The Meaning and Measurement of Explanatory Style

The Meaning and Measurement of Explanatory StyleAuthor(s): Christopher PetersonSource: Psychological Inquiry, Vol. 2, No. 1 (1991), pp. 1-10Published by: Taylor & Francis, Ltd.Stable URL: http://www.jstor.org/stable/1449402 .

Accessed: 14/06/2014 18:51

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to PsychologicalInquiry.

http://www.jstor.org

This content downloaded from 195.78.108.199 on Sat, 14 Jun 2014 18:51:35 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=taylorfrancis

http://www.jstor.org/stable/1449402?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Psychological Inquiry Copyright 1991 by 1991, Vol. 2, No. 1, 1-10 Lawrence Erlbaum Associates, Inc.

TARGET ARTICLE

The Meaning and Measurement of Explanatory Style

Christopher Peterson University of Michigan

Explanatory style is a cognitive personality variable reflecting the way that people habitually explain the causes of bad events. Explanatory style spans three dimensions-internality versus externality, stability versus instability, and globality versus specificity. Although explanatory style has an impressive array of correlates, including depression, achievement, and physical well-being, questions have recently been raised concerning its meaning and measurement. This article identifies and discusses these questions, concluding that explanatory style is a useful construct that deserves further theoretical and empirical attention.

Explanatory style is a cognitive personality variable that reflects how people typically explain the causes of bad events involving themselves (Peterson & Seligman, 1984). Three dimensions of explanatory style are distinguished: (a) internality ("It's me") versus externality, (b) stability ("It's going to last forever") versus instability, and (c) globality ("It's going to undercut everything that I do") versus specificity. Emerging from the reformulated model of learned helplessness as a way of explaining individual differences in the response to uncontrollability (Abramson, Seligman, & Teas- dale, 1978), explanatory style has become a personality construct in its own right (Pervin, 1985). Different strategies for measurement have been developed, and a variety of possible correlates have been explored. Perhaps it was inevitable that this explosion of interest in explanatory style would generate questions and criticisms about this approach to understanding personality.

My purpose in this article is to discuss some of the concerns that have recently appeared in the personality literature about the meaning and measurement of explanatory style. My eventual conclusion will be that explanatory style is useful to personality psychologists, although the notion of explanatory style is still unfinished business. Its flaws seem neither more numerous nor more severe than those inherent in any individual difference with which we concern our- selves. Indeed, the shortcomings of explanatory style are precisely areas in which further theory and research would be productive. It would be unfortunate if recent criticisms of explanatory style had the effect of stifling interest in the construct.

The questions that can be raised about explanatory style roughly divide themselves into conceptual and methodological matters, and I discuss them under these headings. Need- less to say, this is only an approximate distinction, because considerations of meaning bear on considerations of measurement, and vice versa. My attempt is to phrase each question in an even-handed way, and then suggest profitable ways to go about answering it. In some cases, the answer is already available.

Conceptual Questions

Few doubt that people's causal explanations influence their reactions to events (Kelley & Michela, 1980). No one questions that people differ in how they explain similar events (Ickes & Layden, 1978). Even the idea that people show cross-situational consistency in how they think about matters is not challenged too seriously (Mischel, 1968). The basic idea of explanatory style is therefore a sensible one. However, several legitimate questions can be raised about the particular way in which explanatory style is usually regarded.

Why Describe Explanatory Style Along the Dimensions of Internality, Stability, and Globality?

In other words, one can ask why these three dimensions were chosen to describe the ways that particular causal explanations differ (e.g., Anderson, Jennings, & Arnoult, 1988). Once these are chosen as the basic parameters of causal attributions, then they are explanatory style, by definition.

The answer to this question is partly historical and partly theoretical. Explanatory style did not originate in the desire to describe all possible ways in which people's causal attributions differ. Rather, explanatory style was proposed to make sense of the boundary conditions of learned helplessness in people (Abramson et al., 1978). Let me give some back- ground. The original learned helplessness model hypoth- esized simply that exposure to uncontrollable events produced a state of helplessness characterized by motivational, emotional, and cognitive difficulties (Maier & Seligman, 1976; Seligman, 1975). When applied to people, and tested in laboratory experiments, results were found to vary in ways not accommodated within the theory.

In particular, sometimes induced helplessness involved long-lasting difficulties, and sometimes not. Sometimes induced helplessness involved pervasive difficulties, and sometimes not. And sometimes induced helplessness in-



2 PETERSON

volved the loss of self-esteem, and sometimes not. These three sources of contradictory evidence could all be explained by proposing that people's causal attributions for uncontrollability dictated the extent to which helplessness was chronic, pervasive, and undermining of self-esteem.

Consistent with the "person as scientist" metaphor that has dominated attributional theorizing (e.g., Heider, 1958; Kelley, 1973; Weiner, 1986), Abramson et al. (1978) proposed quite logically that stable attributions affect the stability of deficits, global attributions dictate the globality of deficits, and internal attributions influence the degree of self- esteem loss. In sum, anomalies in previous helplessness findings led to the interest in the dimensions of internality, stability, and globality.

Explanatory style came out of an experimental research tradition and thus emphasized the situational determinants of behavior. Explanatory style was a means of accounting for "error variance" vis-a-vis the original helplessness model. Internality, stability, and globality are among the basic dimensions of the universe as seen by many experimental psychologists. Are matters inside or outside the skin, general across time, and general across setting?

However, as explanatory style has emerged in its own right, it has been examined in relationship to matters that do not map into its original purpose of explaining boundary conditions of laboratory-produced helplessness. For instance, in a study by McCormick and Taber (1988), explanatory style was related to chronic gambling. That explanatory style bears an empirical relationship to the severity of gambling is interesting, to be sure, but it is difficult to see how the extent of gambling can be explained in terms of the particular parameters of concern to the helplessness reformulation.

One could examine self-esteem loss in gamblers, perhaps, or the degree to which they lose impulse control at different tracks over varying periods of time. But as explanatory style is applied to matters outside the realm of learned helplessness, it seems unreasonable to let laboratory studies dictate the aspects of the phenomenon worthy of attention.

Why Not Include Other Dimensions of Explanatory Style?

This question follows from the one just discussed. Some investigators are unhappy with the explanatory style construct because it seems to be overlooking an important attributional dimension. Let me repeat that the importance of a parameter depends on the purpose that the researcher has in mind. Just what correlates does one hope to establish with explanatory style? It may well be that the internality, stability, and globality of causes are not the most sensible dimensions of attributions to assess.

Weiner (1986) surveyed various attempts to enumerate the "basic" dimensions of causal attributions. Although Weiner might not agree, the studies he reviewed lead me to conclude that there probably is not a basic set of dimensions invariant across domains and cultures. Thus, the researcher must specify dimensions of explanatory style in relationship to some purpose.

Several researchers have suggested that the controllability of the attributed cause is a critical dimension (e.g., Anderson et al., 1988). Others have suggested that the degree to which people believe they can cope with the consequences of the

bad events is critical (e.g., Wortman & Dintzer, 1978). My response to these suggestions is a "yes, but" reaction.

In some cases, including much of my own research, the intent of using explanatory style has been to explain the extent of helpless behaviors and various proxies. Internality, stability, and globality-for theoretical reasons, as already explained-seem adequate for this purpose. The helplessness model, in its original and reformulated versions, concerns itself with reactions to uncontrollable events. These are not the same as uncontrollable causes of events, nor the same as uncontrollable consequences. There may well be asymmetry, such as uncontrollable events with controllable causes or vice versa (cf. Abramson & Sackeim, 1977).

Learned helplessness research can be criticized for not directly assessing people's perceptions of the uncontrollability of events. Instead, such perceptions are usually inferred from the particular causal explanations that a person offers. A bad event believed to have an internal, stable, and global cause is arguably one that will be regarded as uncontrollable, and vice versa for a bad event explained with an external, unstable, and specific cause. Still, this line of reasoning should be tested more frequently than it has been, although the studies that have been done are supportive (e.g., Peterson, Schwartz, & Seligman, 1981).

In other cases, additional or alternative dimensions might best serve the purpose of a particular study. It is unlikely that a single set of explanatory style dimensions-whether the original triad of internality, stability, and globality or some of the alternatives suggested more recently-will suffice for all research needs. One's choice of dimensions should be a theoretically driven decision, and the fact that different researchers favor the inclusion of different explanatory style dimensions is not, on the face of it, any cause for alarm or confusion.

There have been few attempts to examine all at once the different properties of how we think about events, but several years ago I performed such a study (Peterson, 1986). I asked a group of college students to describe four actual bad events that had happened to them in the recent past along dimensions deemed important by various theorists. Some of these dimensions involved the event itself (undesirable vs. desir- able, unpredictable vs. predictable, uncontrollable vs. controllable, recurring vs. nonrecurring). Other dimensions asked about the "one major cause" of the event in question (internal vs. external, stable vs. unstable, global vs. specific, uncontrollable vs. controllable). Several dimensions asked about the "one major consequence" of the event (stable vs. unstable, global vs. specific, uncontrollable vs. controllable). And the remaining dimensions asked about one's coping with the event (successful vs. unsuccessful, difficult vs. easy).

The results pointed to very simple conclusions. I collapsed these ratings across the four events. In the case of each dimension, there was at least moderate consistency, suggest- ing indeed that one's explanatory style is consistent along many different parameters. Then I factor-analyzed these composite ratings. What emerged was a very clear two-factor solution. Each dimension (except undesirability of event, which had no appreciable variance across subjects) loaded unambiguously on one of two factors.

One factor included the ratings of internality of the cause plus all the ratings of predictability and controllability, cutting across whether the rating in question was made with



MEANING AND MEASUREMENT OF EXPLANATORY STYLE 3

respect to the event, its cause, or its consequence. It can be labeled the Predict and Control way of looking at events. The second factor subsumed all the ratings that involved the magnitude of the event, its cause, and its consequences-as well as the difficulty of coping with it. I have come to label this second factor the Big Deal way of looking at events, perhaps what we regard in everyday life as catastrophizing.

What is the relevance of these findings to explanatory style as defined by the internality, stability, and globality of causes? In the same sample of subjects, I administered a version of the Attributional Style Questionnaire (ASQ; Peter- son et al., 1982) and found that the internality dimension correlated with the Predict and Control factor but not with the Big Deal factor and that the stability and globality dimensions correlated with the Big Deal factor but not with the Predict and Control factor.

This finding is at odds with the learned-helplessness reformulation, which hypothesizes quite clearly that (stability + globality) should predict one's future expectations of controllability (Abramson et al., 1978; Peterson & Seligman, 1984). In another study concerned with how people re- sponded to falling ill, we similarly found that stability and globality of explanatory style did not correlate with expectations that particular health-promoting activities would indeed be effective. The actual cognitive processes linking uncontrollable events to expectations of helplessness and eventually to deficits may well be different from those proposed by the helplessness model (Tennen, 1982). Here is an area where future scrutiny is needed (e.g., Kuhl, 1981; Peter- son, 1978).

These findings have another implication. When a researcher measures one dimension of explanatory style, he or she may in effect tap several others. The subjects included in our studies seem not to segregate their thoughts as neatly as our theoretical disagreements assume. To attribute a particular cause to a bad event is at the same time to think about its aftermath. To think about an event in a given way is to take a stance with respect to its cause. And so on.

Is there a theoretical lesson here? Perhaps we should not be chauvinistic about the particular dimensions on which we choose to focus, because it may be that one or another dimension gets dragged along when we measure what we think is most important. Further, it behooves us to tease out the inter- relationships among different dimensions of explanatory style. Perhaps some dimensions are more primary than others.

Why Form a Composite?

One of the most cogent criticisms of explanatory style is Carver's (1989) recent query into why explanatory style researchers often create a composite of internality, stability, and globality ratings, and then treat this composite as the index of explanatory style. He correctly observed that this makes it impossible to ascertain the roles played by the individual dimensions. He further went on to argue that there may be little theoretical justification for forming such a composite, and thus he recommended against doing so.

On the one hand, individual dimensions of explanatory style have been examined and reported in several published studies (e.g., Peterson & Seligman, 1984, Studies 1 through 5). On the other hand, there has indeed been a recent trend to report only the composite, so Carver's (1989) question must

be addressed. I agree that a composite makes it impossible to examine the specific roles assigned by the helplessness reformulation to particular dimensions (see Peterson & Vil- lanova, 1988). Is there any justification for using a composite besides economy of presentation?'

Several possible answers come to mind. First, many of the outcomes to which explanatory style has been related are not theoretically specific to particular dimensions. Consider depression as a whole, the glut of emotional, physical, motivational, and cognitive symptoms that make up this problem. If we lump these symptoms together, as is done in diagnosing and/or assessing the severity of depression, we have no reason to think that internality should be more or less related to depression than stability or globality. Indeed, a meta-analysis of explanatory style and depression shows that the individual dimensions all relate to this criterion (Sweeney, An- derson, & Bailey, 1986).

Perhaps the use of composites is justified when the outcomes are not specific to particular dimensions. However, this answer undercuts the point made earlier, namely, that a researcher's theoretical purpose should dictate which dimensions of explanatory style are examined in a given study. If outcomes are not specific to particular dimensions, then why use these dimensions at all, either separately or in a composite?

Studies that do focus on differential correlates with the three dimensions of explanatory style have in common the use of outcome measures that are theoretically specific to a given dimension. Eaves and Rush (1984) showed that the stability of explanatory style predicted the chronicity of a depressive reaction. Alloy, Peterson, Abramson, and Selig- man (1984) showed that globality of explanatory style influ- enced the generality of helplessness deficits across different laboratory tasks. Several studies have looked at the relationship between internality of explanatory style and self- esteem (e.g., Ickes & Layden, 1978; Tennen & Herzberger, 1987). In terms of the specific claims of the reformulated helplessness model of depression, these studies are much more informative than those that merely show a correlation between a composite explanatory style score and a measure of depression.

A second answer to the question of why one might form a composite is suggested by Carver (1989) himself. If the various dimensions capture a higher order notion, a so-called latent variable, this would justify their combination. Such a case can be made for explanatory style. As already emphasized, internality, stability, and globality all have something to do with the extent and nature of helplessness deficits. The more of each that is present, the greater should be the deficit in which we are interested.

So, a composite mixes together apples and oranges, but the result can be construed as fruit salad. If you don't like fruit salad, here is another way to look at the latent variable presumably captured by an explanatory style composite, suggested to me by Martin Seligman (personal communication, October 15, 1988). Suppose you are the general manager of a

'Without dwelling on the matter, let me point out that journal editors have sometimes asked me to streamline the data presentation in an article to conserve journal space. I am not blaming these editors for how I decided to proceed, which was often to report only a composite. Rather, I am just observing that what originally seemed an innocuous decision turned out otherwise.



4 PETERSON

baseball team, and you need to promote a shortstop from the minor leagues to take the place of someone who has been injured. For the sake of this analogy, let us assume that there are three attributes of a shortstop that are equally important to you as a general manager: fielding, running, and batting. You have numerous shortstops in your minor-league farm system, and you ask your scouts to rate each one of them- using 7-point scales, of course!-along the three dimensions. Which shortstop is the one you call up to the major leagues?

Obviously the one with the highest total rating gets the call. If you are fortunate, you will find someone who fares well with respect to all three ratings. You do not worry that fielding and batting are different things. Of course they are. You do not worry about whether they correlate or not. You are quite confident that to the degree that a shortstop pos- sesses more of each attribute, he is the one for the major league job. The composite captures something about the essence of "shortstopness" as you conceive it.

Similarly, the composite of explanatory style captures something about the essence of helplessness, with each dimension making its contribution. Someone who scores highly on all three dimensions of explanatory style is more likely to be passive and demoralized than someone who scores low on all three (Peterson & Seligman, 1984).

Internality, stability, and globality ratings are rarely orthogonal to each other. They were proposed on an a priori basis, with no claim that the actual universe of attributions would fall at right angles along these dimensions. Invariably, stability and globality correlate highly with one another, so much so that we suggest that they be regarded as a common factor of hopelessness (Peterson & Seligman, 1985; see also Abramson, Metalsky, & Alloy, 1988, 1989). It may be that a global cause is necessarily a stable one, because it must last long enough to influence outcomes in different domains. Sometimes internality is correlated with stability and globality (e.g., Peterson et al., 1982) and sometimes it is independent (e.g., in the factor-analytic study described earlier-Peterson, 1986), so here the latent-variable argument starts to fall apart.

Regardless of one's interest in fruit salad, baseball play- ers, or an explanatory style composite, one should always look at the ingredients that are combined into these entities. It may well be in a given case that the assumption of a latent variable will be violated. Or perhaps one dimension will prove more critical than the others in predicting an external criterion. The automatic creation of a composite cannot be justified.

What About Explanatory Style for Good Events?

The discussion so far has focused on explanatory style for bad events. However, some research has looked at explanatory style for good events. Two generalizations can be offered about this research (Peterson & Seligman, 1984). First, explanatory style for good events is often independent of explanatory style for bad events. Second, the correlates of explanatory style for good events tend to be the "opposite" of the correlates of explanatory style for bad events and usually less robust. Internal, stable, and global attributions for good events correlate weakly with the absence of depressive symptoms, for instance, as opposed to the stronger

correlations between internal, stable, and global attributions for bad events and the presence of depressive symptoms.

What is the theoretical significance of explanatory style for good events? The reformulated learned helplessness model is silent, because it concerns itself only with how people respond to bad events. But we are still left with the empirical relationships between explanatory style for good events and outcomes like depression. One possible explanation is that explanatory style for good events influences the degree to which we savor our triumphs in life (cf. Weiner, 1986). The positive expectations and good feelings engen- dered by an upbeat explanatory style for good events may buffer us against the depressing effects of loss and disappointment (Taylor & Brown, 1988).

No research of which I am aware bears directly on this explanation. An adequate investigation of explanatory style for good events should measure both positive and negative life events, as well as the good and bad moods that may follow in their wake. I have elsewhere speculated that explanatory style for good events directly influences one's reaction to good events, only indirectly affecting negative outcomes like depression (Peterson & Seligman, 1985; see also Abramson et al., 1989). But it is possible that explanatory style for good events directly affects how we respond to bad events (e.g., "There must be a pony here-look at all the manure").

Research into explanatory style for good events may not yield data as rich as that from research into explanatory style for bad events, simply because people are less "mindful" when thinking about good events (Langer, 1989). Their re- sponses are more likely to be "off the tops of their heads" (Taylor & Fiske, 1978), and people's cliches are not as psy- chologically revealing as the results of their active searches for the causes of bad events (Wong & Weiner, 1981). Still, an interesting direction of inquiry might be to understand why explanatory style for good events is independent of that for bad events. Perhaps the answer lies somewhere in differential socialization of the two.

In some research, a further explanatory style composite is created by subtracting explanatory style scores for good events from explanatory style scores for bad events. This composite is necessarily less reliable than either of its components (granted their independence). At the same time, it usually shows a more robust relationship with external crite- ria than do either of its components (again, granted their independence). We will not understand the significance of this composite until we understand the meaning of explanatory style for good events. Although I have defended the creation of a composite of (internal + stable + global) attributions under some circumstances, a (bad - good) composite cannot be justified until there is a theoretical reason that compels it.

What Is the Relationship Between Explanatory Style and Optimism?

Explanatory style researchers have taken to describing the different ends of the dimension as optimistic versus pessimistic (e.g., Peterson, Seligman, & Vaillant, 1988). Granted that there already exist measures purporting to measure optimism and pessimism (e.g., Scheier & Carver, 1985), what is gained by describing explanatory style in these terms?

There are two related answers to this question. The first




starts with a long joke that the Smothers Brothers used to tell. The gist of the story entails someone who worked in a chocolate factory and fell into a giant vat of chocolate from which he couldn't escape. He screamed "Fire!" at the top of his lungs, over and over until he was rescued. The obvious question is why he screamed "Fire!" to attract attention, and the answer delivered by Tommy Smothers was that if the man had screamed "Chocolate!" no one would have been interested.

Explanatory style has been characterized as optimistic or pessimistic precisely because people pay more attention to the construct with these designations. After all, much of psychology is concerned with constructs suggested by ordi- nary language. This is certainly the case when one is dealing with attributions, a line of work that began with Heider's (1958) analysis of phenomenological causality. "Explanato- ry style" is not exactly jargon, but neither does it reside in the everyday lexicon. Hitting on the description of it as optimistic on the one hand (when bad events are explained with external, unstable, and specific causes) or pessimistic on the other (when bad events are explained with internal, stable, and global causes) was simply a good idea.

But something more than marketing is behind our change in terminology. It is also a sensible idea, because if it is not optimistic to conceive the causes of bad events in exter- nalized and circumscribed ways, then what is it? If it is not pessimistic to locate the causes of misfortune within oneself and project these causes across time and situation, then what is it?

So much for the semantic argument. Consider the empirical evidence. "Pessimistic" explanatory style is associated with depression, lowered expectancies, passivity in the face of challenge, poor achievement, and ill health (e.g., Peterson, 1990; Peterson & Barrett, 1987; Peterson & Bossio, 1989; Peterson & Seligman, 1984, 1987; Seligman & Schulman, 1986). "Optimistic" explanatory style has the opposite correlates. This evidence fleshes out and justifies the characterization of the dimension as pessimistic versus optimistic.

We try to use the terms optimistic and pessimistic as adjec- tives that modify the individual difference of explanatory style, and sometimes we adopt a shorthand way of speaking and refer to optimistic versus pessimistic individuals (i.e., those who have relatively optimistic vs. pessimistic explanatory styles). And sometimes we become careless and turn this dimension into a typology, speaking of optimists and pessimists as if we are studying two discrete groups of individuals. This final usage of our terms is objectionable, because we are not working with a typology. Indeed, we see explanatory style as a dimension; indeed, explanatory style is a bundle of dimensions, as previously explained. In every study we conduct, we find most people bunched up in the middle of our scales-neither optimistic nor pessimistic. The only defense we have is that we are in good company here. For instance, many discussions of locus of control speak of "internals" and "externals" as glibly as we speak of optimists and pessimists. We should all be more careful.

Explanatory style is just one of many related constructs proposed in recent years to reflect individual differences in cognition that presumably mediate how people cope with bad events. Hardiness, self-efficacy, sense of coherence, locus of control, and desire for control are just some of the other members of this family. Peterson and Stunkard (1989) re-

cently reviewed the theoretical literature on perceived control, and labeled such constructs cognates of one another. We also sketched how they might be combined into a generic variable. Another strategy is to be less synthetic and specify how these overlapping notions differ and what each uniquely contributes to our understanding of coping.

Are There Historical or Cultural Limitations on Explanatory Style?

The possibility must be acknowledged that people's ten- dencies to offer causal attributions show historical or cultural boundaries. I doubt that a time or place has ever existed in which individuals do not engage in making sense of what they see, particularly behaviors by themselves or others. But causal explanations-those pointing to antecedent events that covary with phenomena of interest-are obviously but one way to make sense of the world (cf. Pepper, 1942).

Causal attributions may be a particular concern of resi- dents in Western societies during the late 20th century. In other words, causal attributions for behavior may be offered only by people who possess a highly articulated sense of self as distinct from the world, who exalt individuality, and who try to "predict and control" the events that befall them (Baumeister, 1986; van den Berg, 1983; Weisz, Rothbaum, & Blackburn, 1984). Perhaps the popularity of explanatory style as a research topic reflects its particularly good fit with the collective psychology of the United States in the late 20th century.

An informal tabulation I have made over the years seems to support this point. I have published research articles on the "original" helplessness model which concern themselves with the effects of uncontrollable events as well as articles on the reformulated model which focus on explanatory style. When I receive reprint requests from Eastern European and Asian countries, they are usually for the former articles. When I receive reprint requests from Western European countries and the United States, they are usually for the latter articles.

Along these same lines, but less speculatively, we can observe cultural and historical differences with respect to the particular causal attributions that people offer. For instance, Miller (1984) compared the causal attributions for behaviors made by Indians and Americans and found that the former group favored contextual explanations (e.g., roles, norms), whereas the latter group favored dispositional explanations (e.g., traits, attitudes). Her subjects were of different ages, and she found that this cultural difference became more pro- nounced with increasing age. Socialization into a particular culture apparently entails learning characteristic explanations for behavior.

If people in a given time or place do not offer causal explanation, then explanatory style is a meaningless construct for characterizing those people. If people offer causal explanations that differ from those usually made by research subjects in the here and now, then we face several interesting empirical questions. Can these explanations be sensibly described along dimensions of internality, stability, and globality? Do these individuals show a characteristic style? Does this style relate to outcomes as predicted by the reformulated helplessness theory? And so on.

To conclude this discussion of theoretical issues, let me underscore two ideas. First, explanatory style is used for a



6 PETERSON

particular purpose: to explain the parameters of helpless behavior. This dictates the original choice of dimensions and justifies under certain circumstances their combination into a composite. However, research all too often has neglected to investigate the particular predictions of learned helplessness theory. Second, explanatory style as typically conceived is not a necessary condition for the various phenomena to which it is applied. There are many routes, for instance, to failure, depression, and illness. No one has ever regarded explanatory style as the only determinant of these, which means that researchers should explore other ways to explain them. But the success of one investigation need not mean the failure of another. There is plenty of variance to go around, we can be sure.

Am I wallowing in theoretical ecumenism? Not at all, because there may well be theoretical positions that do not prove viable, even my own. What I am urging is a more rigorous approach to psychological research, which sometimes entails faulty reasoning such as the following:

1. According to some theory, variable X predicts a given outcome.

2. Here are data showing that variable Y predicts the outcome.

3. Therefore, variable X does not, and the theory is wrong.

Stated so starkly, the logical error is clear. But actual studies may end up offering conclusions about the validity of a theory without investigating the constructs deemed important by that theory.

The reader may remember the person-situation debate that occupied personality psychology during recent decades. Proponents of one side lambasted the position of the other side with faulty arguments just like the one sketched. I see the same sorts of arguments being raised against helplessness theory in general and explanatory style in particular. The appropriate way to show that explanatory style is not a sensible construct is to measure this variable explicitly and then to find either that it fails to "behave" according to prediction and/or that it is confounded by some other variable.

Methodological Questions

As I stated earlier, it was impossible to discuss the previous conceptual questions without sliding into a discussion of procedures. But here let us make the methodological focus explicit. Questions have been raised not only about the meaning of explanatory style but also about the way that researchers have gone about measuring it. Let me discuss some of these matters, again trying to answer them where possible and point to strategies of resolution where not.

Explanatory style is usually measured with the ASQ (Pe- terson et al., 1982), a self-report questionnaire developed for this purpose. With this questionnaire, subjects are presented with a series of hypothetical events involving themselves (e.g., you go out on a date, and it goes badly). For each event, subjects are asked to imagine it happening to themselves, and then to think about what would be the "one major cause" of this event if it were actually to happen. They write down this cause. Then they rate each cause along three 7- point scales, asking about the internality versus externality of

the cause, the stability versus instability, and the globality versus specificity.

In its original form, the ASQ presented subjects with both good events and bad events. Subject ratings are added within type of event and particular dimensions. As already discussed, a composite can be formed by adding across dimensions.

There also exists a content analysis procedure for assessing explanatory style (the Content Analysis of Verbatim Ex- planations, CAVE; Peterson, Schulman, Castellon, & Selig- man, in press). The CAVE technique scans verbal material of a subject for mention of events involving the individual that have accompanying explanations. These events and attributions are extracted and then rated by judges along dimensions of internality, stability, and globality, as in the ASQ.

Do People Have a Consistent Explanatory Style?

The original version of the ASQ had low internal consistencies for the individual dimensions (Peterson et al., 1982). They were certainly above zero, showing that subjects did evidence a degree of consistency across the different events they explained. But some critics have argued that this means that people do not have an explanatory style, that they are inconsistent with respect to how they explain different events (e.g., Arntz, Gerlsma, & Albersnagel, 1985).

At this point, this issue should be resolved for all whose concern is with contemporary people in the Western world. Peterson and Seligman (1984) noted that the reality of the events that people explain has something to do with the particular causes chosen. What this means is that when we try to assess "style" from just a few events, we run the risk of measuring not a personality characteristic but simply reality. Low reliability should ensue. If we increase the number of events for which attributions are made, then we should come ever closer to measuring a psychological style-presumably by canceling out extraneous determinants of particular attributions.

This is exactly what happens. Peterson and Villanova (1988) described an Expanded Attributional Style Question- naire, for instance, which increases the number of bad events in the original ASQ from 6 to 24, with an impressive increase in the internal consistency of the individual dimensions. These consistencies, in the .7 to .9 range, are thoroughly acceptable as measures of individual differences.

The CAVE technique also yields more satisfactory reli- abilities of explanatory style to the degree that more events as opposed to fewer are used. For instance, Peterson et al. (1988) found much more satisfactory consistency of explanatory style when measures were based on 10 events with attributions than did Peterson, Bettes, and Seligman (1985), who used but 2 events.

This is not supposed to be a dizzying insight. As Peterson and Villanova (1988) observed, the point is really a statistical truism long followed by those concerned with scale development. As long as items show a somewhat positive correlation with each other, the reliability of any scale can be increased simply by adding more items. For some reason, this truism has not filtered into the explanatory style literature, and with some frequency we see the failure of particular measures of explanatory style confused with the nonexistence of the con- strulct they purport to measure.




Another view on individual consistency looks at the stability of explanatory style over time. Presumably, if subjects do not have a consistent explanatory style, this nonexistent style should not show up over time. But in the original description of the ASQ, a decent degree of test-restest reliability over a 5-week period was demonstrated (Peterson et al., 1982). Similar consistency has been reported in children, this time over periods of months (Nolen-Hoeksema, Girgus, & Seligman, 1986). Finally, using the CAVE technique to measure explanatory style from material provided by research subjects, Burns and Seligman (1989) recently found consistency of explanatory style over more than five decades. Such stability, on the face of it, is good evidence for consistency at any given point in time.

So, people's explanations show consistency across different events and stability across time, as shown by statistically significant correlations. The magnitude of these correlations is invariably modest, but no lower than those involving most other personality characteristics. The magnitude of correlations should not be evaluated out of context. Explanatory "style" is as coherent an individual difference as most personality constructs.

Does Explanatory Style Predispose "Spontaneous" Attributions?

Here is another question about explanatory style that has been affirmatively answered in the literature, although perhaps there has not been enough attention to the confirming studies. This question takes on various forms. When the helplessness reformulation was first proposed, the question was raised whether people actually make "spontaneous" causal attributions (e.g., Wortman & Dintzer, 1978).

The success of the CAVE technique in finding causal explanations in a variety of verbal materials shows that people indeed offer attributions; the CAVE technique has been used with psychotherapy sessions, press conferences, political speeches, letters, diaries, sports stories, open-ended ques- tionnaires and interviews, and essays (e.g., Peterson et al., 1985; Peterson, Luborsky, & Seligman, 1983; Peterson et al., 1988; Zullow, Oettingen, Peterson, & Seligman, 1988). Causal attributions abound in this material.

Weiner (1986) reviewed several altogether different research literatures and again found ample evidence that people display spontaneous causal thinking, without prompts from a researcher. Thus, the first doubt about the relationship of explanatory style to particular attributions is laid to rest by observing that people indeed offer attributions in the course of thinking about the events that befall them.

The second form this question takes is to acknowledge that people offer attributions but then to doubt that explanatory style as assessed with the ASQ pertains to them (e.g., Cut- rona, Russell, & Jones, 1984). In other words, attributions about hypothetical events (like those on the ASQ) are unre- lated to attributions about actual events. I have already explained that there will be instances when we should expect these relationships to be attenuated: anytime the "reality" of the event and/or social consensus suggests a particular causal explanation. If one looks only at these sorts of events, then explanatory style will not show a relationship.

The trick is to look elsewhere. Then there is good evidence that explanatory style (as assessed from attributions about hypothetical events) relates to particular attributions about

actual events. Peterson and Villanova (1988) described just such a study. Explanatory style as measured with the Ex- panded ASQ correlated with attributions about actual bad events in someone's life. In this study, some specificity of prediction was shown. Internality of explanatory style correlated with internality of actual attributions to a greater degree than did stability of explanatory style. Stability and globality proved to be more entwined, as our previous discussion would imply.

Metalsky, Halberstadt, and Abramson (1987) showed that explanatory style as measured with the ASQ at the beginning of a school term predicted the actual attributions made by college students for their exam performance some weeks later. Atlas and Peterson (1989) demonstrated that explanatory style measured with the ASQ predicts the attributions made by patrons at a harness-racing track for the outcomes of races on which they lost money.

More generally, several studies have examined the relationship between the ASQ and the CAVE, finding modest convergence between explanatory style assessed in these two ways (see Peterson & Seligman, 1984). Stated another way, attributions about hypothetical events thus relate to attributions about actual events, so long as one gives the personality variable a fair chance to operate. This criticism is reminis- cent of the overly stringent view of "traits" used some years ago by those who wished to argue against their existence. No trait theorist ever expressed the opinion that a trait was evi- dent in every possible situation (Hogan, deSoto, & Solano, 1977). The more constrained a situation, the less likely an individual difference is to manifest itself. With respect to explanatory style, constraints are provided by the actual event that is being explained. Keeping this in mind should help explanatory style researchers find topics where attention to individual differences in the use of attributions should prove fruitful.

Why Are the Correlations Involving Explanatory Style Ostensibly So Low?

The typical correlation with explanatory style is in the ubiquitous .20 to .30 range, and some critics dismiss these as disappointingly low (e.g., Cutrona, 1983). This criticism is not fair, because given variables, whether experimental or correlational, rarely account for more than about 10% of the variance in other variables (e.g., Funder & Ozer, 1983). Let me use this opportunity to mention an important article by Rosenthal and Rubin (1982) about how to interpret the magnitude of correlation coefficients. This point is not suffi- ciently well made within personality psychology, where it is so apt.

Rosenthal and Rubin (1982) begin with the observation that correlation coefficients are not readily grasped by most people who hear about them. One knows they range from - 1.00 through 0.00 to 1.00, of course, but then when we hear that a correlation is "only" .20 or .30, this certainly does not seem very high. When we square the correlation and find that it accounts for "only" 5% to 10% of the variance, this seems even more unsatisfactory. Rosenthal and Rubin proposed that we cast these measures of association in a different form.

Suppose we have a medical treatment that reduces one's risk of dying from a disease from a higher probability to a lower one. This reduction in risk certainly has an intuitive



8 PETERSON

meaning. It is possible to express correlation coefficients in terms of this metric, where Variable 1 is the medical inter- vention and Variable 2 is the reduction of risk. What is the translation of a correlation of .30? Simply the reduction of risk from about .65 to .35: a virtual cutting in half of the likelihood of dying. This correlation seems worth taking seriously.

Let me illustrate how recasting the same data can change the impressiveness of a relationship's magnitude. In a recent study, I showed that explanatory style correlates with the number of days ill that one reports in a month (Peterson, 1988). The correlation coefficient itself appears modest, only .27. But then I divided my sample into the top and bottom half with respect to composite explanatory style. The top half of the sample (subjects with a more pessimistic explanatory style) showed a twofold difference in number of days ill when compared to the bottom half (subjects with a more optimistic explanatory style): 7.89 versus 3.94.

To conclude this section on methodological questions about explanatory style measurement, I suggest that many of the criticisms about the way in which explanatory style research has been conducted take investigators to task for problems that no longer exist. There is good evidence that explanatory style is a consistent individual difference, that it pertains to the particular attributions that people offer in everyday life, and that its relations to external variables are as robust as the correlates to be expected of any personality dimensions.

Conclusion

In this article, I have discussed the conceptual and methodological criticisms of explanatory style that have appeared recently in the personality literature. I have argued against some of these criticisms and agreed with others. My overall conclusion is the one I promised at the outset:

Explanatory style is a useful construct, despite the fact that it is unfinished business. The measurement of explanatory style may be in better shape than the understanding of what the notion means. I therefore suggest that researchers make use of the measures already available, heed the good suggestions about reporting of data suggested by Carver (1989) and others, and get about the business of investigating sub- stantive questions about explanatory style.

Besides issues already raised, the questions that strike me as most interesting about explanatory style include the following:

1. What are its origins? We know much more about the consequences of explanatory style than we do about its ante- cedents. Some preliminary data from my research group suggest that a pessimistic explanatory style may result from failure and disappointment in the immediate past and from harsh and inconsistent socialization in the more distant past (Peterson & Bossio, 1991).

2. Why is explanatory style related to one's physical health? Studies have shown an empirical link between pessimistic explanatory style and poor health operationalized in various ways (Peterson & Seligman, 1987). But these "naked correlations" tell us nothing about the route involved, whether it is biological, psychological, or interpersonal. Perhaps all these routes are plausible, and research

into these possible mediators may give us several empirical answers to the venerable mind-body problem.

3. What happens if we can change someone's explanatory style from pessimistic to optimistic? Seligman et al. (1988) showed that cognitive therapy for depression has the effect of altering someone's explanatory style for the better, and I found the same effect from a program of educational challenge (Peterson, 1990). What remains to be investigated is whether a "born again" optimist reaps the same benefits as people who have had an optimistic explanatory style all along. Perhaps someone who self-consciously offers optimistic attributions is only going through the motions, which means that we must seek explanatory style at a deeper level than researchers have plumbed to date.

4. What about the handful of people who show little consistency in the causal attributions they offer? Such individuals exist, and it is tempting to regard them as flexible in how they explain the events they experience (cf. Taylor, 1983). Perhaps people with no particular explanatory style can be more sensitive to the actual causal texture of the world. Some theorists have argued that it is beneficial for people to be somewhat estranged from the way things actually are, introducing notions like healthy illusions and positive denial (Lazarus, 1983; Taylor & Brown, 1988). These ideas, however intriguing, may be guilty of overstat- ing the matter. Perhaps a flexible point of view is better than a rose-colored one. Indeed, preliminary evidence suggests that attributional "flexibility" is associated with better func- tioning in the world than is an optimistic explanatory style (Silverman, 1989; see also Compas, Forsythe, & Wagner, 1988).

5. Elsewhere I have described explanatory style as a "Velcro" construct (Peterson, 1989), meaning that many things adhere to it, to the constant amazement of those who observe it in operation. Probably the most basic question that can be raised about explanatory style is why it has the array of important correlates that it does (Peterson et al., 1988). Why do people who explain the causes of events in a pessimistic fashion apparently experience more than their share of depressive symptoms, academic failure, deficits in help-seeking, social estrangement, work problems, and poor health?

Explanatory style has come a long way from its origin in the reformulated theory of learned helplessness. What is needed next is another version of the theory to explain which outcomes are most apt to occur for which people in which circumstances. Abramson et al. (1989) provided part of this new theory, by fine-tuning the attributional reformulation of helplessness theory vis-a-vis depression. Their resulting "hopelessness theory" is a detailed process account of how and why explanatory style leads to depression. To my view of things, this is the sort of endeavor to which the analytic energies of psychologists should be directed.

Notes

I thank Scott Bunce and Drew Westen for their comments on a draft of this article.

Christopher Peterson, Department of Psychology, Univer- sity of Michigan, 580 Union Drive, Ann Arbor, MI 48109.




References

Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1988). The hopelessness theory of depression: Does the research test the theory? In L. Y. Abramson (Ed.), Social cognition and clinicalpsychology (pp. 33- 65). New York: Guilford.

Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-based subtype of depression. Psychological Re- view, 96, 358-372.

Abramson, L. Y., & Sackeim, H. A. (1977). A paradox in depression: Uncontrollability and self-blame. Psychological Bulletin, 84, 838- 851.

Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnor- mal Psychology, 87, 49-74.

Alloy, L. B., Peterson, C., Abramson, L. Y., & Seligman, M. E. P. (1984). Attributional style and the generality of learned helplessness. Journal of Personality and Social Psychology, 46, 681-687.

Anderson, C. A., Jennings, D. L., & Amoult, L. H. (1988). Validity and utility of the attributional style construct at a moderate level of specificity. Journal of Personality and Social Psychology, 55, 979-990.

Amtz, A., Gerlsma, C., & Albersnagel, F. A. (1985). Attributional style questioned: Psychometric evaluation of the ASQ in Dutch adolescents. Advances in Behaviour Research and Therapy, 7, 55-89.

Atlas, G. D., & Peterson, C. (1989). Explanatory style among harness racing bettors. Unpublished manuscript, University of Michigan, Ann Arbor.

Baumeister, R. F. (1986). Identity: Cultural change and the struggle for self. New York: Oxford University Press.

Bums, M. O., & Seligman, M. E. P. (1989). Explanatory style across the life span: Evidence for stability over 52 years. Journal of Personality and Social Psychology, 56, 471-477.

Carver, C. S. (1989). How should multifaceted personality constructs be tested?: Issues illustrated by self-monitoring, attributional style, and hardiness. Journal of Personality and Social Psychology, 56, 577- 585.

Compas, B. E., Forsythe, C. J., & Wagner, B. M. (1988). Consistency and variability in causal attribution and coping with stress. Cognitive Ther- apy and Research, 12, 305-320.

Cutrona, C. E. (1983). Causal attributions and perinatal depression. Jour- nal of Abnormal Psychology, 92, 161-172.

Cutrona, C. E., Russell, D., & Jones, R. D. (1984). Cross-situational consistency in causal attributions: Does attributional style exist? Jour- nal of Personality and Social Psychology, 47, 1043-1058.

Eaves, G., & Rush, A. J. (1984). Cognitive patterns in symptomatic and remitted unipolar major depression. Journal of Abnormal Psychology, 93, 31-40.

Funder, D. C., & Ozer, D. J. (1983). Behavior as a function of the situation. Journal of Personality and Social Psychology, 44, 107-112.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Hogan, R., deSoto, C. N., & Solano, C. (1977). Traits, tests, and personality research. American Psychologist, 32, 255-264.

Ickes, W. J., & Layden, M. A. (1978). Attributional styles. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution research (Vol. 2, pp. 119-152). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Kelley, H. H. (1973). The process of causal attribution. American Psychol- ogist, 28, 107-128.

Kelley, H. H., & Michela, J. L. (1980). Attribution theory and research. Annual Review of Psychology, 31, 457-501.

Kuhl, J. (1981). Motivational and functional helplessness: The moderating effect of state versus action orientation. Journal of Personality and Social Psychology, 40, 155-170.

Langer, E. J. (1989). Mindfulness. Reading, MA: Addison-Wesley. Lazarus, R. S. (1983). The costs and benefits of denial. In S. Breznitz

(Ed.), Denial of stress (pp. 1-30). New York: International Univer- sities Press.

Maier, S. F., & Seligman, M. E. P. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology, 105, 3-46.

McCormick, R. A., & Taber, J. I. (1988). Attributional style in patholog- ical gamblers in treatment. Journal of Abnormal Psychology, 97, 368- 370.

Metalsky, G. I., Halberstadt, L. J., & Abramson, L. Y. (1987). Vul- nerability to depressive mood reactions: Toward a more powerful test of

the diathesis-stress and causal mediation components of the reformulated theory of depression. Journal of Personality and Social Psycholo- gy, 52, 386-393.

Miller, J. G. (1984). Culture and the development of everyday social explanation. Journal of Personality and Social Psychology, 46, 961- 978.

Mischel, W. (1968). Personality and assessment. New York: Wiley. Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. P. (1986). Learned

helplessness in children: A longitudinal study of depression, achievement, and explanatory style. Journal of Personality and Social Psy- chology, 51, 435-442.

Pepper, S. C. (1942). World hypotheses. Berkeley: University of Califor- nia Press.

Pervin, L. A. (1985). Personality: Current controversies, issues, and directions. Annual Review of Psychology, 36, 83-114.

Peterson, C. (1978). Learning impairment following insoluble problems: Learned helplessness or altered hypothesis pool? Journal of Experi- mental Social Psychology, 14, 53-68.

Peterson, C. (1986). [Appraisal of bad events]. Unpublished data, Univer- sity of Michigan, Ann Arbor.

Peterson, C. (1988). Explanatory style as a risk factor for illness. Cog- nitive Therapy and Research, 12, 117-130.

Peterson, C. (1989, May). Explanatory style and health. Paper presented at the Attribution-Personality Theory Conference, California School of Professional Psychology, Los Angeles.

Peterson, C. (1990). Explanatory style in the classroom and on the playing field. InS. Graham & V. S. Folkes (Eds.), Attribution theory:Applica- tions to achievement, mental health, and interpersonal conflict (pp. 53-75). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Peterson, C., & Barrett, L. C. (1987). Explanatory style and academic performance among university freshmen. Journal of Personality and Social Psychology, 53, 603-607.

Peterson, C., Bettes, B. A., & Seligman, M. E. P. (1985). Depressive symptoms and unprompted causal attributions: Content analysis. Be- haviour Research and Therapy, 23, 379-382.

Peterson, C., & Bossio, L. M. (1989). Learned helplessness. In R. C. Curtis (Ed.), Self-defeating behaviors (pp. 235-257). New York: Plenum.

Peterson, C., & Bossio, L. M. (1991). Optimism and health. New York: Free Press.

Peterson, C., Luborsky, L., & Seligman, M. E. P. (1983). Attributions and depressive mood shifts. Journal of Abnormal Psychology, 92, 96- 103.

Peterson, C., Schulman, P., Castellon, C., & Seligman, M. E. P. (in press). The explanatory style scoring manual. In C. P. Smith (Ed.), Thematic content analysis for motivation and personality research. New York: Cambridge University Press.

Peterson, C., Schwartz, S. M., & Seligman, M. E. P. (1981). Self-blame and depressive symptoms. Journal of Personality and Social Psycholo- gy, 49, 253-259.

Peterson, C., & Seligman, M. E. P. (1984). Causal explanations as a risk factor for depression: Theory and evidence. Psychological Review, 91, 347-374.

Peterson, C., & Seligman, M. E. P. (1985). The learned helplessness model of depression: Current status of theory and research. In E. E. Beckham & W. R. Leber (Eds.), Handbook of depression: Treatment, assessment, and research (pp. 914-939). Homewood, IL: Dow Jones- Irwin.

Peterson, C., & Seligman, M. E. P. (1987). Explanatory style and illness. Journal of Personality, 55, 237-265.

Peterson, C., Seligman, M. E. P., & Vaillant, G. E. (1988). Pessimistic explanatory style is a risk factor for physical illness: A thirty-five year longitudinal study. Journal of Personality and Social Psychology, 55, 23-27.

Peterson, C., Semmel, A., von Baeyer, C., Abramson, L. Y., Metalsky, G. I., & Seligman, M. E. P. (1982). The Attributional Style Question- naire. Cognitive Therapy and Research, 6, 287-299.

Peterson, C., & Stunkard, A. J. (1989). Personal control and health pro- motion. Social Science and Medicine, 28, 819-828.

Peterson, C., & Villanova, P. (1988). An Expanded Attributional Style Questionnaire. Journal of Abnormal Psychology, 97, 87-89.

Rosenthal, R., & Rubin, D. B. (1982). A simple, general purpose display of magnitude of experimental effect. Journal of Educational Psycholo- gy, 74, 166-169.

Scheier, M. F., & Carver, C. S. (1985). Optimism, coping, and health:



10 PETERSON

Assessment and implications of generalized outcome expectancies. Health Psychology, 4, 219-247.

Seligman, M. E. P. (1975). Helplessness: On depression, development, and death. San Francisco: Freeman.

Seligman, M. E. P., Castellon, C., Cacciola, J., Schulman, P., Luborsky, L., Ollove, M., & Downing, R. (1988). Explanatory style change during cognitive therapy for unipolar depression. Journal of Abnormal Psychology, 97, 13-18.

Seligman, M. E. P., & Schulman, P. (1986). Explanatory style as a predic- tor of productivity and quitting among life insurance agents. Journal of Personality and Social Psychology, 50, 832-838.

Silvernan, R. J. (1989). Attributional style of schizophrenic and depressed outpatients and its role in life adjustment. Unpublished doctoral dissertation, New School for Social Research, New York.

Sweeney, P. D., Anderson, K., & Bailey, S. (1986). Attributional style in depression: A meta-analytic review. Journal of Personality and Social Psychology, 50, 974-991.

Taylor, S. E. (1983). Adjustment to threatening events: A theory of cognitive adaptation. American Psychologist, 38, 1161-1173.

Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193-210.

Taylor, S. E., & Fiske, S. T. (1978). Salience, attention, and attribution: Top of the head phenomena. In L. Berkowitz (Ed.), Advances in experi-

mental social psychology (Vol. 11, pp. 249-288). New York: Academic.

Tennen, H. (1982). A re-view of cognitive mediators of learned helplessness. Journal of Personality, 50, 526-541.

Tennen, H., & Herzberger, S. (1987). Depression, self-esteem, and the absence of self-protective attributional biases. Journal of Personality and Social Psychology, 52, 72-80.

van den Berg, J. H. (1983). The changing nature of man: Introduction to a historical psychology. New York: Norton.

Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer-Verlag.

Weisz, J. R., Rothbaum, F. M., & Blackburn, T. C. (1984). Standing out and standing in: The psychology of control in America and Japan. American Psychologist, 39, 955-969.

Wong, P. T. P., & Weiner, B. (1981). When people ask "why" questions, and the heuristics of attribution search. Journal of Personality and Social Psychology, 40, 649-663.

Wortman, C. B., & Dintzer, L. (1978). Is an attributional analysis of the learned helplessness phenomenon viable?: A critique of the Abram- son-Seligman-Teasdale reformulation. Journal of Abnormal Psychol- ogy, 87, 75-80.

Zullow, H., Oettingen, G., Peterson, C., & Seligman, M. E. P. (1988). Explanatory style and pessimism in the historical record: CAVing LBJ, presidential candidates, and East versus West Berlin. American Psy- chologist, 43, 673-682.



Documents

The Meaning and Measurement of Explanatory Style