Chapter 7

Social Psychology:
Past, Present, and Some Predictions for the Future

Nyla R. Branscombe, University of Kansas and
Russell Spears, University of Amsterdam

Social psychology has always been driven, to some extent, by the prevailing concerns of the day. Even when not explicitly directed by funding agencies and policy makers, social psychologists have drawn their impetus for research from newsworthy social phenomena, such as racism, altruism, and the AIDS crisis. This approach has not only led to a proliferation of applied research implementing established theory in a range of assessment and intervention programs, but it has also led to further theoretical developments. Some of the earliest advances in social psychology on topics, such as group dynamics and social influence, were supported by government funds, around the time of the World War II. Very practical issues, such as how to maintain group morale among the troops (Stouffer, Suchman, DeVinney, Star, & Williams, 1949) and how to persuade homemakers to serve cheaper cuts of meat (Hovland, Janis, & Kelley, 1953), reflected the social concerns of that period.

A balance between the basic and applied wings of the field will need to be maintained if social psychology is to remain healthy, and ongoing changes in society, such as increasing cultural diversity, are likely to influence the direction of such research. Social psychology offers a staggering array of topics that appeal to both seasoned and beginning researchers.

Our own research has been especially directed towards understanding the mechanisms underlying, and potential means of intervening in, socially relevant intergroup conflict situations. Before addressing what we see as useful directions for future work on this vexing problem, we briefly consider the present social context in which social psychology itself is embedded and how that is likely to influence its future research agenda.

The Current Context of Social Psychology

We expect that the current trend of funding research that focuses on the pressing issues of the day will continue. For example, as long as a medical cure for AIDS is far off, the importance of assessing and encouraging behavioral solutions to such health problems is likely to remain a research priority (see Snyder, 1993). That threats to health invite social psychological intervention is equally evident in other areas, such as smoking, drug abuse, and poor dietary habits, cases in which behavioral prevention is better than a medical cure of the ensuing problems. Therefore, social psychology has an important role to play. Societal problems of this sort cost taxpayers billions of dollars every year so investment in research designed to address their prevention may have far reaching economic as well as social consequences. Social psychology can provide insight into how social influences affect our behavior, and how we might intervene. The potential role of psychological shifts in self-definition and social contextual factors have yet to be fully explored, and we think this is a direction that social psychology should and will move. We predict that some of the most exciting research in these areas is yet come.

These examples of various behavioral problems suggest that some of the most serious issues facing humankind are "human-made." Social conflicts that arise from the need to share the planet with people of different creeds and colors seem to increase rather than diminish with the development of "civilization." Conflicts between and within nations are not just political issues; they invite social psychological analyses as a means of contributing to their solution. Social psychology cannot hope to provide all the answers to problems that have strong historical and economic roots, but it has a part to play in helping us understand and shape our world, especially where issues of self, identity, and social interaction play a major role. Obviously, the contributions that social psychologists can make in reducing tensions between and among groups underscores the value of creative research in this important area.

Social psychological research has been traditionally divided into three general topic areas, based on whether the emphasis is on the factors internal to the individual or broader social processes. At the most intrapsychic level, research topics that have been center stage have included self and attribution processes, impression formation, and attitudes. Research at the interpersonal level has focused on attraction and close relationships, prosocial behavior, and aggression. At the intergroup level, research has been aimed at understanding stereotyping and prejudice, social influence processes, and the impact of groups on the individual.

Within this tripartite taxonomy of social psychological research, specialization is both to be expected and functional: the sheer volume of empirical work makes it increasingly difficult to keep up with the latest developments in all parts of the field. This reality encourages young researchers to concentrate early in their careers in a specialized area of social psychology, but not without a cost. Such specialization produces fragmentation in the field; evidence of this can be found in all three areas of social psychology.

If the discipline is to make progress, the principles developed and evidence accumulated in one part of the field should be applicable to and overlap with those investigated in other areas. Because an increasing division into discrete topics thwarts theoretical integration, we consider here only those perspectives that have provided a broader view of the field as a whole. "Grand theories," aimed at explaining all of psychology in terms of a few mechanisms (e.g., behaviorism, Freudian theory), have been out of favor for some time. Mid-range theories that address phenomena within a fairly circumscribed area have been the norm. However, during the past decade, three fairly encompassing theoretical frameworks aimed at integrating a broad selection of social psychological findings have been developed. We briefly review these three different perspectives and point out how they differ from each other in terms of primary research topics and their level of analysis. Then, in the remainder of the chapter, we consider how the most social psychological of these three frameworkstthe social identity and self-categorization perspectivetcan be used further to unify the empirical findings obtained in the field as a whole. We close with a discussion of the processes that we see as requiring further research, and we consider several specific new topics that are increasingly likely to capture investigators' attention in the new millennium.

Unifying Theoretical Approaches to Social Psychology

The View From Below: Cognitive Psychology

One approach that is gaining momentum involves an explicit focus on the potential neural mechanisms underlying social behavior. To some extent, this movement towards an increasingly molecular level of analysis reflects a continuation of the borrowing of models and methods from cognitive psychology that has been ongoing for some time. As described by Fiske and Taylor (1991), the core of research in social cognition during the past two decades has focused on the mental structures and information processing principles employed by the individual perceiver engaged in social judgment tasks. Researchers have and continue to generate an impressive amount of research that describes the structures and processes underlying the individual's conception of self and others (Higgins & Bargh, 1987; Linville, 1987; Markus, Smith, & Moreland, 1985; Smith & Branscombe, 1987; Wyer & Srull, 1986). They have also explored how people employ heuristics when reasoning about social events (Branscombe & Cohen, 1990; Fong, Krantz, & Nisbett, 1986; Kahneman, Slovic, & Tversky, 1982; Schwarz, 1990) as well as how biases in social judgment might arise as a result of cognitive capacity limitations (Fiske & Neuberg, 1990; Gilbert, Pelham, & Krull, 1988; Hamilton & Sherman, 1989). Research stemming from this tradition is increasingly employing connectionist and neural network models in order to develop a "machine code" for phenomena, such as impression formation and social stereotyping (see Smith, 1996). We see this tendency to conceptualize social judgment in terms of cognitive psychologys most recent "bottom up" model as likely to underestimate the social relational and motivational underpinnings of human behavior.

Within the social cognition tradition, researchers have conceptualized emotional and motivational factors primarily as moderators of normal cognitive processes. That is, emotion and motivation have been conceived of as "add-on" factors rather than ones that are integral to all social judgment and behavior. Thus, a variety of investigators have manipulated the mood state of participants by some means that is irrelevant to the task at hand (Bodenhausen, 1990; Fiedler, 1990; Forgas, 1995) in order to assess how social judgment and memory might be affected. For example, after being told that they will be participating in two unrelated studies and that the first study concerns responses to film stimuli, respondents may be shown a comedy, a sad film segment, or a neutral control. Introduced as a separate study on impression formation, the task involves manipulation of the features of the target persons or the conditions under which the judgments are rendered. This method assesses the role of different mood states that may be pre-existing when people are required to make judgments. Within this type of paradigm, moods may influence social processes by altering how the initial information about the target is encoded and what strategy is used to make the judgment. Thus, emotion is expected to limit the cognitive resources available to the individual, resulting in judgments that are more mood-driven and less individuated or responsive to variations in the specific target information provided. The mood state itself can create inaccuracy in person perception by eliciting heuristic means of assessing the likability of the target persons, with perceivers simply assessing their mood state and using that as a basis for judgment (Isen, 1987; Schwarz, 1990; Worth & Mackie, 1987). The relationship of the target persons to the self, or what kind of affect those persons might evoke in the respondent, has not been focused on in this research.

In contrast to mood states, motivation has been conceptualized as the means by which the accuracy of social perception can be enhanced. If motivational factors influence people's willingness to put forth the necessary effort to arrive at a more individuated and less stereotypic impression of others, judgment accuracy may increase (Fiske & Pavelchak, 1986; Showers & Cantor, 1985). The underlying assumption is that focusing on how members of groups are similar to each other is easier than focusing on how individuals differ from one another, with perceived similarity among group members being defined as the essence of social stereotyping. From this perspective, perceived similarity among individual group members is less accurate than perceived differences among members of a group. Social motivations might determine the level of categorization that is employed. For example, whether people are perceived as group members rather than as distinctive individuals can influence categorical judgments. In addition, their judgments might occur in ways that are consistent with the perceivers own position within the social hierarchy. These complications are neglected when motivation and emotion are conceptualized primarily as external moderators of basic social cognition processes rather than as integral to human social perception and evaluation.

The View From Our Past: Evolutionary Psychology

A second broad unifying direction that has come to the fore over the last decade involves attempts to explain social psychological phenomena in terms of evolutionary principles (Buss, 1995; Simpson & Kenrick, 1997). Although this approach can be also characterized as borrowed from biology, it has so far had its greatest impact on topics involving the nature of interpersonal relationships. Specifically, evolutionary psychologists have focused on gender differences in sexual behavior, helping, and aggression. Because men and women are assumed to have faced different adaptation problems because of their differing reproduction roles, residuals of this evolutionary history can be exhibited in ongoing human social behavior. Gender differences that have been central to this theoretical perspective and that have been empirically investigated include number of sexual partners (men report more than women); psychological investment in children (women report more than men); what aspect of infidelity is most distressing (women report greater upset about emotional disloyalty and men report greater concern about sexual disloyalty); who is more likely to behave in a physically aggressive fashion (men more so than women): and who is more likely to help and be helped by strangers versus relatives (men are more likely to display heroic helping of strangers but women take care of family members especially in private settings).

Of course, because any of these effects could be also explained in terms of gender differences in socialization and adult role requirements, any impact of our evolutionary history should be reflected primarily in cross-cultural constants. Some of the existing research has explored such possibilities. For example, Buss (1995) has argued that gender differences in terms of the attributes sought in sexual partners exhibit some cross-cultural generality (men report desiring physically attractive partners as an indicator of reproductive capability and women tend to rate variables related to status as especially desirable in mates). Nevertheless, there is more cross-cultural similarity in the attributes overall that are deemed desirable in a partner by both genders than there are differences. Evolutionary explanations that emphasize consistent differences by gender have difficulty with such similarities. These explanations also struggle with behaviors that are flexible in terms of when and how they are expressed (i.e., those that show considerable context-dependence). We argue that flexibility and sensitivity to context are perhaps the most important hallmarks of human social behavior. We also argue that a theory must be capable of accounting for these characteristics in terms of the psychological processes involved. Evolutionary psychology emphasizes what behaviors are likely to have worked in the context of our ancestors, although the actual nature of that context is much debated. As a result of their focus on the distant past, we are not provided with an explanation of how people adapt to and navigate in the complex and changing social environments that are found in present-day Western technological societies.

A Social Psychological Integrative Perspective

As we see it, the primary problem with attempts at grounding social psychology in neural networks or genetic history is that this neglects more "top down" influences stemming from the social context itself. As a result, a uniquely social psychological level of analysis located in the "here and now" is precluded. Although both of these borrowed frameworks represent theoretical attempts at ordering the proliferation of phenomena and paradigms within social psychology, there may be important limits on the degree to which we can or should rely on other disciplines for explanatory mechanisms. We believe that for a theoretical framework to be maximally useful it should contain uniquely social and psychological mechanisms (see Wicklund, 1990, for a discussion of this issue). For social psychology to make a distinctive contribution, it must acknowledge the fundamentally social nature of mental life and explore how psychological mediation occurs within the individual. Therefore, we turn our attention now to a social psychological perspective that attempts to integrate research within the three areas of social psychologytthe intrapsychic, interpersonal, and intergrouptby employing only social psychological mechanisms.

The theoretical perspective that meets our criteria can be found in an integration of social identity (Tajfel & Turner, 1986) and self-categorization theories (Turner, 1987). Together, they represent a "grand" theory of the interplay of social cognitive and emotional factors in the generation of behavior. Tajfel (1972, 1981) emphasized the importance of avoiding biological reductionism and strictly individualistic explanations for social psychological processes. He underscored the important role that the historical and social context plays. The social identity tradition offers a level of analysis that takes into account the nature of the social structure as well as how it is internalized by the individual. Because this theoretical tradition forms a guiding framework for our review of the different areas of social psychology, we now provide a brief summary of the general principles involved.

Social identity is conceptualized as that part of the person's self definition relating to their membership of a social group (or groups), along with the value and emotional significance that entails (Tajfel, 1978). The concept of self-categorization is closely related to social identity but broader in so far as it also comprises definitions at more unique levels (e.g., "personal identity") and more inclusive levels (e.g. "humankind"; see Turner, 1987). Social identity and self-categorization theories emphasize the links among social contextual factors, how the self is conceptualized, and whether interpersonal or intergroup behavior will be expressed. It does not assume that one level of categorization is more accurate, genuine, or real than another. For example, when interacting with a group of friends in one context, an individual may behave on the basis of his or her individual identity and perceive the self as different from the others present. At another point in time, the persons social identity may determine what actions are undertaken (e.g., as fans of a football team). In that case, even when in the company of the same group of friends at a football game, the individual may feel similar to those same people and typical of "our team." Therefore, all of these processes--perception, evaluation, and action--are solidly rooted in the contextually dependent process of self-definition. From this perspective, decontexualizing an individual will not reveal his or her "true" essence; rather, depending on the context, the person is actually an individual who is different from others and a group member who shares attributes with others.

Social identity theory tries to explain social behavior in terms of the processes of social categorization, social comparison, and social identification. Behavior can be explained as a function of the level at which they categorize themselves in any given context. The type of self-categorization operating in a given context will influence the nature of the comparisons drawn and the emotional significance of others' actions. The degree of emotional significance associated with particular identities will influence responses to others who may be perceived as threatening or supportive of one's social identity. One basic assumption of social identity theory is that people try to maintain a positive sense of themselves as individuals. This objective can be accomplished, theoretically, by either personally differentiating the self from other ingroup members at the individual level, or at the group level by positively differentiating the ingroup from other groups. Self-categorization theory (Turner, 1987) extends social identity theory to create a more encompassing perspective on social processes. In self-categorization theory, the self definition controlling behavior at different levels of inclusiveness ("me" versus "we") depends on the social context and the salience of different types of comparisons in the environment.

The importance of self-definition and subsequent comparisons drawn as a result were illustrated in a recent study majors (Spears, Doosje, & Ellemers, 1997). Psychology students characterized their ingroup as more intelligent when they compared themselves to fine arts students, but preferred to think of themselves in terms of their creativity when comparing themselves with physics. This example makes clear how the definition of self and its attributes is sensitive to the comparative context and the general tendency to view the self positively whenever possible. Our example is also consistent with the operation of some basic human "needs" that have been the focus of much social psychological theorizing of late. These include especially the need to be positively evaluated and the need for attachment or alignment with others (Baumeister & Leary, 1995; Branscombe & Ellemers, 1998; Doosje & Ellemers, 1997; Sedikides, 1993).

The social identity/self-categorization tradition provides us with a relatively simple but powerful theoretical framework for integrating and explaining a variety of social psychological phenomena across the three areas of research. Together they provide a non-reductionistic explanation of intrapsychic processes, interpersonal behavior, and intergroup relations and a theoretical means of integrating this tripartite division. Specifically, self-categorization theory has been used to explain social phenomena, such as self-esteem maintenance, attitude processes, group formation and cohesion, social influence, crowd behavior, and social stereotyping. To the extent that behavior is rooted in the perceivers definition of self, and self-definition is defined in relation to others in the context, we have a truly social psychological way of analyzing social behavior according to general principles that acknowledge diversity of behavioral outcomes across both time and persons. Employment of the central principles in this theoretical tradition allows us to integrate a whole host of empirical findings. We now examine a selection of research findings across all three areas of social psychology and illustrate their fundamentally social nature.

Impact of Self-Categorization and Social Context on Intrapsychic Processes

The social cognition approach has tended to explain social psychological phenomena in information processing terms. When making social judgments, people tend to use simplifying but potentially inaccurate strategies predominantly over systematically processing the information. From this perspective, the social dimension of social cognition simply refers to the fact that we process information about persons rather than implicate any distinctively social influence processes or explanatory principles. Accordingly, intergroup phenomena, such as stereotyping, have been seen as resulting from "normal" information processing biases (Hamilton, 1981). Researchers have conceptualized attitudes more generally in terms of the underlying information- processing mechanisms employed and the type of cognitive structures activated by a task. The main thrust of traditional models of attitudes (e.g., Eagly & Chaiken, 1995; Petty & Cacioppo, 1986) has been aimed at determining the conditions under which people will select one information processing route or another: focus on the "central" characteristics of the persuasive message (e.g., argument strength) or "peripheral" cues (e.g., nature of the source). However, recent research has begun to question the evidence for attitude structural stability and to consider seriously the possibility that attitudes are actually on-line judgments that are constructed differentially as a function of the context (Millar & Tesser, 1992; Schwarz & Sudman, 1996; Wilson & Hodges, 1992).

The tendency to focus on intrapsychic processes has begun to give way to an examination of more communicative and contextual factors implicated in attitude change. For example, researchers have started to examine how impression management concerns can influence message elaboration (e.g., Chen, Shechter, & Chaiken, 1996). The expectation of having to persuade a target on the attitude topic can influence how the contents of the message itself are processed (Nienhuis, 1998). Indeed, one important contextual influence on attitudes involves who the target audience is believed to be. Other people provide a reference point for social comparison and self-definition, and an audience's sensibilities has to be taken into account (Leary & Kowalski, 1992; Noel, Wann, & Branscombe, 1995; Reicher, Spears, & Postmes, 1995). Audiences can both constrain attitude expression and influence how they are constructed initially, especially when the audience is perceived as having the power to judge or affect the outcomes that will be received. For example, when people are motivated to make a good impression on others who have power over them, they are likely to express attitudes that they believe will positively impress their audience, even if that means derogating an outgroup that is not privately perceived as negative. If, however, participants believe that those in power will not learn what attitudes they expressed, then a quite different set of beliefs are reported (Noel et al., 1995). From a self-categorization perspective, both the nature of the self being presented (e.g., personal self versus a particular social identity), and the nature of the audience (e.g., an ingroup or outgroup) can be powerful influences on behavior. Research that fails to consider the purposes individuals may have in communicating a message in a given context will fail to anticipate how variable behavior can be in different contexts, or appreciate when such behavior will or will not reflect underlying attitudes and allegiances (Reicher et al., 1995; Turner, 1991).

Similarly, other basic cognitive processes thought to reflect purely information processing concerns appear to vary depending on how the self is contextually defined. For example, one influential tradition has concerned people's employment of heuristics--simplifying methods of dealing with complex information processing tasks (Kahneman et al., 1982). In the initial demonstration of the operation of the availability heuristic, fresher access to instances from memory is used to make a likelihood judgment. For example, Tversky and Kahneman (1973) showed that after reading a newspaper article about a car crash, participants were more likely to judge the frequency of motor vehicle accidents. They also judged their own chances of befalling one as higher than in a control condition where a neutral text was read. In other words, reading about this accident made available in memory instances of such events that influenced subsequent judgments of their likelihood. In a replication of this research, Stapel, Reicher, and Spears (1994) presented physics students with information about a car crash, but they also varied the link between that information and the participants' own social identities. For half the participants their identity as "physicist" was made salient, whereas for the other half the more inclusive identity as "scientist" was made salient. As a result, how the participants categorized the self varied differed across conditions with the relevance of the available or primed information. Thus, when their scientist identity was made salient, this includes not only physicists but also psychologists. Only when the reported crash was relevant to an ingroup identity (e.g., when the victims were either physicists or psychologists and the salient identity of the participants was as scientists) did respondents display an availability bias and overestimate the likelihood of crashes. When the crash story victims were defined as outgroup members (e.g., the victims were psychologists and the salient identity of the participants was as physicists), then subsequent likelihood judgments were unaffected. In other words, the operation of this seemingly basic cognitive bias stemming from use of the availability heuristic was dependent on the social relation of the relevant stimulus to the self.

The focus on cognitive processing mechanisms has also resulted in a neglect of the role of the communicative context. However, on further analysis, judgments that were once thought to be the due to the operation of heuristics appear to be products of how people interpret information in various social communication settings (e.g., Berndsen, Spears, McGarty, & van der Pligt, 1998; Bless, Strack, & Schwarz, 1993; Branscombe, N'gbala, Kobrynowicz, & Wann, 1997; Hilton & Slugoski, 1997; McGarty & de la Haye, 1997; Stapel, Reicher & Spears, 1995). Thus, what was once thought to be due to the operation of fundamental and unchanging cognitive processes, now appear to be the result of participants attempting to derive social meaning from the materials they are presented with. Recent research re-examining Kahneman and Tversky's (1982) classic demonstration of the simulation heuristic, where ease of imagining a better outcome for an event determines judgments about it, makes this point. In that work, the identical poor outcomes obtained by two individuals--one who puts forth effort in an attempt to make money and one who does nothing--are described in a single scenario to the participants. Consistently, the judge evaluates the acting target more negatively than the target person who did not do anything. This effect has been assumed to be due to the relatively greater ease of mentally simulating the acting target so that a better outcome could be imagined for him than for the non-acting target. Using Kahneman and Tversky's (1982) original materials, N'gbala and Branscombe (1997) showed, however, that differential ability to simulate the two targets was not why the targets were judged differently. Indeed, the two targets were actually equally likely to be mentally simulated. Rather, this judgment "bias" was found to arise out of participants' attempts to understand the situation they were presented with (i.e., two targets individuals who received the same outcome, regardless of their behavior). In order to make sense of the situation, participants directly compared the content of the two targets' behaviors. Given that the poor outcome appeared to be inevitable, participants appear to have concluded that the target who invested energy in what was clearly a losing proposition was a poorer decision-maker than the one who did not invest any energy. Again, the meaning gained and judgment arrived at by these participants in the experimental context they found themselves in did not stem from use of a general heuristic such as simulation; instead it resulted from a fairly simple context-driven comparison of the two targets behaviors.

Recent research using the Ebbinghaus optical illusion has demonstrated how seemingly simple perceptual phenomena, such as object size estimation, can be mediated by social meaning (Stapel & Koomen, 1997). In this classic illusion, a moderate-sized circle surrounded by smaller circles appears to be larger than a comparable moderate-sized circle surrounded by larger circles. This effect has been thought to arise from basic perceptual contrast processes where the context of the surrounding circles influences the subjectively perceived size of the central circle. Building on earlier research showing that the standard effect can be obtained with social stimuli (faces) as well as circles, Stapel and Koomen (1997) showed that the effect only occurred if the face stimuli were defined as emanating from a common social category (e.g., lawyers). When this common category relationship was absent (e.g., the identical stimuli were said to come from multiple social categories), the size-contrast illusion disappeared. In other words, the socially defined relationship of the stimuli to each other facilitated the ostensible "perceptual illusion;" eliminating that social relationship eliminated the illusion as well.

In sum, the assumption underlying much work on heuristics and social judgment is that their use reflects basic cognitive information processing mechanisms. At first sight, their operation would seem to be independent of more obviously social psychological issues of self-definition. However, as we have shown, the social context and the perceivers self definition affect even these }basic processes.” When we turn to more obviously socially influenced judgment processes, such as attributions, the influence of culture and motivation are even more apparent. Indeed, a variety of findings that have been assumed to be universal cannot be generalized to more collectivist cultures which tend to be characterized by different forms of self-definition and social organization (Markus & Kityama, 1991).

One classic example of a phenomenons failure to generalize to collectivist cultures is the case of the "fundamental attribution error," in which people over-attribute the causes of actors' behavior to their internal dispositions compared to plausible situational causes or constraints (Ross, 1977). Although robust and widely replicated in Western cultures (Gilbert & Malone, 1995), this finding of greater personal than situational attributions does not extend to participants in the collectivist culture of India (Miller, 1984). Because people in individualistic cultures are more likely to see others as masters of their own destinies (potentially more than is warranted), they tend to neglect the power of the situation when making attributions for behavior. Moreover, the contents of a given culture are themselves not fixed or constant, but are historically specific (Gergen, 1973; Tajfel 1972; 1981). What it means to be a self, at both the personal and social identity levels, can shift over time. As a result, expectations and beliefs about how social relations are and should be structured must be historically situated. Because the treatment that is expected and is seen as legitimate for the self when categorized at the individual level and the self when categorized as a group member can differ, affective responses to the same events can be expected to vary depending on the social context.

The tendency to attribute positive or beneficial outcomes to some aspect of the self, and the consequences of doing so, depends on the meaning implied and the aspect of the self that is salient in a particular context. To illustrate, Branscombe (1998) asked male and female participants to think about either the positive or negative outcomes that they have received based on their gender group membership. Drawing on the basic social identity theory assumption that people attempt to maintain a positive view of themselves, men and women should be differentially motivated to conceptualize the effects of their group membership in particular ways. Those persons who are members of a powerful social group should be reluctant to, and are likely to find it psychologically uncomfortable to, think about themselves in terms of the benefits or privileges received as a function of that group membership, especially in individualistic cultures. In contrast, powerful group members should find it rewarding to think about the disadvantages that are associated with this group identity, especially if they are relatively localized. The results revealed that men did suffer self-esteem loss following thoughts of group-based privilege compared to thoughts of disadvantage. Thinking about privileges or disadvantages stemming from a subordinate group identity should have quite different effects, with thoughts of disadvantage tending to harm well-being in women compared to men. Focusing on group-based disadvantage in women reduces attributions of personal control and is correlated with depression (see also Kobrynowicz & Branscombe, 1997; Ruggiero & Taylor, 1997). However, among dominant group members, well-being reductions were associated with conceptualizing the self as a group member which, rather than the personal self, was implied to be the critical variable responsible for ones successes. As a result, for men, internal attributions for success appear to be undermined by thoughts about the privileges received based on group membership. In fact, for those who are low in identification with their gender group, thoughts that imply illegitimate receipt of benefits based on ones group membership may result in the experience of group-based guilt. Thus, this research points out how the effects of categorizing one's self in terms of gender depends on the nature of the power relations that exist between the groups and how those are framed in a given context.

Indeed, how people react to categorizing the self as a group member can depend on the nature of the group's history. Differential emotional reactions to an event may occur, depending on how the self is categorized and the degree of identification with the group. That is, although intense emotional responses to an event would not be expected when the self is categorized as a unique individual, elevated emotional responses to the same event may be exhibited when the self is categorized as a group member. Doosje, Branscombe, Spears and Manstead (1998) tested this hypothesis by inducing participants to categorize themselves as members of a group that had historically exploited another group or that had a history of fair treatment toward the other group. Participants also received feedback about whether they personally had or had not displayed prejudice towards members of the other group. Even when participants had nothing to feel guilty about at the personal level, because they believed they had not personally behaved in a prejudicial fashion, knowing their group's history was exploitive, resulted in the induction of feelings of collective guilt. In a second experiment using participants' Dutch national identity, Doosje et al. (1998) showed that the degree of collective guilt experienced mediated the impact of ambiguously presented information about their nation's history on behaviors reflecting a willingness to make reparations to members of their nation's former colony (Indonesia). These results show that people may experience emotional responses, such as guilt, but whether this reaction occurs or not depends on how the self is defined in a particular context. Furthermore, different kinds of emotional experiences can result, depending on whether the self is construed at the personal or the social identity level.

As our review has suggested, the social context and how the self is defined are important determinants of various social judgments and behaviors that are often assumed to be solely a function of information processing mechanisms. The social level influences on cognition that we have described appear to be considerable. Indeed, they may be even greater than those stemming from the impact of human cognitive capacity limitations per se. Social cognition seems to be structured by the definition of the self that is employed, with the context enabling socially meaningful perception rather than simply limiting information processing. As a result, intrapsychic processes can only be artificially separated from the other two areas of social psychology, the interpersonal and the intergroup. We now turn to the role of self-definitional and contextual factors in interpersonal processes.

Interpersonal Processes and the Impact of Differing Self-Definitions

Attraction between individuals has long been assumed to vary as a function of perceived personal similarity (Byrne, Clore, & Smeaton, 1986; Griffin & Sparks, 1990), familiarity (Bornstein, Leone, & Galley, 1987; Moreland & Beach, 1992), and physical attractiveness (Hatfield & Sprecher, 1986). We maintain close relationships when they are perceived to be equitable, instrumentally rewarding, and emotionally satisfying (Hatfield, Traupmann, Sprecher, Utne, & Hay, 1985). These same factors encourage helping others: attributions of deservingness, ability to obtain a variety of rewards (Piliavin, Dovidio, Gaertner, & Clark, 1981), and identification with or empathy for the person in need (Batson, 1987; Smith & Henry, 1996). For the most part, the lack of these factors or the presence of their opposites has been shown to influence the expression of aggression. Specifically, frustration, anger, perceived unfairness, and an inferred intention to bring about an undesirable outcome have all been linked with aggression (Geen, 1990).

Once again, we will not attempt to review all of the research available on interpersonal processes exhaustively. Rather, we will show how some of the themes from social identity and self-categorization can be used to illuminate the underlying processes responsible for some of the major effects that have been observed. Much of the existing social psychological research on interpersonal relations has been rather individualistic in the sense that it conceptualizes social interaction exclusively in terms of what occurs between individuals when they categorize themselves strictly in terms of their individual identity. We will attempt to illustrate, however, that the self-definition of the participant in a given context (e.g., the personal versus social level) can moderate interpersonal process effects.

Research on attraction has primarily emphasized the bonds between individuals that result from factors, such as interpersonal similarity, mutual interdependence, and the like. Yet, from a self-categorization perspective, attraction to others can also occur at the social level. Furthermore, such social attraction is both conceptually and phenomenologically distinct from interpersonal attraction or the attraction between individuals as individuals (Turner, 1987). Perhaps the most obvious example of this distinction comes from research employing "the minimal group paradigm," where people reward ingroup members and discriminate against outgroup members without any personal knowledge of or relationship with the individuals in either group except for group membership per se (Tajfel & Turner, 1986). In fact, in many minimal group studies, the participants' personal friends may be actually categorized as members of the outgroup. Hoggs (1992) research has explicitly examined the distinction between personal and social attraction and their differing consequences. In that research, participants were first categorized into groups with others who, based on bogus pre-testing, were said to dislike some members of their ingroup and prefer individual outgroup members. In other words, group bonds per se were not strong in this condition and evaluation was made on the basis of each individuals personal features. Measures of liking of one form (interpersonal) reflected the features of the individual personalities involved, but favoritism in terms of reward allocations was based on another form of liking (social attraction). As a result, ingroup members who were seen as }interpersonally dislikable" because of their personal characteristics were given greater rewards than were "interpersonally likable outgroup members" when participants were acting on the basis of their social identity. Thus, feelings about and the treatment delivered to others depends on how the self and the target are defined: as individuals or as ingroup and outgroup members.

In a field study of netball players, Hogg and Hains (1997) provided evidence concerning the relative independence of these two forms of attraction. They showed that social attraction (which was related to how prototypical of the group the individual felt and degree of identification with the group) was distinct from personal attraction (which was related to perceived similarity and other interpersonal variables). Indeed we would argue that, in some circumstances, social attraction even may outweigh the impact of interpersonal similarity in evaluation of others. For example, Schmitt and Branscombe (1998) found that men who valued their gender group identity evaluated another man more positively when he matched the group's prototype compared to when he was like themselves personally and was not prototypical of the group. Specifically, men who identified highly with their gender group and whose masculinity was threatened by the experimenter telling them that they were less masculine than the norm of their group, actually derogated someone else who was like themselves and was described as low in masculinity. In contrast, when the target was described as highly masculine (and was therefore personally dissimilar to the participant) but matched the prototype of the group, evaluations were more positive among the high identifiers. Attraction here was driven more by group level protection concerns rather than by interpersonal similarity, which has been heretofore regarded as the main basis of attraction.

Likewise, how people occupying leadership roles are evaluated can depend on how the perceiver defines the self in a given social context. In a series of studies, Platow, Hoar, Reid, Harley and Morrison (1997) showed that in interpersonal contexts people have a preference for fair leaders rather than unfair leaders, as one might expect. However, in intergroup contexts people exhibit a preference for leaders who are unfair (i.e., who are biased against the outgroup). Indeed, leadership endorsement and social influence in both contexts was mediated by the participants level of self-categorization and the extent to which they perceived themselves as members of that group. Such a divergent pattern of findings depending on the judgmental context, reflects the fact that those contexts cue different levels of identity (personal versus group), which evoke their own norms and agendas. In an interpersonal context we are keen to know whether a leader will treat everyone equally and fairly (particularly ourselves). In an intergroup context, however, the person who best supports our own group's interests (against those of the outgroup) may be the one who is seen as best for the group (and oneself). In short, evaluation of leaders, and attraction more generally, is not governed by a fixed set of rules. Who we evaluate positively or negatively depends on the social context the judgment takes place in as well as the identities that these contexts make salient.

The notion that attraction can be socially structured by our group memberships also helps explain some important aspects of helping behavior. There has been a long-standing debate in this literature as to whether "true" altruism exists, or whether the effects that have been obtained can be explained by more self-interested motives. Evidence in favor of altruism has been based on a set of studies illustrating the role of empathy or identification with the person in need. Empathizing with the plight of the person in need, according to Batson (1987), evokes a need to benefit the other rather than helping merely as a means of eliminating ones personal distress in the situation. This research closely parallels the self-categorization notion that only when the other is in some sense categorized as part of an inclusive self-category (e.g., an ingroup) is that person likely to be the beneficiary of helping that is altruistic. Indeed, we would expect that those who are highly identified with a particular group would be more likely to empathize with the plight of another ingroup member or the group as a whole and as a result they should be the most prepared to lend assistance. In our own research (Branscombe, Spears, Ellemers, & Doosje, 1998) we have found that people who feel highly respected by other ingroup members (e.g., they believe the ingroup values them as good group members), choose to invest more of their time helping the group compared to themselves personally. We found that this effect was most prominent when the group itself was a socially devalued one, where the perceived need of the group for the valued member's assistance is especially likely to be high.

As with helping behavior, anger and aggression can depend on how the self and the other are categorized. As Smith (1994) discussed in his critique of the hostility and prejudice literatures, responses to members of devalued social categories are not uniformly negative. In fact, people frequently exhibit distinctly positive responses to outgroup members who stay within circumscribed social roles, particularly when the relationship is defined as interpersonal (see Eagly & Mladinic, 1994; Jackman, 1994). Considerable evidence now exists that women, when categorized by men at the interpersonal level can be very positively evaluated, but when categorized as a member of a competing outgroup they can be treated quite negatively. Similarly, White Americans can express fondness towards Black Americans who occupy domestic or subordinate roles, while at the same time displaying hostility towards those individuals who seek equality for their group. In both of these cases, hostility emerges only at the intergroup level of categorization and when expectancies at that level are violated. Categorization at the interpersonal level in both instances can result in positive evaluations.

Conversely, hostility can be experienced when an individual is categorized at the personal level (e.g., as a function of the individual's dislikable personal attributes). Yet, when the same individual is categorized as a fellow ingroup member, positive evaluations can emerge. Wann and Branscombe (1993) demonstrated that evaluation of an attribute or behavior must be considered in context, according to the interpretation it evokes. An aggressive behavior on the part of a person categorized as a member of one's ingroup (e.g., another University of Kansas basketball fan) was rated positively. The same behavior was displayed by a person categorized as a member of an important competing outgroup (e.g., a University of Missouri team fan) and was evaluated negatively. Thus, shifts in the level of self- and other-categorization can rapidly alter whether aggression occurs as well its intensity. Hostile responses can be based on either an interpersonal or an intergroup categorization of the target.

Contextual and Self-Definitional Processes in Intergroup Relations

The domain of intergroup relations has long been concerned with explaining the all too frequent evidence of intergroup conflict. Early approaches to understanding prejudice tended to be quite individualistic, such as the analysis provided by authoritarian personality researchers (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950). As others have noted (see Billig, 1976), personality-based explanations are ill-equipped for explaining the collective dynamics of intergroup conflict. Indeed, Pettigrew (1958) argued that prejudice was more strongly related to the content of social norms than to individual psychodynamics. More recently, Altemeyer (1987) has reformulated the original analysis of authoritarianism in largely normative terms.

The arrival of the cognitive revolution in social psychology meant, however, that such early attempts to explain prejudice and stereotyping in motivational and normative terms were superseded by accounts that emphasized their emergence as products of everyday information processing biases (Hamilton, 1981). Stereotypes came to be conceptualized as schemas or acquired sets of beliefs about social groups that are stored in an associative memory network (Stangor & Lange, 1994). Like attitudes, they were seen as being fairly fixed mental structures that were resistant to change and that should be relatively immune to contextual factors. Hence, much research energy was invested in describing the content of various prominent social stereotypes (Brigham, 1971; Deaux & Lewis, 1984), as well as in whom and when they are likely to be activated in memory (Devine, 1989). Even when prejudice was considered to be a genuine reflection of socialization in a prejudiced cultural milieu, rather than being due to information processing factors alone, it was regarded as something that we all acquire more or less automatically (Devine, 1989). The view that people are automatically prejudiced, and only override this prejudice by conscious resistance has been recently challenged. In fact, evidence suggests that prejudice itself is not universal and that there are important individual differences (Lepore & Brown, 1997; Wittenbrink, Judd, & Park, 1997). Not only do our identities make a difference to the expression of prejudice, but the observed variability in stereotyping and prejudice means that these phenomena cannot be solely located in cognitive processing universals. Social contextual factors can have an impact on the very meaning assigned to an event and its participants, thereby influencing social behavior.

The metaphor of the "cognitive miser" that has dominated social cognition research in the last two decades was well placed to provide a powerful explanation of social stereotyping. The use of social categories and the stereotypes associated with them were assumed to be a "default setting" in social perception. Therefore, their greatest impact should be observable when people do not have sufficient time or motivation to see people as individuals (Brewer, 1988; Fiske & Taylor, 1991). For this reason, stereotypes might even be seen as "energy saving devices." In support of this conception, Macrae, Milne, and Bodenhausen (1994) found that the processing of stereotype-congruent information was indeed facilitated when the stereotype had been previously primed with an explicit category label. Moreover, compared to a control condition where the stereotype was not primed, additional cognitive resources were freed up as a result of stereotype use, resulting in enhancement of performance on a concurrent task. Thus, from this perspective, although stereotypes might be viewed as dysfunctional at one level, because of the biases that result from their employment, at another level they could be seen as serving a complexity reduction function that facilitates information processing. Stereotype use, according to this perspective, is therefore both understandable and to some extent inevitable.

The view that stereotyping results from the ongoing operation of "normal" cognitive processes implies that the social perceptions underlying group behavior should be stable and relatively insensitive to the social context. However, work within social cognition has begun to question the privileged status of social categories in information processing, and query whether stereotype influences are necessarily automatic (Bargh, 1994; Gilbert & Hixon, 1991) or if their operation are so different from more individuated levels of processing (Kunda & Thagard, 1996). If the meaning of a behavior shifts with the social context and the meaning of a behavior depends on how the person exhibiting it is categorized, then failure to consider both of these aspects will result in an inaccurate portrait of stereotyping processes. Focusing on the architecture and limits of information processing neglects the existing power relations between social groups, which forms an important aspect of the social context, as well as socio-motivational factors that influence how the self is categorized.

New evidence is beginning to emerge that suggests stereotypes may not be fixed cognitive structures (Oakes, Haslam & Turner, 1994; Spears, Oakes, Ellemers, & Haslam, 1997). Rather their content appears to vary as if they are constructed "on-line" according to ongoing social needs (see also Kahneman & Miller, 1986). For example, Haslam, Turner, Oakes, McGarty, and Hayes (1992) showed that the content of a stereotype depends on the comparative frame of reference and the current relations that exist between social groups. When researchers asked Australian participants to characterize Americans, the content of their descriptions (e.g., whether Americans were said to be aggressive or not) varied predictably depending on whether the question was asked before the Gulf War or afterwards. Their answers were also influenced by the inclusion of other countries, which were included in the comparative frame of reference. As the War escalated, Americans were more likely to be stereotyped as aggressive than they had been previously, and this reaction reflected changes in the relationship between the participants own nation and the U.S. In addition, when the comparative context drew attention to the conflict with Iraq, Americans were perceived more negatively than when the frame of reference did not remind participants of the Persian Gulf situation. Such work illustrates that stereotypes may be more akin to communicative devices whose operation varies depending on the relationship between the stereotyper and the target rather than on fixed mental structures.

Researchers have increasingly begun to stress the importance of people's desire to make sense of their world actively rather than simply managing cognitive load as important determinants of stereotyping (Spears & Haslam, 1997; Yzerbyt, Rocher, & Schadron, 1997). To examine the potential role of peoples desire to understand information for stereotyping processes, Yzerbyt et al., (1997) replicated Macrae et al.'s (1994) basic experiment that we described earlier. However, they also added a condition where the person information that was presented for participants to judge was inconsistent with the stereotype prime. In this case, they found that priming a stereotype actually consumed cognitive resources (rather than conserving them) as perceivers tried to make sense of and resolve the inconsistency. In a complex social world then, stereotype use may require interpretative effort and, as a result, it may be more cognitively costly than has been previously supposed. If so, conserving cognitive resources cannot be the primary underlying factor in stereotyping. Based on these data it would appear that either economy or inefficiency in information processing can occur as a result of stereotype use, and both appear to be by-products of meaning-making in a given context (Spears & Haslam, 1997).

An important social dimension to stereotyping neglected by an exclusively cognitive focus is the fact that stereotypic images are shared and communicated. If stereotypes were not socially shared, they would be of little social consequence. However, as a result of their socially consensual nature, the process by which they come to be promulgated is of central concern (Tajfel, 1981, 1982). The knowledge that other group members share one's views may be an important means by which stereotypic views are validated. Social influence processes appear to play a critical role in stereotype acquisition (Hardin & Higgins, 1996; Haslam, 1997). Therefore, attachment to a social group and the sharing of the groups perspective may be important inputs in the structuring of social cognition itself. Analysis of social relations operating in a particular context therefore seems to be essential to understanding both the stability and variability of intergroup perceptions. This consideration necessitates an examination of the role of self-definition and social context for intergroup perception.

In contrast to the cognitive miser approach to stereotyping, social identity theory provides a socio-motivational explanation of prejudice and discrimination (Tajfel & Turner, 1986). Research that has tested the various tenets of social identity theory (see Ellemers, 1993, for a review) has delineated important variables that moderate people's intergroup responses. Specifically, prejudice is more likely to be exhibited when a salient social identity is threatened by either negative social comparisons or by a threat to group distinctiveness, and by persons who are highly identified with the specific group that is threatened (Branscombe & Wann, 1994; Doosje & Ellemers, 1997). In fact, some degree of group identification is necessary for group-based behavior to occur. Research shows that high identifiers react in fundamentally different ways than low identifiers. For example, Wann and Branscombe (1990) showed that fans who identified strongly with their team (the "die-hard" fans) were more likely to stick with their team even when it was threatened by defeat or a poor record compared to low identifiers (the "fair-weather" fans). Furthermore, when an important social identity is threatened, high identifiers are consistently more likely than low identifiers to see both groups as more homogeneous (Doosje, Ellemers, & Spears, 1995), perceive themselves as more prototypical of the ingroup (Spears et al., 1997), and reject deviant ingroup members (Branscombe, Wann, Noel, & Coleman, 1993). High identifiers are thus more likely to embrace their group, particularly when it needs their support, whereas low identifiers adopt a more instrumental and individualistic stance, distancing themselves from the group when it suits their personal interests or when it might represent a threat to their personal identity. The correspondence between such cognitive effects as self-stereotyping and perceived group homogeneity which are used to preserve the ingroups distinctiveness with indicators of group behavior such as intergroup reward allocations as we have outlined, once again emphasizes the close link between social cognition and intergroup processes. Such cognitive changes facilitate ingroup cohesiveness among those who are highly identified. These changes also encourage the use of group-level strategies, such as collective action, to change the intergroup relational status quo.

More often than not, social psychologists have been tempted to reduce other aspects of group behavior to individualistic processes that fail to capture the social nature of the phenomenon under study. Perhaps the classic case of this involves the explanation of crowd behavior in terms of deindividuation theory (Festinger, Pepitone, & Newcomb, 1952; Zimbardo, 1969; Diener, 1980; Prentice-Dunn & Rogers, 1989). This theoretical tradition draws directly on the earlier writings of Le Bon (1895/1995) who described the individual in the group or crowd as being reduced to an "inferior form of evolution" where all normative restraints on behavior were stripped away. Although deindividuation theory has been repeatedly reformulated, this basic anti-social view of the individual in the crowd has remained a central feature. According to the most contemporary account, deindividuation is a state of diminished private self-awareness coupled with arousal caused by immersion in the group (Prentice-Dunn & Rogers, 1989). This state has been assumed to lead to disinhibited behavior and violations of social norms.

In a critique of the deindividuation explanation of crowd behavior, Reicher et al. (1995) provided an alternative account based on social identity theory (the social identity model of deindividuation effects or the "SIDE" model). According to this view, anonymity and immersion in the group do not result in a loss of the self (as argued by classical deindividuation theory), but serve to enhance the salience of social identity at the expense of personal identity. As a result of such an identity shift, conformity to group norms is enhanced rather than diminished. However, this approach makes a distinction between sensitivity to the local social norms that are active in a given research context and more generic societal norms (which define aggressiveness as generally antisocial). This distinction between what was normatively appropriate in the experimental contexts employed in the original deindividuation research and broader social norms was not taken into account.

Consider the classic deindividuation paradigm, where participants were dressed in hoods to make them feel anonymous. They were then required by the experimenter to administer electric shocks to another participant who was actually a confederate (Zimbardo, 1969). It was assumed that administering shocks is characteristic of aggression and is, therefore, anti-normative. However, this interpretation fails to take into account the meaning from the participants viewpoint, given the contextual demands that required them to administer shocks to another person. Indeed, an equally well-known line of research on obedience to authority, where participants were also required to administer shocks to another person, has been used as proof of compliance to an authority figure (Milgram, 1974) and not as evidence of socially unregulated behavior. The local normative demands that were operative in deindividuation studies appear to be rather different than those that might be found in other social contexts. To explore this possibility, in a meta-analysis of the all the relevant deindividuation studies, independent raters rated both the local norms governing the experimental setting as well as more general societal norms that might apply (Postmes & Spears, 1998). The evidence obtained overwhelmingly supported the normative analysis based on the SIDE model: conformity to the specific social norms operating in prior deindividuation studies best accounted for the outcomes that had been obtained. Thus, people's interpretation of the situation they are faced with is guided by contextually derived social norms.

Group norms have been postulated to play an important role in moderating intergroup behavior such as ingroup favoritism (Turner, 1987). There is a close link between social identification and conformity to group norms. In a series of studies, Jetten, Spears, and Manstead (1996) manipulated ingroup and outgroup norms orthogonally to create a discrimination or fairness norm. Results showed that group members conformed to the ingroup norm, even when it dictated fairness to the outgroup. This finding suggests that group norms have the power to moderate the extent to which ingroup bias occurs. Indeed, people who identified highly with their group were particularly likely to conform to their group's norm, whether it was one of discrimination or fairness (Jetten, Spears, & Manstead, 1997). A major question that might profitably be pursued in future research is how "fairness" social norms can be more broadly created. According to this analysis, such norms may well be critical if we hope to reduce social discrimination.

In existing models of persuasion and attitude change, intrapsychic processes have been emphasized more strongly than has the communicative context. In the dual process models of persuasion that we described earlier, the group membership of the source was conceptualized as a peripheral cue that can lead to weak and ephemeral social influence. By contrast, cognitive elaboration was expected to operate on the central arguments of the message. Likewise, classical perspectives on social influence subscribe to the view of the group as exerting an external pressure to comply ("normative influence"), which does not necessarily lead to internalized or "true" influence (Deutsch & Gerard, 1955). However, the self-categorization approach to social influence accords the group a more central and influential role because its norms can be central to how we define ourselves. Therefore, knowing what groups people identify with can help us understand what norms people are likely to have internalized (Turner, 1991). In this sense, the group is not peripheral but is central to our understanding of when and how social groups exert an influence on the individual.

As our selective summary of research on intergroup relations has illustrated, the operation of the same social principles can be observed as when interpersonal and intrapsychic processes are examined. Our overview emphasized the importance of the social context, how the self is categorized, and the implications for social behavior that follow from conformity to various types of social group norms. Social psychology has become sufficiently mature to generate its own theoretical explanations that are uniquely social, and the integrative framework provided by social identity and self-categorization theories offers a useful set of principles for understanding seemingly widely disparate social phenomena. Indeed, rather than social psychological phenomena originating in basic cognitive processes or capacity limitations, the research on intrapsychic processes that we discussed suggests that many apparently basic cognitive and perceptual processes may themselves be socially mediated by level of self-definition and social contextual factors. Research in all three of the major research areas of social psychology has provided evidence for the operation of these fundamentally social influences. Using the principles stemming from social identity and self-categorization theories allowed us to survey social psychology in a fairly integrated and interrelated fashion, as a distinctive and unitary discipline. Although there have been other candidates for integration and unification, either "from below" (neuroscience, connectionism) or from "our past" (evolutionary psychology), by their very nature they fail to do full justice to the social psychological nature of the phenomena investigated.

Into the New Millennium

All three of the different unifying theoretical approaches to social psychology that we discussed (connectionism and information processing, evolutionary theory, social identity and self-categorization) are likely to receive increasing research attention as we move into the new millennium, in part because of their ability to order seemingly disparate findings. As the methods and models within cognitive psychology become more sophisticated and plausible neural pathways are developed, some social psychologists will increasingly pursue this route. Similarly, although evolutionary approaches to social psychology are as yet in their infancy, the indications are that their influence will grow as advocates try to push the limits of this level of explanation as far as possible and provide further links to the biological sciences. Because we believe that the richness of social behavior is inherently influenced by people's social identities and the agendas dictated by those groups' norms, we expect that research illustrating how behavior is shaped by a host of economic, cultural, and historical conditions will increase as well. Such effects, as they emerge, are unlikely to be reducible to the micro-mechanisms of the neural network or the macro-mechanisms of evolutionary selection pressures.

The challenges of social psychology in the new millenium should be especially attractive to scientists who have the capacity to synthesize findings from diverse theoretical approaches. Graduate students looking for an academic home in social psychology will need to be able to use theory in a fluid and integrative manner to bring the best elements of relevant theories to bear on the social puzzles before us.

We have chosen to focus on one important representative of a social psychological level of analysis, but there are also other equally "social" candidates that we have neglected and that may gain in influence in the future. In particular, social psychological approaches, such as the discursive and social constructionist traditions which have hitherto been somewhat separated from mainstream experimental research, should continue to thrive (e.g., Gergen, 1991; Potter & Wetherell, 1987). These research traditions have diverse roots, notably in ethnomethodology within sociology, and speech act theory and semiotics within linguistics and philosophy. There are also close connections between constructionism and the more "sociological" branch of social psychology, the "symbolic interactionist" tradition. Social constructionist approaches tend to focus on how social reality is constructed through linguisitic and discourse processes. In its more radical form, this view questions whether there is indeed a reality "beyond the text" or outside of linguistic constructions (postmodernist philosophy has been very influential here). In methodological terms, this tradition has been critical of experimental means of imposing a particular construction of reality on participants, and for assuming that it can access some underlying truth beyond the understandings and discourses of the participants involved. In response, mainstream social psychologists have criticized those in the social constructionist and discourse camps for the relativism and indeterminacy of its position and for questioning the value of quantitative methods of assessing underlying causal relations (Spears, 1997). These criticisms of the experimental method have doubtless contributed to their current lack of impact on mainstream social psychological research.

The issue of how meaning is socially constructed has, for the most part, been neglected in mainstream social psychology and it may well become an important problem around which integrative progress can occur. Some attempts are already evident. Research within the social cognition tradition has begun to reveal the value of incorporating constructionist principles in ways that use rather than oppose standard scientific paradigms. Research on "construal" processes (Griffin & Ross, 1991) has shown that the objects of perception and judgment should not be taken as givens. Indeed, the way in which we construe (or "construct") an object, has important consequences for judgment and behavior. For example Asch (1952) showed that interpretation of and agreement with the statement "a little rebellion now and then is a good thing" was dramatically affected by whether its source was cited as Thomas Jefferson or V.I. Lenin. Such differential interpretations are socially shared and socially constructed as well as cognitively construed. In this sense, social constructionism has the capacity to add an even more social dimension to social psychology. Similarly, recent research examining communication principles and "conversational norms" has highlighted their importance for intrapsychic processes such as attribution (Hilton & Slugoski, in press). Thus, many of our existing experimental results may be the product of how people make sense of the experimental context according to the pragmatic rules of communication, rather than reflecting the direct and unmediated products of cognition. Sensitivity to such processes will help us not only better distinguish artifact from reality, but also to understand better the communicative dimensions of social reality.

In terms of the social identity and self-categorization traditions that have formed our main thematic focus, there is also scope for integration with a more constructionist perspective. There is a tendency for these theoretical traditions to take the definition of self and social contexts as "givens" whereas these too are to some extent negotiated and contested in social discourse. For example, whether we view a rioting crowd as a "mad mob" or as a group victimized by police, is a matter of how we construct this event. These differing constructions are influenced by self definition and social context (Reicher, 1995). How we see the event is dependent on whether we define ourselves as peaceful demonstrators who are being attacked by the police, or as neutral bystanders who are witnessing an intergroup conflict. In fact, such bystanders may construct the event differently than those who merely read about it in newspapers which present it through the eyes of journalists who use certain discourses to define the crowd in terms akin to those used by Le Bon. In short, there may be multiple possible constructions of the same event, with each capable of influencing our evaluation of it. Greater recognition of the constructed nature of social categories and situations may provide one of the missing links in self-categorization research in particular, and in mainstream social psychology in general.

Developments That Will Further Link the Three Social Psychological Research Areas

As we described already, degree of identification with one's group is an important predictor of group behavior when the relevant identity is salient. We argued that participants in minimal groups studies were attempting to achieve a positively distinct social identity. However, research has increasingly begun to look at the role of intragroup dynamics as contributors to intergroup behavior. Factors that can moderate the occurrence of intergroup behavior include the individual's personal status or ability (McFarland & Buehler, 1995; Seta & Seta, 1996), the personal self-esteem level of the individual (Long & Spears, 1997), and the degree of perceived respect received from other ingroup members (Smith & Tyler, 1997; Tyler, Degoey, & Smith, 1996). Research linking intragroup dynamics and intergroup behavior reveals that depending on one's position within a group (be it peripheral or central), behavior can reflect either intragroup concerns or more group-level purposes (Branscombe et al., 1998; Noel et al., 1995). An outwardly appearing identical behavior (discrimination against an outgroup) can be either personally instrumental--engaged in as a means of currying favor for the self with other ingroup members, or it can be enacted in the service of the group's needs out of genuine commitment to the group and its goals. Although these studies illustrate that intergroup behavior can be quite complex and can derive from intragroup as well as intergroup agendas, additional research examining how contextual factors alter their relative weight is needed.

We also perceive a need for additional research linking traditonal intrapsychic topics, such as attribution with interpersonal and intergroup processes. Attribution theory has focused primarily on how individuals themselves, as isolated agents, construct explanations for events. However, others may exert important influences on the nature of the explanation for an outcome that is ultimately deemed to be the most plausible one. Research concerning normative influences on acceptance of group-based explanations for events (i.e., discrimination), and how they come to be socially validated or not, could greatly expand our understanding of the development and promulgation of social ideologies. Furthermore, because individuals can and do categorize themselves at different identity levels depending on the nature of the social context, the explanations that they are willing to accept for the same outcome might vary accordingly. Consider the attributional dilemma of a job candidate who applies for a position but is not selected. Are there normative supports available for this person to consider the possibility that group-based discrimination may have played a role in producing this outcome, or is such an explanation seen as a socially undesirable one to voice? Does the degree of social influence on such explanatory processes depend on whether an ingroup member or an outgroup member suggests that discrimination may have played a causal role in producing the outcome? Does the explanation that is suggested by others encourage consideration of how "people like us" have been historically treated by "people like them," or does it discourage such possibilities by pointing to more individualistic factors (Smith & Spears, 1996)? Because quite different behaviors may be dictated depending on who suggests what type of explanation best accounts for a given outcome, the potential impact of such normative factors deserves additional research attention in the future.

Applied or Social Issue Topics That Will be Increasingly Pursued
Opportunities to contribute in meaningful ways to understanding group phenomena linking these three approaches are plentiful. We will suggest many important areas that could serve as dissertation topics for graduate students or the start of important research streams for beginning researchers.

In the foregoing sections we have tried to show how intergroup conflict and discrimination are not necessarily inevitable. We believe that appropriate group norms, ideologies, and values may ultimately help to keep the peace. With the changing structure of the American population in terms of increasing cultural diversity and increasingly global interactions, the possibilities for interacting with different sorts of people have soared (Gergen, 1991). Such interactions mean that wider and more diverse types of social contacts and influences can be expected, providing for both the possibility of greater integration based on diversity and acceptance of cultural differences, although it also raises the potential for increased intergroup tension.

The social identity tradition in social psychology is one perspective that has encouraged the discipline to take a new look at more macro-level social concerns such as cultural diversity, multiculturalism, and the nature of power. Research has already begun to examine how cultural factors can influence social perception and behavior. The roles that power and status play via memberships in dominant and devalued social groups is also likely to become an increasingly important research topic (Branscombe & Ellemers, 1998; Fiske, 1993). We are seeing a new interest in the psychological experience and perspective of members of devalued social groups (Crocker & Major, 1989; Steele, 1997) and the response strategies available to them when they face social discriminatory treatment. Thus, we predict that research will increasingly employ members of devalued groups as participants. The tenability of the existing assumption that all people, regardless of their place in the social structure, respond similarly to the social conditions they find themselves in is beginning to be questioned. To take one recent example, Branscombe, Schmitt, and Harvey (in press) distinguished between the psychological responses of African-Americans who differ in terms of how pervasive they perceive prejudice to be, and how those are likely to differ from dominant group members who perceive themselves to be victims of discrimination. Historically disenfranchised group members are more likely to perceive the outgroups discrimination as stable and pervasive, and this perception encourages such persons to turn toward their minority group as a means of protecting their well-being. Because the US is becoming an increasingly ethnically diverse society, such variations in willingness to define the self in terms of a minority group membership and the social interactional consequences of that choice should receive greater attention in future research.

The face of the population is changing in other important ways that are likely to have consequences for social psychological research in the future. As a consequence, creative students will have ample problems to study in this area. One demographic change that is likely to exert a widespread social impact concerns the age profile of the population. As a result of increases in life expectancy, an increasing proportion of the population is elderly, with those over 85 being the fastest growing group (Hansson, 1989). As the number of older Americans grows, stereotyping and intergroup conflict research based on age groups will also come increasingly to the fore. In fact, the National Science Foundation has recently announced a new grant initiative aimed at integrating existing social psychological research with aging issues. Although age group membership is a particularly interesting example of a group conflict situation because it is the only one where we rotate through the different categories (see Snyder & Miene, 1994), the possibility of intergenerational conflict is nevertheless quite real. Increased dependence by this growing segment of the population on public funds from social security and Medicare, combined with political organizations that present the elderly as taking resources from the young reflects the need for research that can assist in the management of this social change. Furthermore, a focus on aging could act as an impetus for work on multiple category memberships. As Hansson (1989) has noted, the aging experience is fundamentally dependent on other social group memberships. It is primarily White Americans, rather than minority group members, who constitute the bulk of the elderly population, especially those who are relatively well-off financially. Likewise, the number of elderly women is growing at a much faster rate than the number of men. Thus, existing group conflicts based on ethnicity and gender may be played out in terms of stereotyping and prejudice against the elderly. Therefore, solutions will require a more solid understanding of how one kind of social category membership can influence the ongoing impact of another.

More generally, we see an important need for research in the future on the role of multiple group memberships. We currently have only a minimal understanding of what influences when people will perceive others in terms of one or another of their group identities, or if the processes differ for intersecting category memberships. Furthermore, identities are rarely experienced as a constant across time, and how some come to gain importance while others decrease has received little attention. As people navigate their way through life, they may encounter circumstances in which desirable new identities can be added (Ethier & Deaux, 1990), and the loss of others must be coped with. As Breakwell's (1986) pioneering work on people facing unemployment, divorce, and immigration illustrates, the psychological impact of such profound identity changes depends on the degree to which they are voluntary or externally imposed. In addition, the degree to which people anticipate being accepted into and receiving positive treatment based on their new group membership is likely to be critical for adjustment.

Perhaps because social psychology has been historically wedded to the undergraduate as research participant (see Sears, 1987), we have not as yet fully appreciated how group commitments may shift across the lifespan. Addressing additional questions such as why people might choose to exhibit negatively evaluated identities with visible markers (e.g., tattoos, body piercings) deserves attention. When such markers of group membership will come to be widely accepted throughout the culture and when they will remain indicators of "fringe" status could be used to investigate broader questions concerning the processes of social diffusion and social change. Similarly, gaining an understanding of what circumstances and why people might attempt to hide their group membership (e.g., homosexuals, signs of aging) will require us to seek research participants outside the academy. As a starting place, we suggest that some people may choose to align themselves with social groups that are negatively evaluated because they reject the standards of the "mainstream" (potentially as a result of feeling rejected by it), and they wish to convey that they are not "one of them." Other people may move toward an alternative group identity because they do identify with that group's norms and values, and they wish to publicly express their alignment with it. The possibility that differing social motivations may be crucial for joining different social groups or in different members of the same group cannot be assessed without research on actual members of such groups.

Just as important as the social and demographic changes that are ongoing in society are the technological changes that accompany them. These changes will provide countless research ideas for enterprising students. For example, increasing automation of production and so forth makes possible alternative ways of organizing work and leisure, as well as altering the structure of the relations between people. Perhaps the most significant change in this area in the last 10 years, and one that will continue to develop, concerns the new communications technologies such as e-mail and widespread access to the internet. Increasing numbers of people are interacting with one another by means of text-based computer-mediated communication and standard on-line video links are not far off. Once again social psychological research can make a contribution to our understanding of the effects that these new technologies can exert on social interaction and organization.

The proliferation of communications media means that we are potentially inundated with information on a scale not before known, leading to what one author has referred to as the "saturated self" (Gergen, 1991). Many people may experience the infinite choices provided by these media as overwhelming and find the shifting skills required to manage these technologies as psychologically stressful. More positively, the range of choices provided by these media, and the Internet in particular, means that people can channel and tailor them to their own work, consumer, and entertainment needs. The day when the mass media are as varied as the groups and individuals that they address is not too far around the corner. The "agenda setting" influence of the mass media may increasingly disappear if individuals can set their own agendas to suit their existing political preferences and tastes. Such flexibility may, however, come at the cost of social fragmentation if individuals increasingly withdraw from the real social world into their virtual ones.

However, these virtual worlds do offer new ways of being and relating to others. In cyberspace, there are new possibilities afforded for identity construction when freed from the constraints of time, distance, and personal appearance. The Internet provides a new medium for interpersonal contact that is no less intimate than face-to-face communication and sometimes is more so (Lea & Spears, 1995). Indeed, these new communications media may be experienced by some as personally liberating in so far as they allow for increasing control of our information consumption and our interpersonal relations.

Nevertheless, there is also no guarantee that access to these technologies will be equally distributed and that existing power and status gaps may be reproduced in terms of differential access, with the opportunities provided by them exacerbating the gap between the haves and the have nots. Moreover, even for those persons who do have access, there is some evidence that features of the Internet (anonymity and isolation) can paradoxically reinforce social boundaries rather than break them down. Although some researchers have argued that the relative absence of social cues can undermine status differences and lead to more equalized and democratic participation (e.g., Kiesler, Siegel, & McGuire, 1984), this view is not universally shared. Research in the tradition of the social identity model of deindividuation effects that we discussed earlier suggests that such interaction may be more bounded by social context and relational factors than previously thought. Thus, according to this model, the visual anonymity associated with such media can reinforce the impact of social identities, social norms, and the operation of existing power relations, compared to more face-to-face communication where individual differences are more salient (Spears & Lea, 1994). Simply because people interact with each other less directly than in face-to-face communication, does not render an analysis in terms of social psychological principles any less relevant or applicable.

Although predicting the future of any human endeavor can be a risky business, we have pointed to what we believe will be important directions for social psychological research in the future. By first examining important existing threads of research, within an integrative framework, we were able to project those strands forward to provide a peek at what the social psychological quilt might look like in the future. What can be confidently predicted is increasing theoretical integration across the three traditional areas of social psychology in terms of common mechanisms and a continuing concern with addressing applied social problems as they emerge. As such, students capable of sophisticated theory skills will not only have a variety of opportunities from which to choose, but also may be able to make a substantial impact on the serious social problems that we will face in the next century.

References

Adorno, T.W., Frenkel-Brunswik, E., Levinson, D.J., & Sanford, R.N. (1950). The authoritarian personality. New York: Harper.

Altemeyer, B. (1987). Enemies of freedom: Understanding right-wing authoritarianism. San Francisco, CA: Jossey-Bass.

Asch, S. (1952). Social psychology. New York: Prentice Hall.

Bargh, J. (1994). The four horsemen of automaticity: Awareness, intention, efficiency, and control in social cogntion. In R.S. Wyer, Jr., & T.K. Srull (Eds.). Handbook of social cognition (2nd ed.) (pp.1-40). Hillsdale, NJ: Erlbaum.

Batson, C.D. (1987). Prosocial motivation: Is it ever truly altruistic? In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 65-122). New York: Academic Press.

Baumeister, R. (1991). The meanings of life. New York: Guilford Press.

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529.

Berndsen, M., Spears, R., McGarty, C., & van der Pligt, J. (1998). Dynamics of differentiation: Similarity as a precursor and product of stereotype formation. Journal of Personality and Social Psychology, 74, 1451-1463.

Billig, M. (1976). Social psychology and intergroup relations. London: Academic Press.

Bless, H., Strack, F., & Schwarz, N. (1993). The informative functions of research procedures: Bias and the logic of conversation. European Journal of Social Psychology, 23, 149-166.

Bodenhausen, G.V. (1990). Stereotypes as judgmental heuristics: Evidence of circadian variations in discrimination. Psychological Science, 1, 319-322.

Bornstein, R.F., Leone, D.R., & Galley, D.J. (1987). The generalizability of subliminal mere exposure effects: Influence of stimuli perceived without awareness on social behavior. Journal of Personality and Social Psychology, 53, 1070-1079.

Branscombe, N.R. (1998). Thinking about one's gender group's privileges or disadvantages: Consequences for well-being in women and men. British Journal of Social Psychology, 37, 167-184.

Branscombe, N.R., & Cohen, B.M. (1990). Motivation and complexity levels as determinants of heuristic use in social judgment. In J.P. Forgas (Ed.), Emotion and social judgment (pp. 145-160). Oxford: Pergamon Press.

Branscombe, N.R., & Ellemers, N. (1998). Coping with group-based discrimination: Individualistic versus group-level strategies. In J.K. Swim & C. Stangor (Eds.), Prejudice: The target's perspective (pp. 243-266). New York: Academic Press.

Branscombe, N.R., N'gbala, A., Kobrynowicz, D., & Wann, D.L. (1997). Self and group protection concerns influence attributions but they are not determinants of counterfactual mutation focus. British Journal of Social Psychology, 36, 387-404.

Branscombe, N.R., Schmitt, M.T., & Harvey, R.D. (in press). Perceiving pervasive discrimination among African-Americans: Implications for group identification and well-being. Journal of Personality and Social Psychology.

Branscombe, N.R., Spears, R., Ellemers, N., & Doosje, B. (1998). Ask not what your group can do for you, ask what you can do for it: Respect and prestige as determinants of group affiliation and behavior. Manuscript submitted for publication.

Branscombe, N.R., & Wann, D.L. (1994). Collective self-esteem consequences of outgroup derogation when a valued social identity is on trial. European Journal of Social Psychology, 24, 641-657.

Branscombe, N.R., Wann, D.L., Noel, J.G., & Coleman, J. (1993). In-group or out-group extremity: Importance of the threatened social identity. Personality and Social Psychology Bulletin, 19, 381-388.

Breakwell, G.M. (1986). Coping with threatened identities. New York: Methuen.

Brewer, M. B. (1988). A dual process model of impression formation. In T. K. Srull & R. S. Wyer (Eds.) Advances in social cognition (Vol. 1, pp. 1-36). Hillsdale, NJ: Erlbaum.

Brigham, J.C. (1971). Ethnic stereotypes. Psychological Bulletin, 76, 15-33.

Buss, D.M. (1995). Psychological sex differences: Origins through sexual selection. American Psychologist, 50, 164-168.

Byrne, D., Clore, G.L., & Smeaton, G. (1986). The attraction hypothesis: Do similar attitudes affect anything? Journal of Personality and Social Psychology, 51, 1167-1170.

Chen, S., Shechter, D., & Chaiken, S. (1996). Getting at the truth or getting along: Accuracy and impression motivated heuristic and systematic processing. Journal of Personality and Social Psychology, 71, 262-275.

Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of stigma. Psychological Review, 96, 608-630.

Deaux, K., & Lewis, L.L. (1984). Structure of gender stereotypes: Interrelationships among components and gender label. Journal of Personality and Social Psychology, 46, 991-1004.

Deutsch, M. & Gerard, H.B. (1955). A study of normative and informational influences upon individual judgement. Journal of Abnormal and Social Psychology, 51, 629-36.

Devine, P.G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18.

Diener, E. (1980). Deindividuation: The absence of self-awareness and self-regulation in group members. In P.Paulus (Ed.), The psychology of group influence (pp. 209-242). Hillsdale, NJ: Erlbaum.

Doosje, B., Branscombe, N.R., Spears, R., & Manstead, A.S.R. (1998). Guilty by association: When one's group has a negative history. Journal of Personality and Social Psychology, 75, 872-886.

Doosje, B., & Ellemers, N. (1997). Stereotyping under threat: The role of group identification. In R. Spears, P.J. Oakes, N. Ellemers, & S.A.Haslam (Eds.), The social psychology of stereotyping and group life (pp. 257-272). Oxford, UK: Blackwell.

Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psychology, 31, 410-436.

Eagly, A.H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace Jovanovich.

Eagly, A.H., & Mladinic, A. (1994). Are people prejudiced against women? Some answers from research on attitudes, gender stereotypes, and judgements of competence. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 1-35). Chichester, UK: John Wiley & Sons.

Ellemers, N. (1993). The influence of socio-structural variables on identity management strategies. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 27-58). Chichester, UK: John Wiley & Sons.

Ethier, K.A., & Deaux, K. (1990). Hispanics in ivy: Assessing identity and perceived threat. Sex Roles, 22, 427-440.

Festinger, L., Pepitone, A., & Newcomb, T. (1952). Some consequences of deindividuation in a group. Journal of Abnormal and Social Psychology, 47, 382-389.

Fiedler, K. (1990). Mood-dependent selectivity in social cognition. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 1, pp. 1-32). New York: Wiley.

Fiske, S. T. (1993). Controlling other people: The impact of power on stereotyping. American Psychologist, 48, 621-628.

Fiske, S.T., & Neuberg, S.E. (1990). A continuum of impression formation from category-based to individuating processes: Influences of information and motivation on attention and interpretation. In M.P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 1-74). New York: Academic Press.

Fiske, S.T., & Pavelchak, M.A. (1986). Category-based versus piecemeal-based affective responses: Developments in schema-triggered affect. In R.M. Sorrentino & E.T. Higgins (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2, pp. 167-203). New York: Guilford Press.

Fiske, S.T., & Taylor, S.E. (1991). Social cognition (2nd Ed.). New York: McGrawHill.

Fong, G.T., Krantz, D.H., & Nisbett, R.E. (1986). The effects of statistical training on thinking about everyday problems. Cognitive Psychology, 18, 253-292.

Forgas, J.P. (1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin, 117, 39-66.

Geen, R. (1990). Human aggression. Pacific Grove, CA: Brooks/Cole.

Gergen, K.J. (1973). Social psychology as history. Journal of Personality and Social Psychology, 26, 309-320.

Gergen, K.J. (1991). The saturated self: Dilemmas of identity in contemporary life. New York: Basic Books.

Gilbert, D.T. & Hixon, J.G. (1991). The trouble of thinking: Activation and application of stereotypic beliefs. Journal of Personality and Social Psychology, 60, 509-517.

Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117, 21-38.

Gilbert, D.T., Pelham, B.W., & Krull, D.S. (1988). On cognitive busyness: When person perceivers meet persons perceived. Journal of Personality and Social Psychology, 54, 733-739.

Griffin, D.W., & Ross, L. (1991). Subjective construal, social inference, and human misunderstanding. In: Zanna, M.P. (Ed.), Advances in experimental social psychology (Vol. 24, pp. 319-359). San Diego, CA: Academic Press.

Griffin, E., & Sparks, G.G. (1990). Friends forever: A longitudinal exploration of intimacy in same-sex friends and platonic pairs. Journal of Social and Personal Relationships, 7, 29-46.

Hamilton, D. L. (1981). Stereotyping and intergroup behavior: Some thoughts on the cognitive approach. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 333-354). Hillsdale, NJ: Erlbaum.

Hamilton, D.L., & Sherman, S.J. (1989). Illusory correlations: Implications for stereotype theory and research. In D. Bar-Tal, C.F. Graumann, A.W. Kruglanski, & W. Stroebe (Eds.), Stereotypes and prejudice: Changing conceptions (pp. 59-82). New York: Springer-Verlag.

Hansson, R.O. (1989). Old age: Testing the parameters of social psychological assumptions. In S. Spacapan & S. Oskamp (Eds.), The social psychology of aging (pp. 25-51). Newbury Park, CA: Sage.

Hardin, C., & Higgins, E.T. (1996). Shared reality: How social verification makes the subjective objective. In R.M. Sorrentino & E.T. Higgins (Eds.), Handbook of motivation and cognition: Foundations of social behavior (pp. 28-84). New York: Guilford.

Haslam, S.A. (1997). Stereotyping and social influence: Foundations of stereotype consensus. In R. Spears, P.J. Oakes, N. Ellemers, & S.A.Haslam (Eds.), The social psychology of stereotyping and group life (pp. 119-143). Oxford, UK: Blackwell.

Haslam, S.A., Turner, J.C., Oakes, P.J., McGarty, C., & Hayes, B.K. (1992). Context-dependent variation in social stereotyping 1: The effects of intergroup relations as mediated by social change and frame of reference. European Journal of Social Psychology, 22, 3-20.

Hatfield, E., & Sprecher, S. (1986). Mirror, mirror: The importance of looks in everyday life. Albany, NY: State University of New York Press.

Hatfield, E., Traupmann, J., Sprecher, S., Utne, M., & Hay, J. (1985). Equity and intimate relations: Recent research. In W. Ickes (Ed.), Compatible and incompatible relationships. (pp. 91-118). New York: Springer-Verlag.

Higgins, E.T., & Bargh, J.A. (1987). Social cognition and social perception. In M.R. Rosenweig & L.W. Porter (Eds.), Annual review of psychology (Vol. 38, pp. 369-425). Palo Alto, CA: Annual Reviews.

Hilton, D., & Slugoski, B.R. (in press). Judgement and decision-making in social context: Discourse processes and rational inferences. In T. Connolly (Ed.), Judgment and decision making: An interdisciplinary reader. Cambridge, UK: Cambridge University Press.

Hogg, M.A. (1992). The social psychology of group cohesiveness: From social attraction to social identity. Hemel Hempstead, UK: Harvester Wheatsheaf.

Hogg, M.A., & Hains, S.C. (1997). Intergroup relations and group solidarity: Effects of group identification and social beliefs on depersonalized attraction. Journal of Personality and Social Psychology, 70, 295-309.

Hoveland, C.I., Janis, I.L., & Kelley, H.H. (1953). Communication and persuasion: Psychological studies in opinion change. New Haven, CT: Yale University Press.

Isen, A.M. (1987). Positive affect, cognitive processes, and social behavior. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 203-253). New York: Academic Press.

Jackman, M.R. (1994). The velvet glove: Paternalism and conflict in gender, class, and race relations. Berkeley, CA: University of California Press.

Jetten, J., Spears, R., & Manstead, A.S.R. (1996). Intergroup norms and intergroup discrimination: Disti