811.

51.

Swann

W. B.

Jr. (1996). Self-traps: The elusive quest for self-esteem. New York, NY: Freeman.

52.

Swim

J. K.

Johnston

Pearson

N. B.

(2009). Daily experiences with heterosexism: Relations between heterosexist hassles and psychological well-being. Journal of Social and Clinical Psychology, 28, 597-629.

53.

Swim

J. K.

Pearson

N. B.

Johnston

(2007). Daily encounters with heterosexism: A week in the life of lesbian, gay and bisexual individuals. Journal of Homosexuality, 53, 18-31.

54.

Taylor

S. E.

Mettee

D. R.

(1971). When similarity breeds contempt. Journal of Personality and Social Psychology, 20, 75-81.

55.

Thiessen

Young

R. K.

Delgado

(1997). Social pressures for assortative mating. Personality and Individual Differences, 22, 157-164.

56.

Turner

J. C.

(2010). Social categorization and the self-concept: A social cognitive theory of group behavior. In Postmes

Branscombe

(Eds.) Rediscovering Social Identity (pp. 243-272). New York, NY, US: Psychology Press.

57.

Yalom

(1980). Existential psychotherapy. New York, NY: Basic Books.