How do consumers reconcile conflicting motives for social group identification and individual uniqueness? Four studies demonstrate that consumers simultaneously pursue assimilation and differentiation goals on different dimensions of a single choice: they assimilate to their group on one dimension (by conforming on identity-signaling attributes such as brand) while differentiating on another dimension (distinguishing themselves on uniqueness attributes such as color). Desires to communicate social identity lead consumers to conform on choice dimensions that are strongly associated with their group, particularly in identity-relevant consumer categories such as clothing. Higher needs for uniqueness lead consumers to differentiate within groups by choosing less popular options among those that are associated with their group. By examining both between- and within-group levels of comparison and using multidimensional decisions, this research provides insight into how multiple identity motives jointly influence consumer choice.
People often behave similarly to those around them—they adopt the music their friends listen to and buy the latest clothing trends to help them fit in. Indeed, conformity is one of the oldest topics in psychology and consumer research (Asch 1955; Burnkrant and Cousineau 1975; Sherif 1936), and choosing the same thing as other in-group members facilitates the communication of desired social identities (Berger and Heath 2007; Escalas and Bettman 2005). At the same time, people also want to be different. They purchase shirts with distinctive logos to set them apart from the masses or wear designer suits when they want to stand out for an important interview (Snyder and Fromkin 1980; Tian, Bearden, and Hunter 2001). How do these conflicting motives for similarity and difference combine to drive consumer choice?
Social influences on assimilation and differentiation are well documented, but they have mostly been examined in separate research streams (Hornsey and Jetten 2004). Further, research has artificially forced these motives into opposition. By studying contexts in which people are only given the option to select what someone else picked (assimilation) or something different (differentiation), prior work suggests that consumers must trade off between these two motives and that only one motive can prevail in any single choice (Mason, Conrey, and Smith 2007). Therefore, although it is well established that people often assimilate to or differentiate from the behavior of others, less is known about whether and how consumers reconcile these competing tendencies.
We propose that consumers can satisfy desires for assimilation and differentiation within a single choice context by satisfying different motives on different choice dimensions. They may select a product that allows them to communicate a desired social identity (e.g., a brand preferred by an in-group), while also differentiating within the group (e.g., a less popular product from that brand). By studying both individual and group levels of comparison and using a multidimensional dependent measure, we demonstrate that people do not simply assimilate or differentiate but often do both simultaneously.
Assimilation and Differentiation
People often assimilate to the behaviors of others (Asch 1955; Bearden, Netemeyer, and Teel 1989; Burnkrant and Cousineau 1975; Sherif 1936). Conformity may be due to informational or normative influence (Deutsch and Gerard 1955), and being similar to others supports the human need for validation (Brewer 1991; Snyder and Fromkin 1980). People also tend to behave similarly to aspiration groups (Englis and Solomon 1995) and make choices that are consistent with positive reference groups to construct or express desired identities (Berger and Heath 2007, 2008; Escalas and Bettman 2003, 2005). For example, if Harley-Davidson motorcycles are associated with tough guys, then people who want to seem tough may buy that brand. Or if hybrid cars are a signal of environmentally conscious people, then people who want to seem green may purchase a Toyota Prius.
Conversely, there are also countervailing pressures for differentiation (Maslach 1974; Snyder and Fromkin 1980; Vignoles, Chryssochoou, and Breakwell 2000). People want to be at least somewhat unique (see Lynn and Snyder 2002 for a review) and being too similar to others can generate a negative emotional reaction (Snyder and Fromkin 1980). People with higher needs for uniqueness prefer products that are more scarce or differentiated (Lynn and Harris 1997; Tian et al. 2001). And situational factors can activate people's desires to make different choices or distinguish themselves from those around them (Ariely and Levav 2000; Fishbach, Ratner, and Zhang 2011; Maimaran and Wheeler 2008). Differentiation may also be driven by the symbolic meaning of consumption; consumers often diverge from the behavior of out-group members to avoid communicating undesired identities (Berger and Heath 2007, 2008; Berger and Rand 2008; White and Dahl 2006, 2007).
But while some research has recognized motives for assimilation, and other research has recognized motives for differentiation, these motives have mostly been examined in separate research streams (Hornsey and Jetten 2004). Therefore, little empirical work has actually examined how people integrate these motives. Further, prior work has taken a one-dimensional view of similarity or differentiation using either binary choice (e.g., people must select the same product as another person or a different one) or a continuum of low to high similarity (Mason et al. 2007). For example, people are often forced to either conform and do the same thing as others, or differentiate and do something different. Because these studies require that people trade off between the two motives, they do not allow for the possibility that both can be satisfied simultaneously through a single choice.
The little work that has attempted to reconcile these two motives has focused on how these competing motives can be achieved through group-level behavior. Optimal distinctiveness theory argues that people satisfy these opposing needs through contrasting social identities, so that “the need for deindividuation is satisfied within in-groups, while the need for distinctiveness is met through intergroup comparisons” (Brewer 1991, 477). When distinctiveness is threatened, people may describe themselves as more similar to other in-group members, for example, because it heightens differences from out-group members (Pickett, Bonner, and Coleman 2002). Along similar lines, although not explicitly focused on drives for similarity and differentiation, research on divergence and the meaning of consumption has also examined how group-level comparisons can satisfy different identity motives (Berger and Heath 2007, 2008; Berger and Rand 2008; White and Dahl 2006, 2007). By converging with the choices of similar others (e.g., a jock dressing like the jocks) and diverging from the choices of dissimilar others (e.g., jocks dressing differently than punks), consumption gains symbolic meaning as a marker of group membership. Thus, according to prior work, people satisfy assimilation motives within groups and differentiation motives between them.
By focusing on assimilation within groups and differentiation between them, existing perspectives often overlook the fact that differentiation also occurs within groups. Bikers may tend to wear leather, but one biker may wear a leather jacket, whereas another may wear a leather vest. Similarly, Goths may tend to wear black, but one Goth may wear a black T-shirt, whereas another may wear a black trench coat. This suggests that intergroup comparisons alone may not be sufficient in satisfying needs for distinctiveness.
Further, because prior research has studied these motives independently, it has difficulty explaining much of actual consumer behavior. Work on uniqueness, for example, suggests that people want to be at least slightly different but says little about how that difference is enacted when faced with multiple differentiating options (Maslach 1974; Snyder and Fromkin 1980; Vignoles et al. 2000). Imagine that black Chevy cars are popular among someone's friends. If this person wanted to be unique, there are many ways he could do it. He could select the same brand but a different color (red Chevy), a different brand but the same color (black Ford), or a different brand and color altogether. Any of these choices could provide differentiation, and thus uniqueness theories alone provide little guidance on what this person would choose. Yet casual observation suggests that people do not choose among such options randomly. Groups of friends can often be seen wearing different options from the same brands, for example.
The Current Research
This article develops an integrative perspective explaining how similarity and difference combine to drive consumer choice. Real choice involves multiple product dimensions (e.g., brand and color), and we propose that these different attributes enable consumers to simultaneously satisfy desires to both assimilate and feel unique. In particular, we suggest that consumers resolve competing identity motives at different levels of a single choice—they conform to their in-group on one dimension of choice while differentiating on another.
Importantly, which particular product attributes foster assimilation versus differentiation should depend on their relevance to identity-signaling, that is, how strongly they communicate group membership. Brands often signal group identities (Escalas and Bettman 2003, 2005; Muniz and O'Guinn 2001; White and Dahl 2007). Polo and Abercrombie, for example, tend to be associated with preppy college fraternities, whereas Vans and Quiksilver tend to be associated with skateboarders. Consequently, if wearing a certain brand (e.g., Polo) is a good signal of a particular social group, then someone who wants to communicate that identity while also feeling unique may buy a shirt from that brand but select a particularly unique color (e.g., orange).
While identity-signaling motives lead people to choose in ways that are similar to or different from groups, we suggest that uniqueness motives will lead them to choose varying degrees of differentiation from members of their in-group. Consumers can thus make choices that simultaneously allow them to conform to desired reference groups on an attribute of choice that signals identity (e.g., brand), while differentiating from in-group members on a uniqueness attribute (e.g., color) to satisfy needs for uniqueness.
In situations where other choice dimensions are stronger signals of social identity, however, the effects may differ. Take fashion, where a new color is en vogue every season and multiple brands carry a variation of this trend. If purple is the color of the season, fashionistas may converge to wear that color, but those with higher needs for uniqueness may differentiate themselves on attributes that have less identity-signaling value in that context (e.g., the cut of clothing or potentially even the brand). Thus, the exact product attributes on which consumers assimilate versus differentiate from the in-group will depend on the particular context but will also be driven by which attributes are more or less signaling-relevant. In choosing this way, consumers are able to simultaneously signal their social identity and satisfy desires for uniqueness through a single consumer purchase.
Affiliation motives will drive preferences on choice dimensions associated with desired social identities. People will conform on identity-signaling attributes and choose items that strongly signal membership to an in-group.
Uniqueness motives will drive preferences at the within-group level. Higher needs for uniqueness or situations that activate uniqueness motives will lead people to differentiate themselves on uniqueness attributes and choose less popular items among in-group options.
Four studies test these hypotheses. They demonstrate that people tend to choose options preferred by in-group members on dimensions that are linked to their social identities (studies 1–4), and that this is driven by desires for other people to associate them with those groups (studies 2 and 3). Desires for uniqueness, in contrast, influence choices at the within-group level; higher needs for uniqueness (studies 1, 2, and 4) or situations that activate drives for uniqueness (study 3) lead people to make differentiating choices among group-associated options. By studying both group and individual levels of comparison and using a multidimensional dependent measure, we show how people do not simply assimilate or differentiate but simultaneously do both on different dimensions of choice.
Study 1: Everyday Clothing Choices
Do consumers' real everyday choices allow them to simultaneously communicate both their social identities and their uniqueness relative to others in their group? Study 1 provides a preliminary test of our hypotheses by examining clothing choices in a field setting. We took pictures of what people from two groups wore on a usual day and then showed them to observers to address two key questions. First, we examine whether observers can use people's clothing choices to accurately guess to which social group they belong. Second, we examine whether these same choices simultaneously express individual uniqueness, such that observers view people with higher needs for uniqueness as more differentiated in their in-group.
This study consisted of two parts: a field data collection and an online survey. Fifty-four students participated in the field portion in exchange for $5; 35 of these participants from the field portion returned to participate in the online study along with 28 new participants, for a total of 63 students who participated in the online study in exchange for $10.
Field Data Collection
The study was conducted at a private northeastern university where most juniors and seniors belong to one of 10 co-ed eating clubs. In addition to providing a venue where students eat their meals, each club's house also functions as a social gathering place for its 100–200 members. The eating clubs are geographically close (located on the same street) but often carry distinct social identities (e.g., athletic, liberal, southern, etc.).
Male and female members of three eating clubs were recruited to participate in this study. At two of the clubs (referred to as clubs A and B to preserve anonymity), students were asked to participate as they approached the club for dinner, and a photo was taken of each participant who agreed to participate (club A: 9 males, 17 females; club B: 11 males, 17 females). Participants were dressed in casual, everyday attire with no visible eating club names or logos, and there were many similarities in the clothing choices of the two groups. For example, almost all of the men wore shorts, and about half the women in each club did as well. Importantly, however, there were also some differences: many club A members dressed in athletic or preppy attire, whereas club B members favored a more hipster or alternative style.
Students from the third club (club C) were recruited as a control group for the online study.
Three days later, an online survey was sent to participants from all three clubs (63 responded: 35 of the original participants from clubs A and B and 28 new participants from control club C). First, these 63 participants (hereafter referred to as “observers”) were shown the photos and asked to indicate whether each photographed person belonged to club A or B (actual club names were used in the survey). To minimize the possibility that observers would correctly identify photographed people because they recognized people they knew, each photo was retouched to blur out both the person's face and the background (i.e., only their clothes were shown; fig. 1).
Second, observers were shown the same photos—this time grouped by club—and asked to rate how unique each photographed person's style was compared to other people in his or her club (1 = not at all unique; 7 = very unique). A mean uniqueness rating was calculated for each photo (excluding an individual's rating of his or her own photo) to be used in subsequent analyses.
Finally, the 35 returning members of clubs A and B completed the Consumer Need for Uniqueness scale (Cronbach's α = .95; Tian et al. 2001). This provided a trait measure of each individual's motive to achieve differentiation through consumer goods. Need for uniqueness scores did not differ between clubs A and B (t(33) = 1.32, p > .19).
Given our interest in how consumers simultaneously satisfy different motives, we analyzed how well people's clothing choices communicated both group affiliation and individual uniqueness.
First, results indicated that people's clothing choices successfully communicated their social identities. Each observer's responses were scored to determine what percentage of photographed people they accurately categorized into the correct club (we assumed that observers from clubs A and B accurately categorized their own photo and omitted this in the analysis). The average score was 85%, showing that observers were very good at categorizing people to their correct social groups (t(62) = 30.99, p < .001 vs. chance). While one might worry that this accuracy could be driven by members recognizing fellow club members (despite having their faces blurred), this was not the case. Even people who did not belong to either focal club (control club C) showed great accuracy (average accuracy score of 80%, t(27) = 20.32, p < .001 vs. chance).
Second, clothing choices also successfully communicated desires for uniqueness. Even though they only had access to a single clothing choice example for each photographed person, observers rated individuals with higher needs for uniqueness as having more unique styles relative to others in their in-group (r(33) = .35, p < .04).
A final test of whether these motives can be achieved simultaneously comes from examining the relationship between need for uniqueness and the accuracy of social categorization. One might argue that satisfying one motive comes at the cost of the other. For example, while people with higher needs for uniqueness might dress in ways that communicate their desire for differentiation, doing so might make them be more likely to be miscategorized into the wrong social group. But this was not the case. There was no significant relationship between accuracy of social categorization and need for uniqueness scores (r(33) = −.01, p > .96) or accuracy of social categorization and uniqueness ratings by observers (r(33) = −.15, p > .37). Thus, more unique individuals were just as likely to be recognized as members of their respective clubs as less unique individuals. This provides further evidence that satisfying one motive need not come at the expense of the other, and that real everyday choices can simultaneously communicate identity at both levels.
By using real everyday choices in a naturalistic setting, study 1 provides preliminary evidence that consumers choose in ways that can simultaneously communicate both social identity and uniqueness. Everyday clothing choices not only effectively signaled social identities, allowing observers to accurately categorize people into their respective social groups, but also simultaneously conveyed individual desires for uniqueness, allowing choosers to communicate their desires for differentiation. Further, the fact that achieving one motive did not come at the cost of the other supports the notion that these motives can act in concert.
The results of study 1 provide initial support for our theory, and the following studies use more controlled paradigms to shed light on the motives behind such choices and rule out alternative explanations.
Study 2: The Role of Identity-Signaling
Study 2 examines how various identity motives influence different levels of consumer choice. By experimentally manipulating the social group associated with different options (i.e., in-group or out-group), we simultaneously test how social identity motives and individual desires for distinction combine to drive choices.
Many aspects of consumer choice can communicate identity, but past research demonstrates a particularly strong association between social identities and brands (Escalas and Bettman 2003, 2005; Muniz and O'Guinn 2001; White and Dahl 2007). For example, research has shown that consumers form stronger connections with brands that are used by members of an in-group. Building on this association, study 2 uses brands as markers of social meaning and examines whether consumer choice on this dimension (e.g., choosing a Chevy over a BMW) is driven by desires to signal particular social identities. In particular, people should be more likely to choose a brand when it is strongly linked to an in-group (a group to which they belong) as opposed to an out-group (a group to which they do not belong).
We also conduct two ancillary tests to provide further support for our theory. First, we examine whether the tendency to choose group-associated brands is driven by how much people want to be associated with that group—the more people want to be associated with a particular group identity, the more likely they should be to select a group-associated brand. Second, we examine whether these effects are moderated by the identity-relevance of the choice domain. Certain product domains are more commonly used in the communication of identity (e.g., cars and clothes as opposed to dish soap and bike lights; Berger and Heath 2007), and if these effects are really about communicating social identity, then they should be stronger in identity-relevant domains.
Our theory also suggests that choice should simultaneously satisfy desires for differentiation. Products are distinguished not only by different brands (e.g., Chevy or BMW), but also by different options within those brands (e.g., a black or red Chevy, or a BMW 3-series or 5-series). Consequently, choosing a less popular style or color from the brand preferred by in-group members should allow participants to construct and communicate desired social identities while also allowing those with higher needs for uniqueness to differentiate themselves.
One hundred thirty-two students participated in this study on a computer as part of a larger set of experiments in exchange for financial payment.
Depending on condition, participants were first asked to identify either an in-group or out-group using instructions adapted from prior work (Escalas and Bettman 2005). Participants in the in-group [out-group] condition read: “We would like you to write in the name of a small, tightly knit social group that you [do not] belong to and [do not] feel a part of. You should feel you are [not] this type of person and that you [do not] fit in with these people. This group should be quite specific (so much smaller than say your high school class or all engineering students).” Participants identified groups such as athletic teams, student councils, and fraternities. We also measured desires for association by asking participants, “How much would you want other people to associate you with this group?” (1 = not at all; 7 = a great deal).
Next, participants made choices in 10 familiar consumer categories (e.g., cars, sunglasses, and toothpaste). In each category, participants were asked which of four self-generated options they preferred. Two of the products (A1 and A2) were from one brand (brand A) and two (B3 and B4) were from another brand (brand B). Participants were asked to imagine that they had a general idea about the preferences of people in the group they had specified, and that out of 100 group members, 60 preferred product A1, 17 preferred A2, 17 preferred B3, and 6 preferred B4. We provided one example (i.e., 60 group members might prefer a silver BMW, 17 might prefer a black BMW, 17 might prefer a silver Mercedes, and 6 might prefer a black Mercedes) and asked them to think of brands and products relevant to the group they listed when making their choices. Importantly, the preferences were deliberately distributed so that brand A was more strongly linked to the in-group than brand B (77% of the in-group preferred brand A). Moreover, they were also distributed so that there was an option to choose a popular product (A1 or B3) or a differentiating product (A2 or B4) from each brand.
Finally, participants completed the Consumer Need for Uniqueness scale (Cronbach's α = .93; Tian et al. 2001). There was no effect of the manipulation on need for uniqueness scores (t(130) = 0.21, p > .83).
Similar to study 1, we analyzed our data at different levels (in this case, brand and product choice) to test the effects of both identity-signaling and uniqueness motives on choice.
Identity-Signaling Choice: Influence of Reference Group
First, we examined choice at the group-signaling level. A mixed effects binary logistic regression (with a random intercept to control for repeated measures) revealed that people were more likely to choose the reference-group-associated brand (brand A) when the reference group was an in-group as opposed to out-group (β = 1.18, SE = 0.19, t(1,318) = 6.10, p < .001). Whereas people in the out-group condition chose an option from the group-associated brand 47% of the time, this jumped to 72% in the in-group condition.
Further, when individual participants' need for uniqueness scores and the need for uniqueness and reference group interaction were included in the model, the effect of reference group on choice remained significant (p < .005), whereas the effects of need for uniqueness and its interaction were not significant (both p > .16).
Identity-Signaling Choice: Mediation by Desires to Be Associated with Reference Group
To provide further evidence that this difference between conditions is driven by desires to signal group identity, we examined whether the effect was mediated by participants' desires to have other people associate them with the group they listed. Participants in the in-group condition reported stronger desires to be associated with the reference group listed than those in the out-group condition (Min = 5.34 vs. Mout = 1.81; β = 1.76, SE = 0.11, t(131) = 15.73, p < .001). Further, when both reference group condition and desires to be associated with the group were included in the earlier model predicting brand choice, the effect of association ratings was significant (β = 0.15, SE = 0.07, t(1,317) = 2.03, p < .05), and the effect of reference group condition was reduced (β = 0.64, SE = 0.32, t(1,317) = 1.98, p < .05). A significant mediation effect was confirmed by generating a confidence interval of the indirect effect, which did not include zero (95% confidence interval [CI] = 0.02–0.51 using the Monte Carlo method for assessing mediation; Bauer, Preacher, and Gil 2006; Selig and Preacher 2008). This underscores the notion that choices at the brand level were driven by people's desires to communicate their social identity to others. Further, when included in each step of the mediation, need for uniqueness and the need for uniqueness by reference group interaction were not significant (all p > .19), and the overall mediation pattern was unchanged (95% CI = 0.02–0.51).
Identity-Signaling Choice: Moderation by Identity Relevance of Consumer Category
Finally, to further test that identity-signaling motives were driving choice at the brand level, we examined whether choice was moderated by the identity relevance of the consumer category. A separate set of participants (N = 138) rated how effectively each of the 10 consumer categories signaled identity (1 = not at all; 5 = a great deal). Mean ratings were calculated to form a continuous identity-relevance measure for each consumer category. For example, cars and apparel were seen as more identity-relevant, while electronics and household goods were seen as less identity-relevant. We next constructed a mixed effects binary logistic regression to predict choice of the group-associated brand A. The independent variables in this model were reference group, category identity-relevance (as a continuous measure), reference group by identity-relevance interaction (to test our hypothesized moderation), and a random intercept to control for repeated measures.
A main effect of category identity-relevance (β = −0.84, SE = 0.13, t(1,316) = −6.59, p < .001) was qualified by the predicted reference group type by identity-relevance interaction (β = 0.41, SE = 0.20, t(1,316) = 2.04, p < .05). To illustrate this interaction, we dichotomized the continuous identity-relevance variable using a median split and conducted separate mixed effects binary logistic regressions for low and high identity-relevant categories; choice of brand A was modeled using reference group as the independent variable and a random intercept to control for repeated measures. For low identity-relevance categories, the odds of in-group participants choosing the group-associated brand A were 2.54 times that of out-group participants (β = 0.93, SE = 0.26, t(658) = 3.54, p < .001). However, the difference in odds nearly doubled to 4.75 when participants were choosing in high identity-relevant categories (β = −1.56, SE = 0.26, t(658) = 6.00, p < .001). Therefore, the tendency for people to choose an in-group associated brand and avoid an out-group associated brand was stronger in consumer categories where choice is more likely to be seen as a signal of identity (see fig. 2 for results displayed using median splits of identity-relevance).
Differentiating Choice: Influence of Need for Uniqueness
We also examined choice at the product level. Because we are interested in studying how people simultaneously differentiate within their group, we examined the product choices made by participants conditional on having chosen an option from the brand strongly linked to the reference group (i.e., between products A1 and A2 from brand A). Importantly, if our conceptualization is correct, then the identity of the reference group linked to the brand should moderate the effect. Need for uniqueness should have a stronger influence on choice of products from the reference-group-associated brand when that group is an in-group (as opposed to out-group); for participants referencing an in-group, higher needs for uniqueness should be positively associated with choice of the differentiating product. To test this, we conducted a mixed effects binary logistic regression with reference group type, need for uniqueness score, and group type by need for uniqueness interaction predicting choice of product A1 versus A2 (a random intercept controlled for repeated measures).
Consistent with our theorizing, a main effect of group type (β = −2.90, SE = 0.94, t(773) = −3.08, p < .003) was qualified by a significant group type by need for uniqueness interaction (β = 0.97, SE = 0.38, t(773) = 2.54, p < .02). Specifically, among people in the in-group condition, those with higher needs for uniqueness were more likely to choose the less popular product A2 (β = 0.73, SE = 0.26, t(440) = 2.85, p < .006). There was no corresponding relationship between need for uniqueness and product choice among those who referenced an out-group (p > .44). Further supporting our hypotheses, the effect of need for uniqueness on choices among in-group associated options was not mediated by desires to be associated with the group, as the confidence interval of the indirect effect crossed zero (95% CI = −0.17 to 0.09 using the Monte Carlo method for assessing mediation).
These results demonstrate that motives for uniqueness influence choice at a within-group level. Among people referencing an in-group who had chosen a brand A (group-associated) option, those with higher needs for uniqueness were more likely to choose the less popular product A2 (preferred by fewer in-group members) than those with lower needs for uniqueness. This was not the case among people who referenced an out-group, however, as they should not feel a need to differentiate within a group to which they do not belong.
Results of study 2 provide further support for our hypotheses about how various identity motives combine to drive consumer choice. In this case, brands were signals of identity, and choice at the brand level was driven by desires to signal social identity. People were more likely to choose reference-group–associated options (brand A) when that group was an in-group (vs. out-group), and this was mediated by desires to be associated with the reference group. Further, these effects were stronger in identity-relevant domains, consistent with the notion that choice was driven by desires to communicate identity.
Needs for uniqueness did not influence choice at the brand level, but at the product level they influenced choice among the products from the in-group-associated brand. Specifically, among those referencing an in-group, people with higher needs for uniqueness were more likely to choose the product preferred by fewer group members.
One might wonder whether within-group differentiation occurred only because between-group differentiation was not sufficiently salient. The choice task used in this study provided only in-group preferences for consideration, which may not have adequately highlighted between-group differences or may have been inferred as a general majority preference. Prior work on optimal distinctiveness suggests that salient out-group comparisons should satisfy psychological needs for differentiation (Brewer 1991). When between-group contrasts are heightened, people's desires for uniqueness could be satisfied by the fact that their in-group is different from an out-group, and this may remove any effects of needs for differentiation from other in-group members through choice. Ancillary data, however, suggest that this is not the case.
In a follow-up study, participants (N = 33) identified both an in-group and an out-group and completed a choice task similar to study 2 across six identity-relevant consumer categories. In this case, however, they were told to imagine that 70 people from their in-group preferred product A1 and 30 people preferred A2, while 70 people from their out-group preferred B3 and 30 people preferred B4. By providing options associated with both an in-group and an out-group, we intended to heighten the salience of between-group comparisons, thus providing an opportunity to differentiate by contrasting against an out-group. Participants also completed the Consumer Need for Uniqueness scale (Cronbach's α = .91; Tian et al. 2001). Results corroborate the findings of study 2. Brand A options were chosen 88% of the time, and need for uniqueness was not associated with brand choice (β = −0.38, SE = 0.84, t(196) = −0.45, p > .65). Furthermore, need for uniqueness significantly predicted product choice within the brand linked to the in-group (β = 0.88, SE = 0.33, t(172) = 2.68, p < .009); those higher in need for uniqueness were more likely to choose the product preferred by fewer in-group members. These results suggest that needs for uniqueness still exert an influence on choice, even when psychological contrasts to out-groups can be made. They also show that while people with higher needs for uniqueness may be willing to select options that are less linked to their own group, they are unlikely to select options linked with other groups; rather, they tend to differentiate within the options associated with their in-group.
Study 3: Manipulating Drives for Distinction
To provide further evidence that uniqueness motives are underlying choice at the within-group level, study 3 manipulates rather than measures them. We expose some participants to images that prime uniqueness (Maimaran and Wheeler 2008) and use a similar choice task to study 2, in which brands were strong markers of social meaning.
Consistent with study 2, we predict that identity-signaling motives should again lead people to select options from the brand linked to their in-group (vs. out-group), and this should be driven by how much they wish others to associate them with the group. However, the priming manipulation should affect which product they select from that brand: those primed with uniqueness should be more likely to select the differentiating product from the in-group associated brand.
One hundred seventy students participated in this study on a computer as part of a larger set of experiments in exchange for financial payment. They were randomly assigned to a condition in a 2 (prime: uniqueness vs. control) by 2 (group type: in-group vs. out-group) between subjects design.
First, following study 2, participants specified either an in-group or out-group and rated how much they wanted to be associated with that group.
Next, we primed half the participants with uniqueness (adapted from Maimaran and Wheeler 2008). These individuals were asked to look at eight pictures and identify the number of circles and squares in each image. Each picture contained an array of shapes in which all the shapes were the same except one (e.g., ◻○◻◻◻◻◻◻◻◻). Exposure to such figures has been shown to increase uniqueness seeking behavior by making uniqueness motives more accessible (Maimaran and Wheeler 2008). Control participants did not complete the priming task.
Finally, participants were presented with the choice task from study 2. To simplify the design, they were asked only to make choices from six identity-relevant consumer categories (e.g., cars, shirts, sunglasses, etc.). Choices were analyzed using an approach similar to study 2.
Identity-Signaling Choice: Influence of Reference Group
Consistent with study 2, referencing an in-group (vs. an out-group) increased the odds of choosing an option from the group-associated brand (brand A). A mixed effects binary logistic regression with reference group type, prime, and their interaction (with a random intercept to control for repeated measures) predicting brand choice showed only a main effect of group type: people chose the group-associated brand (brand A) only 35% of the time when it was preferred by an out-group, but this nearly doubled to 62% of the time when it was preferred by an in-group (β = 1.38, SE =0.39, t(1,016) = 3.55, p < .001). Neither the uniqueness prime, nor its interaction, affected brand choice (both p > .45).
Identity-Signaling Choice: Mediation by Desires to Be Associated with Reference Group
As in study 2, results again demonstrated the mediating effect of desires to be associated with the reference group on brand choice. The confidence interval of the indirect effect did not include zero (95% CI = 0.13–1.10 using the Monte Carlo method for assessing mediation). The uniqueness prime and the prime by reference group interaction were not significant when included in each step of the mediation (all p > .46) and the overall mediation pattern remained significant (95% CI = 0.13–1.11).
Differentiating Choice: Influence of Uniqueness Prime
Next, we examined how the uniqueness prime influenced product choices made by participants, conditional on having chosen an option from the brand strongly linked to the reference group (i.e., brand A). We conducted a mixed effects binary logistic regression with reference group type, uniqueness prime, and their interaction predicting choice of product A1 versus A2 (a random intercept controlled for repeated measures).
The pattern of results was consistent with study 2. An effect of group type (β = 2.13, SE = 0.48, t(486) = 4.34, p < .001) was qualified by the predicted uniqueness prime by reference group interaction (β = −1.33, SE = 0.63, t(486) = −2.12, p < .04; fig. 3). Among people who referenced an in-group, the uniqueness prime increased the choice of the less popular product A2 (β = −0.81, SE = 0.35, t(305) = −2.30, p < .03). There was no corresponding effect of prime in the out-group condition (β = 0.54, SE = 0.57, t(181) = 0.95, p > .34).
By manipulating drives for distinction rather than measuring them, the results of study 3 underscore the findings of study 2. People chose in ways that simultaneously allowed them to communicate both social identity and uniqueness. In this case, brands signaled identity and desires to be associated with particular social identities again drove assimilation at the brand level. At the same time, activating drives for differentiation, this time through a situational prime, drove differentiation among in-group linked options.
Study 4: Manipulating Dimensions for Differentiation
Studies 2 and 3 used brands as a signal of social identity and products as a means of differentiation, but as we noted in the introduction, this may not always be the case. In any given season, certain product styles or colors are in fashion, and multiple brands may carry their own version of this trend. In such instances, product choice may signal social identity (e.g., sneakers vs. dress shoes), and the brand one chooses may provide in-group differentiation (e.g., Keds vs. Converse sneakers).
Study 4 tests this possibility by manipulating which dimension of choice—product or brand—is seen as a means to assimilate to or differentiate within one's in-group. If our theorizing is correct, people with higher needs for uniqueness should still choose to differentiate themselves within their in-group, but a priming task should shift which dimension they use (product or brand). Priming brands as a signal of social identity and products as a means for in-group differentiation should lead people with higher needs for uniqueness to prefer the less popular product from the group-associated brand. In contrast, priming people to think of product type as a signal of identity and brands as a means for differentiation should lead them to prefer to differentiate themselves by choosing the group-associated product but from a less popular brand.
One hundred sixty-three students participated in this study on a computer as part of a larger set of experiments in exchange for financial payment. They were randomly assigned to either the product differentiation or brand differentiation prime condition in a two-factor between subjects design.
First, participants specified an in-group using the same instructions as previous studies (there was no out-group condition in this study).
Second, they were presented with a sorting task designed to highlight either brands or product types as a point of differentiation within one's in-group. All participants were asked to “Consider Mike, a member of an on-campus group Gamma.” Participants primed to think of products as a uniqueness attribute were then told that Mike uses the same brand as other Gammas but a different product, while participants primed to think of brands as a uniqueness attribute were told that Mike uses the same product as other Gammas but a different brand.
Specifically, participants in the product differentiation condition were told that “Gamma members typically drive BMWs. Most Gammas drive BMW sports cars. Mike also drives a BMW, but he drives a BMW SUV.” Thus, participants in this condition were primed to think of products as providing within-group differentiation. After reading these instructions, participants were given a photo sorting task that involved separating different options from the same brand. They were presented with 10 photos of automobiles: 5 BMW sports cars and 5 BMW SUVs; for each photo, participants indicated whether the automobile would be preferred by Mike or other members of Gamma.
In contrast, participants in the brand differentiation condition were told that “Gamma members typically drive sports cars. Most Gammas drive BMW sports cars. Mike also drives a sports car, but he drives a Lexus sports car.” Thus, participants in this condition were primed to think of brands as a uniqueness attribute. They then completed a similar photo sorting task, but in this condition they separated different brands that made the same type of car. They were shown photos of 5 BMW sports cars and 5 Lexus sports cars and asked to indicate whether the automobile would be preferred by Mike or other members of Gamma.
Thus, the key difference between the two conditions was which dimension—brands or product types—was a uniqueness attribute that would provide within-group differentiation.
Third, participants were given a choice task similar to the one used in study 3—this time, choosing among three options. They were asked to imagine that out of 100 people in their reference group, 60 preferred product A1, 20 preferred product A2 (a different product type from the same brand), and 20 preferred product B1 (the same product type from a different brand). Note that products A2 and B1 were equally less popular (both preferred by 20% of people) which should appeal to those higher in needs for uniqueness. However, we hypothesized that the sorting task would influence preference between the two options that provided some differentiation (product A2 vs. B1)—thinking of brands as a signal of identity and products as a means of differentiation would increase preference for product A2, whereas thinking of products as a signal of identity and brands as a means of differentiation would increase preference for product B1.
Finally, participants completed the Consumer Need for Uniqueness scale (Cronbach's α = .94; Tian et al. 2001). There was no effect of prime on need for uniqueness scores (t(161) = 0.11, p > .91).
The data were analyzed using a mixed effects multinomial logistic regression with condition, need for uniqueness, and the two-way interaction as predictors (a random intercept controlled for repeated measures). The overall model revealed a significant effect of the prime (F(2, 970) = 3.92, p < .03), need for uniqueness (F(2, 970) = 8.73, p < .001, and a marginally significant interaction (F(2, 970) = 2.42, p = .09).
As predicted, the prime significantly influenced the choice proportions of the two potentially differentiating options (products A2 and B1; β = −2.35, SE = 1.06, t(970) = −2.22, p < .03; fig. 4). More specifically, the differentiating product from brand A (A2) was chosen more often when people were primed to think of brands as a signal of social identity and products as a uniqueness attribute (30%) than when they were primed to think of products as a signal of social identity and brands as a uniqueness attribute (18%). Conversely, the differentiating brand for product 1 (B1) was chosen more often when people were primed to think of products as a signal of social identity and brands as a uniqueness attribute (36%) than when they were primed to think of brands as a signal of social identity and products as a uniqueness attribute (26%). A mixed effects binary logistic regression with condition predicting choice of product B1 (vs. A1 and A2) showed a significant effect of prime condition (β = −0.59, SE = 0.25, t(976) = −2.35, p < .02). When need for uniqueness and the prime by need for uniqueness interaction were included in the model, the effect of the prime remained significant (p < .04), the effect of uniqueness was significant (p < .02), and the interaction was not significant (p > .11).
Second, we again found that desires for uniqueness drove choice of less popular options. Among those primed to think of products as a uniqueness attribute, higher needs for uniqueness increased the odds of choosing product A2 over A1 (β = 1.00, SE = 0.28, t(476) = 3.62, p < .001). Similarly, among those primed to think of brands as a uniqueness attribute, higher needs for uniqueness increased the odds of choosing product B1 over A1 (β = 0.79, SE = 0.31, t(494) = 2.54, p < .02).
Study 4 again illustrates that desires for differentiation lead people to choose less popular options relative to others in their in-group. However, consistent with our theoretical position about the meaning of choice dimensions, the way they chose was moderated by manipulating which dimension of choice—product or brand—people viewed as relevant to signaling and uniqueness. When primed to think of products as a uniqueness attribute and brands as the group signal, people with higher needs for uniqueness were more likely to choose less popular product options from the group-associated brand (product A2). The reverse was found when people were primed to think of brands as a uniqueness attribute and products as a group signal—people higher in needs for uniqueness were more likely to choose the less popular brand option of the group-associated product (product B1). Therefore, study 4 provides evidence that situational cues or consumption meaning can alter which choice dimensions are better signals of social identity or uniqueness. And consistent with the prior studies, people tended to conform on dimensions they perceived to be a signal of group identity, and differentiated among group-associated options to satisfy desires for uniqueness.
This article integrates research on assimilation, differentiation, and the meaning of consumption to illustrate how people can simultaneously reconcile identity-signaling and uniqueness motives. Previous research has typically studied these motives in isolation or from a one-dimensional perspective. We combine these various research streams and examine different dimensions of choice to gain deeper insight into identity processes, as well as how these processes combine to drive consumer choice.
Four studies demonstrate that by using different choice dimensions, people are able to simultaneously satisfy motives for both identity-signaling and uniqueness within a single choice. As shown in study 1, people's everyday clothing choices allow them to simultaneously be recognized as a member of their social group and express their individual desires for uniqueness relative to other in-group members. People tend to assimilate with in-group choices on dimensions that strongly signal their social identities (studies 2–4). Moreover, this increased choice is mediated by desires to be associated with their group (studies 2 and 3) and moderated by the identity relevance of the consumer category (study 2). At the same time, desires for differentiation tend to play out at a within-group level of choice. Individuals with higher needs for uniqueness (study 2) or primed with uniqueness (study 3) are more likely to choose a less popular product option from the brand linked to their in-group. Finally, situational cues and the meanings attached to consumption choices can alter the dimensions on which people choose to assimilate and differentiate (study 4).
This research highlights the value of a more multidimensional view of consumer choice and contributes to the literature in several ways. First, while prior perspectives have suggested that people may assimilate or differentiate from others, they have often focused on either the group or individual level, but not both. Further, they have tended to look at only one dimension of choice (e.g., choosing the same brand or a different brand) or use a single continuous dependent measure (e.g., asking people to rate their perceived similarity to other members of a group). Real choice, however, is much more complex, and explicitly allowing for this complexity provides a richer understanding of the nuances that drive consumer behavior. By studying both group and individual levels of comparison and using a multidimensional dependent measure, we are able to show that people do not simply assimilate or differentiate but can simultaneously do both on different dimensions.
Second, our perspective provides insight into which specific choice dimensions may be used for assimilation versus differentiation. Beyond reflecting general motives to be similar or different, certain choice dimensions may acquire symbolic meaning as markers of group identity, and these meanings may then come to shape choice. Brands are often seen as signals of social identity. Consequently, people often converge to their in-group preferences on this signaling attribute while differentiating themselves on a uniqueness attribute (e.g., color). However, when other attributes are more relevant to communicating group identity (e.g., wearing a certain color), then these effects may reverse, with people converging on color and using other attributes to differentiate themselves (as in study 4).
Third, the results suggest that uniqueness motives mainly drive choice within groups, rather than between them. While more empirical work is certainly necessary to examine this issue in greater detail, it seems that higher needs for uniqueness drive people to select more differentiated options within their in-group rather than leading them to select options outside their group. Thus, future work might test how between-group differentiation may be conceptually and practically different than within-group differentiation (also see Hornsey and Jetten 2004). Research might also examine whether and when one motive may supersede the other, either in terms of the degree of influence on choice or the sequence in which the two motives are considered in the decision-making process. While our work shows that both motives can be satisfied simultaneously through a single consumer choice, the order in which each motive is considered and the dimensions of choice evaluated may or may not differ across individuals and situations.
Fourth, the theoretical implications of this research extend beyond the consumer choice literature to inform the social psychology of identity more broadly. Theories of conformity, social identity, and uniqueness have a long and rich history in psychology, and this article contributes to understanding how these related literatures can be woven together. Our research provides insight into decision-making and behavior when there are tensions between motives of assimilation and differentiation, even in situations that may not involve consumption. For example, an employee may desire to be both an integrated team member and have a unique role in the organization. Similarly, elected politicians and their loyal constituents may wish to both toe the party line and voice their individual opinions. In such situations, we may observe expressions and behaviors that broadly communicate affiliation with one's group (e.g., advocating support for a piece of legislation) while also asserting individuality (e.g., emphasizing the importance of a unique component of the legislation).
Finally, our findings shed light on how consumers may navigate complex choice environments in which multiple internal or external drivers may influence a single choice. Laboratory research is often criticized for the parsimony of its experimental designs. While such approaches are valuable in isolating, understanding, and convincingly demonstrating a specific effect, these insights usually come at the expense of real-world relevance. Although an effect may be observed in the lab, it can be difficult to abstract implications to complex or noisy situations in which multiple forces are at play (Staw 2010). In this article, we have demonstrated one way people can integrate and simultaneously satisfy multiple motives in a single choice—by satisfying each motive on a different dimension. Our results are even more compelling in this regard because the two motives we studied are not only different but are in opposition. While we do not claim that our laboratory studies fully replicate everyday life, we have captured at least one additional level of complexity through our multidimensional dependent variable. Moreover, we have provided converging evidence by observing everyday choices in a natural setting (study 1).
Directions for Future Research and Marketing Implications
As with most research, there are a number of intriguing directions for future study. One is examining how these motivations for assimilation and differentiation extend cross-culturally. While existing research has found that European Americans prefer uniqueness more than East Asians (Aaker and Schmitt 2001; Kim and Markus 1999), this finding says little about how such differentiation is actually enacted. One could achieve greater differentiation by joining smaller groups, distinguishing oneself from other in-group members, or differentiating one's group more from out-groups. Furthermore, research suggests that choice may not be as strongly linked to identity in all cultural contexts (Kim and Drolet 2003; Savani, Markus, and Conner 2008; Stephens, Markus, and Townsend 2007). Examining the degree to which these motivations exist in various cultural contexts, as well as how they combine to drive choice, may provide insight not only into differentiation itself but also the communication of identity across cultures and the integration of multiple motives more broadly.
These findings also have important marketing implications. Creating multiple product options may not only generate better fit with consumer preferences (Lancaster 1990), but also allow consumers to differentiate themselves. Even though different colored iPods are functionally identical, for example, the proliferation of colors allows consumers to see themselves as differentiated, even though they are making essentially the same choice (Pronin, Berger, and Molouki 2007). Other brands offer seemingly endless ways for consumers to differentiate themselves; for example, programs such as NIKEiD and Trek Project One let consumers customize the materials, styles, and colors of their shoes and bikes, resulting in a unique yet branded product. Future research may help to determine if particular attributes can better communicate social identity or more effectively satisfy uniqueness motives.
In summary, this research illustrates one way that people integrate different identity motives through consumer choice. Opposing desires to signal social identity and uniqueness can be resolved by making strategic choices on different choice dimensions: consumers may conform on dimensions that are associated with their in-group and simultaneously differentiate by making a more distinct choice among group-associated options. Our findings also illuminate the complexity of how people balance different motives when making choices, and the benefits of acknowledging such complexity when designing choice stimuli. Finally, while research in identity-signaling has typically focused on contrasting in-groups and out-groups, we direct our attention to the individuals who form these groups to demonstrate how group and individual influences combine to drive consumer choice.