Journal of Educational Psychology 1979, Vol. 71, No. 1,3-25

A Theory of Motivation for Some Classroom Experiences

Bernard Weiner University of California, Los Angeles

A theory of motivation based upon attributions of causality for success and failure is offered. The heart of the theory consists of an identification of the dimensions of causality and the relation of these underlying properties of causes to psychological consequences. Three central causal dimensions have been discerned: stability, locus, and control; these dimensions, respectively, are linked with expectancy change, esteem-related emotions, and interperso- nal judgments. Within achievement-related contexts, this theory is pertinent to a diverse array of phenomena and topics, including self-esteem mainte- nance, achievement-change programs, reinforcement schedules, hopelessness, sources of emotion, helping, evaluation, and liking. The range of the theory is further demonstrated by applications to hyperactivity, mastery, parole deci- sions, loneliness and affiliation, and depression. It appears that a general theory of motivation is under development that has important implications for the understanding of classroom thought and behavior.

The attributional approach to classroom certain laws, while other notions remain motivation and experience has proven ex- unchanged. ceedingly rich. In this article I examine the particular attributional path I have followed and document its richness by outlining a few The Search for Causes of the empirical and theoretical relations A central assumption of attribution that appear to be conclusive. The extensity theory, which sets it apart from pleasure- of the theoretical network suggests that a pain theories of motivation, is that the search general theory of motivation is under de- for understanding is the (or a) basic "spring velopment; I also address the issue of theo- of action." This does not imply that hu- retical breadth here. mans are not pleasure seekers, or that they Some of the thoughts expressed in this never bias information in the pursuit of he- article have been voiced in previous reviews donic goals. Rather, information seeking (Weiner, 1972,1974,1976). With each op- and veridical processing are believed to be portunity to take stock of where we are, some normative, may be manifested in spite of a ideas become more firmly fixed, others are conflicting pleasure principle, and, at the discarded and new presumptions take their least, comprehension stands with hedonism place, some earlier evidence grows in stature, among the primary sources of motivation and other prior data require reinterpreta- (see W. Meyer, Folkes, & Weiner, 1976). tion. There certainly is some advantage to In a school setting the search for under- the dictum of publish and perish, which al- standing often leads to the attributional lows one to convey his or her ideas in a single, question of "Why did I succeed or fail?" or, self-contained, and final package. Like more specifically, "Why did I flunk math?" most others, however, I communicate my or "Why did Mary get a better mark on this thoughts as they evolve, and prior ques- exam than me?" But classrooms are envi- tionable truths give way to new, equally un- ronments for the satisfaction of motivations other than achievement. Thus, attribu- This article was written while the author was sup- tional questions also might pertain to, for ported by Grant MH 25687-04 from the National In- example, interpersonal acceptance or rejec- stitute of Mental Health. Requests for reprints should be sent to Bernard tion, such as "Why doesn't Johnny like me?" Weiner, Department of Psychology, University of Cal- However, for the time being attention will be ifornia, Los Angeles, California 90024. centered upon achievement concerns.

Copyright 1979 by the American Psychological Association, Inc. 0022-0663/79/7101-0003$00.75 3 BERNARD WEINER Among the unknowns of this attributional the only perceived determinants of success analysis is a clear statement of when people or failure, or even that they were the most ask "why" questions. It has been demon- salient ones in all achievement situations. strated that this search is more likely given In later work (e.g., Weiner, 1974; Weiner, failure (rejection) than success (acceptance) Russell, & Lerman, 1978), we explicitly in- (Folkes, 1978). Furthermore, it is plausible dicated that factors such as mood, fatigue, to speculate that unexpected events are more illness, and bias could serve as necessary likely to lead to "why" questions than ex- and/or sufficient reasons for achievement pected events (Lau & Russell, Note 1) and performance. Research restricting causality that subjective importance also will influ- to the four causes given above at times might ence the pursuit of knowledge. Finally, it give rise to false conclusions. For example, has been demonstrated that during task assume that one is testing the hedonic bias performance "failure-oriented" or "helpless" notion that success primarily is self-ascribed. students especially tend to supply attribu- By not including help from others, for ex- tions (Diener & Dweck, 1978). Diener and ample, among the alternative causes, the Dweck also intimate that a subset of stu- hedonic bias hypothesis might be supported dents, called "mastery-oriented," do not because the given external causes (task dif- engage in attribution making. However, I ficulty and luck) do not adequately capture suspect that attributional inferences often the phenomenology of the subject. are quite retrospective, summarize a number In the last few years intuition has given of experiences, take place below a level of way to empirical studies attempting to immediate awareness, and are intimately identify the perceived causes of success and tied with self-esteem and self-concept. failure. At least four investigations of aca- Thus, I believe that attributions are supplied demic attributions (Elig & Frieze, 1975; by the mastery-oriented children as well, Frieze, 1976; Bar-Tal & Darom, Note 2; although not necessarily during or immedi- Cooper & Burger, Note 3) have been con- ately following all task performances. ducted (there undoubtedly are many more Our initial statement regarding the per- unknown to me), and there have been a ceived causes of success and failure (Weiner, number of studies that examine attributions Frieze, Kukla, Reed, Rest, & Rosenbaum, outside of the classroom context (e.g., work 1971) was guided by Heider (1958) as well as experiences and athletics). The methodol- our own intuitions. We postulated that in ogies of the classroom inquiries have minor achievement-related contexts the causes variations, with students or teachers stating perceived as most responsible for success and the causes of success or failure at real or im- failure are ability, effort, task difficulty, and agined events, and judging themselves or luck. That is, in attempting to explain the others. The responses are then categorized prior success or failure at an achievement- and tabulated. related event, the individual assesses his or Cooper and Burger (Note 3) provide a her level of ability, the amount of effort that concise summary of the data from three of was expended, the difficulty of the task, and the studies (see Table 1). It is evident that the magnitude and direction of experienced ability, effort (both typical and immediate), luck. We assumed that rather general and task difficulty are among the main per- values are assigned to these factors and that ceived causes of achievement performance. the task outcome is differentially ascribed to Thus, the prior intuitions of Heider (1958) the causal sources. In a similar manner, and my colleagues and me were not incor- future expectations of success and failure rect. In addition, Table 1 shows that others would then be based upon one's perceived (teachers, students, and family), motivation level of ability in relation to the perceived (attention and interest), and what Cooper difficulty of the task (labeled by Heider as and Burger label as acquired characteristics can), as well as an estimation of the intended (habits and attitudes) and physiological effort and anticipated luck. processes (mood, maturity, and health) In listing the four causes reported above comprise the central determinants of success we did not intend to convey that they were and failure. Luck is not included with the A THEORY OF MOTIVATION

Table 1 Summary of Previous Coding Systems (Adapted From Cooper & Burger, Note 3)

Frieze (1976) Bar-Tal & Darom (Note 2) Cooper & Burger (Note 3)

Ability Ability Academic ability Stable effort Effort during test Physical and emotional ability Immediate effort Preparation at home Previous experience Task Interest in the subject matter Habits Other person Difficulty of the test Attitudes Mood Difficulty of material Self-perceptions Luck Conditions in the home Maturity Other Typical effort Effort in preparation Attention Directions Instruction Task Mood Family Other students Miscellaneous dominant causes but could be prominent on 1971). Heider also clearly acknowledged the specific occasions, particularly in career or distinction between a naive psychology and athletic accomplishments (see Mann, a scientific psychology. He stated, 1974). In sum, there are a myriad of perceived There is no prior reason why the causal description causes of achievement events. In a cross- [scientific language] should be the same as the phe- cultural study it was even reported that pa- nomenal description [naive language], though, of course, tience (Greece and Japan) and tact and unity the former should adequately account for the latter. (India) are perceived as causes of success and (Heider, 1958, p. 22). failure (Triandis, 1972). But there is a I now turn from the layperson's perception rather small list from which the main causes of causality to the scientific language that is repeatedly are selected. Furthermore, imposed on these causes. In this article I within this list ability and effort appear to be completely neglect the process by which the most salient and general of the causes. causal beliefs are reached, although this is That is, outcomes frequently depend upon the most common problem in the attribu- what we can do and how hard we try to do it. tional field and is what is meant by the at- A clear conceptual analysis of only ability tribution process (see Kelley, 1967, 1971; and effort would greatly add to our knowl- Weiner, 1974). This void is left so that full edge, given an attributional perspective. space can be devoted to the psychological Before moving on to this conceptual for- consequences of perceived causality, the mulation, it should be recognized that Table topic most central to my concerns. 1 presents only a description of the perceived reasons for success and failure in achieve- ment settings. Although attribution theory Dimensions of Causality often is referred to as a naive conception, using the language of the person on the Inasmuch as the list of conceivable causes street, it also has been appreciated that of success and failure is infinite, it is essential science has to go beyond mere phenomeno- to create a classification scheme or a taxon- logy. That is, order must be imposed using omy of causes. In so doing, similarities and scientific terminology that may not be part differences are delineated and the underly- of the logic of the layperson. This is implicit ing properties of the causes are identified. in, for example, the work of Kelley (1967, This is an indispensable requirement for the 6 BERNARD WEINER construction of an attributional theory of ble. Effort and attention may be augmented motivation. or decreased from one episode to the next, The prior theoretical analyses of Rotter while mood is conceived as a temporary (1966) and Heider (1958) were available to state. However, as indicated previously, the serve as our initial guides in this endeavor. perceived properties of a cause can vary. Rotter and his colleagues proposed a one- For example, mood might be thought of as a dimensional classification of causality. temporary state or as a permanent trait. In Causes either were within (internal) or out- addition, experimenters can alter the per- side (external to) the person. In a similar ceived properties of a cause. For example, manner, Heider (1958) as well as de Charms although difficulty level of a task generally (1968), Deci (1975), and many others have is considered a stable characteristic (Weiner articulated an internal-external classifica- et al., 1971), Valle and Frieze (1976) por- tion of causality. Rotter labeled this di- trayed task difficulty as unstable by an- mension locus of control, whereas in the choring this concept to assigned sales terri- present context locus is conceived as a tory, which could be shifted for any sales- backward-looking belief and therefore is person. At times task difficulty is classified referred to as locus of causality. Indeed, I as stable, while the experimental manipu- contend that the concepts of locus and con- lation strongly suggests that subjects would trol must be separated. perceive this factor as unstable (see Riemer, The causes listed in Table 1 can be readily 1975). catalogued as internal or external to the in- Still a third dimension of causality that dividual. From the perspective of the stu- was identified by Heider and later incorpo- dent, the personal causes include ability, rated into the achievement domain by Ros- effort, mood, maturity, and health, while enbaum (1972) was labeled intentionality. teacher, task, and family are among the ex- Causes such as effort or the bias of a teacher ternal sources of causality. But the relative or supervisor were categorized as intentional, placement of a cause on this dimension is not whereas ability, the difficulty of the task, invariant over time or between people. For mood, and so on were specified by Rosen- example, health might be perceived as an baum to be unintentional. internal ("I am a sickly person") or as an In prior writings this distinction was ac- external ("The 'flu bug' got me") cause of cepted (e.g., Weiner, 1974, 1976). But fol- failure. Inasmuch as attribution theory lowing a suggestion of Litman-Adizes (Note deals with phenomenal causality, such per- 4), it is now apparent that Rosenbaum (1972) sonal interpretations must be taken into mislabeled this dimension. Rosenbaum account. That is, the taxonomic placement argued that the dimension of intentionality of a cause depends upon its subjective is needed to differentiate, for example, mood meaning. Nonetheless, in spite of possible from effort. Both of these are internal and individual variation, there is general agree- unstable causes, yet intuitively they are quite ment when distinguishing causes as internal distinct. Rosenbaum invoked the intent or external. dimension to describe this difference, with A second dimension of causality, which we mood classified as unintentional and effort have come to perceive as increasingly im- classified as intentional. However, it seems portant, is labeled stability (Weiner et al., that the dimension Rosenbaum had identi- 1971). The stability dimension defines fied was that of control. Failure attributed causes on a stable (invariant) versus unstable to a lack of effort does not signify that there (variant) continuum. Again Heider (1958) was an intent to fail. Intent connotes a de- served as our guide, for he contrasted dis- sire, or want. Rather, effort differs from positional and relatively fixed characteristics mood in that only effort is perceived as such as ability with fluctuating factors such subject to volitional control. Hence, I pro- as effort and luck. Examining Table 1, pose that a third dimension of causality ca- ability, typical effort, and family would be tegorizes causes as controllable versus un- considered relatively fixed, while immediate controllable. effort, attention, and mood are more unsta- Causes theoretically can be -classified A THEORY OF MOTIVATION Table 2 Causes of Success and Failure, Classified According to Locus, Stability, and Controllability

Internal External

Controllability Stable Unstable Stable Unstable

Uncontrollable Ability Mood Task difficulty Luck Controllable Typical effort Immediate effort Teacher bias Unusual help from others within one of eight cells (2 levels of locus X intent from control concerns a criminal who 2 levels of stability X 2 levels of control). does not want to commit a crime but cannot Among the internal causes, ability is stable control the compulsion. Criminal justice and uncontrollable; typical effort is stable also accepts the possibility of control without and controllable; mood, fatigue, and illness intent, as in negligence. are unstable and uncontrollable; and tem- Still another possible dimension of cau- porary exertion is unstable and controllable. sality, identified by Abramson, Seligman, Among the external causes, task difficulty is and Teasdale (1978), has been labeled stable and uncontrollable; teacher bias may globality. The global versus specific ends be perceived as stable and controllable; luck of this dimension capture the concept of is unstable and uncontrollable; and unusual stimulus generalization (while stability ex- help from others is unstable and controllable presses temporal generalization). For ex- (see Table 2). ample, one's ability may be perceived as Some problems with this classification task-specific ("I failed because I am poor at scheme remain unsolved, particularly among math") or as a general trait influencing the external causes. For example, can an performance in a wide variety of settings ("I external cause be perceived as controllable? failed because I am dumb"). The answer to this question depends on how The dimensions of causality introduced far back one goes in a causal inference chain above were derived from a logical examina- as well as whether controllability assumes tion of perceived causes. More recently, a only the perspective of the actor, which is not number of investigators have employed the case in Table 2 (e.g., teacher bias may be techniques such as factor analysis or multi- controllable from the vantage point of the dimensional scaling to discover the dimen- teacher, but not given the perspective of the sions of causality (e.g., J. Meyer, 1978; pupil). These questions, as well as the pro- Passer, 1977; Michela, Peplau, & Weeks, posed independence of the dimensions, are Note 5). In the inceptive study by Passer, difficult issues for future thought and re- male and female subjects rated the similarity search. of the causes of either success or failure. Although the main dimensions of causal- Eighteen causes were presented in all pos- ity in achievement-related contexts may sible pairs to the subjects. The similarity have been identified, other dimensions are judgments provided the input for a multi- likely to emerge with further analysis and dimensional scaling procedure. This will raise additional problems about the in- method is akin to a cluster analysis and de- dependence of the dimensions. Intention picts the underlying judgment dimensions. may be one of these dimensions and logically Passer found two clear dimensions of could be separable from control (although causality: (a) a locus dimension, anchored causes are certain to correlate highly on these at the internal end with causes such as bad two dimensions). A causal statement re- mood and no self-confidence and at the ex- garding a neglected homework assignment ternal extreme with causes such as bad illustrating the separation of intent from teacher and hard exam; and (b) an inten- control is "I wanted to study, but could not tional-unintentional dimension (which I will control myself from going out." A con- call controllable-uncontrollable), anchored ceptually similar example disassociating at the controllable end with causes such as BERNARD WEINER never studies hard and lazy, and at the un- yielded results fully supporting the logical controllable extreme with nervous and bad analysis. For the scientist these dimensions mood. The findings reported by Passer are second-order concepts (Schvitz, 1967, p. (1977) were similar for males and females in 59); they are concepts used by attribution both the success and failure scaling solu- theorists to organize the causal concepts of tions. the layperson. The proposed third dimension of causal- ity, stability, was not displayed. Never- Consequences of Causal Properties theless, Passer's results are encouraging in that two of the three dimensions that had I turn now from the dimensions of cau- been presumed did emerge, and other un- sality to the consequences or the implica- anticipated dimensions which had not been tions of these dimensions for thought and part of the logical analysis did not appear. action. I contend that each of the three di- The datajeported by Michela et al. (Note mensions of causality has a primary psy- 5) were equally promising. Although they chological function or linkage, as well as a were concerned with the causes of loneliness, number of secondary effects. The primary two familiar dimensions emerged in their relation of the stability dimension is to the study—stability and locus. There was some magnitude of expectancy change following indication that control also appeared in the success or failure. The locus dimension of data, although it did not come through as an causality has implications for self-esteem, independent dimension and was more evi- one of the emotional consequences of dent among the internal causes. This achievement performance; affect also is a suggests that perhaps control cannot be secondary association for casual stability. paired with externality. The dimensional linkages with expectancy The investigation by J. Meyer (1978) and affect (value) integrate attribution provides the best evidence for the dimen- theory with expectancy-value formulations sions portrayed in Table 2. Meyer gave of motivation as outlined by Atkinson subjects information relevant to the judg- (1964), Lewin (1935), and others (see ment of the causes of success and failure, Weiner, 1972, 1974), although this unifica- such as past history and social norms (Kel- tion is not examined in this article. Finally, ley, 1967). The subjects then rated nine perceived control by others relates to help- possible causes of the outcomes, including ing, evaluation, and liking. The theory thus ability, effort, task difficulty, luck, mood, addresses both self- and other-perception and teacher. A factor analysis of these rat- and intra- as well as interpersonal behavior. ings yielded the three dimensions suggested The locus and control dimensions have a in Table 2. number of secondary effects that also will be It therefore appears that what dimensions very briefly considered. emerge in part depends on the empirical procedure that is used. Given a multidi- Stability mensional scaling method where subjects rate the similarity of the causes, the dimen- The primary conceptual linkage of the sions generated by the logical analysis may stability dimension with expectancy of suc- not be identical to those emerging with the cess was first explored by Weiner et al. empirical procedure. For example, as shown (1971) and has not greatly changed since that in the Passer (1977) data, a naive person may time (see Weiner, 1972,1974,1976). I now not spontaneously recognize that mood, luck, more fully perceive the implications of this and effort are similar because they are un- association, other secondary linkages with stable, and thus a stability dimension of causal stability have been uncovered, and causality will not be evident. On the other the empirical data have grown in clarity. hand, factor-analytic procedures are not But the following discussion is consistent subject to this limitation, and as J. Meyer with prior statements and is partially re- (1978) has demonstrated, this procedure has dundant with these earlier writings. A THEORY OF MOTIVATION Research in the attributional domain has at a block-design task, with different proven definitively that causal ascriptions subjects in the six experimental conditions. for past performance are an important de- Following the success trial(s), expectancy of terminant of goal expectancies. For exam- success and causal ascriptions were assessed. ple, failure that is ascribed to low ability or Expectancy of future success was deter- to the difficulty of a task decreases the ex- mined by having subjects indicate "how pectation of future success more than failure many of the next ten similar designs you that is ascribed to bad luck, mood, or a lack believe that you will successfully complete" of immediate effort. In a similar manner, (Weiner et al., 1976, p. 61). To assess per- success ascribed to good luck or extra exer- ceptions of causality, subjects were required tion results in lesser increments in the sub- to mark four rating scales that were identical jective expectancy of future success at that with respect to either the stability or locus task than does success ascribed to high dimensional anchors but differing along the ability or to the ease of the task. More alternate dimension. Specifically, one at- generally, expectancy shifts after success and tribution question was, "Did you succeed on failure are dependent upon the perceived this task because you are always good at stability of the cause of the prior outcome; these kinds of tasks, or because you tried ascription of an outcome to stable factors especially hard on this particular task?" produces greater typical shifts in expectancy "Always good" and "tried hard," the anchors (increments in expectancy after success and on this scale, are identical on the locus of decrements after failure) than do ascriptions causality dimension (internal), but they to unstable causes. Stated somewhat dif- differ in perceived stability, with ability a ferently, if one attains success (or failure) stable attribute and effort an unstable cause. and if the conditions or causes of that out- In a similar manner, judgments were made come are perceived as remaining unchanged, between "lucky" and "tried hard" (unstable then success (or failure) will be anticipated causes differing in locus), "these tasks are with a greater degree of certainty. But if the always easy" and "lucky" (external causes conditions or causes are subject to change, differing in stability), and "always good" and then there is some doubt that the prior out- "always easy" (stable causes differing in come will be repeated. locus). Thus, the judgments permitted a direct test of the locus versus stability in- Empirical Evidence terpretation of expectancy change. Expectancy estimates were examined A large number of research investigations separately for each of the causal judgments. support the above theoretical contentions The data revealed that within both the in- (e.g., Fontaine, 1974; McMahan, 1973; J. ternal and the external causes, expectancy Meyer, 1978; Ostrove, 1978; Rosenbaum, increments were positively associated with 1972; Valle, 1974; Valle & Frieze, 1976; the stability of the ascription; that is, there Weiner, Nierenberg, & Goldstein, 1976; W. were higher expectancies given ability and Meyer, Note 6; Pancer & Eiser, Note 7). In task ease ascriptions than given effort or luck the Weiner et al. (1976) investigation, it was attributions. Contrasting locus of causality demonstrated that expectancy changes are differences within either the stable or the related to the dimension of stability and are unstable ascriptions disclosed that the dis- not associated with the locus of causality. parate causal locus groups did not differ in This is an important finding, not only be- their expectancies of success. cause two attributional dimensions are discriminated, but also because a vast com- Locus of Control Controversy peting literature relates expectancy changes to the dimension of locus (see Weiner et al., One of my disappointments has been that 1976, for a review). investigators associated with social learning Weiner et al. (1976) gave subjects either 0, theory and locus of control have failed to 1,2,3,4, or 5 consecutive success experiences recognize or admit the stability-expectancy 10 BERNARD WEINER linkage and the existence of other dimen- Weiner, 1971). Hence, Valle and Frieze sions of perceived causality. Some re- (1976) conclude, searchers (e.g., Lefcourt, von Baeyer, Ware, & Cox, Note 8) are incorporating the stabil- There is some value for the difference between the ini- ity dimension into perceived causality scales. tial expectations and the actual outcome that will But this is in contrast with the position of maximally change a person's predictions for the future. other investigators. For example, Phares If the difference is greater than this point, the outcome will be attributed to unstable factors to such a great (1978) states, extent that it will have less influence on the person's future predictions, (p. 581) At the present time there does not appear to be a con- vincing body of data supporting the utility of adding the These ideas have important implications stability dimension .... Even should the addition of stability find support in laboratory studies of expec- for the maintenance of one's self-concept tancy changes, it is not at all clear that.. . [broader] and for attributional change programs (see demonstrations of utility will be forthcoming, (p. Weiner, 1974,1976). For example, assume 270) that an individual with a high self-concept of ability believes that he or she has a high In opposition to this statement, the liter- probability of success at a task. It is prob- ature associating stability with expectancy able that failure then would be ascribed to change is unequivocal, and the findings unstable causes such as luck or mood, which generalize outside of the laboratory as well may not reduce the subsequent expectancy as beyond the achievement domain (as will of success and sustains a high ability self- be documented later). It may indeed be concept. On the other hand, success would that the concept of locus of control has great be ascribed to ability, which increases the utility; my modest hope is that individuals subsequent expectancy (certainty) of success in this area will acknowledge some of the and confirms one's high self-regard. The prior shortcomings in their conceptual converse analysis holds given a low self- analysis of expectancy shifts at skill and concept of ability and a low expectancy of chance tasks and in their limited approach success: Success would be ascribed to un- to causality (for a fuller discussion of these stable factors, and failure to low ability. issues, see Weiner et al., 1976). These attributions result in the preservation of the initial self-concept (see Ames, 1978; Formal Analysis and Self-Concept Fitch, 1970; Gilmore & Minton, 1974; Ickes Maintenance & Layden, 1978). In addition, the above analysis suggests that in change programs McMahan (1973) and Valle and Frieze involving expectancies or self-concept the (1976) have developed formal models of ex- perceived causes of performance must be pectancy shifts based upon the concept of altered, and a modification in self-perception causal stability. Valle and Frieze postulate would have to involve a gradual process that predictions of expectancies (P) are a (Valle & Frieze, 1976). function of the initial expectancy (E) plus In one research investigation guided by the degree to which outcomes (O) are at- the above reasoning, Ames, Ames, and Gar- tributed to stable causes (S): rison (1977) had children of high or low social status in the classroom attribute causality for positive and negative interpersonal out- P = f|E + 0 [f(S)]j. comes. For example, the children were given situations such as, "Suppose you meet In addition, Valle and Frieze (1976) also note a new student at school and you become that the perceived causes of success and friends quickly"; or "Imagine you ask failure are related to the initial expectancy someone to play with you after school, but of success. It has been clearly documented they say they cannot play." The children that unexpected outcomes lead to unstable then attributed causality for each situation attributions, particularly luck (Feather, either to an internal, external, or mutual 1969; Feather & Simon, 1971; Frieze & cause. The data indicated that given nega- A THEORY OF MOTIVATION 11 tive interpersonal outcomes, high-social- to unstable causes only given the chance in- status children made greater use of external structions. causal ascriptions, and given positive inter- A related notion is that information gen- personal outcomes, they made more internal erating lack of effort ascriptions for failure attributions than the low-social-status pu- also should result in response maintenance pils. (see Rest, 1976). There are data in the in- frahuman experimental literature that may be interpreted as supporting this hypothesis. Resistance to Extinction and Lawrence and Festinger (1962), marshalling Achievement Change evidence to support their cognitive disso- nance explanation of extinction, report that resistance to extinction is positively related The stability concept is generalizable to to the effortfulness of a response. Our the body of psychological literature con- analysis suggests that when great exertion is cerning experimental extinction (see Rest, required to attain a reward the salience of 1976). Experimental extinction often is effort as the cause of goal attainment is defined as the cessation of a previously in- augmented. Thus, the expectancy of reward strumental response following the perma- following nonattainment of the goal should nent withholding of the reward. It is rea- be comparatively unchanged and extinction sonable to presume that when a response is prolonged. With repeated nonreward, perceived as no longer instrumental to goal however, the ascription shifts from effort to attainment, the organism will cease making ability and/or task difficulty, thus decreasing that response. Hence, any attribution that expectancy and producing extinction. maximizes the expectation that the response These ideas have more than just a passing will not be followed by the goal should fa- relevance to educational practices. Many of cilitate extinction. On the other hand, at- the burgeoning achievement-chance pro- tributions that minimize goal expectancy grams make direct or indirect use of attri- decrements after nonreward should retard butional principles. These programs often extinction. attempt to induce students to attribute their As discussed above, the stability or in- failures to a lack of effort, which is both un- stability of the perceived causal factors in- stable and under volitional control (see An- fluences the expectancy that the outcome of drews & Debus, 1978; Chapin & Dyck, 1976; an action might change in the future. Dweck, 1975). This goal is expressly es- Therefore, I suggest that resistance to ex- tablished for "failure-oriented" children who tinction is a function of attributions to the apparently ascribe their failures to a lack of causal dimension of stability during the pe- ability, which is a stable and uncontrollable riod of nonreinforcement. Ascriptions of cause (see Diener & Dweck, 1978). Pre- nonreinforcement to bad luck, lack of im- sumably, inasmuch as effort can be increased mediate effort, or other unstable causes are volitionally, ascriptions of nonattainment of hypothesized to minimize expectancy dec- a goal to lack of effort will result in the sus- rements and result in slower extinction than taining of hope and increased persistence attributions of nonattainment of a goal to toward the goal. On the other hand, since perceived stable factors, such as teacher bias, ability is stable and not subject to volitional high task difficulty, or lack of ability. Rest control, ascription of nonattainment of a goal (1976) has presented strong evidence con- to low ability results in giving up and the firming these hypotheses. Inasmuch as cessation of goal-oriented behavior. random reinforcement schedules elicit un- In sum, it is suggested that the relations stable causal attributions (Weiner et al., between diverse independent variables (re- 1971), they also should (and do) increase ward schedules, effortfulness of the response, resistance to extinction. In a similar man- and certain attributional biases) and the ner, chance rather than skill instructions also dependent variables of resistance to extinc- increase resistance to extinction (Phares, tion or persistence in goal-related behavior 1957), presumably because failure is ascribed are mediated by perceptions of causality: 12 BERNARD WEINER Reinforcement schedule; Response effortfulness; (observation)^ Resistance to Person characteristics extinction (attributional biases)

Causal attribution for Stability of the Expectancy of future reward and nonreward attribution goal attainment

(Inference concerning attributional mediation)

Locus drance from students or family, will pre- sumably generate great anger and hostility. In contrast with the rather stable beliefs In this event, externality is positively related about causal stability, our thoughts con- to emotional intensity. Thus, the position cerning locus of causality have fluctuated expressed in Weiner et al. (1971) cannot be greatly. A temporary resolution is proposed correct (see Weiner, 1977; Weiner et al., here which is a synthesis of our previous 1978). antithetical positions and better accounts for We therefore initiated a series of studies the complexity of human affective re- to determine the relation between attribu- sponses. tion and affect (Weiner et al., 1978; Weiner, Initially, Weiner et al. (1971) postulated Russell, & Lerman, in press). In our first that locus of causality is related to the af- investigation, subjects were given a scenario fective consequences of success and failure. that depicted a success or failure experience Emotional reactions were believed to be at an exam, along with a causal attribution maximized given internal attributions for for that outcome (e.g., Joan failed because success and failure and minimized given she did not have the ability). The subjects external attributions. Thus, for example, then reported the affects that they surmised pride and shame, the alleged dominant af- would be experienced in this situation. fects in achievement situations (Atkinson, About 100 affects for success and 150 for 1964; McClelland, Atkinson, Clark, & Low- failure were provided, with responses made ell, 1953), would be most experienced given on rating scales indicating the intensity with personal responsibility for success and fail- which the affects would be experienced. ure, as opposed to instances in which exter- There were two general findings of inter- nal factors such as luck or others were per- est. First, there was a set of outcome-de- ceived as the causal agents. This postulated pendent, attribution-independent affects relation seemed intuitively reasonable, was that represented broad positive or negative consistent with Atkinson's (1964) formula- reactions to success and failure, regardless tions concerning the incentive value of suc- of the "why" of the outcome. Given success, cess and failure, and found support in a va- feelings of pleasure, happiness, satisfaction, riety of research investigations. Because a goodness, and so on were reported as equally detailed account of this position recently was experienced in the disparate attribution presented in this journal (Weiner, 1977), I conditions. In a similar manner, given will not discuss it in any further detail. failure, there were a number of outcome- Subsequently, it became evident that it is linked emotions, such as feeling uncheerful, incorrect to presume an invariant positive displeased, and upset. The outcome-de- relation between internality and the mag- pendent affects for both success and failure nitude of emotional reactions in achievement were reported as the ones that would be most settings. For example, failure ascribed to intensely experienced. others, such as the bias of a teacher or hin- But for both success and failure there were A THEORY OF MOTIVATION 13 Table 3: Percentage of Respondents Stating a Particular Emotion for Success, as a Function of the Attribution for Success

Causal attribution

Unstable Stable Emotion Ability effort effort Personality Others Luck

Competence 30 12 20 19 5 2 Confidence 20 19 18 19 14 4 Contentment 4 4 12 0 7 2 Excitement 3 9 8 11 16 6 Gratefulness 9 1 4 8 43 14 Guilt 1 3 0 3 2 18 Happiness 44 43 43 38 46 48 Pride 39 28 39 43 21 8 Relief 4 28 16 11 13 26 Satisfaction 19 24 16 14 9 0 Surprise 7 16 4 14 4 52 Thankfulness 0 1 0 0 18 4 many emotions discriminably related to affective reactions in achievement contexts. specific attributions. Given success, the Rather, emotions appeared to be either unique attribution-affect linkages were the outcome or attributionally generated, following: ability-competence and confi- without any intervening dimensional dence; typical effort-relaxation; immediate placement. effort-activation; others-gratitude; per- Additional evidence, however, has re- sonality-conceit; and luck-surprise. That sulted in a synthesis of our prior antithetical is, if one perceived that success was caused stances. In a recent study (Weiner, Russell, by ability, then competence and confidence & Lerman, in press) subjects recreated a were reported as intensely experienced; if "critical incident" in their lives in which they one succeeded because of help from others, succeeded (or failed) at an academic exam then the dominant reported affect was because of ability, typical effort, immediate gratitude; and so on. In a similar manner, effort, help (or hindrance) from others, for failure, the attribution-affect linkages personality, or luck. They then listed three were the following: ability-incompetence; affects they experienced in this situation. effort-guilt and shame; personality-resig- Table 3 includes only the emotions that were nation; others-aggression; and luck-surprise reported for success by more than 10% of the (see Weiner et al, 1978). respondents for any particular attribution. It is of interest to point out that at times The table shows the percentage of subjects causal attributions yield opposite reactions in all the attribution conditions reporting for success and failure, as would be expected these relatively shared experiences. given diametric outcomes (respectively, The data in Table 3 are consistent with competence versus incompetence given our previous findings. The most dominant ability attributions; gratitude versus affect, happiness, is expressed regardless of aggression for attributions to others). But the reason for the success. In addition to at times the same emotion accompanies both this outcome-linked emotion, there are sig- positive and negative outcomes (surprise nificant attribution-affect linkages. These given a luck attribution); and given still other associations are as follows: ability-compe- ascriptions, such as typical or immediate tence and pride; other people-gratefulness effort, the emotions that accompany success and thankfulness; stable effort-content- (respectively, relaxation and activation) are ment; personality-pride; and luck-surprise, unrelated to the failure-tied affects (guilt relief, and guilt (the linkages are based on and shame). comparisons within an emotion but across These data suggested we should reject the attributions). supposition that locus of causality mediates The failure data also revealed systematic 14 BERNARD WEINER patterns. There were significant outcome- found that the affects of depression, apathy, linked emotions including disappointment, and resignation were reported primarily as well as attribution-affect associations given internal and stable attributions for consistent with prior research: ability- failure (lack of ability, lack of typical effort, incompetence and resignation; effort-guilt; personality deficit). This suggests that only other people-anger; and luck-surprise. attributions conveying that events will not Additional analyses of these data also change in the future beget feelings of help- demonstrated that causal dimensions play lessness, giving up, and depression. Perhaps an essential role in affective life. Given in- the control dimension also plays a role in ternal attributions for success (ability, effort, generating these particular emotions. personality), the affects pride, competence, Hence, the dimensions of causality relate confidence, and satisfaction were reported to different sets of emotions. more frequently than they were given ex- In another research investigation sup- ternal attributions (others, luck). Internal porting a stability-emotion union, Arkin and ascriptions for failure generated the emo- Maruyama (1979) assessed students' attri- tions of guilt and resignation. In sum, par- butions for their success or failure at a college ticular affects clustered with the internal class. In addition, anxiety associated with causes. Reanalysis of Weiner et al. (1978) school performance was measured. It was revealed virtually identical results. found that among successful students, the It therefore appears that in achievement stability of their attributions was negatively situations there are (at least) three sources correlated with anxiety. That is, when of affect. First, there are emotions tied di- success was ascribed to stable causes, stu- rectly to the outcome. One feels "good" dents reported relatively little anxiety. On given success and "bad" given failure, re- the other hand, among the unsuccessful gardless of the reason for the outcome. students, attributional stability and anxiety These probably are the initial and strongest correlated positively; most fear was reported reactions. Second, accompanying these when failure was perceived as likely to recur general feelings are more distinct emotions, in the future. such as gratitude or hostility if success or failure, respectively, is due to others, surprise Cognition-Emotion Sequence in when the outcome is due to luck, and so on. Achievement Contexts Third, the affects that are associated with self-esteem, such as competence, pride, and On the basis of the above discussion, I shame, are mediated by self-ascriptions. suggest that in achievement-related contexts Many emotional reactions are shared given (and, in particular, school settings), the actor success due to ability or effort, the two progresses through something like the fol- dominant internal attributions. It therefore lowing cognition-emotion scenarios: may be that the central self-esteem emotions 1. "I just received a D in the exam. That that facilitate or impede subsequent is a very low grade." (This generates feel- achievement performance are dimensionally ings of being frustrated and upset.) "I re- linked, referred by the actor to him- or her- ceived this grade because I did not try hard self. Some affects thus seem to be mediated enough" (followed by feelings of shame and by the locus dimension, but in a manner guilt). "There really is something lacking much more complex than was originally in me, and it is permanent" (followed by low posited. It is likely that these dimension- self-esteem or lack of worth and hopeless- linked affects have the greatest longevity and ness). most significance for the individual. 2. "I just received an A on the exam. That is a very high grade" (generating hap- Stability and Affect piness and satisfaction). "I received this grade because I worked very hard during the In addition to the locus-affect linkage, entire school year" (producing contentment there also is a relation between causal sta- and relaxation). "I really do have some bility and emotions. Weiner et al. (1978) positive qualities, and will continue to have A THEORY OF MOTIVATION 15 them in the future" (followed by high self- Rather, it seems that the experiential state esteem and feelings of self-worth, as well as of an origin and correlated behaviors are optimism for the future). exhibited because of the perceived personal control of the situation, or the belief that Some Thoughts About Feelings causality is both internal and controllable. Thus, the discussion of the secondary link- Psychology is completing two movements ages with locus is postponed until the pre- that have relatively neglected the study of sentation of the control dimension of cau- affect. The first is the behavioristic period, sality. which denied verbal report data; the second is the cognitive movement, which focuses on Control intellective structures. In contrast to these periods, I believe that psychologists and Attribution theory as formulated by educators now will turn to the study of af- Heider (1958), Jones and Davis (1965), and fect. Kelley (1967) primarily concerns person- At present many of the investigations of perception, or inferences about the inten- affect in the schools measure some global tions and dispostions of others. But thus far feeling state such as "satisfaction." But for in this article I have only been concerned the study of emotions greater differentiation with self-perception. I believe that one of must be allowed. For example, one might the main contributions of our work has been speculate that differential classroom "at- the adaptation of some principles of social mospheres" provide the opportunity for the perception for the construction of a theory experiencing of disparate emotions. Per- of motivation that has the individual as the haps settings that promote internal ascrip- unit of analysis. tions maximize positive or negative self- In the discussion of the implications of images and feelings of pride and shame. On causal dimensions, self- and other-percep- the other hand, environments that permit tion were not distinguished. Considering more student interaction enhance feelings changes in the expectancy of success, the such as gratitude and anger, inasmuch as same cause-effect logic pertaining to causal attributions of success and failure to others stability should hold when considering are promoted. Overall satisfaction ratings oneself or others. The discussion of affect mask distinctions between, for example, also is equally applicable to both the self and pride and gratitude. It is time that closer others, although of course, the emotional attention was paid to affective life in the experiences are limited to the self and in- classroom. ferred about others. But if success or failure is perceived as being due to certain causes, Secondary Linkages then particular affective experiences should follow. Because of the vast literature in the locus The following examination of the dimen- of control area, it might be anticipated that sion of control centers upon inferences about causal locus is directly linked with many others and how beliefs about another's re- psychological reactions in addition to es- sponsibility for success and failure influence teem-related affects. This indeed is likely an actor's reactions toward that person. to be the case. For example, it has been re- The reactions examined are helping, evalu- ported that locus of control relates positively ation, and sentiments. to behaviors such as information seeking and to experiences such as feeling like an "origin" (de Charms, 1968). In most of this research, Helping however, the concepts of locus and control are united. It is not reasonable to expect Ickes and Kidd (1976), guided by Weiner individuals who attribute failure to a lack of et al. (1971) and Rosenbaum (1972), pro- ability, which is internal but uncontrollable, posed an attributional analysis of helping to seek out information or feel like origins. behavior. A number of investigators prior 16 BERNARD WEINER to Ickes and Kidd (1976) had established tional control, whereas illness is not. When that the tendency to help is influenced by the a failure is perceived as controllable, then perceived cause of the need for aid (e.g., help is withheld; the persons presumably Berkowitz, 1969; Ickes, Kidd, & Berkowitz, should help themselves. For this reason, it 1976; Piliavin, Rodin, & Piliavin, 1969; is much easier to raise charity funds for Schopler & Matthews, 1965). The majority battered children or blindness than for al- of these experiments concluded that help is coholism centers. more likely when the perceived cause of the Guided by the prior research of Barnes, need is an environmental barrier, as opposed Ickes, and Kidd (Note 9), Simon and Weiner to being internal to the person desirous of (in press) applied these ideas to one instance aid. For example, Berkowitz (1969) re- of altruism in the classroom—lending class ported that individuals are more inclined to notes to an unknown classmate. In this in- help an experimental subject when the ex- vestigation, two themes were created for a perimenter, caused a delay in the subject's student's failure to take class notes. One response, in contrast with a condition in theme involved a professor, and the second which the subject is perceived as personally concerned an employer. In the professor responsible for falling behind in the experi- theme, the student always (stable) or some- ment. times (unstable) did not take notes because In their review, Ickes and Kidd (1976) of something about himself (internal) or argued that this locus of control explanation something about the professor (external). of helping confounds the causal dimensions Either he was unable to take good notes of locus and intentionality (which I again will (uncontrollable) or he did not try (control- call controllability). They suggest that in lable), while the professor either was unable the study conducted by Berkowitz (1969), to give a clear lecture or did not try. Thus, the causal ascription to the experimenter is for example, an internal, stable, and uncon- both external and uncontrollable (from the trollable cause was that the student never perspective of the actor), whereas an attri- was able to take good notes (low ability), bution to the subject's own mismanagement while an external, unstable, and uncontrol- is internal to the actor and is perceived by lable cause was that the professor at times the potential helper as under volitional could not give a clear lecture. Each story control. Hence, two dimensions of causality within the eight possible causal combina- are confounded, and it is impossible to de- tions (2 levels of stability X 2 levels of locus termine which of the two causal dimensions X 2 levels of control) elaborated the basic is responsible for the differential help giving. scenario. The second theme involved a work Ickes and Kidd, in contrast with Berkowitz, situation in which the student did not have suggest that it is the controllable aspect of the notes because he (or the boss) always the perceived cause, and not the locus, that (sometimes) was responsible for his coming mediated the disparate help giving. The late to school, which could (could not) have reader should note how similar this analysis been avoided. is to the one pertaining to expectancy shifts Following each causal statement the in skill and chance tasks. Both contro- subjects rated the likelihood of lending their versies point out that the locus of control notes to the student. Judgments were made literature has been plagued by an inadequate on a 10-point scale anchored at the extremes analysis of perceived causality. Further- with "definitely would lend my notes" and more, what is required is research that sep- "definitely would not lend my notes." arates the various causal dimensions. The mean helping judgments for four Other data support the Ickes and Kidd conditions (2 levels of locus X 2 levels of (1976) interpretation of helping behavior. control) are shown in Table 4. Stability did For example, Piliavin et al. (1969) found that not effect the judgments and thus is ignored there is a bias to aid an ill person in distress in the analysis. Table 4 reveals that helping as opposed to helping a drunk. According is reported to be relatively equal and rea- to the above argument, this is because sonably high in all conditions except when drunkenness is perceived as subject to voli- the cause is internal and controllable, in A THEORY OF MOTIVATION 17 Table 4 moral feelings—trying to attain a socially valued goal Mean Likelihood of Helping as a Function of is something that one "ought" to do. Second, rewarding Perceived Locus of Causality and and punishing effort is instrumental to changing be- Controllability havior, inasmuch as effort is believed to be subject to volitional control. On the other hand, ability is per- ceived as nonvolitional and relatively stable and thus Locus of should be insensitive to external control attempts. causality Controllable Uncontrollable (Weiner, 1977, p. 508) Internal 3.13 6.74 Thus, both the moral and control aspects External 7.35 6.98 of evaluation were considered. But it was Note. Data are from Simon and Weiner (in press). Higher not realized that evaluation is conceptually numbers indicate greater likelihood of note lending. similar to behaviors and feelings such as help giving, altruism, liking, and blame. That is, which case aid is unlikely to be given. That there is a pervasive influence of perceived is, if the student did not try to take notes controllability or personal responsibility on (professor theme) or could have avoided interpersonal judgments in achievement- being absent (employer theme), then help is related contexts, including how students are withheld. The findings concerning the in- graded. fluence of intent information on moral judgments and criminal justice support this Sentiments line of reasoning (see Carroll & Payne, 1976, 1977, discussed later in this article). Investigations linking liking to percep- tions of controllability primarily have been conducted in the area of loneliness (see Evaluation Peplau, Russell, & Heim, in press). Michela, Some of the early experimental work Peplau, & Weeks (Note 5) found that per- conducted by me and my colleagues was sons lonely for reasons thought to be con- undertaken to promote the distinction be- trollable (e.g., does not try to make friends) tween various causes of success and failure. are liked less than individuals lonely for In particular, we attempted to provide evi- uncontrollable reasons (e.g., no opportunity dence that ability and effort should be dis- to meet people). In addition, when a lonely tinguished, although both are internal in person puts forth effort to make friends, that locus of causality. person is liked and elicits sympathy (Wimer In one reference experiment that was & Peplau, Note 10). In contrast, if it is be- employed, subjects were asked to pretend lieved that the lonely individual is respon- that they were teachers and were to provide sible for his or her plight, then sympathy is evaluative "feedback" to their pupils (e.g., not forthcoming, and respondents indicate Eswara, 1972; Kaplan & Swant, 1973; Rest, they would avoid such persons. I assume Nierenberg, Weiner, & Heckhausen, 1973; that this pattern of results will also be evi- Weiner & Kukla, 1970; Weiner & Peter, dent in achievement-related contexts. 1973). The pupils were characterized in Surely a teacher will not particularly like a terms of effort, ability, and performance on student who does not try, and failure per- an exam. The data from these investiga- ceived as due to lack of effort does not elicit tions conclusively demonstrated that effort sympathy. is of greater importance than ability in de- termining reward and punishment. High Self-Perception of Control effort was rewarded more than high ability given success, and lack of effort was pun- While perceived control in others relates ished more than lack of ability given failure. to interpersonal judgments, self-perceptions To explain these findings, I stated, of control have quite a different array of consequences. These intrapersonal effects There appear to be two reasons for the discrepancy appear to be vast, ranging from experiential between ability and effort as determinants of reward states, such as feeling as an origin (de and punishment. First, effort attributions elicit strong Charms, 1968) and perceiving freedom of 18 BERNARD WEINER

Antecedent conditions — *• Perceived causes * Causal dimension ' Primary effects — » Other consequences

Ability Performance intensity

Effort (typical) Persistence and immediate) affects

Others (students, Choice family, teacher) judgments Etc.

Motivation (atten- Intentionality tion, interest) Etc. Globality

Figure 1. Partial representation of an attributional theory of motivation. choice (Steiner, 1970), to specific behaviors, Theoretical Range such as information search (see Rotter, 1966) and normal functioning rather than learning, The theory rather sketchily conveyed in cognitive, and motivational deficits that are Figure 1 has been shown to be relevant to postulated to accompany the loss of control many classroom-related thoughts and ac- (Seligman, 1975). This is a complex subject tions. The topics already examined in this matter in need of systematic examination article include the perceived reasons for and synthesis that goes well beyond the success and failure, expectancy change, scope of our present knowledge. self-concept maintenance, achievement change programs, reinforcement schedules, Summary hopelessness, sources of emotion, self-es- teem, helping, evaluation, and liking. Still A variety of sources of information (not other achievement-related topics have been discussed here) are used to reach causal in- demonstrated to be encompassed within this ferences in achievement-related contexts. attributional conception (see Weiner, 1974, The perceived causes of success and failure 1976). The breadth of the phenomena in- primarily are ability and effort but also in- corporated within our attributional frame- clude a small number of other salient factors work intimates that a general theory of mo- such as home environment and teacher, and tivation is being constructed. In the re- a countless host of idiosyncratic factors. mainder of this article I document other These causes can be comprised within three areas to which the theory is applicable. primary dimensions of causality: stability, Some of the theoretical extensions are ger- locus, and control. There also are an unde- mane to the school setting, while other topics termined number of subordinate causal di- are of interest to an audience of educational mensions, including perhaps intentionality psychologists* primarily because they dem- and globality. The three main dimensions, onstrate the range of the conception. respectively, are linked to expectancy changes, esteem-related affects, and inter- personal judgments (decisions about help- Hyperactivity and Psychostimulants ing, evaluation, and sentiments). In addi- tion, there are secondary linkages between Whalen and Henker (1976) have outlined the causal dimensions and psychological an attributional analysis of the effects of effects: Stability relates to depression-type drug treatment for hyperactive children. affects, and control is associated with par- They contend that when hyperactivity is ticular feeling states and behaviors. The combated with a drug, the belief is conveyed dimension-consequence linkages influence to both the child and his or her parents that motivated behaviors such as persistence the cause of hyperactivity is a physiological and choice. This theory is depicted in dysfunction. Hence, the involved individ- Figure 1. uals are not responsible for or in control of A THEORY OF MOTIVATION 19 the maladaptive behavior that is exhibited. placed 5-year-olds in an experimental room Because this physiological deficit is per- containing two machines. The machines ceived as an uncontrollable cause, neither each had colored lights and movable handles. the child nor the parents need feel guilty or For one machine (A), the onset of the lights blame themselves for the aberrant behavior. was preprogrammed by the experimenter. That is, the shift in perceived causality from The lights in the alternate machine (B) went "lack of effort" minimizes self-blame, low on or off only when the handle was moved self-esteem, and negative evaluations from beyond a certain point. Thus, although others. This appears to be a beneficial and both machines stimulated the viewer per- an unanticipated side effect of the treatment ceptually, the children were the producers or technique. the cause of the stimulation only with ma- On the other hand, Whalen and Henker chine B. (1976) also state that "the reputed physio- The subjects in this experiment were free logical dysfunctions used to explain the to spend their time with either machine. failures of hyperactive children are fre- The experimenters recorded various indexes quently viewed as stable and relatively un- of choice or preference, such as the time responsive to behavior change effects" (p. spent with each machine and verbal reports 1123). Thus, the perception of fixed cau- of liking. Both observational and self-report sation might lead to "demoralization about data revealed that the children strongly problem solutions.... and interferes with preferred machine B over machine A. effective coping" (Whalen & Henker, 1976, These findings have been replicated by p. 1124). Weiner, Kun, and Weiner (in press). In sum, again this is an analysis of a psy- From the theoretical perspective shown in chological phenomenon from the perspective Figure 1, the experiment by Nuttin (1973) shown in Figure 1. Individuals utilize in- illustrates a temporal sequence involving the formation (treatment technique) to infer use of information, inferences concerning causation about an event (hyperactivity). locus of causality, positive affect, and some The perceived cause (a genetic deficit) is behavioral consequences of emotional states. perceived as uncontrollable and stable. This That is, on the basis of the observed covar- minimizes certain negative affects and un- iation between their own actions and the favorable evaluations (beneficial effects) but onset of the lights in machine B, the children also weakens the perceived possibility of infer that they are personally responsible recovery (a harmful consequence). These (ability and effort) for the stimulation from two factors, in turn, influence the long-range that machine. Self-attribution for the out- influence of the treatment (negatively, ac- come increases positive esteem-related af- cording to Whalen and Henker, 1976, inas- fects, and the augmented affect increases the much as they perceive expectancy to be the probability of engaging in the action again as more potent determinant of long-term be- well as increasing "liking" about playing with havioral change). the machine. This interpretation is applicable to an- Mastery other developmental study that has not been conceptualized as involving mastery-type behavior. Watson (1966, 1967) demon- The labels mastery and competence are strated that 8-week-old infants can learn an prominent among the writings of many instrumental response (a head turn) to in- psychologists (e.g., Nissen, 1954; White, crease stimulation (the movement of a mo- 1959). However, in my opinion systematic bile). He also reported that infants in the experimental work elucidating these alleged instrumental response condition apparently motivators of behavior has not been con- displayed more instances of positive affect ducted. An investigation by Nuttin (1973), (smiling and cooing) than children in a con- described as demonstrating "causality dition in which the mobile movement was pleasure," could provide an important ex- controlled by the experimenter. This again perimental paradigm for this area. Nuttin suggests the following temporal sequence: 20 BERNARD WEINER

Specific case Judgment of / "Punishment" information about goodness—badness \. severity the crime and the I likeableness criminal, e.g., ••Tresponsibility crime description, prior record y ) Parole decision .^ prison term * N criminality General knowledge (beliefs) about crime, criminals, P'Risk" or any other rele- Judgment of / recidivism vant knowledge future acts S. expectations

Figure 2. An attributional framework for the parole decision process (in this case, an attributional analysis of perceptions of crime and criminals). (From Carroll and Payne, 1977, p. 200. Copyright 1977 by Plenum Press, Inc. Reprinted by permission.) response-outcome covariation - perceived process as conceptualized by Carroll and internal causation - positive affects of com- Payne (1976, 1977). The figure indicates petence and pride - choice. That is, the that the decision maker is provided with a enhanced positive affect and subsequent variety of information about the criminal, performance of the instrumental response the crime, and other pertinent facts. This are mediated by perceptions of self-respon- information is combined and synthesized, sibility (perhaps the control dimension also yielding attributions about the cause of the plays a role here). crime. The causal attributions, in turn, in- The underlying premise of this inter- fluence judgments about deserved punish- pretation of Watson's (1966,1967) research ment and social risk, which are believed to be is that affect and choice can be used to infer the basis for the final parole decision. cognitive processes (perceptions of causali- Carroll and Payne (1976), after reviewing ty). It may seem far-fetched to draw the an extensive literature, contend that the inference that 8-week-olds have the cogni- parole decision process is tive capacities to make causal deductions. However, it also may be that a differentia- based on a simple two-part model. In the first part, the tion between the self and the environment primary concern of the decision maker is to make the punishment fit the crime .... At the second part. .. has developed by that age, and that primitive the primary concern ... is with parole risk, i.e., the inferences about locus and control can be probability that the person being considered for release made using proprioceptive feedback infor- will again violate the laws of society, (p. 15) mation. If this interpretation has any va- lidity, then Watson perhaps has identified According to Figure 2, crimes that are as- the existence of attempts at mastery among cribed to internal and/or intentional (con- very young infants. Note also that one may trollable) factors (e.g., personality charac- consider the contribution of the Watson in- teristics, evil intents) should result in vestigations from a light somewhat different harsher evaluation (punishment) than from what is usual for psychologists—the crimes attributed to external and/or unin- observation of the behavior is of interest tentional (uncontrollable) causes (e.g., eco- primarily because it tells us something about nomic conditions, bad friends, etc.). In ad- the contents of the mind! dition, the risk associated with parole should depend on the stability of the perceived Parole Decisions cause of the crime. If, for example, the crime is attributed to some fixed personality A parole decision is a complex judgment trait, then the decision maker will expect in which causal attributions play a major that a crime again will be committed if the role. Figure 2 depicts the parole decision prisoner is paroled. On the other hand, if A THEORY OF MOTIVATION 21 the cause of the crime has been or can be al- and failure and found them to be similar to tered (e.g., economic conditions have im- the causes of achievement success and fail- proved, a job can be found, etc.), then the ure. As already indicated, Michela et al. criminal will be perceived as a good parole (Note 5) used scaling procedures to discover risk. the dimensions of the causes of social failure Given the above analysis, a criminal is and found them to be similar to the dimen- least likely to be paroled if the cause of the sions uncovered in achievement contexts. crime is perceived as internal and/or con- The question that then remains is whether trollable but stable ("He is an evil person"). the attributional dimensions in the affiliative Conversely, parole will tend to be granted domain relate to psychological factors in the when the crime is perceived as caused by same manner as in the achievement domain. external and/or noncontrollable and unsta- Research reveals that is indeed the case (see ble factors (e.g., prior economic conditions). Peplau et al., in press). Stability relates to The remaining causal combinations should the perceived probability of remaining lonely fall between these extremes in terms of pa- in the future, locus is associated with es- role probability. teem-related affects, and as previously Carroll and Payne have furnished support stated, control is linked with liking and for these hypotheses, examining professional sympathy toward the lonely person (also see parole decision makers and the judgments Folkes, 1978). of college students when given simulated criminal cases. They find, for example, that perceptions of the locus, stability, and con- Depression and Learned Helplessness trollability of causes significantly relate to perceived responsibility for the crime, like- In accordance with the trend in loneliness lihood of recidivism, likability, prison term, research, recent explanations of depression and the purpose of the sentence. have focused upon the cognitive, rather than In sum, according to Carroll and Payne the affective, aspects of this disorder (e.g., (1976, 1977) the parole decision procedure Beck, 1976). The work of Seligman (1975), is conceptually identical to the perceived captured under the label of learned help- sequence of events in the achievement do- lessness, has been especially influential. I main: Antecedent information is processed, will not dwell upon Seligman's use of this a causal judgment is reached, and the cause construct or the supporting empirical evi- is placed within the locus, stability, and in- dence in this context. Rather, my goal is to tentionality (control) dimensions. This convey the pertinence of the learned help- influences evaluation and expectancy, which lessness literature to the attributional model are the main determinants of the parole de- depicted in Figure 1 (see Abramson et al., cision. 1978; Weiner & Litman-Adizes, in press). Learned helplessness communicates the Affiliation and Loneliness belief that there is no perceived association between responding and environmental It has been reasoned that in our culture outcomes. That is, the actor believes that two sources of motivation are most domi- the likelihood of an event is independent of nant: achievement and social recognition what he or she does. The belief in help- (or, in Freud's more general terms, Arbeit lessness is alleged to produce deficits in and Liebe). Hence, affiliative motivation motivation and learning, negative affect, and is a natural area to turn toward in the de- a syndrome that has been labeled "depres- velopment of a theory of motivation. sion." An attributional analysis of affiliative As this work has progressed from infra- motivation guided by the theory shown in human to human research, it has become Figure 1 conceives of loneliness as a social evident that it also is essential to consider failure (Gordon, 1976; Stein & Bailey, 1973). why actions and outcomes are perceived to Hanusa (Note 11) and Heim (Note 12) ex- be independent. For example, Klein, Fen- amined the perceived causes of social success cil-Morse, and Seligman (1976) found that 22 BERNARD WEINER only individuals making internal attribu- role decisions, loneliness and affiliation, and tions for response-outcome independence depression. The data strongly suggest that exhibited aspects of the learned helplessness a general conception of motivation, as well syndrome. Attributions of response-out- as a particular method of psychological come independence to external factors did analysis, is evolving. not produce any learning deficits. In a similar manner, Tennen and Eller (1977) Reference Notes found learned helplessness only under con- ditions that promote low ability attributions 1. Lau, R. R., & Russell, D. Attributions in the sports pages: A field test of some current hypotheses in for prior lack of control. attribution research. Unpublished manuscript, Partially because of these data, Abramson University of California, Los Angeles, 1978. et al. (1978) adopted an attributional 2. Bar-Tal, D., & Darom, E. Causal perceptions of framework for helplessness. I have ex- pupils' success or failure by teachers and pupils: tracted the following from the Abramson et A comparison. Unpublished manuscript, Uni- versity of Tel-Aviv, Israel, 1977. al. (1978) manuscript and pieced together 3. Cooper, H. M., & Burger, J. M. Internality, sta- new paragraphs to illustrate their think- bility, and personal efficacy: A categorization of ing. free response academic attributions. Unpublished manuscript, University of Missouri—Columbia, 1978. Our reformulated hypothesis makes a major new set of 4. Litman-Adizes, T. An attributional model of de- predictions. The helpless individual first finds out that pression: Laboratory and clinical investigations. certain outcomes and responses are independent, then Unpublished manuscript, University of California, he makes an attribution about the cause's. This attri- Los Angeles, 1977. bution determines the chronicity, generality, and in- 5. Michela, J., Peplau, L. A.,& Weeks, D. Perceived tensity of the deficits. Depressed people seem to make dimensions and consequences of attributions for more global, stable, and possibly internal attributions loneliness. Unpublished manuscript, University about the cause of their helplessness and as a conse- of California, Los Angeles, 1978. quence show more general, chronic, and intense deficits 6. Meyer, W. U. Selbstanwortlichkeit und Le- than nondepressed people. istungsmotivation. Unpublished doctoral dis- Depression occurs when an individual expects that the sertation, Ruhr Universitat, Bochum, West Ger- probability of a highly preferred outcome is low and he many, 1970. expects that he is helpless to increase it. If the attri- 7. Pancer, S. M., & Eiser, J. R. Expectations, aspi- butions for the present state of affairs are to stable and rations, and evaluations as influenced by another's global factors, the future will look dark to the individual. attributions for success and failure. Paper pre- He expects that he will find himself helpless again and sented at the 83rd Annual Meeting of the American again. This is what is usually meant by "hopelessness." Psychological Association, Chicago, September Another implication of the formulation is that indi- 1975. viduals will show the greatest loss of self-esteem when 8. Lefcourt, H. M., von Baeyer, C. L., Ware, E. E., & they make internal, global, and stable attributions for Cox, D. J. The multidimensional-multiattribu- their failures. (Abramson et al., 1978) tional causality scale: The development of a goal-specific locus of control scale. Unpublished In sum, it is argued that depressed indi- manuscript, University of Waterloo, Ontario, viduals attempt to make sense out of per- Canada, 1978. ceived evidence that their responses do not 9. Barnes, R. D., Ickes, W. J., & Kidd, R. F. Effects of perceived intentionality and stability of an- affect outcomes. A cause is determined other's dependency on helping behavior: A field which often is classified as stable, internal, experiment. Unpublished manuscript, University and global. This leads to a low expectancy of Wisconsin—Madison, 1977. of success across a wide array of environ- 10. Wimer, S. W., & Peplau, L. A. Determinants of ments and a heightened negative affect (loss reactions to lonely others. Paper presented at the 58th Annual Meeting of the Western Psychological of self-esteem), which are sufficient precur- Association, San Francisco, April 1978. sers of depression. 11. Hanusa, B. H. An extension of Weiner's attribu- tion approach to social situations: Sex differences in social situations. Paper presented at the 46th Conclusion Annual Meeting of the Eastern Psychological As- sociation, New York, April 1975. 12. Heim. M. Sex differences in causal attributions I have selectively reviewed the extensive for achievement in social tasks. Paper presented literature outside of the achievement do- at the 46th Annual Meeting of the Eastern Psy- main, including hyperactivity, mastery, pa- chological Association, New York, April 1975. A THEORY OF MOTIVATION 23 References ment in relation to ability, effort, and performance. Journal of Social Psychology, 1972,87, 139-140. Feather, N. T. Attribution of responsibility and va- Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. lence of success and failure in relation to initial con- Learned helplessness in humans: Critique and re- fidence and task performance. Journal of Person- formulation. Journal of Abnormal Psychology, ality and Social Psychology, 1969,13, 129-144. 1978,87, 49-74. Feather, N. T., & Simon, J. G. Causal attributions for Ames, C. Children's achievement attributions and success and failure in relation to expectations of self-reinforcement: Effects of self-concept and success based upon selective or manipulative control. competitive reward structure. Journal of Educa- Journal of Personality, 1971,39, 527-541. tional Psychology, 1978, 70, 345-355. Fitch, G. Effects of self-esteem, perceived perfor- Ames, R., Ames, C., & Garrison, W. Children's causal mance, and choice on causal attributions. Journal ascriptions for positive and negative interpersonal of Personality and Social Psychology, 1970, 16, outcomes. Psychological Reports, 1977, 41, 595- 311-315. 602. Folkes, V. S. Causal communication in the early Andrews, G. R., & Debus, R. L. Persistence and causal stages of affiliative relationships. Unpublished perception of failure: Modifying cognitive attribu- doctoral dissertation, University of California, Los tions. Journal of Educational Psychology, 1978, 70, Angeles, 1978. 154-166. Fontaine, G. Social comparison and some determi- Arkin, R. M., & Maruyama, G. M. Attribution, affect, nants of expected personal control and expected and college exam performance. Journal of Educa- performance in a novel task situation. Journal of tional Psychology, 1979, 71, 85-93. Personality and Social Psychology, 1974, 29, 487- Atkinson, J. W. An introduction to motivation. 496. Princeton, N.J.: Van Nostrand, 1964. Frieze, I. H. Causal attributions and information Beck, A. T. Cognitive therapy and the emotional seeking to explain success and failure. Journal of disorders. New York: International Universities Research in Personality, 1976,10, 293-305. Press, 1976. Frieze, I. H., & Weiner, B. Cue utilization and attri- Berkowitz, L. Resistance to improper dependency butional judgments for success and failure. Journal relationships. Journal of Experimental Social of Personality, 1971,39, 591-605. Psychology, 1969,5, 283-294. Gilmore, T. M., & Minton, H. L. Internal versus ex- Carroll, J. S., & Payne, J. W. The psychology of the ternal attributions of task performanace as a function parole decision process: A joint application of at- of locus of control, initial confidence and success- tribution theory and information processing psy- failure outcome. Journal of Personality, 1974,42, chology. In J. S. Carroll & J. W. Payne (Eds.), Cog- 159-174. nition and social behavior. Hillsdale, N.J.: Erl- Gordon, S. Lonely in America. New York: Simon & baum, 1976. Schuster, 1976. Carroll, J. S., & Payne, J. W. Judgments about crime Heider, F. The psychology of interpersonal relations. and the criminal: A model and a method for inves- New York: Wiley, 1958. tigating parole decision. In B. D. Sales (Ed.), Pros- Ickes, W. J., & Kidd, R. F. An attributional analysis of pectives in law and psychology. Vol. 1: The criminal helping behavior. In J. H. Harvey, W. J. Ickes, & R. justice system. New York: Plenum Press, 1977. F. Kidd (Eds.), New directions in attribution re- Chapin, M., & Dyck, D. G. Persistence in children's search (Vol. 1). Hillsdale, N.J.: Erlbaum, 1976. reading behavior as a function of N length and attri- Ickes, W. J., Kidd, R. F., & Berkowitz, L. Attributional bution retraining. Journal of Abnormal Psychology, determinants of monetary help-giving. Journal of 1976,85, 511-515. Personality, 1976, 44, 163-178. de Charms, R. Personal causation. New York: Ac- Ickes, W. J., & Layden, M. A. Attributional styles. In ademic Press, 1968. J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New Deci, E. L. Intrinsic motivation. New York: Plenum, directions in attribution research (Vol. 2). Hillsdale, 1975. N.J.: Erlbaum, 1978. Diener, C. I., & Dweck, C. A. An analysis of learned Jones, E. E., & Davis, K. E. From acts to dispositions: helplessness: Continuous changes in performance, The attribution process in person perception. In L. strategy, and achievement cognitions following fail- Berkowitz (Ed.), Advances in experimental social ure. Journal of Personality and Social Psychology, psychology (Vol. 2). New York: Academic Press, 1978,36, 451-462. 1965. Dweck, C. S. The role of expectations and attributions Kaplan, R. M., & Swant, S. G. Reward characteristics in the alleviation of learned helplessness. Journal of appraisal of achievement behavior. Representa- of Personality and Social Psychology, 1975, 31, tive Research in Social Psychology, 1973, 4(2), 11- 674-685. 17. Elig, T. W., & Frieze, I. H. A multidimensional scheme Kelley, H. H. Attribution theory in social psychology. for coding and interpreting perceived causality for In D. Levine (Ed.), Nebraska Symposium on Moti- success and failure events: The Coding Scheme of vation: 1967. Lincoln: University of Nebraska Perceived Causality (CSPC). JSAS Catalog of Se- Press, 1967. lected Documents in Psychology, 975,5, 313. (Ms. Kelley, H. H. Attribution in social interaction. No. 1069) Morristown, N.J: General Learning Press, 1971. Eswara, H. S. Administration of reward and punish- Klein, D. C., Fencil-Morse, E., & Seligman, M. E. P. 24 BERNARD WEINER Learned helplessness, depression, and the attribution Riemer, B. S. Influence of causal beliefs on affect and of failure. Journal of Personality and Social Psy- expectancy. Journal of Personality and Social chology, 1976,33, 508-516. Psychology, 1975, 31, 1163-1167. Lawrence, D. H., & Pestinger, L. Deterrents and re- Rosenbaum, R. M. A dimensional analysis of the inforcement. Stanford, Calif.: Stanford University perceived causes of success and failure. Unpub- Press, 1962. lished doctoral dissertation, University of California, Lewin, K. A dynamic theory of personality. New Los Angeles, 1972. York: McGraw-Hill, 1935. Rotter, J. B. Generalized expectancies for internal Mann, L. On being a sore loser: How fans react to versus external control of reinforcement. Psycho- their team's failure. Australian Journal of Psy- logical Monographs, 1966,80(1, Whole No. 609). chology, 1974, 26, 37-47. Schopler, J., & Matthews, M. W. The influence of the McClelland, D. C., Atkinson, J. W., Clark, R. A., & perceived causal locus of partner's dependence of the Lowell, E. L. The achievement motive. New York: use of interpersonal power. Journal of Personality Appleton-Century-Crofts, 1953. and Social Psychology, 1965,2, 609-612. McMahan, I. D. Relationships between causal attri- Schiitz, A. Collected papers. I. The problem of social butions and expectancy of success. Journal of Per- reality. The Hague: Martinus Nijhoff, 1967. sonality and Social Psychology, 1973, 28, 108-114. Seligman, M. E. P. Helplessness: On depression, Meyer, J. P. Dimensions of causal attribution for development, and death. San Francisco: Freeman, success and failure: A multivariate investigation. 1975. Unpublished doctoral dissertation, University of Simon, P., & Weiner, B. Dimensions of causal attri- Western Ontario, London, Canada, 1978. butions related to helping judgments. Personality Meyer, W. U., Folkes, V. S., & Weiner, B. The per- and Social Psychology Bulletin, in press. ceived information value and affective consequences Stein, A. H., & Bailey, M. M. The socialization of of choice behavior and intermediate difficulty task achievement orientation in females. Psychological selection. Journal of Research in Personality, 1976, Bulletin, 1973,80, 345-366. 10, 410-423. Steiner, I. D. Perceived freedom. In L. Berkowitz Nissen, H. W. The nature of drive as innate determi- (Ed.), Advances in experimental social psychology nant of behavioral organization. In M. R. Jones (Vol. 5). New York: Academic Press, 1970. (Ed.), Nebraska Symposium on Motivation: 1954. Tennen, H., & Eller, S. J. Attributional components Lincoln: University of Nebraska Press, 1954. of learned helplessness and facilitation. Journal of Nuttin, J. R. Pleasure and reward in motivation and Personality and Social Psychology, 1977, 35, 265- learning. In D. Berlyne (Ed.), Pleasure, reward, 271. preference. New York: Academic Press, 1973. Triandis, H. The analysis of subjective culture. New Ostrove, N. Expectations for success on effort-deter- York: Wiley-Interscience, 1972. mined tasks as a function of incentive and perfor- Valle, V. A. Attributions of stability as a mediator in mance feedback. Journal of Personality and Social the changing of expectations. Unpublished doctoral Psychology, 1978, 36, 909-916. dissertation, University of Pittsburgh, 1974. Passer, M. W. Perceiving the causes of success and Valle, V. A., & Frieze, I. H. Stability of causal attri- failure revisited: A multidimensional scaling ap- butions as a mediator in changing expectations for proach. Unpublished doctoral dissertation, Uni- success. Journal of Personality and Social Psy- versity of California, Los Angeles, 1977. chlology, 1976, 33, 579-587. Peplau, L. A., Russell, D., & Heim, M. An attributional Watson, J. S. The development and generalization of analysis of loneliness. In I. Frieze, D. Bar-Tal, & J. "contingency awareness" in early infancy: Some Carroll (Eds.), Attribution theory: Applications to hypotheses. Merrill-Palmer Quarterly, 1966, J2(2), social problems. San Francisco: Jossey-Bass, in 123-135. press. Watson, J. S. Memory and "contingency analysis" in Phares, E. J. Expectancy changes in skill and chance infant learning. Merrill-Palmer Quarterly, 1967, situations. Journal of Abnormal and Social Psy- 13(1), 55-76. chology, 1957,54, 339-342. Weiner, B. Theories of motivation: Fram mechanism Phares, E. J. Locus of control. In H. London & J. E. to cognition. Chicago: Rand McNally, 1972. Exner, Jr. (Eds.), Dimensions of personality. New Weiner, B. (Ed.). Achievement motivation and at- York: Wiley, 1978. tribution theory. Morristown, N.J.: General Piliavin, I. M., Rodin, J., & Piliavin, J. A. Good sa- Learning Press, 1974. maritanism: An underground phenomenon? Weiner, B. An attributional approach for educational Journal of Personality and Social Psychology, 1969, psychology. In L. Shulman (Ed.), Review of research 13, 289-299. in education (Vol. 4). Itasca, 111.: Peacock, 1976. Rest, S. Schedules of reinforcement: An attributional Weiner, B. Attribution and affect: Comments on analysis. In J. H. Harvey, W. J. Ickes, & R. F. Kidd Sohn's critique. Journal of Educational Psychology, (Eds.), New directions in attribution research (Vol. 1977,69, 506-511. 1). Hillsdale, N.J.: Erlbaum, 1976. Weiner, B., Frieze, I. H., Kukla, A., Reed, L., Rest, S., Rest, S., Nierenberg, R., Weiner, B., & Heckhausen, H. & Rosenbaum, R. M. Perceiving the causes of suc- Further evidence concerning the effects of percep- cess and failure. Morristown, N.J.: General tions of effort and ability on achievement evaluation. Learning Press, 1971. Journal of Personality and Social Psychology, 1973, Weiner, B., & Kukla, A. An attributional analysis of 28, 187-191. achievement motivation. Journal of Personality A THEORY OF MOTIVATION 25

and Social Psychology, 1970,15, 1-20. velopmental Psychology, 1973, 9, 290-309. Weiner, B., Kun, A., & Weiner, M. B. The development Weiner, B., Russell, D., & Lerman, D. Affective con- of mastery, emotions, and morality from an attribu- sequences of causal ascriptions. In J. H. Harvey, W. tional perspective. In W. A. Collins (Ed.), Minnesota J. Ickes, & R. F. Kidd (Eds.), New directions in at- symposium on child development (Vol. 13). tribution research (Vol. 2). Hillsdale, N.J.: Erl- Hillsdale, N.J.: Erlbaum, in press. baum, 1978. Weiner, B., & Litman-Adizes, T. An attributional, Weiner, B., Russell, D., & Lerman, D. The cognition- expectancy-value analysis of learned helplessness and emotion process in achievement-related contexts. depression. In J. Garber and M. E. P. Seligman Journal of Personality and Social Psychology, in (Eds.), Human control. New York: Academic press. Press, in press. Whalen, C. K., & Henker, B. Psychostimulants and Weiner, B., Nierenberg, R., & Goldstein, M. Social children: A review and analysis. Psychological learning (locus of control) versus attributional (causal Bulletin, 1976,83, 1113-1130. stability) interpretations of expectancy of success. White, R. W. Motivation reconsidered: The concept Journal of Personality, 1916,44, 52-68. of competence. Psychological Review, 1959, 66, Weiner, B., & Peter, N. A cognitive-developmental 297—333 analysis of achievement and moral judgments. De- Received August 11,1978 •