Entitativity and Group Perception

Prejudice, stereotypes and discrimination refer to attitudes, beliefs and behavior towards other people based on their association with a particular social group. In studying these aspects of social behavior, the focus is most commonly on race, ethnicity, religion or gender. This is not surprising given their prominence in historical conflicts and social movements. They also are core aspects of social identity for most people. Yet social groups of this type only represent some of the many groups in which we participate, to which we belong, or with which we identify. There are many types of groups in our social environment. But what exactly is meant by a social group and what determines whether we view a set of individuals as constituting a group or simply a mass of distinct individuals?

Lickel et al. (2000) approached these questions through a set of studies in the United States and Poland. Participants were asked to rate 40 different types of groups using descriptions such as "members of a family," "members of a professional sports team," "Women," "students at a university," and "people in line at the bank." Each collection of people was rated on a variety of measures, including: (1) importance of the group to its members; (2) the amount of interaction among it members; (3) whether members of the group shared common goals; (4) whether they shared common outcomes; (5) the similarity among group members; (6) its typical duration; (7) the degree to which members can enter or leave the group voluntarily, that is, permeability; and (8) group size. Also assessed was the extent to which the group description was viewed as actually qualifying as a group. Ratings were obtained on a continuous nine-point scale ranging from "not a group at all" to "very much a group." It was thought that this measure would quantify perceptions of a group's entitativity--a concept first introduced by Donald Campbell (1958)--which Lickel et al. define as the "degree to which a collection of persons are perceived as being bonded together into a coherent unit" (pg. 224).

A number of different analyses were conducted. One focus was on which of the measures best predicted the degree of entitativity of the groups. The strongest predictor was found to be the degree of interaction among group members, but group importance, common goals, common outcomes, and member similarity were also directly related to entitativity ratings. In contrast, increased group size and permeability of group boundaries were more modestly, and inversely, related to entitativity. These findings suggest that aggregates of people are more likely to be viewed as a group as they become more coordinated and focused on common goals and outcomes. These attributes often characterize groups with explicit rules or prevalent implicit norms and with specific roles for coordinating the actions of its members. Hierarchies with status and power differentials also tend to characterize highly entitative groups. Hamilton, Sherman and Lickel (1998) argue that such organization and structure is a key factor contributing to the perception of a group as being a unified, coherent, and behaviorally consistent entity.

Under conditions of high entitativity it becomes more common to speak of the group as an entity in and of itself. We might, for example, refer to a workgroup as having objectives or a reputation. A sports team has a record of wins and losses shared by all members of the team, and it may be viewed collectively as being relatively competent or incompetent. There may be good and bad athletes on the team but the team has a record that all share regardless of its members' individual abilities. Similarly, a corporation may be said to have a particular culture and display certain traits, such as being socially conscious, collegial, or even playful in nature. Thus, one can conceptualize the collective entity as having inherent properties, even though these may be different than the individual characteristics of particular group members. It has been proposed that as groups appear more entitative in nature, processing and storage of information by perceivers will come to mimic what occurs when impressions are formed of individual persons rather than showing the differences normally found in individual and group studies of impression formation (see Hamilton & Sherman, 1996).

Lickel et al. (200) also asked participants to sort the presented group descriptions into piles of distinct types based on their perceived similarity to each other. Although there were slight differences across analyses, there was much agreement, despite the use of different techniques. There was also a very high degree of consistency across the different studies and subject populations. The final study of the series, and later presentation by Hamilton and his colleagues (Hamilton, Sherman, Crump & Spencer-Rodgers, 2009), endorsed a solution with four basic types of groups. These were labeled (1) intimacy groups, (2) task groups, (3) social categories, and (4) loose associations.

The first of these types included such groups as family members and friends. These were seen as small groups having high degrees of interaction, of long duration, high importance and impermeable boundaries. Task groups included committees and work groups. They were also viewed as small groups with shared goals, but having a shorter life span and more permeable boundaries than intimacy groups. Social categories are what we normally think of when talking about prejudice, such as racial, ethnic or religious groups. They are large in size but with relatively little interaction among members in the broader sense and which are long lasting and impermeable. Finally, loose associations included cases where there was only a minimal bond among people, such as people sharing a common interest, or little shared other than proximity or other superficial similarity. Examples of the latter would be people queuing up at a bank or bus stop. Entitativity was rated most highly for intimacy groups, followed by task groups, social categories, and finally loose associations.

There are some groups that do not fit so obviously into one of these four types. For example, political parties are often viewed as social categories rather than task groups but they possess elements of both. They are large in size like many social categories. But they have less rigid borders and people are free to enter or leave the party at will. They are generally long lasting but not compared to ethnic, racial or religious groups. Some parties have arose only for a limited number of political campaigns and died out shortly thereafter--a quality more typical of task groups. Like task groups political parties have an organizational structure, shared objectives, and common outcomes in terms of their success in elections. However, high levels of interaction occur primarily among a subset of the party's members and the activity level is seasonal, varying as a function of particular campaign cycles.

Closely related to these groups are "conservatives" and "liberals." These groups are very similar in many ways to political parties in that they favor particular political and ideological positions. Like party members, those falling within these groups may be active or detached with respect to the political process and may enter or leave the category at will, at least in their own minds. There is no formal declaration by an individual as to membership in these groups and no formal organizational structure in this sense. As such they appear to be even more removed from the task group category than political parties. Yet all of these politically oriented groups seem to straddle the boundary between task groups and social categories. Are they large task groups or social categories with permeable boundaries and relatively shorter durations?

It is interesting to note that in a study thought to scale social categories along a dimension of entitativity, "Republicans" and "liberals" were among the groups scoring highest on this metric (Haslam, Rothschild & Ernst, 2000). Unfortunately, due to the emphasis on social categories rather than all types of groups, the range of entitativity across those groups sampled in this research was severely limited. Many of those factors which promote perceptions of entitativity are not generally associated with social categories. Members of social categories exhibit relatively low levels of interaction among themselves. There is little coordination and organization within the group. To some degree, those comprising a social category may share a common fate as a result of discrimination or common historical treatment within a society. They may also share norms or experiences as a result of a common cultural heritage. However, social categories represent very large aggregates where interaction and common outcomes, in a more proximal sense, is unlikely between any two arbitrarily chosen members. This may be one reason why social categories are typically viewed as only moderate in entitativity.

Rothbart and Taylor (1999) argue that perceivers act as though social categories reflect natural kinds rather than socially constructed classifications. Such categories are thought to have inductive potential in the sense that once a person is classified as an exemplar of the category, it is possible to infer other characteristics thought to be common to its members. There is an assumed similarity or homogeneity among a group's members which derives from an underlying essence within each of its members. Once classified as an instance of the category based on readily apparent surface attributes, the perceiver will come to infer other less apparent, more distal, and possibly more abstract attributes. For example, observing such surface attributes as body shape, hair style, clothing, or others that might distinguish between men and women, other underlying attributes, such as nurturance or aggressiveness, may be inferred. Such is the basis for stereotyping and overgeneralization in social perception. An initial categorization of someone as belonging to a category results in the activation of many other associated beliefs.

Rothbart and Taylor also posit that social categories are construed by perceivers as unalterable, with impermeable borders, and whose membership is not freely chosen by members. It is thought that one cannot readily shed one's race or ethnicity, for example. Of course, such beliefs may not be based in fact but still be psychologically meaningful. What is the race of someone who is a mixture of European and African lineage? Is there a single essence or a combination of different essences, and if the latter, does the concept any longer serve a useful function? Such cases of mixed identities argue against the essentialist position and point to the social origin of such categories.

Another example is presented by Rothbart and Taylor in their discussion of how Jews have been treated throughout history. They note that Jews were once viewed as a group of people who had simply not chosen to believe in Christianity and that those who converted were treated the same as other Christians. Later, during the Spanish Inquisition, converted Jews were treated as a separate, and persecuted, subgroup of Christians. Still later, within Nazi Germany, Jews were considered a separate race entirely, defined by parental lineage. Thus, different considerations challenge the assumptions underlying essentialist belief systems. Rothbart and Taylor argue not for this untenable philosophical position but for the notion that people adopt this perspective in their perceptions and thinking about other people.

The tendency to view social categories as natural kinds with inherent characteristics shared among members may explain why stereotyping is so commonplace towards these groups despite the modest entitativity associated with social categories more generally. Hamilton et al. (2009) offer another explanation. They suggest that perceived entitativity may combine with strong contrast effects among opposing social categories (i.e., a large metacontrast ratio or high levels of intragroup similarity combined with large intergroup differences) and thereby foster stereotyping towards only moderately entitative groups. Yet another reason perhaps is the greater number of attributes and the richness of descriptions associated with social categories compared to task groups (Spencer-Rodgers, Hamilton & Sherman, 2007, study 1), although this may be the outcome of stereotyping rather than a basis of its origin.

Hamilton et al. (2009) argue that with increasing entitativity the group emerges as a unified entity and its members become largely interchangeable elements within it. The authors summarize findings showing that there is greater difficulty distinguishing between the members of more highly entitative groups. In one series of studies (Crawford, Sherman & Hamilton, 2002), for example, it was found that traits spontaneously inferred from the behaviors of one group member were later attributed erroneously to other members of the same group and that these errors were most evident when entitativity was high rather than low. These findings suggest that members of the highly unified groups came to be seen as equivalent and undifferentiated. As a result of such intragroup confusion, perceptions of homogeneity within the group and stereotyping towards it may be promoted.

Spencer-Rodgers et al. (2007) view entitativity or the property of "groupness" as an essential aspect of group perception and a necessary condition for stereotyping. In their words, "a collection of individuals must be perceived as having a core nature or a set of common features that connect and bind the members together before stereotypic attributes can be ascribed to the group as a whole.. one must view a group as a meaningful social unit before one can hold a stereotype about the group" (pg. 386). In their research (study 2) they demonstrate that entitativity mediates stereotypic judgments for both task groups and social categories.

But is there is single dimension of entitativity that applies to all types of groups and which has common perceptual cues in all instances? Brewer, Hond and Li (2004) suggest that there are two different implicit theories of group entitativity with each having different cues and units of analysis. One meaning of entitativity is based on a theory of essence in which the perceiver attends to cues of intragroup similarity or homogeneity of personality. Other cues to entitativity include the degree of behavioral consistency across individuals and across time for the group overall. There is a search for clear boundaries that enable the group to be distinguished from others. This corresponds to the type of group coherence described by Rothbart and Taylor (1999).

The other meaning of entitativity is based on a theory of agency in which common goals define the group. The focus of attention in this case is on heterogeneity among different members within the group (e.g., different roles) and dynamic changes signifying group development and goal attainment. With respect to intergroup behavior there is an attempt to understand the group's relationship to other groups and the group's place within the larger encompassing social structure. This distinction between two versions of the entitativity concept will hopefully clarify what has been a rather confusing area of research.

Brewer et al. (2004) place some important limiting conditions on perceptions of these two forms of entitativity: "Attribute similarity alone does not correspond to high entitativity, unless the attributes have been essentialized. Similarity of goals and purposes alone does not correspond to high entitativity, unless the group is perceived as acting in concert on their common goals" (pg. 29). One question that arises in light of these two different conceptions of entitativity is how the two types combine in the process of group perception. What happens when both are heightened as could occur when members of a social category appear to become mobilized to action or social protest? Examples would include recent demonstrations by Latinos over the Arizona immigration law or African Americans' protests and boycotts during the civil rights marches of the 1960s. Do these events give rise to a new distinct subgroup within the larger social category, or do such actions generalize to the larger collective from which it has emerged? Under what conditions might the actions of a group of activists be associated with one particular social group despite a mixture of support from other groups as well? Another issue is whether any changes occur to group stereotypes in terms of their content and strength during and following such dynamic changes in activity and coordination among the members of a social category. Social change involves dynamic changes in group cohesiveness, unity, and the crystallization of shared goals. Consequently, changes in perceived group entitativity would be expected to accompany such changes along with changes in group identification and stereotyping among by members of both ingroups and outgroups.