Judith A Howard & Danielle Kane. Encyclopedia of Sociology. 2nd edition, Volume 1, Macmillan Reference USA, 2001.
Attribution is a cognitive process that entails linking an event to its causes. Attribution is one of a variety of cognitive inferences that are included within social cognition, which is one of several theoretical models within social psychology. Social cognition has been the most dominant social psychological perspective within psychology since the 1960s, and this is evident in the popularity of research on attribution. In the mid-1970s, as much as 50 percent of the articles in major social psychology journals concerned attributional processes, in part because attribution theory is relevant to the study of person perception, event perception, attitude change, the acquisition of self-knowledge, and a host of applied topics including therapeutic interventions, close relationships, legal and medical decision making, and so forth. Although the proportion of published research that focused on this topic declined during the 1980s, attribution remains one of the more popular fields of social psychological research.
An attribution is an inference about why an event occurred. More generally, “attribution is a process that begins with social perception, progresses through a causal judgment and social inference, and ends with behavioral consequences” (Crittenden 1983, p. 426). Although most theories of and research on attribution focus on causal inference, empirical research has dealt with attributions not only of cause but also of blame and responsibility. Although these types of attributions are closely related, they are not conceptually identical. Furthermore, because personality characteristics constitute a major category of potential causes of behavior, attributions about individuals’ traits (both one’s own traits and those of others) have received explicit theoretical attention.
Major Theories of Attribution
Even though attribution has been one of the most popular social psychological research topics in the social sciences, only a few theories of attribution have been developed. The study of attribution began with Fritz Heider’s (1958) original attempt to provide a systematic, conceptual explanation of “naive” psychology. Heider maintained that people strive to understand, predict, and control events in their everyday lives in much the same way as scientists do in their professional lives. On the basis of observation, individuals form theories about their social worlds, and new observations then serve to support, refute, or modify these theories. Because people act on the basis of their beliefs, Heider argued that it is important to understand this layperson’s psychology. Although Heider did not develop an explicit theory of attribution, he did assert several principles that have guided all subsequent theorizing on this topic.
Primary among these principles is the notion that people are inclined to attribute actions to stable or enduring causes rather than to transitory factors. Heider also stressed the importance of distinguishing unintentional from intentional behavior, a distinction that has been particularly influential in theories of the attribution of responsibility. He identified environmental and personal factors as two general classes of factors that produce action and hypothesized that an inverse relationship exists between these two sets of causes. He also suggested that the “covariational principle” is fundamental to attribution: An effect is attributed to a factor that is present when the effect is present and to a factor that is absent when the effect is absent. Heider’s early analyses of social perception represent a general conceptual framework about common sense, implicit theories people use in understanding events in their daily lives. The two most influential theories of attribution are based on Heider’s work but go beyond it in the development of more systematic statements about attributional processes.
Covariational model. Harold Kelley’s (1967, 1973) covariational model of attribution addresses the question of whether a given behavior is caused by an actor or, alternatively, by an environmental stimulus with which the actor engages. According to this model, the attribution of cause is based on three types of information: consensus, distinctiveness, and consistency. Consensus refers to the similarity between the actor’s behavior and the behavior of other people in similar circumstances. Distinctiveness refers to the generality of the actor’s behavior: Does she or he behave in this way toward stimuli in general, or is the behavior specific to this stimulus? Consistency refers to the actor’s behavior toward this stimulus across time and modality. There are many possible combinations of these three types of information, but Kelley makes explicit predictions about just three. The combination of high consensus, high distinctiveness, and high consistency supports an attribution to the environmental stimulus, whereas a profile of low consensus, low distinctiveness, and high consistency supports an attribution to the actor. When the behavior is inconsistent, regardless of the level of consensus or distinctiveness, an attribution to circumstances is predicted.
Empirical tests of Kelley’s model have focused either on the effects of a particular type of information or, more in keeping with his formulation, on the effects of particular patterns of information (McArthur 1972). In an innovative analysis, Miles Hewstone and Jos Jaspars (1987) proposed a different logic (although one consistent with Kelley’s model), suggesting that potential attributers consider whether different causal loci are necessary and sufficient conditions for the occurrence of an effect. They conclude that the notion of causality is flexible and thus assert that there may be some advantage to conceiving of situation-specific notions of causality. More recent work has examined the universality and external validity of Kelley’s model. Irina Anderson and Geoffrey Beattie (1998), for example, analyzed actual conversations between men and women talking about rape. They found that men tended to use the reasoning outlined in Kelley’s model by making reference to consensus, distinctiveness, and consistency, as well as by using these types of information to formulate attributions for behavior. Women, on the other hand, made less use of these variables and introduced the variable of “foreseeability” into their analyses.
Correspondent inference. The theory of correspondent inference (Jones and Davis 1965; Jones and McGillis 1976) addresses the attribution of personality traits to actors on the basis of their behavior and focuses on attributions about persons in greater depth than does Kelley’s covariational model. These two theories thus address different questions. Kelley asks: When do we attribute an event to an actor or to some stimulus in the environment? Edward Jones asks: When do we attribute a trait to an actor on the basis of her or his behavior? The theory of correspondent inference focuses more narrowly on the actor but also yields more information about the actor in that it specifies what it is about the actor that caused the behavior. Jones and his coauthors predict that two factors guide attributions: (1) the attributer’s prior expectancies for behavior, specifically, expectancies based either on knowledge of earlier behaviors of the actor (target-based) or on the actor’s social category memberships (category-based), and (2) the profile of effects that follow from the behavioral choices available to the actor.
Edward Jones and Daniel McGillis propose that expectancies determine the degree of confidence with which a particular trait is attributed; the lower the expectancy of behavior, the more confident the attribution. The profile of effects helps the attributer identify what trait might have produced the behavior in question. Noncommon effects—effects that follow from only one of the behavioral options—provide information about the particular disposition. The fewer the noncommon effects, the clearer the attribution. Thus, behavior that contradicts prior expectancies and a profile of behavioral choices with few noncommon effects combine to maximize the possibility of attributing a disposition to the actor (a correspondent inference). Empirical research generally has supported these predictions.
These two models share some attributional principles. Expectancy variables (target-based and category-based) are analogous to Kelley’s types of information. Although the predicted effects of consensus information and its analogue, category-based expectancies, are compatible, the predicted effects of consistency and distinctiveness information and their analogue, target-based expectancies, present some incompatibilities. This contradiction has been evaluated both conceptually and empirically (Howard and Allen 1990).
Attribution in achievement situations. Bernard Weiner (1974) and his colleagues have applied attributional principles in the context of achievement situations. According to this model, we make inferences about an individual’s success on the basis of the individual’s ability to do the task in question, how much effort is expended, how difficult the task is, and to what extent luck may have influenced the outcome. Other possible causal factors have since been added to this list. More important perhaps is Weiner’s development of, first, a structure of causal dimensions in terms of which these causal factors can be described and, second, the implications of the dimensional standing of a given causal factor (Weiner, Russell, and Lerman 1978). The major causal dimensions are locus (internal or external to the actor), stability or instability, and intentionality or unintentionality of the factor. Thus, for example, ability is internal, stable, and unintentional. The stability of a causal factor primarily affects judgments about expectancies for future behavior, whereas locus and intentionality primarily affect emotional responses to behavior. This model has been used extensively in educational research and has guided therapeutic educational efforts such as attribution retraining. Cross-cultural research has explored the cultural generalizability of these models. Paul Tuss, Jules Zimmer, and Hsiu-Zu Ho (1995) and Donald Mizokawa and David Rickman (1990) report, for instance, that Asian and Asian American students are more likely to attribute academic failure and success to effort than are European American students, who are more likely to attribute performance to ability. European American students are also more likely to attribute failure to task difficulty. Interestingly, as Asian Americans spend more time in the United States, they place less emphasis on the role of effort in performance.
Attribution biases. The theoretical models described above are based on the assumption that social perceivers follow the dictates of logical or rational models in assessing causality. Empirical research has demonstrated, not surprisingly, that there are systematic patterns in what has been variously conceived of as bias or error in the attribution process (Ross 1977). Prominent among these is what has been called the “fundamental attribution error,” the tendency of perceivers to overestimate the role of dispositional factors in shaping behavior and to underestimate the impact of situational factors. One variant of this bias has particular relevance for sociologists. This is the general tendency to make inadequate allowance for the role-based nature of much social behavior. That is, perceivers fail to recognize that behavior often derives from role memberships rather than from individual idiosyncrasy. Again, cross-cultural research has called into question the generalizability of this bias. Joan Miller (1984) shows that the tendency to attribute behavior to persons is markedly more prominent in the United States, among both adults and children, whereas a tendency to attribute behavior to situational factors is more prominent among Indian Hindus, both adults and children, calling into question the “fundamentalness” of this attributional pattern.
A second systematic pattern is the actor-observer difference, in which actors tend to attribute their own behavior to situational factors, whereas observers of the same behavior tend to attribute it to the actor’s dispositions. A third pattern concerns what have been called self-serving or egocentric biases, that is, attributions that in some way favor the self. According to the false consensus bias, for example, we tend to see our own behavioral choices and judgments as relatively common and appropriate whereas those that differ from ours are perceived as uncommon and deviant. There has been heated debate about whether these biases derive from truly egotistical motives or reflect simple cognitive and perceptual errors.
Methodological and Measurement Issues
The prevalent methodologies and measurement strategies within attributional research have been vulnerable to many of the criticisms directed more generally at social cognition and to some directed specifically at attribution. The majority of attributional research has used structured response formats to assess attributions. Heider’s original distinction between person and environmental cause has had a major influence on the development of these structured measures. Respondents typically are asked to rate the importance of situational and dispositional causes of events. These ratings have been obtained on ipsative scales as well as on independent rating scales. Ipsative measures pose these causes as two poles on one dimension; thus, an attribution of cause to the actor’s disposition is also a statement that situational factors are not causal. This assumed inverse relationship between situational and dispositional causality has been rejected on conceptual and empirical grounds. In more recent studies, therefore, respondents assign each type of causality separately. (Ipsative measures are appropriate for answering some questions, however, such as whether the attribution of cause to one actor comes at the expense of attribution to another actor or to society.) In theory, then, both dispositional and situational variables could be identified as causal factors.
The breadth of these two categories has also been recognized as a problem. Dispositional causes may include a wide variety of factors such as stable traits and attitudes, unstable moods and emotions, and intentional choices. Situational cause is perhaps an even broader category. It is quite possible that these categories are so broad as to render a single measure of each virtually meaningless. Thus, researchers often include both general and more specific, narrower responses as possible choices (e.g., choices of attributing blame to an assailant or to the situation might be refined to the assailant’s use of a weapon, physical size, and psychological state, on the one hand, and the location, time of day, and number of people nearby, on the other).
Structured measures of attributions are vulnerable to the criticism that the categories of causes presented to respondents are not those they use in their everyday attributions. Recognizing this limitation, a few researchers have used open-ended measures. Comparative studies of the relative utility of several different types of measures of causal attributions conclude that scale methods perform somewhat better in terms of their inter-test validity and reliability, although open-ended measures are preferable when researchers are exploring causal attributions in new situations. Some researchers have attempted to overcome some of the limitations of existing scales; Curtis McMillen and Susan Zuravin (1997), for example, have developed and refined “Attributions of Responsibility” and “Blame Scales” for use in clinical research that may be useful for other kinds of attributional situations.
The great majority of attribution studies use stimuli of highly limited social meaning. Generally, the behavior is represented with a brief written vignette, often just a single sentence. Some researchers have shifted to the presentation of visual stimuli, typically with videotaped rather than actual behavioral sequences, in order to ensure comparability across experimental conditions. Recognizing the limitations of brief, noncontextualized stimuli, a few researchers have begun to use a greater variety of more extended stimuli including newspaper reports and published short stories. Most of these stimuli, including the videotaped behavioral sequences, rely heavily on language to convey the meaning of behavior. Conceptual attention has turned recently to how attribution relies on language and to the necessity of considering explicitly what that reliance means. Some researchers (Anderson and Beattie 1998; Antaki and Leudar 1992) have attempted to use more naturalistic approaches, incorporating into the study the analysis of spontaneous conversations among study participants.
When Do We Make Attributions?
Long after attribution had attained its popularity in social psychology, a question that perhaps should have been raised much earlier began to receive attention: When do we make attributions? To what extent are the attributions in this large body of research elicited by the experimental procedures themselves? This is a question that can be directed to any form of social cognition. It is particularly relevant, however, to attribution. Most people, confronted with a form on which they are to answer the question “Why?,” do so. There is no way of knowing, within the typical experimental paradigm, whether respondents would make attributions on their own. In a sense there are two questions: Do people make attributions spontaneously, and if they do, under what circumstances do they do so?
In response to the first question, Weiner (1985) has marshaled impressive evidence that people do indeed make attributions spontaneously. Inventive procedures have been developed for assessing the presence of attributional processing that is not directly elicited. This line of evidence has dealt almost entirely with causal attributions. Research suggests that trait attributions may be made spontaneously much more often than causal attributions. In response to the second question, a variety of studies suggest that people are most likely to make attributions when they encounter unexpected events or events that have negative implications for them.
Sociological Significance of Attribution
Attribution is a cognitive process of individuals; much of the extant research on attribution is, accordingly, highly individualistic. In the 1980s, however, researchers began to pay increasing attention to the sociological relevance of attribution. The process of attribution itself is fundamentally social. Attribution occurs not only within individuals but also at the interpersonal, intergroup, and societal levels. Moreover, the process of attribution may underlie basic sociological phenomena such as labeling and stratification.
Interpersonal attribution. At the interpersonal level, attribution is basic to social interaction. Interpersonal encounters are shaped in many ways by attributional patterns. Behavioral confirmation, or self-fulfilling prophecies, illustrate the behavioral consequences of attribution in social interaction; attribution of specific characteristics to social actors creates the expectancies that are then confirmed in behavior. Considering attribution at this interpersonal level demonstrates the importance of different social roles and perspectives (actors vs. observers) as well as how attribution is related to evaluation. The self is also important to the attribution process; the evidence for attributional egotism (self-esteem enhancing attributional biases), self-presentation biases, and egocentrism is persuasive. Attributions also affect social interaction through a widespread confirmatory attribution bias that leads perceivers to conclude that their expectancies have been confirmed in social interaction.
Research has challenged the universality of egocentric biases by examining differences in attributional styles according to race, class, and gender. Several studies (Reese and Brown 1995; Wiley and Crittenden 1992; Broman 1992; Andrews and Brewin 1990) have shown that women are less likely than men to make self-serving dispositions. In a study of academics’ accounts of their success in the profession, Mary Glenn Wiley and Kathleen Crittenden (1992) argue that women explain their success in a more modest manner in order to preserve a feminine identity, at the expense of their professional identities. This attributional style may make it more likely for women to blame themselves for various types of negative situations. Consistent with this reasoning, Bernice Andrews and Chris Brewin (1990) found that female victims of marital violence tended to blame themselves for their experience of violence. Childhood experiences of physical or sexual abuse increased women’s chances of characterological self-blame when, as adults, they found themselves in abusive relationships. (Other research, however, notes that the relationship between attributions and adjustment is complex and not always so straightforward; see McMillen and Zuravin 1997.)
Race and class characteristics also create distinct attributional patterns; research on these factors has been more likely than research on gender to consider possible interactive effects on attributions. In a study conducted by Fathali Moghaddam and colleagues (1995) on attributional styles of whites, blacks, and Cubans in Miami, for example, middle-class black respondents were more likely to blame negative outcomes on discrimination than were lower-class blacks. Lower-class whites were the only group to attribute failure to themselves personally.
The great preponderance of research on social interaction has been based on relationships between strangers in experimental contexts, which may seem to undermine the claim that attribution is significant for interpersonal interaction. The best evidence of this significance, then, is the increasingly large body of research on the role of attribution in the formation, maintenance, and dissolution of close relationships. There is substantial evidence that attributions are linked to relationship satisfaction and behaviors such as conflict resolution strategies. There is also evidence that distressed and nondistressed couples make differing attributions for significant events in their relationships; these patterns may actually serve to maintain marital distress among troubled couples, thus ultimately influencing marital satisfaction. Attributions also play an important role in relationship dissolution. Attributions are a critical part of the detailed accounts people provide for the dissolution of their relationships, and these accounts go beyond explanation to rationalize and justify the loss of relationships.
Research on interpersonal attribution extends the intrapersonal approach in several ways. When people who interact have substantial knowledge of and feelings about each other, the attribution process involves evaluation as well as cognition. Issues of communication, and hence potential changes in preexisting attributions for recurrent relationship events, also become salient at the interpersonal level. There is very little research on how attributions change through interaction and relationships, but this is clearly a significant topic. Attributions at the interpersonal level also entail greater concern with accountability for action; causality at this level also raises issues of justification.
Intergroup attribution. Intergroup attribution refers to the ways in which members of different social groups explain the behavior of members of their own and other social groups. At this level, social categorization has a direct impact on attribution. Studies using a variety of subjects from different social groups and often different countries show consistent support for an ingroup-serving attributional pattern, for example, a tendency toward more dispositional attributions for positive as opposed to negative behavior for ingroup actors. The evidence for the converse pattern, more dispositional attributions for negative as opposed to positive behavior for outgroup actors, is not as strong. (Moreover, these patterns are stronger in dominant than in dominated groups.) Social desirability biases can counteract this pattern; Steven Little, Robert Sterling, and Daniel Tingstrom (1996), for example, found that white respondents held black actors less responsible for participating in a bar fight than did black respondents. The authors suggest that this finding may be due to the desire of the white sample—undergraduates from a suburban area—to appear racially progressive.
Parallel studies of a group’s success and failure show a consistent pattern of ingroup protection. Outgroup failure is attributed more to lack of ability than is ingroup failure. Effects of group membership on attributions about success are not as strong. Interestingly, there is also some evidence of outgroup-favoring and/or ingroup-derogating attributions among widely recognized lower-status, dominated groups such as migrant labor populations (Hewstone 1989). A third form of evidence of intergroup attribution is provided by studies of attributions about social positions occupied by existing groups. In general, these studies, like those cited above, show higher ratings of ingroup-serving as opposed to outgroup-serving attributions.
Societal attribution. At the societal level, those beliefs shared by the members of a given society form the vocabulary for social attributions. The concept of social representations, which has its origins in Durkheim’s concept of “representation collectives,” was developed by Serge Moscovici (1976) to represent how knowledge is shared by societal members in the form of common-sense theories about that society. Social representations are intimately connected to the process of attribution. Not only are explanation and accountability part of a system of collective representations, but such representations determine when we seek explanations. Social representations serve as categories that influence the perception and processing of social information; moreover, they underscore the emphasis on shared social beliefs and knowledge. Social representations are useful in interpreting research on laypersons’ explanations of societal events such as poverty and wealth, unemployment, and racial inequality. Poverty tends to be attributed to individualistic factors, for example, whereas unemployment tends to be attributed to societal factors. Not surprisingly, these patterns may be qualified by attributors’ own class backgrounds; although middle-class people attribute poverty more often to internal factors, those who are themselves poor attribute poverty more often to external factors such as governmental policies (Singh 1989). Gender and racial stereotyping can also shape attributions. Cynthia Willis, Marianne Hallinan, and Jeffrey Melby (1996), for example, found that respondents with a traditional sex-role orientation showed a favorable bias toward the male perpetrator in domestic violence situations. When the female victim was African American and married, both egalitarians and traditionalists were less likely to attribute blame to the man.
Moving away from an emphasis on normative stereotyping, research has considered possible effects of counter-stereotyping on attributional patterns, and of attributions on changes in stereotypes. Portrayals of structurally subordinate groups that counter stereotypical expectations can increase the perceived credibility of members of those groups (Power, Murphy, and Coover 1996). Moreover, an attribution of counterstereotypic behavior to dispositional factors of clearly typical outgroup members can modify outgroup stereotypes (Wilder, Simon, and Faith 1996).
Cross-cultural research illustrates another aspect of societal attributions. This work compares the extent and type of attributional activity across cultures. Although there has been some support for the applicability of Western models of attribution among non-Western cultures, most of this research has demonstrated in a variety of ways the cultural specificity of particular patterns of attribution (Bond 1988). As noted above, for example, Joan Miller (1984) provides cross-cultural empirical evidence that the fundamental attribution error, the tendency to attribute cause more to persons than to situations, is characteristic of Western but not non-Western societies. Attributions about the self also vary across cultures; members of some non-Western cultures attribute performance more to effort than to ability; the opposite pattern has been found in the United States. Cultural patterns also shape responsibility attributions; V. Lee Hamilton and Shigeru Hagiwara (1992) found that U.S. (and male) respondents were more likely to deny responsibility for their own inappropriate behaviors than were Japanese (and female) respondents, who were more likely to apologize for the behaviors. Hamilton and Hagiwara argue that the Japanese (and women) were more concerned with maintaining the quality of the interaction between the accuser and the accused.
It may be useful to identify several sociological phenomena to which theories of attribution are relevant. A number of scholars have suggested integrating attribution theory and labeling theory (Crittenden 1983). Attribution occurs within an individual; labels are applied by a group. Judith Howard and Randy Levinson (1985) offer empirical evidence that the process of applying a label is directly analogous to attribution. They report that the relationships between attribution information and jury verdicts are consistent with predictions based on a labeling perspective.
Richard Della Fave (1980) demonstrates the importance of attribution for understanding a key but neglected aspect of stratification, namely, how it is that stratification systems become legitimated and accepted by those disadvantaged as well as by those advantaged by those systems. He draws heavily on attribution theory in developing a theory of legitimation and in identifying possible sources of delegitimation.
Attribution is a significant social process that ranges widely from cognitive processes to collective beliefs. The field is still imbalanced; more work has been done at the intrapersonal and interpersonal levels. There is evidence for both intergroup and societal attributions, however, and research at these two levels is steadily increasing (Hewstone 1989). Recent research has demonstrated connections across areas as diverse as social cognition, social interaction, intergroup relations, and social representations; these connections provide increasing evidence of the importance of attribution for sociological phenomena. Indeed, in the next decade it may be that the fruits of attribution theories will be evident more in research on these other topics, than in research on attribution alone.