Journal of Educational Psychology1979, Vol. 71, No. 1,3-25

A Theory of Motivation for Some Classroom Experiences

Bernard WeinerUniversity of California, Los Angeles

A theory of motivation based upon attributions of causality for success andfailure is offered. The heart of the theory consists of an identification of thedimensions of causality and the relation of these underlying properties ofcauses to psychological consequences. Three central causal dimensions havebeen 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 pertinentto 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 theoryis further demonstrated by applications to hyperactivity, mastery, parole deci-sions, loneliness and affiliation, and depression. It appears that a generaltheory of motivation is under development that has important implicationsfor the understanding of classroom thought and behavior.

The attributional approach to classroommotivation and experience has proven ex-ceedingly rich. In this article I examine theparticular attributional path I have followedand document its richness by outlining a fewof the empirical and theoretical relationsthat appear to be conclusive. The extensityof the theoretical network suggests that ageneral theory of motivation is under de-velopment; I also address the issue of theo-retical breadth here.

Some of the thoughts expressed in thisarticle have been voiced in previous reviews(Weiner, 1972,1974,1976). With each op-portunity to take stock of where we are, someideas become more firmly fixed, others arediscarded and new presumptions take theirplace, some earlier evidence grows in stature,and other prior data require reinterpreta-tion. There certainly is some advantage tothe dictum of publish and perish, which al-lows one to convey his or her ideas in a single,self-contained, and final package. Likemost others, however, I communicate mythoughts as they evolve, and prior ques-tionable truths give way to new, equally un-

This article was written while the author was sup-ported by Grant MH 25687-04 from the National In-stitute of Mental Health.

Requests for reprints should be sent to BernardWeiner, Department of Psychology, University of Cal-ifornia, Los Angeles, California 90024.

certain laws, while other notions remainunchanged.

The Search for Causes

A central assumption of attributiontheory, which sets it apart from pleasure-pain theories of motivation, is that the searchfor understanding is the (or a) basic "springof action." This does not imply that hu-mans are not pleasure seekers, or that theynever bias information in the pursuit of he-donic goals. Rather, information seekingand veridical processing are believed to benormative, may be manifested in spite of aconflicting pleasure principle, and, at theleast, comprehension stands with hedonismamong the primary sources of motivation(see W. Meyer, Folkes, & Weiner, 1976).

In a school setting the search for under-standing often leads to the attributionalquestion of "Why did I succeed or fail?" or,more specifically, "Why did I flunk math?"or "Why did Mary get a better mark on thisexam than me?" But classrooms are envi-ronments for the satisfaction of motivationsother than achievement. Thus, attribu-tional questions also might pertain to, forexample, interpersonal acceptance or rejec-tion, such as "Why doesn't Johnny like me?"However, for the time being attention will becentered 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 attributionalanalysis is a clear statement of when peopleask "why" questions. It has been demon-strated that this search is more likely givenfailure (rejection) than success (acceptance)(Folkes, 1978). Furthermore, it is plausibleto speculate that unexpected events are morelikely to lead to "why" questions than ex-pected events (Lau & Russell, Note 1) andthat subjective importance also will influ-ence the pursuit of knowledge. Finally, ithas been demonstrated that during taskperformance "failure-oriented" or "helpless"students especially tend to supply attribu-tions (Diener & Dweck, 1978). Diener andDweck also intimate that a subset of stu-dents, called "mastery-oriented," do notengage in attribution making. However, Isuspect that attributional inferences oftenare quite retrospective, summarize a numberof experiences, take place below a level ofimmediate awareness, and are intimatelytied with self-esteem and self-concept.Thus, I believe that attributions are suppliedby the mastery-oriented children as well,although not necessarily during or immedi-ately following all task performances.

Our initial statement regarding the per-ceived causes of success and failure (Weiner,Frieze, Kukla, Reed, Rest, & Rosenbaum,1971) was guided by Heider (1958) as well asour own intuitions. We postulated that inachievement-related contexts the causesperceived as most responsible for success andfailure are ability, effort, task difficulty, andluck. That is, in attempting to explain theprior success or failure at an achievement-related event, the individual assesses his orher level of ability, the amount of effort thatwas expended, the difficulty of the task, andthe magnitude and direction of experiencedluck. We assumed that rather generalvalues are assigned to these factors and thatthe task outcome is differentially ascribed tothe causal sources. In a similar manner,future expectations of success and failurewould then be based upon one's perceivedlevel of ability in relation to the perceiveddifficulty of the task (labeled by Heider ascan), as well as an estimation of the intendedeffort and anticipated luck.

In listing the four causes reported abovewe did not intend to convey that they were

the only perceived determinants of successor failure, or even that they were the mostsalient ones in all achievement situations.In later work (e.g., Weiner, 1974; Weiner,Russell, & Lerman, 1978), we explicitly in-dicated that factors such as mood, fatigue,illness, and bias could serve as necessaryand/or sufficient reasons for achievementperformance. Research restricting causalityto the four causes given above at times mightgive rise to false conclusions. For example,assume that one is testing the hedonic biasnotion that success primarily is self-ascribed.By not including help from others, for ex-ample, among the alternative causes, thehedonic bias hypothesis might be supportedbecause the given external causes (task dif-ficulty and luck) do not adequately capturethe phenomenology of the subject.

In the last few years intuition has givenway to empirical studies attempting toidentify the perceived causes of success andfailure. At least four investigations of aca-demic attributions (Elig & Frieze, 1975;Frieze, 1976; Bar-Tal & Darom, Note 2;Cooper & Burger, Note 3) have been con-ducted (there undoubtedly are many moreunknown to me), and there have been anumber of studies that examine attributionsoutside of the classroom context (e.g., workexperiences and athletics). The methodol-ogies of the classroom inquiries have minorvariations, with students or teachers statingthe causes of success or failure at real or im-agined events, and judging themselves orothers. The responses are then categorizedand tabulated.

Cooper and Burger (Note 3) provide aconcise summary of the data from three ofthe studies (see Table 1). It is evident thatability, effort (both typical and immediate),and task difficulty are among the main per-ceived causes of achievement performance.Thus, the prior intuitions of Heider (1958)and my colleagues and me were not incor-rect. In addition, Table 1 shows that others(teachers, students, and family), motivation(attention and interest), and what Cooperand Burger label as acquired characteristics(habits and attitudes) and physiologicalprocesses (mood, maturity, and health)comprise the central determinants of successand failure. Luck is not included with the

A THEORY OF MOTIVATION

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

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

AbilityStable effortImmediate effortTaskOther personMoodLuckOther

AbilityEffort during testPreparation at homeInterest in the subject matterDifficulty of the testDifficulty of materialConditions in the home

Academic abilityPhysical and emotional abilityPrevious experienceHabitsAttitudesSelf-perceptionsMaturityTypical effortEffort in preparationAttentionDirectionsInstructionTaskMoodFamilyOther studentsMiscellaneous

dominant causes but could be prominent onspecific occasions, particularly in career orathletic accomplishments (see Mann,1974).

In sum, there are a myriad of perceivedcauses of achievement events. In a cross-cultural study it was even reported that pa-tience (Greece and Japan) and tact and unity(India) are perceived as causes of success andfailure (Triandis, 1972). But there is arather small list from which the main causesrepeatedly are selected. Furthermore,within this list ability and effort appear to bethe most salient and general of the causes.That is, outcomes frequently depend uponwhat we can do and how hard we try to do it.A clear conceptual analysis of only abilityand effort would greatly add to our knowl-edge, given an attributional perspective.

Before moving on to this conceptual for-mulation, it should be recognized that Table1 presents only a description of the perceivedreasons for success and failure in achieve-ment settings. Although attribution theoryoften is referred to as a naive conception,using the language of the person on thestreet, it also has been appreciated thatscience has to go beyond mere phenomeno-logy. That is, order must be imposed usingscientific terminology that may not be partof the logic of the layperson. This is implicitin, for example, the work of Kelley (1967,

1971). Heider also clearly acknowledged thedistinction between a naive psychology anda scientific psychology. He stated,

There is no prior reason why the causal description[scientific language] should be the same as the phe-nomenal description [naive language], though, of course,the former should adequately account for the latter.(Heider, 1958, p. 22).

I now turn from the layperson's perceptionof causality to the scientific language that isimposed on these causes. In this article Icompletely neglect the process by whichcausal beliefs are reached, although this isthe most common problem in the attribu-tional field and is what is meant by the at-tribution process (see Kelley, 1967, 1971;Weiner, 1974). This void is left so that fullspace can be devoted to the psychologicalconsequences of perceived causality, thetopic most central to my concerns.

Dimensions of Causality

Inasmuch as the list of conceivable causesof success and failure is infinite, it is essentialto create a classification scheme or a taxon-omy of causes. In so doing, similarities anddifferences are delineated and the underly-ing properties of the causes are identified.This is an indispensable requirement for the

6 BERNARD WEINER

construction of an attributional theory ofmotivation.

The prior theoretical analyses of Rotter(1966) and Heider (1958) were available toserve as our initial guides in this endeavor.Rotter and his colleagues proposed a one-dimensional classification of causality.Causes either were within (internal) or out-side (external to) the person. In a similarmanner, Heider (1958) as well as de Charms(1968), Deci (1975), and many others havearticulated an internal-external classifica-tion of causality. Rotter labeled this di-mension locus of control, whereas in thepresent context locus is conceived as abackward-looking belief and therefore isreferred to as locus of causality. Indeed, Icontend that the concepts of locus and con-trol must be separated.

The causes listed in Table 1 can be readilycatalogued as internal or external to the in-dividual. From the perspective of the stu-dent, the personal causes include ability,effort, mood, maturity, and health, whileteacher, task, and family are among the ex-ternal sources of causality. But the relativeplacement of a cause on this dimension is notinvariant over time or between people. Forexample, health might be perceived as aninternal ("I am a sickly person") or as anexternal ("The 'flu bug' got me") cause offailure. Inasmuch as attribution theorydeals with phenomenal causality, such per-sonal interpretations must be taken intoaccount. That is, the taxonomic placementof a cause depends upon its subjectivemeaning. Nonetheless, in spite of possibleindividual variation, there is general agree-ment when distinguishing causes as internalor external.

A second dimension of causality, which wehave come to perceive as increasingly im-portant, is labeled stability (Weiner et al.,1971). The stability dimension definescauses on a stable (invariant) versus unstable(variant) continuum. Again Heider (1958)served as our guide, for he contrasted dis-positional and relatively fixed characteristicssuch as ability with fluctuating factors suchas effort and luck. Examining Table 1,ability, typical effort, and family would beconsidered relatively fixed, while immediateeffort, attention, and mood are more unsta-

ble. Effort and attention may be augmentedor decreased from one episode to the next,while mood is conceived as a temporarystate. However, as indicated previously, theperceived properties of a cause can vary.For example, mood might be thought of as atemporary state or as a permanent trait. Inaddition, experimenters can alter the per-ceived properties of a cause. For example,although difficulty level of a task generallyis considered a stable characteristic (Weineret al., 1971), Valle and Frieze (1976) por-trayed task difficulty as unstable by an-choring this concept to assigned sales terri-tory, which could be shifted for any sales-person. At times task difficulty is classifiedas stable, while the experimental manipu-lation strongly suggests that subjects wouldperceive this factor as unstable (see Riemer,1975).

Still a third dimension of causality thatwas identified by Heider and later incorpo-rated into the achievement domain by Ros-enbaum (1972) was labeled intentionality.Causes such as effort or the bias of a teacheror supervisor were categorized as intentional,whereas ability, the difficulty of the task,mood, and so on were specified by Rosen-baum to be unintentional.

In prior writings this distinction was ac-cepted (e.g., Weiner, 1974, 1976). But fol-lowing a suggestion of Litman-Adizes (Note4), it is now apparent that Rosenbaum (1972)mislabeled this dimension. Rosenbaumargued that the dimension of intentionalityis needed to differentiate, for example, moodfrom effort. Both of these are internal andunstable causes, yet intuitively they are quitedistinct. Rosenbaum invoked the intentdimension to describe this difference, withmood classified as unintentional and effortclassified as intentional. However, it seemsthat the dimension Rosenbaum had identi-fied was that of control. Failure attributedto a lack of effort does not signify that therewas an intent to fail. Intent connotes a de-sire, or want. Rather, effort differs frommood in that only effort is perceived assubject to volitional control. Hence, I pro-pose that a third dimension of causality ca-tegorizes causes as controllable versus un-controllable.

Causes theoretically can be -classified

A THEORY OF MOTIVATION

Table 2Causes of Success and Failure, Classified According to Locus, Stability, and Controllability

Internal External

Controllability Stable Unstable Stable Unstable

UncontrollableControllable

AbilityTypical effort

MoodImmediate effort

Task difficultyTeacher bias

LuckUnusual help

from others

within one of eight cells (2 levels of locus X2 levels of stability X 2 levels of control).Among the internal causes, ability is stableand uncontrollable; typical effort is stableand controllable; mood, fatigue, and illnessare unstable and uncontrollable; and tem-porary exertion is unstable and controllable.Among the external causes, task difficulty isstable and uncontrollable; teacher bias maybe perceived as stable and controllable; luckis unstable and uncontrollable; and unusualhelp from others is unstable and controllable(see Table 2).

Some problems with this classificationscheme remain unsolved, particularly amongthe external causes. For example, can anexternal cause be perceived as controllable?The answer to this question depends on howfar back one goes in a causal inference chainas well as whether controllability assumesonly the perspective of the actor, which is notthe case in Table 2 (e.g., teacher bias may becontrollable from the vantage point of theteacher, but not given the perspective of thepupil). These questions, as well as the pro-posed independence of the dimensions, aredifficult issues for future thought and re-search.

Although the main dimensions of causal-ity in achievement-related contexts mayhave been identified, other dimensions arelikely to emerge with further analysis andwill raise additional problems about the in-dependence of the dimensions. Intentionmay be one of these dimensions and logicallycould be separable from control (althoughcauses are certain to correlate highly on thesetwo dimensions). A causal statement re-garding a neglected homework assignmentillustrating the separation of intent fromcontrol is "I wanted to study, but could notcontrol myself from going out." A con-ceptually similar example disassociating

intent from control concerns a criminal whodoes not want to commit a crime but cannotcontrol the compulsion. Criminal justicealso accepts the possibility of control withoutintent, as in negligence.

Still another possible dimension of cau-sality, identified by Abramson, Seligman,and Teasdale (1978), has been labeledglobality. The global versus specific endsof this dimension capture the concept ofstimulus generalization (while stability ex-presses temporal generalization). For ex-ample, one's ability may be perceived astask-specific ("I failed because I am poor atmath") or as a general trait influencingperformance in a wide variety of settings ("Ifailed because I am dumb").

The dimensions of causality introducedabove were derived from a logical examina-tion of perceived causes. More recently, anumber of investigators have employedtechniques such as factor analysis or multi-dimensional scaling to discover the dimen-sions of causality (e.g., J. Meyer, 1978;Passer, 1977; Michela, Peplau, & Weeks,Note 5). In the inceptive study by Passer,male and female subjects rated the similarityof the causes of either success or failure.Eighteen causes were presented in all pos-sible pairs to the subjects. The similarityjudgments provided the input for a multi-dimensional scaling procedure. Thismethod is akin to a cluster analysis and de-picts the underlying judgment dimensions.

Passer found two clear dimensions ofcausality: (a) a locus dimension, anchoredat the internal end with causes such as badmood and no self-confidence and at the ex-ternal extreme with causes such as badteacher and hard exam; and (b) an inten-tional-unintentional dimension (which I willcall controllable-uncontrollable), anchoredat the controllable end with causes such as

BERNARD WEINER

never studies hard and lazy, and at the un-controllable extreme with nervous and badmood. The findings reported by Passer(1977) were similar for males and females inboth the success and failure scaling solu-tions.

The proposed third dimension of causal-ity, stability, was not displayed. Never-theless, Passer's results are encouraging inthat two of the three dimensions that hadbeen presumed did emerge, and other un-anticipated dimensions which had not beenpart of the logical analysis did not appear.

The datajeported by Michela et al. (Note5) were equally promising. Although theywere concerned with the causes of loneliness,two familiar dimensions emerged in theirstudy—stability and locus. There was someindication that control also appeared in thedata, although it did not come through as anindependent dimension and was more evi-dent among the internal causes. Thissuggests that perhaps control cannot bepaired with externality.

The investigation by J. Meyer (1978)provides the best evidence for the dimen-sions portrayed in Table 2. Meyer gavesubjects information relevant to the judg-ment of the causes of success and failure,such as past history and social norms (Kel-ley, 1967). The subjects then rated ninepossible causes of the outcomes, includingability, effort, task difficulty, luck, mood,and teacher. A factor analysis of these rat-ings yielded the three dimensions suggestedin Table 2.

It therefore appears that what dimensionsemerge in part depends on the empiricalprocedure that is used. Given a multidi-mensional scaling method where subjectsrate the similarity of the causes, the dimen-sions generated by the logical analysis maynot be identical to those emerging with theempirical procedure. For example, as shownin the Passer (1977) data, a naive person maynot spontaneously recognize that mood, luck,and effort are similar because they are un-stable, and thus a stability dimension ofcausality will not be evident. On the otherhand, factor-analytic procedures are notsubject to this limitation, and as J. Meyer(1978) has demonstrated, this procedure has

yielded results fully supporting the logicalanalysis. For the scientist these dimensionsare second-order concepts (Schvitz, 1967, p.59); they are concepts used by attributiontheorists to organize the causal concepts ofthe layperson.

Consequences of Causal Properties

I turn now from the dimensions of cau-sality to the consequences or the implica-tions of these dimensions for thought andaction. I contend that each of the three di-mensions of causality has a primary psy-chological function or linkage, as well as anumber of secondary effects. The primaryrelation of the stability dimension is to themagnitude of expectancy change followingsuccess or failure. The locus dimension ofcausality has implications for self-esteem,one of the emotional consequences ofachievement performance; affect also is asecondary association for casual stability.The dimensional linkages with expectancyand affect (value) integrate attributiontheory with expectancy-value formulationsof motivation as outlined by Atkinson(1964), Lewin (1935), and others (seeWeiner, 1972, 1974), although this unifica-tion is not examined in this article. Finally,perceived control by others relates to help-ing, evaluation, and liking. The theory thusaddresses both self- and other-perceptionand intra- as well as interpersonal behavior.The locus and control dimensions have anumber of secondary effects that also will bevery briefly considered.

Stability

The primary conceptual linkage of thestability dimension with expectancy of suc-cess was first explored by Weiner et al.(1971) and has not greatly changed since thattime (see Weiner, 1972,1974,1976). I nowmore fully perceive the implications of thisassociation, other secondary linkages withcausal stability have been uncovered, andthe empirical data have grown in clarity.But the following discussion is consistentwith prior statements and is partially re-dundant with these earlier writings.

A THEORY OF MOTIVATION

Research in the attributional domain hasproven definitively that causal ascriptionsfor past performance are an important de-terminant of goal expectancies. For exam-ple, failure that is ascribed to low ability orto the difficulty of a task decreases the ex-pectation of future success more than failurethat is ascribed to bad luck, mood, or a lackof immediate effort. In a similar manner,success ascribed to good luck or extra exer-tion results in lesser increments in the sub-jective expectancy of future success at thattask than does success ascribed to highability or to the ease of the task. Moregenerally, expectancy shifts after success andfailure are dependent upon the perceivedstability of the cause of the prior outcome;ascription of an outcome to stable factorsproduces greater typical shifts in expectancy(increments in expectancy after success anddecrements after failure) than do ascriptionsto unstable causes. Stated somewhat dif-ferently, if one attains success (or failure)and if the conditions or causes of that out-come are perceived as remaining unchanged,then success (or failure) will be anticipatedwith a greater degree of certainty. But if theconditions or causes are subject to change,then there is some doubt that the prior out-come will be repeated.

Empirical Evidence

A large number of research investigationssupport the above theoretical contentions(e.g., Fontaine, 1974; McMahan, 1973; J.Meyer, 1978; Ostrove, 1978; Rosenbaum,1972; Valle, 1974; Valle & Frieze, 1976;Weiner, Nierenberg, & Goldstein, 1976; W.Meyer, Note 6; Pancer & Eiser, Note 7). Inthe Weiner et al. (1976) investigation, it wasdemonstrated that expectancy changes arerelated to the dimension of stability and arenot associated with the locus of causality.This is an important finding, not only be-cause two attributional dimensions arediscriminated, but also because a vast com-peting literature relates expectancy changesto the dimension of locus (see Weiner et al.,1976, for a review).

Weiner et al. (1976) gave subjects either 0,1,2,3,4, or 5 consecutive success experiences

at a block-design task, with differentsubjects in the six experimental conditions.Following the success trial(s), expectancy ofsuccess and causal ascriptions were assessed.Expectancy of future success was deter-mined by having subjects indicate "howmany of the next ten similar designs youbelieve that you will successfully complete"(Weiner et al., 1976, p. 61). To assess per-ceptions of causality, subjects were requiredto mark four rating scales that were identicalwith respect to either the stability or locusdimensional anchors but differing along thealternate dimension. Specifically, one at-tribution question was, "Did you succeed onthis task because you are always good atthese kinds of tasks, or because you triedespecially hard on this particular task?""Always good" and "tried hard," the anchorson this scale, are identical on the locus ofcausality dimension (internal), but theydiffer in perceived stability, with ability astable attribute and effort an unstable cause.In a similar manner, judgments were madebetween "lucky" and "tried hard" (unstablecauses differing in locus), "these tasks arealways easy" and "lucky" (external causesdiffering in stability), and "always good" and"always easy" (stable causes differing inlocus). Thus, the judgments permitted adirect test of the locus versus stability in-terpretation of expectancy change.

Expectancy estimates were examinedseparately for each of the causal judgments.The data revealed that within both the in-ternal and the external causes, expectancyincrements were positively associated withthe stability of the ascription; that is, therewere higher expectancies given ability andtask ease ascriptions than given effort or luckattributions. Contrasting locus of causalitydifferences within either the stable or theunstable ascriptions disclosed that the dis-parate causal locus groups did not differ intheir expectancies of success.

Locus of Control Controversy

One of my disappointments has been thatinvestigators associated with social learningtheory and locus of control have failed torecognize or admit the stability-expectancy

10 BERNARD WEINER

linkage and the existence of other dimen-sions of perceived causality. Some re-searchers (e.g., Lefcourt, von Baeyer, Ware,& Cox, Note 8) are incorporating the stabil-ity dimension into perceived causality scales.But this is in contrast with the position ofother investigators. For example, Phares(1978) states,

At the present time there does not appear to be a con-vincing body of data supporting the utility of adding thestability dimension . . . . Even should the addition ofstability find support in laboratory studies of expec-tancy changes, it is not at all clear tha t . . . [broader]demonstrations of utility will be forthcoming, (p.270)

In opposition to this statement, the liter-ature associating stability with expectancychange is unequivocal, and the findingsgeneralize outside of the laboratory as wellas beyond the achievement domain (as willbe documented later). It may indeed bethat the concept of locus of control has greatutility; my modest hope is that individualsin this area will acknowledge some of theprior shortcomings in their conceptualanalysis of expectancy shifts at skill andchance tasks and in their limited approachto causality (for a fuller discussion of theseissues, see Weiner et al., 1976).

Formal Analysis and Self-ConceptMaintenance

McMahan (1973) and Valle and Frieze(1976) have developed formal models of ex-pectancy shifts based upon the concept ofcausal stability. Valle and Frieze postulatethat predictions of expectancies (P) are afunction of the initial expectancy (E) plusthe degree to which outcomes (O) are at-tributed to stable causes (S):

P = f|E + 0 [f(S)]j.

In addition, Valle and Frieze (1976) also notethat the perceived causes of success andfailure are related to the initial expectancyof success. It has been clearly documentedthat unexpected outcomes lead to unstableattributions, particularly luck (Feather,1969; Feather & Simon, 1971; Frieze &

Weiner, 1971). Hence, Valle and Frieze(1976) conclude,

There is some value for the difference between the ini-tial expectations and the actual outcome that willmaximally change a person's predictions for the future.If the difference is greater than this point, the outcomewill be attributed to unstable factors to such a greatextent that it will have less influence on the person'sfuture predictions, (p. 581)

These ideas have important implicationsfor the maintenance of one's self-conceptand for attributional change programs (seeWeiner, 1974,1976). For example, assumethat an individual with a high self-conceptof ability believes that he or she has a highprobability of success at a task. It is prob-able that failure then would be ascribed tounstable causes such as luck or mood, whichmay not reduce the subsequent expectancyof success and sustains a high ability self-concept. On the other hand, success wouldbe ascribed to ability, which increases thesubsequent expectancy (certainty) of successand confirms one's high self-regard. Theconverse analysis holds given a low self-concept of ability and a low expectancy ofsuccess: Success would be ascribed to un-stable factors, and failure to low ability.These attributions result in the preservationof the initial self-concept (see Ames, 1978;Fitch, 1970; Gilmore & Minton, 1974; Ickes& Layden, 1978). In addition, the aboveanalysis suggests that in change programsinvolving expectancies or self-concept theperceived causes of performance must bealtered, and a modification in self-perceptionwould have to involve a gradual process(Valle & Frieze, 1976).

In one research investigation guided bythe above reasoning, Ames, Ames, and Gar-rison (1977) had children of high or low socialstatus in the classroom attribute causalityfor positive and negative interpersonal out-comes. For example, the children weregiven situations such as, "Suppose you meeta new student at school and you becomefriends quickly"; or "Imagine you asksomeone to play with you after school, butthey say they cannot play." The childrenthen attributed causality for each situationeither to an internal, external, or mutualcause. The data indicated that given nega-

A THEORY OF MOTIVATION 11

tive interpersonal outcomes, high-social-status children made greater use of externalcausal ascriptions, and given positive inter-personal outcomes, they made more internalattributions than the low-social-status pu-pils.

Resistance to Extinction andAchievement Change

The stability concept is generalizable tothe body of psychological literature con-cerning experimental extinction (see Rest,1976). Experimental extinction often isdefined as the cessation of a previously in-strumental response following the perma-nent withholding of the reward. It is rea-sonable to presume that when a response isperceived as no longer instrumental to goalattainment, the organism will cease makingthat response. Hence, any attribution thatmaximizes the expectation that the responsewill not be followed by the goal should fa-cilitate extinction. On the other hand, at-tributions that minimize goal expectancydecrements after nonreward should retardextinction.

As discussed above, the stability or in-stability of the perceived causal factors in-fluences the expectancy that the outcome ofan action might change in the future.Therefore, I suggest that resistance to ex-tinction is a function of attributions to thecausal dimension of stability during the pe-riod of nonreinforcement. Ascriptions ofnonreinforcement to bad luck, lack of im-mediate effort, or other unstable causes arehypothesized to minimize expectancy dec-rements and result in slower extinction thanattributions of nonattainment of a goal toperceived stable factors, such as teacher bias,high task difficulty, or lack of ability. Rest(1976) has presented strong evidence con-firming these hypotheses. Inasmuch asrandom reinforcement schedules elicit un-stable causal attributions (Weiner et al.,1971), they also should (and do) increaseresistance to extinction. In a similar man-ner, chance rather than skill instructions alsoincrease resistance to extinction (Phares,1957), presumably because failure is ascribed

to unstable causes only given the chance in-structions.

A related notion is that information gen-erating lack of effort ascriptions for failurealso should result in response maintenance(see Rest, 1976). There are data in the in-frahuman experimental literature that maybe interpreted as supporting this hypothesis.Lawrence and Festinger (1962), marshallingevidence to support their cognitive disso-nance explanation of extinction, report thatresistance to extinction is positively relatedto the effortfulness of a response. Ouranalysis suggests that when great exertion isrequired to attain a reward the salience ofeffort as the cause of goal attainment isaugmented. Thus, the expectancy of rewardfollowing nonattainment of the goal shouldbe comparatively unchanged and extinctionprolonged. With repeated nonreward,however, the ascription shifts from effort toability and/or task difficulty, thus decreasingexpectancy and producing extinction.

These ideas have more than just a passingrelevance to educational practices. Many ofthe burgeoning achievement-chance pro-grams make direct or indirect use of attri-butional principles. These programs oftenattempt to induce students to attribute theirfailures to a lack of effort, which is both un-stable and under volitional control (see An-drews & Debus, 1978; Chapin & Dyck, 1976;Dweck, 1975). This goal is expressly es-tablished for "failure-oriented" children whoapparently ascribe their failures to a lack ofability, which is a stable and uncontrollablecause (see Diener & Dweck, 1978). Pre-sumably, inasmuch as effort can be increasedvolitionally, ascriptions of nonattainment ofa goal to lack of effort will result in the sus-taining of hope and increased persistencetoward the goal. On the other hand, sinceability is stable and not subject to volitionalcontrol, ascription of nonattainment of a goalto low ability results in giving up and thecessation of goal-oriented behavior.

In sum, it is suggested that the relationsbetween diverse independent variables (re-ward schedules, effortfulness of the response,and certain attributional biases) and thedependent variables of resistance to extinc-tion or persistence in goal-related behaviorare mediated by perceptions of causality:

12 BERNARD WEINER

Reinforcement schedule;Response effortfulness;Person characteristics

(attributional biases)

(observation)^ Resistance toextinction

Causal attribution forreward and nonreward

Stability of theattribution

Expectancy of futuregoal attainment

(Inference concerning attributional mediation)

Locus

In contrast with the rather stable beliefsabout causal stability, our thoughts con-cerning locus of causality have fluctuatedgreatly. A temporary resolution is proposedhere which is a synthesis of our previousantithetical positions and better accounts forthe complexity of human affective re-sponses.

Initially, Weiner et al. (1971) postulatedthat locus of causality is related to the af-fective consequences of success and failure.Emotional reactions were believed to bemaximized given internal attributions forsuccess and failure and minimized givenexternal attributions. Thus, for example,pride and shame, the alleged dominant af-fects in achievement situations (Atkinson,1964; McClelland, Atkinson, Clark, & Low-ell, 1953), would be most experienced givenpersonal responsibility for success and fail-ure, as opposed to instances in which exter-nal factors such as luck or others were per-ceived as the causal agents. This postulatedrelation seemed intuitively reasonable, wasconsistent with Atkinson's (1964) formula-tions concerning the incentive value of suc-cess and failure, and found support in a va-riety of research investigations. Because adetailed account of this position recently waspresented in this journal (Weiner, 1977), Iwill not discuss it in any further detail.

Subsequently, it became evident that it isincorrect to presume an invariant positiverelation between internality and the mag-nitude of emotional reactions in achievementsettings. For example, failure ascribed toothers, such as the bias of a teacher or hin-

drance from students or family, will pre-sumably generate great anger and hostility.In this event, externality is positively relatedto emotional intensity. Thus, the positionexpressed in Weiner et al. (1971) cannot becorrect (see Weiner, 1977; Weiner et al.,1978).

We therefore initiated a series of studiesto determine the relation between attribu-tion and affect (Weiner et al., 1978; Weiner,Russell, & Lerman, in press). In our firstinvestigation, subjects were given a scenariothat depicted a success or failure experienceat an exam, along with a causal attributionfor that outcome (e.g., Joan failed becauseshe did not have the ability). The subjectsthen reported the affects that they surmisedwould be experienced in this situation.About 100 affects for success and 150 forfailure were provided, with responses madeon rating scales indicating the intensity withwhich the affects would be experienced.

There were two general findings of inter-est. First, there was a set of outcome-de-pendent, attribution-independent affectsthat represented broad positive or negativereactions to success and failure, regardlessof the "why" of the outcome. Given success,feelings of pleasure, happiness, satisfaction,goodness, and so on were reported as equallyexperienced in the disparate attributionconditions. In a similar manner, givenfailure, there were a number of outcome-linked emotions, such as feeling uncheerful,displeased, and upset. The outcome-de-pendent affects for both success and failurewere reported as the ones that would be mostintensely experienced.

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 ofthe Attribution for Success

Causal attribution

Emotion

CompetenceConfidenceContentmentExcitementGratefulnessGuiltHappinessPrideReliefSatisfactionSurpriseThankfulness

Ability

30204391

4439

41970

Unstableeffort

12194913

43282824161

Stableeffort

201812840

43391616

40

Personality

19190

1183

38431114140

Others

5147

1643

246211394

18

Luck

2426

1418488

260

524

many emotions discriminably related tospecific attributions. Given success, theunique attribution-affect linkages were thefollowing: ability-competence and confi-dence; typical effort-relaxation; immediateeffort-activation; others-gratitude; per-sonality-conceit; and luck-surprise. Thatis, if one perceived that success was causedby ability, then competence and confidencewere reported as intensely experienced; ifone succeeded because of help from others,then the dominant reported affect wasgratitude; and so on. In a similar manner,for failure, the attribution-affect linkageswere the following: ability-incompetence;effort-guilt and shame; personality-resig-nation; others-aggression; and luck-surprise(see Weiner et al, 1978).

It is of interest to point out that at timescausal attributions yield opposite reactionsfor success and failure, as would be expectedgiven diametric outcomes (respectively,competence versus incompetence givenability attributions; gratitude versusaggression for attributions to others). Butat times the same emotion accompanies bothpositive and negative outcomes (surprisegiven a luck attribution); and given still otherascriptions, such as typical or immediateeffort, the emotions that accompany success(respectively, relaxation and activation) areunrelated to the failure-tied affects (guiltand shame).

These data suggested we should reject thesupposition that locus of causality mediates

affective reactions in achievement contexts.Rather, emotions appeared to be eitheroutcome or attributionally generated,without any intervening dimensionalplacement.

Additional evidence, however, has re-sulted in a synthesis of our prior antitheticalstances. In a recent study (Weiner, Russell,& Lerman, in press) subjects recreated a"critical incident" in their lives in which theysucceeded (or failed) at an academic exambecause of ability, typical effort, immediateeffort, help (or hindrance) from others,personality, or luck. They then listed threeaffects they experienced in this situation.Table 3 includes only the emotions that werereported for success by more than 10% of therespondents for any particular attribution.The table shows the percentage of subjectsin all the attribution conditions reportingthese relatively shared experiences.

The data in Table 3 are consistent withour previous findings. The most dominantaffect, happiness, is expressed regardless ofthe reason for the success. In addition tothis outcome-linked emotion, there are sig-nificant attribution-affect linkages. Theseassociations are as follows: ability-compe-tence and pride; other people-gratefulnessand thankfulness; stable effort-content-ment; personality-pride; and luck-surprise,relief, and guilt (the linkages are based oncomparisons within an emotion but acrossattributions).

The failure data also revealed systematic

14 BERNARD WEINER

patterns. There were significant outcome-linked emotions including disappointment,as well as attribution-affect associationsconsistent with prior research: ability-incompetence and resignation; effort-guilt;other people-anger; and luck-surprise.

Additional analyses of these data alsodemonstrated that causal dimensions playan essential role in affective life. Given in-ternal attributions for success (ability, effort,personality), the affects pride, competence,confidence, and satisfaction were reportedmore frequently than they were given ex-ternal attributions (others, luck). Internalascriptions for failure generated the emo-tions of guilt and resignation. In sum, par-ticular affects clustered with the internalcauses. Reanalysis of Weiner et al. (1978)revealed virtually identical results.

It therefore appears that in achievementsituations there are (at least) three sourcesof affect. First, there are emotions tied di-rectly to the outcome. One feels "good"given success and "bad" given failure, re-gardless of the reason for the outcome.These probably are the initial and strongestreactions. Second, accompanying thesegeneral feelings are more distinct emotions,such as gratitude or hostility if success orfailure, respectively, is due to others, surprisewhen the outcome is due to luck, and so on.Third, the affects that are associated withself-esteem, such as competence, pride, andshame, are mediated by self-ascriptions.Many emotional reactions are shared givensuccess due to ability or effort, the twodominant internal attributions. It thereforemay be that the central self-esteem emotionsthat facilitate or impede subsequentachievement performance are dimensionallylinked, referred by the actor to him- or her-self. Some affects thus seem to be mediatedby the locus dimension, but in a mannermuch more complex than was originallyposited. It is likely that these dimension-linked affects have the greatest longevity andmost significance for the individual.

Stability and Affect

In addition to the locus-affect linkage,there also is a relation between causal sta-bility and emotions. Weiner et al. (1978)

found that the affects of depression, apathy,and resignation were reported primarilygiven internal and stable attributions forfailure (lack of ability, lack of typical effort,personality deficit). This suggests that onlyattributions conveying that events will notchange in the future beget feelings of help-lessness, giving up, and depression. Perhapsthe control dimension also plays a role ingenerating these particular emotions.Hence, the dimensions of causality relateto different sets of emotions.

In another research investigation sup-porting a stability-emotion union, Arkin andMaruyama (1979) assessed students' attri-butions for their success or failure at a collegeclass. In addition, anxiety associated withschool performance was measured. It wasfound that among successful students, thestability of their attributions was negativelycorrelated with anxiety. That is, whensuccess was ascribed to stable causes, stu-dents reported relatively little anxiety. Onthe other hand, among the unsuccessfulstudents, attributional stability and anxietycorrelated positively; most fear was reportedwhen failure was perceived as likely to recurin the future.

Cognition-Emotion Sequence inAchievement Contexts

On the basis of the above discussion, Isuggest that in achievement-related contexts(and, in particular, school settings), the actorprogresses through something like the fol-lowing cognition-emotion scenarios:

1. "I just received a D in the exam. Thatis a very low grade." (This generates feel-ings of being frustrated and upset.) "I re-ceived this grade because I did not try hardenough" (followed by feelings of shame andguilt). "There really is something lackingin me, and it is permanent" (followed by lowself-esteem or lack of worth and hopeless-ness).

2. "I just received an A on the exam.That is a very high grade" (generating hap-piness and satisfaction). "I received thisgrade because I worked very hard during theentire school year" (producing contentmentand relaxation). "I really do have somepositive qualities, and will continue to have

A THEORY OF MOTIVATION 15

them in the future" (followed by high self-esteem and feelings of self-worth, as well asoptimism for the future).

Some Thoughts About Feelings

Psychology is completing two movementsthat have relatively neglected the study ofaffect. The first is the behavioristic period,which denied verbal report data; the secondis the cognitive movement, which focuses onintellective structures. In contrast to theseperiods, I believe that psychologists andeducators now will turn to the study of af-fect.

At present many of the investigations ofaffect in the schools measure some globalfeeling state such as "satisfaction." But forthe study of emotions greater differentiationmust be allowed. For example, one mightspeculate that differential classroom "at-mospheres" provide the opportunity for theexperiencing of disparate emotions. Per-haps settings that promote internal ascrip-tions maximize positive or negative self-images and feelings of pride and shame. Onthe other hand, environments that permitmore student interaction enhance feelingssuch as gratitude and anger, inasmuch asattributions of success and failure to othersare promoted. Overall satisfaction ratingsmask distinctions between, for example,pride and gratitude. It is time that closerattention was paid to affective life in theclassroom.

Secondary Linkages

Because of the vast literature in the locusof control area, it might be anticipated thatcausal locus is directly linked with manypsychological reactions in addition to es-teem-related affects. This indeed is likelyto be the case. For example, it has been re-ported that locus of control relates positivelyto behaviors such as information seeking andto experiences such as feeling like an "origin"(de Charms, 1968). In most of this research,however, the concepts of locus and controlare united. It is not reasonable to expectindividuals who attribute failure to a lack ofability, which is internal but uncontrollable,to seek out information or feel like origins.

Rather, it seems that the experiential stateof an origin and correlated behaviors areexhibited because of the perceived personalcontrol of the situation, or the belief thatcausality is both internal and controllable.Thus, the discussion of the secondary link-ages with locus is postponed until the pre-sentation of the control dimension of cau-sality.

Control

Attribution theory as formulated byHeider (1958), Jones and Davis (1965), andKelley (1967) primarily concerns person-perception, or inferences about the inten-tions and dispostions of others. But thus farin this article I have only been concernedwith self-perception. I believe that one ofthe main contributions of our work has beenthe adaptation of some principles of socialperception for the construction of a theoryof motivation that has the individual as theunit of analysis.

In the discussion of the implications ofcausal dimensions, self- and other-percep-tion were not distinguished. Consideringchanges in the expectancy of success, thesame cause-effect logic pertaining to causalstability should hold when consideringoneself or others. The discussion of affectalso is equally applicable to both the self andothers, although of course, the emotionalexperiences are limited to the self and in-ferred about others. But if success or failureis perceived as being due to certain causes,then particular affective experiences shouldfollow.

The following examination of the dimen-sion of control centers upon inferences aboutothers and how beliefs about another's re-sponsibility for success and failure influencean actor's reactions toward that person.The reactions examined are helping, evalu-ation, and sentiments.

Helping

Ickes and Kidd (1976), guided by Weineret al. (1971) and Rosenbaum (1972), pro-posed an attributional analysis of helpingbehavior. A number of investigators prior

16 BERNARD WEINER

to Ickes and Kidd (1976) had establishedthat the tendency to help is influenced by theperceived cause of the need for aid (e.g.,Berkowitz, 1969; Ickes, Kidd, & Berkowitz,1976; Piliavin, Rodin, & Piliavin, 1969;Schopler & Matthews, 1965). The majorityof these experiments concluded that help ismore likely when the perceived cause of theneed is an environmental barrier, as opposedto being internal to the person desirous ofaid. For example, Berkowitz (1969) re-ported that individuals are more inclined tohelp an experimental subject when the ex-perimenter, caused a delay in the subject'sresponse, in contrast with a condition inwhich the subject is perceived as personallyresponsible for falling behind in the experi-ment.

In their review, Ickes and Kidd (1976)argued that this locus of control explanationof helping confounds the causal dimensionsof locus and intentionality (which I again willcall controllability). They suggest that inthe study conducted by Berkowitz (1969),the causal ascription to the experimenter isboth external and uncontrollable (from theperspective of the actor), whereas an attri-bution to the subject's own mismanagementis internal to the actor and is perceived bythe potential helper as under volitionalcontrol. Hence, two dimensions of causalityare confounded, and it is impossible to de-termine which of the two causal dimensionsis responsible for the differential help giving.Ickes and Kidd, in contrast with Berkowitz,suggest that it is the controllable aspect ofthe perceived cause, and not the locus, thatmediated the disparate help giving. Thereader should note how similar this analysisis to the one pertaining to expectancy shiftsin skill and chance tasks. Both contro-versies point out that the locus of controlliterature has been plagued by an inadequateanalysis of perceived causality. Further-more, what is required is research that sep-arates the various causal dimensions.

Other data support the Ickes and Kidd(1976) interpretation of helping behavior.For example, Piliavin et al. (1969) found thatthere is a bias to aid an ill person in distressas opposed to helping a drunk. Accordingto the above argument, this is becausedrunkenness is perceived as subject to voli-

tional control, whereas illness is not. Whena failure is perceived as controllable, thenhelp is withheld; the persons presumablyshould help themselves. For this reason, itis much easier to raise charity funds forbattered children or blindness than for al-coholism centers.

Guided by the prior research of Barnes,Ickes, and Kidd (Note 9), Simon and Weiner(in press) applied these ideas to one instanceof altruism in the classroom—lending classnotes to an unknown classmate. In this in-vestigation, two themes were created for astudent's failure to take class notes. Onetheme involved a professor, and the secondconcerned an employer. In the professortheme, the student always (stable) or some-times (unstable) did not take notes becauseof something about himself (internal) orsomething about the professor (external).Either he was unable to take good notes(uncontrollable) or he did not try (control-lable), while the professor either was unableto give a clear lecture or did not try. Thus,for example, an internal, stable, and uncon-trollable cause was that the student neverwas able to take good notes (low ability),while an external, unstable, and uncontrol-lable cause was that the professor at timescould not give a clear lecture. Each storywithin the eight possible causal combina-tions (2 levels of stability X 2 levels of locusX 2 levels of control) elaborated the basicscenario. The second theme involved a worksituation in which the student did not havethe notes because he (or the boss) always(sometimes) was responsible for his cominglate to school, which could (could not) havebeen avoided.

Following each causal statement thesubjects rated the likelihood of lending theirnotes to the student. Judgments were madeon a 10-point scale anchored at the extremeswith "definitely would lend my notes" and"definitely would not lend my notes."

The mean helping judgments for fourconditions (2 levels of locus X 2 levels ofcontrol) are shown in Table 4. Stability didnot effect the judgments and thus is ignoredin the analysis. Table 4 reveals that helpingis reported to be relatively equal and rea-sonably high in all conditions except whenthe cause is internal and controllable, in

A THEORY OF MOTIVATION 17

Table 4Mean Likelihood of Helping as a Function ofPerceived Locus of Causality andControllability

Locus ofcausality

InternalExternal

Controllable

3.137.35

Uncontrollable

6.746.98

Note. Data are from Simon and Weiner (in press). Highernumbers indicate greater likelihood of note lending.

which case aid is unlikely to be given. Thatis, if the student did not try to take notes(professor theme) or could have avoidedbeing absent (employer theme), then help iswithheld. The findings concerning the in-fluence of intent information on moraljudgments and criminal justice support thisline of reasoning (see Carroll & Payne, 1976,1977, discussed later in this article).

Evaluation

Some of the early experimental workconducted by me and my colleagues wasundertaken to promote the distinction be-tween various causes of success and failure.In particular, we attempted to provide evi-dence that ability and effort should be dis-tinguished, although both are internal inlocus of causality.

In one reference experiment that wasemployed, subjects were asked to pretendthat they were teachers and were to provideevaluative "feedback" to their pupils (e.g.,Eswara, 1972; Kaplan & Swant, 1973; Rest,Nierenberg, Weiner, & Heckhausen, 1973;Weiner & Kukla, 1970; Weiner & Peter,1973). The pupils were characterized interms of effort, ability, and performance onan exam. The data from these investiga-tions conclusively demonstrated that effortis of greater importance than ability in de-termining reward and punishment. Higheffort was rewarded more than high abilitygiven success, and lack of effort was pun-ished more than lack of ability given failure.To explain these findings, I stated,

There appear to be two reasons for the discrepancybetween ability and effort as determinants of rewardand punishment. First, effort attributions elicit strong

moral feelings—trying to attain a socially valued goalis something that one "ought" to do. Second, rewardingand punishing effort is instrumental to changing be-havior, inasmuch as effort is believed to be subject tovolitional control. On the other hand, ability is per-ceived as nonvolitional and relatively stable and thusshould be insensitive to external control attempts.(Weiner, 1977, p. 508)

Thus, both the moral and control aspectsof evaluation were considered. But it wasnot realized that evaluation is conceptuallysimilar to behaviors and feelings such as helpgiving, altruism, liking, and blame. That is,there is a pervasive influence of perceivedcontrollability or personal responsibility oninterpersonal judgments in achievement-related contexts, including how students aregraded.

Sentiments

Investigations linking liking to percep-tions of controllability primarily have beenconducted in the area of loneliness (seePeplau, Russell, & Heim, in press). Michela,Peplau, & Weeks (Note 5) found that per-sons lonely for reasons thought to be con-trollable (e.g., does not try to make friends)are liked less than individuals lonely foruncontrollable reasons (e.g., no opportunityto meet people). In addition, when a lonelyperson puts forth effort to make friends, thatperson is liked and elicits sympathy (Wimer& Peplau, Note 10). In contrast, if it is be-lieved that the lonely individual is respon-sible for his or her plight, then sympathy isnot forthcoming, and respondents indicatethey would avoid such persons. I assumethat this pattern of results will also be evi-dent in achievement-related contexts.Surely a teacher will not particularly like astudent who does not try, and failure per-ceived as due to lack of effort does not elicitsympathy.

Self-Perception of Control

While perceived control in others relatesto interpersonal judgments, self-perceptionsof control have quite a different array ofconsequences. These intrapersonal effectsappear to be vast, ranging from experientialstates, such as feeling as an origin (deCharms, 1968) and perceiving freedom of

18 BERNARD WEINER

Antecedent conditions — *• Perceived causes *

Ability

Effort (typical)and immediate)

Others (students,family, teacher)

Motivation (atten-tion, interest)

Etc.

Causal dimension

Intentionality

Globality

' Primary effects

affects

judgments

— » Other consequences

Performance intensity

Persistence

Choice

Etc.

Figure 1. Partial representation of an attributional theory of motivation.

choice (Steiner, 1970), to specific behaviors,such as information search (see Rotter, 1966)and normal functioning rather than learning,cognitive, and motivational deficits that arepostulated to accompany the loss of control(Seligman, 1975). This is a complex subjectmatter in need of systematic examinationand synthesis that goes well beyond thescope of our present knowledge.

Summary

A variety of sources of information (notdiscussed here) are used to reach causal in-ferences in achievement-related contexts.The perceived causes of success and failureprimarily are ability and effort but also in-clude a small number of other salient factorssuch as home environment and teacher, anda countless host of idiosyncratic factors.These causes can be comprised within threeprimary dimensions of causality: stability,locus, and control. There also are an unde-termined number of subordinate causal di-mensions, including perhaps intentionalityand globality. The three main dimensions,respectively, are linked to expectancychanges, esteem-related affects, and inter-personal judgments (decisions about help-ing, evaluation, and sentiments). In addi-tion, there are secondary linkages betweenthe causal dimensions and psychologicaleffects: Stability relates to depression-typeaffects, and control is associated with par-ticular feeling states and behaviors. Thedimension-consequence linkages influencemotivated behaviors such as persistenceand choice. This theory is depicted inFigure 1.

Theoretical Range

The theory rather sketchily conveyed inFigure 1 has been shown to be relevant tomany classroom-related thoughts and ac-tions. The topics already examined in thisarticle include the perceived reasons forsuccess and failure, expectancy change,self-concept maintenance, achievementchange programs, reinforcement schedules,hopelessness, sources of emotion, self-es-teem, helping, evaluation, and liking. Stillother achievement-related topics have beendemonstrated to be encompassed within thisattributional conception (see Weiner, 1974,1976). The breadth of the phenomena in-corporated within our attributional frame-work intimates that a general theory of mo-tivation is being constructed. In the re-mainder of this article I document otherareas to which the theory is applicable.Some of the theoretical extensions are ger-mane to the school setting, while other topicsare of interest to an audience of educationalpsychologists* primarily because they dem-onstrate the range of the conception.

Hyperactivity and Psychostimulants

Whalen and Henker (1976) have outlinedan attributional analysis of the effects ofdrug treatment for hyperactive children.They contend that when hyperactivity iscombated with a drug, the belief is conveyedto both the child and his or her parents thatthe cause of hyperactivity is a physiologicaldysfunction. Hence, the involved individ-uals are not responsible for or in control of

A THEORY OF MOTIVATION 19

the maladaptive behavior that is exhibited.Because this physiological deficit is per-ceived as an uncontrollable cause, neitherthe child nor the parents need feel guilty orblame themselves for the aberrant behavior.That is, the shift in perceived causality from"lack of effort" minimizes self-blame, lowself-esteem, and negative evaluations fromothers. This appears to be a beneficial andan unanticipated side effect of the treatmenttechnique.

On the other hand, Whalen and Henker(1976) also state that "the reputed physio-logical dysfunctions used to explain thefailures of hyperactive children are fre-quently viewed as stable and relatively un-responsive to behavior change effects" (p.1123). Thus, the perception of fixed cau-sation might lead to "demoralization aboutproblem solutions.... and interferes witheffective coping" (Whalen & Henker, 1976,p. 1124).

In sum, again this is an analysis of a psy-chological phenomenon from the perspectiveshown in Figure 1. Individuals utilize in-formation (treatment technique) to infercausation about an event (hyperactivity).The perceived cause (a genetic deficit) isperceived as uncontrollable and stable. Thisminimizes certain negative affects and un-favorable evaluations (beneficial effects) butalso weakens the perceived possibility ofrecovery (a harmful consequence). Thesetwo factors, in turn, influence the long-rangeinfluence of the treatment (negatively, ac-cording to Whalen and Henker, 1976, inas-much as they perceive expectancy to be themore potent determinant of long-term be-havioral change).

Mastery

The labels mastery and competence areprominent among the writings of manypsychologists (e.g., Nissen, 1954; White,1959). However, in my opinion systematicexperimental work elucidating these allegedmotivators of behavior has not been con-ducted. An investigation by Nuttin (1973),described as demonstrating "causalitypleasure," could provide an important ex-perimental paradigm for this area. Nuttin

placed 5-year-olds in an experimental roomcontaining two machines. The machineseach had colored lights and movable handles.For one machine (A), the onset of the lightswas preprogrammed by the experimenter.The lights in the alternate machine (B) wenton or off only when the handle was movedbeyond a certain point. Thus, althoughboth machines stimulated the viewer per-ceptually, the children were the producers orthe cause of the stimulation only with ma-chine B.

The subjects in this experiment were freeto spend their time with either machine.The experimenters recorded various indexesof choice or preference, such as the timespent with each machine and verbal reportsof liking. Both observational and self-reportdata revealed that the children stronglypreferred machine B over machine A.These findings have been replicated byWeiner, Kun, and Weiner (in press).

From the theoretical perspective shown inFigure 1, the experiment by Nuttin (1973)illustrates a temporal sequence involving theuse of information, inferences concerninglocus of causality, positive affect, and somebehavioral consequences of emotional states.That is, on the basis of the observed covar-iation between their own actions and theonset of the lights in machine B, the childreninfer that they are personally responsible(ability and effort) for the stimulation fromthat machine. Self-attribution for the out-come increases positive esteem-related af-fects, and the augmented affect increases theprobability of engaging in the action again aswell as increasing "liking" about playing withthe machine.

This interpretation is applicable to an-other developmental study that has not beenconceptualized as involving mastery-typebehavior. Watson (1966, 1967) demon-strated that 8-week-old infants can learn aninstrumental response (a head turn) to in-crease stimulation (the movement of a mo-bile). He also reported that infants in theinstrumental response condition apparentlydisplayed more instances of positive affect(smiling and cooing) than children in a con-dition in which the mobile movement wascontrolled by the experimenter. This againsuggests the following temporal sequence:

20 BERNARD WEINER

Specific caseinformation aboutthe crime and thecriminal, e.g.,crime description,prior record

General knowledge(beliefs) aboutcrime, criminals,or any other rele-vant knowledge

Judgment ofgoodness—badness

Judgment offuture acts

/ "Punishment"\. severity

I likeableness••Tresponsibility

y ) Parole decision.̂ prison term

* N criminality

P'Risk"/ recidivismS. expectations

Figure 2. An attributional framework for the parole decision process (in this case, an attributional analysisof perceptions of crime and criminals). (From Carroll and Payne, 1977, p. 200. Copyright 1977 by PlenumPress, Inc. Reprinted by permission.)

response-outcome covariation - perceivedinternal causation - positive affects of com-petence and pride - choice. That is, theenhanced positive affect and subsequentperformance of the instrumental responseare mediated by perceptions of self-respon-sibility (perhaps the control dimension alsoplays a role here).

The underlying premise of this inter-pretation of Watson's (1966,1967) researchis that affect and choice can be used to infercognitive processes (perceptions of causali-ty). It may seem far-fetched to draw theinference that 8-week-olds have the cogni-tive capacities to make causal deductions.However, it also may be that a differentia-tion between the self and the environmenthas developed by that age, and that primitiveinferences about locus and control can bemade using proprioceptive feedback infor-mation. If this interpretation has any va-lidity, then Watson perhaps has identifiedthe existence of attempts at mastery amongvery young infants. Note also that one mayconsider the contribution of the Watson in-vestigations from a light somewhat differentfrom what is usual for psychologists—theobservation of the behavior is of interestprimarily because it tells us something aboutthe contents of the mind!

Parole Decisions

A parole decision is a complex judgmentin which causal attributions play a majorrole. Figure 2 depicts the parole decision

process as conceptualized by Carroll andPayne (1976, 1977). The figure indicatesthat the decision maker is provided with avariety of information about the criminal,the crime, and other pertinent facts. Thisinformation is combined and synthesized,yielding attributions about the cause of thecrime. The causal attributions, in turn, in-fluence judgments about deserved punish-ment and social risk, which are believed to bethe basis for the final parole decision.

Carroll and Payne (1976), after reviewingan extensive literature, contend that theparole decision process is

based on a simple two-part model. In the first part, theprimary concern of the decision maker is to make thepunishment fit the crime .... At the second part. ..the primary concern . . . is with parole risk, i.e., theprobability that the person being considered for releasewill again violate the laws of society, (p. 15)

According to Figure 2, crimes that are as-cribed to internal and/or intentional (con-trollable) factors (e.g., personality charac-teristics, evil intents) should result inharsher evaluation (punishment) thancrimes attributed to external and/or unin-tentional (uncontrollable) causes (e.g., eco-nomic conditions, bad friends, etc.). In ad-dition, the risk associated with parole shoulddepend on the stability of the perceivedcause of the crime. If, for example, thecrime is attributed to some fixed personalitytrait, then the decision maker will expectthat a crime again will be committed if theprisoner is paroled. On the other hand, if

A THEORY OF MOTIVATION 21

the cause of the crime has been or can be al-tered (e.g., economic conditions have im-proved, a job can be found, etc.), then thecriminal will be perceived as a good parolerisk.

Given the above analysis, a criminal isleast likely to be paroled if the cause of thecrime is perceived as internal and/or con-trollable but stable ("He is an evil person").Conversely, parole will tend to be grantedwhen the crime is perceived as caused byexternal and/or noncontrollable and unsta-ble factors (e.g., prior economic conditions).The remaining causal combinations shouldfall between these extremes in terms of pa-role probability.

Carroll and Payne have furnished supportfor these hypotheses, examining professionalparole decision makers and the judgmentsof college students when given simulatedcriminal cases. They find, for example, thatperceptions of the locus, stability, and con-trollability of causes significantly relate toperceived responsibility for the crime, like-lihood of recidivism, likability, prison term,and the purpose of the sentence.

In sum, according to Carroll and Payne(1976, 1977) the parole decision procedureis conceptually identical to the perceivedsequence of events in the achievement do-main: Antecedent information is processed,a causal judgment is reached, and the causeis placed within the locus, stability, and in-tentionality (control) dimensions. Thisinfluences evaluation and expectancy, whichare the main determinants of the parole de-cision.

Affiliation and Loneliness

It has been reasoned that in our culturetwo sources of motivation are most domi-nant: achievement and social recognition(or, in Freud's more general terms, Arbeitand Liebe). Hence, affiliative motivationis a natural area to turn toward in the de-velopment of a theory of motivation.

An attributional analysis of affiliativemotivation guided by the theory shown inFigure 1 conceives of loneliness as a socialfailure (Gordon, 1976; Stein & Bailey, 1973).Hanusa (Note 11) and Heim (Note 12) ex-amined the perceived causes of social success

and failure and found them to be similar tothe causes of achievement success and fail-ure. As already indicated, Michela et al.(Note 5) used scaling procedures to discoverthe dimensions of the causes of social failureand found them to be similar to the dimen-sions uncovered in achievement contexts.

The question that then remains is whetherthe attributional dimensions in the affiliativedomain relate to psychological factors in thesame manner as in the achievement domain.Research reveals that is indeed the case (seePeplau et al., in press). Stability relates tothe perceived probability of remaining lonelyin the future, locus is associated with es-teem-related affects, and as previouslystated, control is linked with liking andsympathy toward the lonely person (also seeFolkes, 1978).

Depression and Learned Helplessness

In accordance with the trend in lonelinessresearch, recent explanations of depressionhave focused upon the cognitive, rather thanthe affective, aspects of this disorder (e.g.,Beck, 1976). The work of Seligman (1975),captured under the label of learned help-lessness, has been especially influential. Iwill not dwell upon Seligman's use of thisconstruct or the supporting empirical evi-dence in this context. Rather, my goal is toconvey the pertinence of the learned help-lessness literature to the attributional modeldepicted in Figure 1 (see Abramson et al.,1978; Weiner & Litman-Adizes, in press).

Learned helplessness communicates thebelief that there is no perceived associationbetween responding and environmentaloutcomes. That is, the actor believes thatthe likelihood of an event is independent ofwhat he or she does. The belief in help-lessness is alleged to produce deficits inmotivation and learning, negative affect, anda syndrome that has been labeled "depres-sion."

As this work has progressed from infra-human to human research, it has becomeevident that it also is essential to considerwhy actions and outcomes are perceived tobe independent. For example, Klein, Fen-cil-Morse, and Seligman (1976) found that

22 BERNARD WEINER

only individuals making internal attribu-tions for response-outcome independenceexhibited aspects of the learned helplessnesssyndrome. Attributions of response-out-come independence to external factors didnot produce any learning deficits. In asimilar manner, Tennen and Eller (1977)found learned helplessness only under con-ditions that promote low ability attributionsfor prior lack of control.

Partially because of these data, Abramsonet al. (1978) adopted an attributionalframework for helplessness. I have ex-tracted the following from the Abramson etal. (1978) manuscript and pieced togethernew paragraphs to illustrate their think-ing.

Our reformulated hypothesis makes a major new set ofpredictions. The helpless individual first finds out thatcertain outcomes and responses are independent, thenhe makes an attribution about the cause's. This attri-bution determines the chronicity, generality, and in-tensity of the deficits. Depressed people seem to makemore global, stable, and possibly internal attributionsabout the cause of their helplessness and as a conse-quence show more general, chronic, and intense deficitsthan nondepressed people.

Depression occurs when an individual expects that theprobability of a highly preferred outcome is low and heexpects that he is helpless to increase it. If the attri-butions for the present state of affairs are to stable andglobal factors, the future will look dark to the individual.He expects that he will find himself helpless again andagain. This is what is usually meant by "hopelessness."Another implication of the formulation is that indi-viduals will show the greatest loss of self-esteem whenthey make internal, global, and stable attributions fortheir failures. (Abramson et al., 1978)

In sum, it is argued that depressed indi-viduals attempt to make sense out of per-ceived evidence that their responses do notaffect outcomes. A cause is determinedwhich often is classified as stable, internal,and global. This leads to a low expectancyof success across a wide array of environ-ments and a heightened negative affect (lossof self-esteem), which are sufficient precur-sers of depression.

Conclusion

I have selectively reviewed the extensiveliterature outside of the achievement do-main, including hyperactivity, mastery, pa-

role decisions, loneliness and affiliation, anddepression. The data strongly suggest thata general conception of motivation, as wellas a particular method of psychologicalanalysis, is evolving.

Reference Notes

1. Lau, R. R., & Russell, D. Attributions in the sportspages: A field test of some current hypotheses inattribution research. Unpublished manuscript,University of California, Los Angeles, 1978.

2. Bar-Tal, D., & Darom, E. Causal perceptions ofpupils' success or failure by teachers and pupils:A comparison. Unpublished manuscript, Uni-versity of Tel-Aviv, Israel, 1977.

3. Cooper, H. M., & Burger, J. M. Internality, sta-bility, and personal efficacy: A categorization offree response academic attributions. Unpublishedmanuscript, University of Missouri—Columbia,1978.

4. Litman-Adizes, T. An attributional model of de-pression: Laboratory and clinical investigations.Unpublished manuscript, University of California,Los Angeles, 1977.

5. Michela, J., Peplau, L. A.,& Weeks, D. Perceiveddimensions and consequences of attributions forloneliness. Unpublished manuscript, Universityof California, Los Angeles, 1978.

6. Meyer, W. U. Selbstanwortlichkeit und Le-istungsmotivation. Unpublished doctoral dis-sertation, Ruhr Universitat, Bochum, West Ger-many, 1970.

7. Pancer, S. M., & Eiser, J. R. Expectations, aspi-rations, and evaluations as influenced by another'sattributions for success and failure. Paper pre-sented at the 83rd Annual Meeting of the AmericanPsychological Association, Chicago, September1975.

8. Lefcourt, H. M., von Baeyer, C. L., Ware, E. E., &Cox, D. J. The multidimensional-multiattribu-tional causality scale: The development of agoal-specific locus of control scale. Unpublishedmanuscript, University of Waterloo, Ontario,Canada, 1978.

9. Barnes, R. D., Ickes, W. J., & Kidd, R. F. Effectsof perceived intentionality and stability of an-other's dependency on helping behavior: A fieldexperiment. Unpublished manuscript, Universityof Wisconsin—Madison, 1977.

10. Wimer, S. W., & Peplau, L. A. Determinants ofreactions to lonely others. Paper presented at the58th Annual Meeting of the Western PsychologicalAssociation, San Francisco, April 1978.

11. Hanusa, B. H. An extension of Weiner's attribu-tion approach to social situations: Sex differencesin social situations. Paper presented at the 46thAnnual Meeting of the Eastern Psychological As-sociation, New York, April 1975.

12. Heim. M. Sex differences in causal attributionsfor achievement in social tasks. Paper presentedat the 46th Annual Meeting of the Eastern Psy-chological Association, New York, April 1975.

A THEORY OF MOTIVATION 23

References

Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D.Learned helplessness in humans: Critique and re-formulation. Journal of Abnormal Psychology,1978,87, 49-74.

Ames, C. Children's achievement attributions andself-reinforcement: Effects of self-concept andcompetitive reward structure. Journal of Educa-tional Psychology, 1978, 70, 345-355.

Ames, R., Ames, C., & Garrison, W. Children's causalascriptions for positive and negative interpersonaloutcomes. Psychological Reports, 1977, 41, 595-602.

Andrews, G. R., & Debus, R. L. Persistence and causalperception of failure: Modifying cognitive attribu-tions. Journal of Educational Psychology, 1978, 70,154-166.

Arkin, R. M., & Maruyama, G. M. Attribution, affect,and college exam performance. Journal of Educa-tional Psychology, 1979, 71, 85-93.

Atkinson, J. W. An introduction to motivation.Princeton, N.J.: Van Nostrand, 1964.

Beck, A. T. Cognitive therapy and the emotionaldisorders. New York: International UniversitiesPress, 1976.

Berkowitz, L. Resistance to improper dependencyrelationships. Journal of Experimental SocialPsychology, 1969,5, 283-294.

Carroll, J. S., & Payne, J. W. The psychology of theparole decision process: A joint application of at-tribution theory and information processing psy-chology. In J. S. Carroll & J. W. Payne (Eds.), Cog-nition and social behavior. Hillsdale, N.J.: Erl-baum, 1976.

Carroll, J. S., & Payne, J. W. Judgments about crimeand the criminal: A model and a method for inves-tigating parole decision. In B. D. Sales (Ed.), Pros-pectives in law and psychology. Vol. 1: The criminaljustice system. New York: Plenum Press, 1977.

Chapin, M., & Dyck, D. G. Persistence in children'sreading behavior as a function of N length and attri-bution retraining. Journal of Abnormal Psychology,1976,85, 511-515.

de Charms, R. Personal causation. New York: Ac-ademic Press, 1968.

Deci, E. L. Intrinsic motivation. New York: Plenum,1975.

Diener, C. I., & Dweck, C. A. An analysis of learnedhelplessness: Continuous changes in performance,strategy, and achievement cognitions following fail-ure. Journal of Personality and Social Psychology,1978,36, 451-462.

Dweck, C. S. The role of expectations and attributionsin the alleviation of learned helplessness. Journalof Personality and Social Psychology, 1975, 31,674-685.

Elig, T. W., & Frieze, I. H. A multidimensional schemefor coding and interpreting perceived causality forsuccess and failure events: The Coding Scheme ofPerceived Causality (CSPC). JSAS Catalog of Se-lected Documents in Psychology, 975,5, 313. (Ms.No. 1069)

Eswara, H. S. Administration of reward and punish-

ment in relation to ability, effort, and performance.Journal of Social Psychology, 1972,87, 139-140.

Feather, N. T. Attribution of responsibility and va-lence of success and failure in relation to initial con-fidence and task performance. Journal of Person-ality and Social Psychology, 1969,13, 129-144.

Feather, N. T., & Simon, J. G. Causal attributions forsuccess and failure in relation to expectations ofsuccess based upon selective or manipulative control.Journal of Personality, 1971,39, 527-541.

Fitch, G. Effects of self-esteem, perceived perfor-mance, and choice on causal attributions. Journalof Personality and Social Psychology, 1970, 16,311-315.

Folkes, V. S. Causal communication in the earlystages of affiliative relationships. Unpublisheddoctoral dissertation, University of California, LosAngeles, 1978.

Fontaine, G. Social comparison and some determi-nants of expected personal control and expectedperformance in a novel task situation. Journal ofPersonality and Social Psychology, 1974, 29, 487-496.

Frieze, I. H. Causal attributions and informationseeking to explain success and failure. Journal ofResearch in Personality, 1976,10, 293-305.

Frieze, I. H., & Weiner, B. Cue utilization and attri-butional judgments for success and failure. Journalof Personality, 1971,39, 591-605.

Gilmore, T. M., & Minton, H. L. Internal versus ex-ternal attributions of task performanace as a functionof locus of control, initial confidence and success-failure outcome. Journal of Personality, 1974,42,159-174.

Gordon, S. Lonely in America. New York: Simon &Schuster, 1976.

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

Ickes, W. J., & Kidd, R. F. An attributional analysis ofhelping behavior. In J. H. Harvey, W. J. Ickes, & R.F. Kidd (Eds.), New directions in attribution re-search (Vol. 1). Hillsdale, N.J.: Erlbaum, 1976.

Ickes, W. J., Kidd, R. F., & Berkowitz, L. Attributionaldeterminants of monetary help-giving. Journal ofPersonality, 1976, 44, 163-178.

Ickes, W. J., & Layden, M. A. Attributional styles. InJ. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), Newdirections in attribution research (Vol. 2). Hillsdale,N.J.: Erlbaum, 1978.

Jones, E. E., & Davis, K. E. From acts to dispositions:The attribution process in person perception. In L.Berkowitz (Ed.), Advances in experimental socialpsychology (Vol. 2). New York: Academic Press,1965.

Kaplan, R. M., & Swant, S. G. Reward characteristicsof appraisal of achievement behavior. Representa-tive Research in Social Psychology, 1973, 4(2), 11-17.

Kelley, H. H. Attribution theory in social psychology.In D. Levine (Ed.), Nebraska Symposium on Moti-vation: 1967. Lincoln: University of NebraskaPress, 1967.

Kelley, H. H. Attribution in social interaction.Morristown, N.J: General Learning Press, 1971.

Klein, D. C., Fencil-Morse, E., & Seligman, M. E. P.

24 BERNARD WEINER

Learned helplessness, depression, and the attributionof failure. Journal of Personality and Social Psy-chology, 1976,33, 508-516.

Lawrence, D. H., & Pestinger, L. Deterrents and re-inforcement. Stanford, Calif.: Stanford UniversityPress, 1962.

Lewin, K. A dynamic theory of personality. NewYork: McGraw-Hill, 1935.

Mann, L. On being a sore loser: How fans react totheir team's failure. Australian Journal of Psy-chology, 1974, 26, 37-47.

McClelland, D. C., Atkinson, J. W., Clark, R. A., &Lowell, E. L. The achievement motive. New York:Appleton-Century-Crofts, 1953.

McMahan, I. D. Relationships between causal attri-butions and expectancy of success. Journal of Per-sonality and Social Psychology, 1973, 28, 108-114.

Meyer, J. P. Dimensions of causal attribution forsuccess and failure: A multivariate investigation.Unpublished doctoral dissertation, University ofWestern Ontario, London, Canada, 1978.

Meyer, W. U., Folkes, V. S., & Weiner, B. The per-ceived information value and affective consequencesof choice behavior and intermediate difficulty taskselection. Journal of Research in Personality, 1976,10, 410-423.

Nissen, H. W. The nature of drive as innate determi-nant of behavioral organization. In M. R. Jones(Ed.), Nebraska Symposium on Motivation: 1954.Lincoln: University of Nebraska Press, 1954.

Nuttin, J. R. Pleasure and reward in motivation andlearning. In D. Berlyne (Ed.), Pleasure, reward,preference. New York: Academic Press, 1973.

Ostrove, N. Expectations for success on effort-deter-mined tasks as a function of incentive and perfor-mance feedback. Journal of Personality and SocialPsychology, 1978, 36, 909-916.

Passer, M. W. Perceiving the causes of success andfailure revisited: A multidimensional scaling ap-proach. Unpublished doctoral dissertation, Uni-versity of California, Los Angeles, 1977.

Peplau, L. A., Russell, D., & Heim, M. An attributionalanalysis of loneliness. In I. Frieze, D. Bar-Tal, & J.Carroll (Eds.), Attribution theory: Applications tosocial problems. San Francisco: Jossey-Bass, inpress.

Phares, E. J. Expectancy changes in skill and chancesituations. Journal of Abnormal and Social Psy-chology, 1957,54, 339-342.

Phares, E. J. Locus of control. In H. London & J. E.Exner, Jr. (Eds.), Dimensions of personality. NewYork: Wiley, 1978.

Piliavin, I. M., Rodin, J., & Piliavin, J. A. Good sa-maritanism: An underground phenomenon?Journal of Personality and Social Psychology, 1969,13, 289-299.

Rest, S. Schedules of reinforcement: An attributionalanalysis. In J. H. Harvey, W. J. Ickes, & R. F. Kidd(Eds.), New directions in attribution research (Vol.1). Hillsdale, N.J.: Erlbaum, 1976.

Rest, S., Nierenberg, R., Weiner, B., & Heckhausen, H.Further evidence concerning the effects of percep-tions of effort and ability on achievement evaluation.Journal of Personality and Social Psychology, 1973,28, 187-191.

Riemer, B. S. Influence of causal beliefs on affect andexpectancy. Journal of Personality and SocialPsychology, 1975, 31, 1163-1167.

Rosenbaum, R. M. A dimensional analysis of theperceived causes of success and failure. Unpub-lished doctoral dissertation, University of California,Los Angeles, 1972.

Rotter, J. B. Generalized expectancies for internalversus external control of reinforcement. Psycho-logical Monographs, 1966,80(1, Whole No. 609).

Schopler, J., & Matthews, M. W. The influence of theperceived causal locus of partner's dependence of theuse of interpersonal power. Journal of Personalityand Social Psychology, 1965,2, 609-612.

Schiitz, A. Collected papers. I. The problem of socialreality. The Hague: Martinus Nijhoff, 1967.

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

Simon, P., & Weiner, B. Dimensions of causal attri-butions related to helping judgments. Personalityand Social Psychology Bulletin, in press.

Stein, A. H., & Bailey, M. M. The socialization ofachievement orientation in females. PsychologicalBulletin, 1973,80, 345-366.

Steiner, I. D. Perceived freedom. In L. Berkowitz(Ed.), Advances in experimental social psychology(Vol. 5). New York: Academic Press, 1970.

Tennen, H., & Eller, S. J. Attributional componentsof learned helplessness and facilitation. Journal ofPersonality and Social Psychology, 1977, 35, 265-271.

Triandis, H. The analysis of subjective culture. NewYork: Wiley-Interscience, 1972.

Valle, V. A. Attributions of stability as a mediator inthe changing of expectations. Unpublished doctoraldissertation, University of Pittsburgh, 1974.

Valle, V. A., & Frieze, I. H. Stability of causal attri-butions as a mediator in changing expectations forsuccess. Journal of Personality and Social Psy-chlology, 1976, 33, 579-587.

Watson, J. S. The development and generalization of"contingency awareness" in early infancy: Somehypotheses. Merrill-Palmer Quarterly, 1966, J2(2),123-135.

Watson, J. S. Memory and "contingency analysis" ininfant learning. Merrill-Palmer Quarterly, 1967,13(1), 55-76.

Weiner, B. Theories of motivation: Fram mechanismto cognition. Chicago: Rand McNally, 1972.

Weiner, B. (Ed.). Achievement motivation and at-tribution theory. Morristown, N.J.: GeneralLearning Press, 1974.

Weiner, B. An attributional approach for educationalpsychology. In L. Shulman (Ed.), Review of researchin education (Vol. 4). Itasca, 111.: Peacock, 1976.

Weiner, B. Attribution and affect: Comments onSohn's critique. Journal of Educational Psychology,1977,69, 506-511.

Weiner, B., Frieze, I. H., Kukla, A., Reed, L., Rest, S.,& Rosenbaum, R. M. Perceiving the causes of suc-cess and failure. Morristown, N.J.: GeneralLearning Press, 1971.

Weiner, B., & Kukla, A. An attributional analysis ofachievement motivation. Journal of Personality

A THEORY OF MOTIVATION 25

and Social Psychology, 1970,15, 1-20.Weiner, B., Kun, A., & Weiner, M. B. The development

of mastery, emotions, and morality from an attribu-tional perspective. In W. A. Collins (Ed.), Minnesotasymposium on child development (Vol. 13).Hillsdale, N.J.: Erlbaum, in press.

Weiner, B., & Litman-Adizes, T. An attributional,expectancy-value analysis of learned helplessness anddepression. In J. Garber and M. E. P. Seligman(Eds.), Human control. New York: AcademicPress, in press.

Weiner, B., Nierenberg, R., & Goldstein, M. Sociallearning (locus of control) versus attributional (causalstability) interpretations of expectancy of success.Journal of Personality, 1916,44, 52-68.

Weiner, B., & Peter, N. A cognitive-developmentalanalysis of achievement and moral judgments. De-

velopmental Psychology, 1973, 9, 290-309.Weiner, B., Russell, D., & Lerman, D. Affective con-

sequences of causal ascriptions. In J. H. Harvey, W.J. Ickes, & R. F. Kidd (Eds.), New directions in at-tribution research (Vol. 2). Hillsdale, N.J.: Erl-baum, 1978.

Weiner, B., Russell, D., & Lerman, D. The cognition-emotion process in achievement-related contexts.Journal of Personality and Social Psychology, inpress.

Whalen, C. K., & Henker, B. Psychostimulants andchildren: A review and analysis. PsychologicalBulletin, 1976,83, 1113-1130.

White, R. W. Motivation reconsidered: The conceptof competence. Psychological Review, 1959, 66,297—333

Received August 11,1978 •