Berger & Heath (2008) Who drives divergence?

Berger J & Heath C (2008) Who drives divergence? Identity signaling, outgroup dissimilarity, and the abandonment of cultural tastes. Journal of Personality and Social Psychology 95:593-607.

Why do people differentiate themselves from others? Why do groups diverge?

This paper proposes an identity-signaling approach to divergence. Instead of assuming people diverge to better understand their place in the social environment or reduce their own internal uncertainty about who they are, they suggest that people diverge to ensure that others understand who they are. In particular, people often diverge to avoid sending undesired identity signals to others.

We suggest that people often abandon identity-relevant cultural tastes that are adopted by outgroup members. Further, our identity-signaling perspective predicts that people will be more likely to abandon a taste when it is adopted by dissimilar others.

Previous explanations of divergence (and why they fall flat):

• Divergence drive by low-status others
• Based on the "trickle down theory of fashion" that suggests that people adopt from those higher in status -- Fashions are initiated by higher status people and imitated by people with lower status. But once the lower strata have adopted the fashion, the signal value is lost, and high-status people abandon the taste.
• However, many examples of taste abandonment are not easily classified in terms of a unidimensional status ordering. For example, teenagers reject catchphrases once their (higher status) parents adopt them, and racial groups work hard to differentiate themselves. Further, some fashion originates among the lower classes or marginalized groups. This suggests that divergence is driven by more than status. 
• Divergence driven by disliked others
• People increase their association with groups they view positively, and thus it follows that they might be driven to decrease association with others they view negatively. Indeed, people use musical tastes to reinforce and shift boundaries between themselves and groups they dislike. If a disliked other (or group) likes a certain cultural taste, people should be more likely to dislike that taste themselves.
• However, people also often diverge from liked groups. Newly minted professors like their friends who are still graduate students, but they no longer want to dress like those friends. 
• People also sometimes adopt or accentuate negative traits as a way to achieve differentiation (e.g., Polish students consider it important that others correctly identify their national identity, and listed typical traits of Poles such as quarrelsome, vulgar, and disorderly). This, divergence is driven by more than just liking. 
• Divergence driven by similar others
• A great deal of literature, including work on uniqueness, optimal distinctiveness, and intergroup differentiation would predict that people diverge from others who are similar. These argue that people have a drive to be unique or differentiated, and thus that people differentiate themselves when they feel overly similar. 
• However, people also often diverge from groups that are dissimilar. Los Angeles residents may keep wearing a clothing style that is adopted by people from San Francisco, but they may be more likely to abandon it if adopted by residents of Des Moines. In these situations, groups diverge more strongly when the adopting groups are dissimilar. Because uniqueness theory and ODT predict greater divergence from people who seem to be overly similar to one’s own group and because work examining intergroup differentiation on behavioral measures finds greater differentiation from similar others , these perspectives are less suited to explain why people might abandon tastes that are adopted by others who are dissimilar.

An Identity-Signaling Approach to Divergence

• Cultural tastes act as signals or markers within the social communication system, communicating aspects of people (e.g., the social groups to whom they belong or other preferences they hold) to others. Attitudes and products can serve a symbolic function, and people infer things about others based on their cultural tastes -- e.g. a Birkenstock wearer supports liberal politics or that a Harley Davidson rider prefers beer to wine.
• Tastes communicate identity through their association with groups, or types of individuals, but if they are adopted by outsiders, the taste may lose its ability to signal desired characteristics effectively. If lots of tough people ride Harley motorcycles, then Harleys may come to signal that rugged identity. But if suburban accountants start riding Harleys in an attempt to seem tough, the meaning of the taste may shift. It will likely become diluted and lose its value in distinguish- ing between groups, or it may come to signal entirely different characteristics (“wannabe” tough guys). Original taste holders may then abandon the taste to avoid signaling an undesired, or unclear, identity. [MEAT OF THE THEORY]
• It is plausible that people diverge not because they are in the process of defining themselves but because they already have a clear self-definition and want to ensure it is correctly communicated to others. 
• People may abandon tastes adopted by other social groups to avoid signaling their characteristics.
• PREDICTION: Identity-signaling approach suggests that the cost of misidentification should lead people to diverge most strongly when their cultural tastes are adopted by others who are dissimilar.
• Misidentification is costly because it leads people to miss out on desired interactions and interaction partners and to get stuck interacting with undesired partners.
• For example, wearing a T-shirt emblazoned with the name of a heavy metal rock band may facilitate interactions with people who like heavy metal music. But if fashionistas start wearing such T-shirts because they look good with black jackets, the T-shirts will no longer be an effective signal. Consequently, true heavy metal fans who wear the T-shirts may be ignored by desired mates and potential friends. Heavy metal fans may stop approaching others wearing the T-shirts because they are now unsure whether the wearer cares about guitar riffs or Prada’s spring collection. 
• PREDICTION: Since dissimilar others are less likely to share identity-relevant traits, people should be more motivated to diverge when dissimilar others adopt their tastes. 
• If Stanford University students wear Reef flip-flops to signal their casual, outdoorsy approach to life, then adoption of Reefs by similar students at the University of California—Santa Cruz probably would not have much effect because Santa Cruz students share a similar casual, outdoorsy approach. But if a dissimilar group like business executives adopts the flip-flops, then that footwear loses its ability to clearly signal the original identity-relevant traits. 

Experimental Studies

They performed 7 studies that support their theory. 

Study summaries: 

1. Costs of misidentification. In surveys, Stanford students preferred not to be mistaken for being a member of groups they thought to be dissimilar to them, and these factors were correlated. They controlled for how much individuals liked/disliked those groups, and this was not a significant factor. 
2. Taste abandonment - don't think of me as one of them. Stanford students in a jocky dorm were sold a $1 Livestrong bracelet, as was another dorm across campus as a control (to test the effect of boredom). A week later, members of a dorky academic dorm were also sold the bracelet, and these students tended to interact heavily with members of the jocky dorm, in classes and the like. Another week later, 32% of the jocks but only 6% of the control dorm members had stopped wearing the bracelets. 
3. The role of public vs. private behavior. In contrast to ODT, Identity signaling does not require that private differentiation accompany public divergence. Undergrads were exposed to information that linked junk food with similar others (other undergrads), dissimilar others (grad students), or neither. Later, participants made real food choices that included health/junk food pairs (e.g., apple vs. brownie, V8 vs. Hawaiian Punch) either in private or in full view of other participants; in the latter case it was made salient that the others were forming impressions of them. When choices were made in private, there was no statistical difference in the choice of junk food between the similar and dissimilar groups. When choices were made in public, individuals who had read that dissimilar groups consume more junk food were significantly less likely to choose junk food. 
• Studies 4-7 examine a wider range of similarity:
4.   Abandonment based on outgroup similarity, status, and liking. Question to Stanford students: You and your friends have a unique catchphrase that you use. How likely would you be to increase/decrease your use if each of 15 social groups started using the phrase? Subjects said they would decrease use for 14/15 of the groups. Performed multiple regression to test how the results were effected by the other groups' similarity, liking, and status. Both dissimilarity and disliking predicted decrease in adoption to about the same degree. Status was not significant.
5. Divergence to avoid confusion with dissimilar others. Participants rated the same groups as in Study 4 on 3 measures: how they would change their taste (use of catchphrase) if the group adopted it, outgroup similarity, and their desire to avoid being confused as a member. The predicted relationships were tested using a hierarchical linear model. People were more likely to diverge from dissimilar outgroups. Outgroup dissimilarity also predicted people's desire to avoid being confused as a member. Thus, the desire to avoid being misidentified mediated the effect of outgroup dissimilarity on taste divergence. 
6. Groups, domains, and abandonment. Identity signaling predicts that people should be most likely to diverge from dissimilar others in domains such as clothing or music that people regularly use to signal identity. Students were given 1 of 3 social groups that varied in similarity, and 10 different taste domains that varied in identity-relatedness. The 3 groups were equally liked but differed in similarity to Stanford undergrads: business execs (low), inner city teens (moderate), and Princeton undergrads (high). They were asked to imagine that members of another social group had started to copy their tastes in a particular domain. They then indicated what percentage of that group would have to adopt the taste for them to abandon it. The prediction was that the similarity of the adopting group should influence divergence, but only in domains that people use to communicate identity. This prediction was well supported by the data (Fig 4). However, it is interesting that people diverged slightly more from the similar than from the moderately similar outgroup. NOTE: It seems likely that Stanford students don't expect to be mistaken for inner city youth, but could be seen as future businessmen. 
7. Outgroups, domains, and abandonment. Participants from a national Internet sample (N=163, mean age = 32 yrs) listed a group they considered an outgroup ("a social group you do not consider yourself a member of"). Participants were then asked how they would respond if that group adopted their tastes in different domains (you and a friend have some taste, not imagine this outgroup's members started adopting the same taste). Also rated similarity, liking, status, and cost of being confused with that group. Identity-signaling theory predicts (a) that people would diverge more from dissimilar groups, (b) that these effects should be more pronounced in domains people use to communicate identity, and (c) that the effects should be mediated by the cost of misidentification. These effects were all borne out in the results. Moreover, when outgroup similarity and cost of misidentification were both entered into regression predicting taste change, cost of misidentification was a significant predictor, and similarity dropped to insignificant. 

Discussion – Other Influences on Divergence:

• Although the main finding was that people diverged more from dissimilar groups, there may be cases, beyond those we have discussed in this article, in which people will diverge from similar others for signaling reasons, particularly when the likelihood of misidentification is high. Consequently, members of the two groups may be particularly prone to diverge from one another in situations where correct recognition is desired, but confusion is likely.
• Context matters. The perceived similarity of a given group can vary on the basis of context. However, regardless of how similar people perceive a particular group, the role that similarity plays in divergence should be the same: People should diverge more from groups they see as dissimilar.
• Other factors beyond similarity may influence divergence. People should abandon tastes adopted by stigmatized groups to avoid sending undesired identity signals. By the same token, identity signaling predicts that people will continue using tastes adopted by aspiration groups or other social groups they want to look like.