A Cognitive (Attribution)-Emotion-Action Model of Motivated … · 2017-07-10 · A Cognitive...

Journal of Personality and Social Psychology1980, Vol. 39, No. 2, 186-200

A Cognitive (Attribution)-Emotion-Action Model of MotivatedBehavior: An Analysis of Judgments of Help-Giving

Bernard WeinerUniversity of California, Los Angeles

Six experiments examined the relations of causal attributions and affect to judg-ments of help-giving. The first experiment considered the influence of threedimensions of causality (locus, stability, and control) on judgments concerningthe lending of class notes. Ratings of help were lowest when the cause of theneed was internal to the actor and was controllable (e.g., lack of effort). It wassuggested that ascriptions to internal controllable factors maximize negativeaffect (disgust and anger) and promote avoidance behavior. On the other hand,attributions to uncontrollable factors (e.g., ability or teacher shortcomings) wereanticipated to generate positive affect (sympathy) and give rise to approachbehavior (help). The next five experiments examined these hypotheses and anattributional model of helping, using a simulational, judgment paradigm withboth correlational and experimental designs. These investigations utilized sce-narios describing a drunk or a disabled individual in need of aid (from Piliavin,Rodin, & Piliavin), The data suggested a temporal sequence of attribution-affect-action in which attributions guide our feelings, but emotional reactionsprovide the motor and direction for behavior. Comparisons and contrasts weremade between this approach to helping and other conceptions.

Many behavioral sequences appear to be ini-tiated following a causal ascription for anevent. For example, in an achievement-relatedcontext, an individual may succeed at an exam.The success is then attributed to help fromothers, such as a classmate who lent him orher the class notes. Attribution of success toothers gives rise to gratitude (Weiner, Russell,& Lerman, 1978, 1979) and seems likely topromote actions instrumental to the mainte-nance of the relationship, such as the purchaseof a gift. In a similar manner, in an affiliativecontext, assume that an individual attemptingto establish a dating relationship is rejected.This rebuff could be attributed to an aspectof the self, such as an aversive personality

This research was supported by Grant MH2S687-OSfrom the National Institute of Mental Health and bya grant from the Spencer Foundation. Appreciationis expressed to Paul Simon and Kathy Sredl fortheir help.

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

characteristic or an unbecoming physical ap-pearance. A self-ascription for rejection en-genders a number of negative esteem-relatedaffects and "hurt feelings" (Folkes, 1978)that may in turn initiate actions anticipatedto be instrumental to the avoidance of suchaffects, such as not appearing at a party.

In the above scenarios, following the per-ception of an event a cognition (attribution)- emotion - action temporal sequence is sug-gested in which causal ascriptions produce af-fect (although there are other sources of af-fect; see Weiner et al., 1978, 1979), and emo-tions, in turn, provide the motor and directionfor behavior (see Tomkins, 1963). Thus, asequential organization between the tripartitedivision within psychology of thought, feel-ing, and action is proposed.

The postulated motivational ordering isdocumented here by analyzing a situation ofhelp-giving first investigated by Piliavin,Rodin, and Piliavin (1969). In the studyconducted by Piliavin et al., an individual (aconfederate) falls in a subway. In one con-

Copyright 1980 by the American Psychological Association, Inc. 0022-3S14/80/3902-0186$00.75

186

ANALYSIS OF JUDGMENTS OF HELP-GIVING 187

dition the confederate appears to be drunk(carrying a bottle and smelling of alcohol),whereas in a second condition he seems to bedisabled (carrying a black cane). Piliavin et al.related bystander help to the perceived causeof falling and a number of other variables.For the present purposes, the two experimentalconditions described above are the only fac-tors of importance.

Guided by the proposed model of moti-vated behavior, it is presumed that the per-ception of an event (falling) gives rise to asearch for causation (although this may be asecondary appraisal, following reflexive ap-proach or avoidance reactions and a primaryemotional appraisal such as fear or startle;see Lazarus, 1968). The reasons for falling,in this case, are made evident by the experi-mental manipulations of drunkenness and ill-ness. These attributions are then subject tofurther causal analysis, with the ascriptionsplaced within particular causal dimensions thatdescribe the basic properties of causes. Threedimensions of causality tentatively have beenidentified: locus, stability, and controllability(see Weiner, 1979). In the present context,the locus and controllability dimensions seemto be of greatest importance. It is reasonedthat illness is perceived as not subject topersonal control, whereas the individual is be-lieved to be personally responsible for beingdrunk. These opposing construals are hypothe-sized to give rise to differential affects, herelabeled pity and sympathy (toward the dis-abled person) and disgust or anger (towardthe drunk). These affects, in turn, respectivelybeget approach versus avoidance behavior orhelp versus neglect. Hence, the sequence ofmotivated behavior is depicted as:

rather than affect, is the immediate cause ofaction:

falling -\

(observation fromPiliavin et al.)

help (or neglect)

causal analysis(inference)

Other motivational sequences also may behypothesized. For example, it might be pre-sumed that a causal attribution gives rise toboth affect and behavior so that thought,

falling causal analysis •

affect

behavior

Given this model, falling might be perceived,for example, as resulting from an internallycontrollable cause, and this attribution wouldgenerate both disgust and neglect.

The following series of studies examinesthe attributional and the affective determi-nants of helping judgments (and, I believe,helping behavior). In Experiment 1, threedimensions of causality are related to judg-ments of helping within an academic context.Experiment 2 then moves to a different socialsetting, with subjects making judgments aboutcausal dimensions and additionally reportingwhat their feelings would be in the situationscreated by Piliavin et al. (1969). In thismanner the presupposed drunkenness - in-ternal-control - disgust and illness - lack-of-personal-control - sympathy linkages wereexamined. In Experiment 3, following the con-firmation of the proposed linkages, the degreeof personal control, the positive and negativeemotional reactions of sympathy and disgust,and the perceived likelihood of aid were rated,intercorrelated, and contrasted between thedrunk and the ill conditions. In addition, par-tial correlational analyses determined the ef-fects of perceived control on helping ratingswith affect statistically partialed out, and theinfluence of affect on judged behavior whenperceived control was held constant.

The next three investigations were experi-mental rather than correlational, in that thevariables hypothesized as mediating betweenfalling and help-giving (i.e., attributions andaffect) were manipulated rather than assessed.In Experiment 4, the two polarities of per-ceived personal control were factorially com-bined with the drunk-ill variable; in Experi-ment 5, the two types of affect were factoriallycombined with the two perceived causes;finally, Experiment 6 employed a more com-plete design (two causes of falling by twopolarities of perceived control by two typesof affect). In all the investigations the de-

188 BERNARD WEINER

pendent variable was the subjects' judgmentsabout whether or not they would give help ifthey were present in the situation.

As was just indicated, in this research thedependent variable was judgment of aid, notactual help-giving. The implicit starting pointof this endeavor is the belief that "role-enact-ment strategies . . . can help us to spread abroad net, to generate hypotheses, and tobuild heuristic models of human social be-havior" (Cooper, 1976, p. 609). It is not as-sumed that the rated likelihoods of aid andactual help-giving are isomorphic; individualssurely are not aware of all the variables thateither facilitate or hinder helping behavior,and they may incorrectly believe that certainvariables will influence their actions, whenin reality the variables are unimportant. Inthe situations presented here, however, it isnot unreasonable to presume that the judg-ment data in the various experimental condi-tions would correspond to the rank orderingof actual aid proffered, if all else were equalbetween conditions.

Experiment 1

As already indicated, three dimensions ofcausality have been identified: locus (whetherthe cause is internal or external to the actor);stability (whether the cause is perceived astemporary or permanent); and controllability(whether or not the cause is subject to per-sonal influence—see Weiner, 1979). A numberof investigators have suggested that aid isdetermined, in part, by the perceived locus ofcausality, with help more likely when the per-ceived cause of a need is an environmentalbarrier, as opposed to being internal to theperson desirous of aid (e.g., Berkowitz, 1969;Ickes, Kidd, & Berkowitz, 1976; Schopler &Matthews, 1965). For example, Berkowitz re-ported that individuals were more inclinedto help an experimental subject when the ex-perimenter caused a delay in the subject's re-sponse, as opposed to a condition in whichthe subject was perceived as responsible forfalling behind in the experiment.

Ickes and Kidd (1976) have pointed outthat the locus of causality explanation sug-gested by Berkowitz confounds the dimensionsof locus and intentionality (which I will label

control; see Weiner, 1979). They suggest thatin the Berkowitz (1969) investigation, thecausal ascription to the experimenter both isexternal to the subject and is uncontrollableby him or her, whereas an attribution to thesubject's own mismanagement is internal tothe subject and is perceived as controllable.Hence, two dimensions of causality are con-founded, and it is impossible to concludewhich of the two causal dimensions is respon-sible for the differential help-giving. Ickes andKidd (1976), in contrast to Berkowitz(1969), contend that it is the controllableaspect of the perceived cause, and not thelocus, that mediated the disparate help-giving.

In Experiment 1, the three-dimensionalclassification of causes was applied to an in-stance of helping behavior (lending class notesto an unknown classmate; from Barnes, Ickes,& Kidd, 1979). The causes of the need foraid selected represented eight possible causalconditions (two levels of locus by two levelsof stability by two levels of control). This de-sign allowed the determination of which di-mension^) of causality mediate judgmentsconcerning help-giving. Determination of thecausal dimension(s) relevant to helping judg-ments provides a necessary foundation for thesubsequent investigations.

Method

Subjects were 15 male and IS female studentsenrolled in introductory psychology at the Univer-sity of California, Los Angeles, participating tofulfill a course requirement. The subjects judged thelikelihood of help-giving in 16 situations. Two dif-ferent themes ("professor" and "employer"), eachgenerating eight judgments, were used, with thestories presented in two random orders within atheme. The order of the themes was counterbal-anced. These themes were selected after extensivepilot studies revealed that the specific stories withineach theme were classified in the experimenter-designated three-dimensional placement by more than90% of the judges.

Both themes involved an unknown student askingto borrow the notes from the last class period (fromBarnes et al., 1979). The "professor" theme con-veyed that the student always (stable) or sometimes(unstable) did not take notes because of somethingabout himself (internal) or something about theprofessor (external). Either he was unable to takegood notes (uncontrollable) or did not try (con-trollable), whereas the professor either was unableto give a clear lecture or did not try. For example,


an internal, stable, and uncontrollable cause wasthat the student was never able to take good notes(low ability), whereas an external, stable, and un-controllable cause was that the professor was neverable to give a clear lecture. Each specific storywithin the eight conditions elaborated this basicscenario. The "employer" theme stated that the stu-dent did not have the notes because he (or theboss) always (or sometimes) was responsible forhis coming late to school, which could (could not)have been avoided.

Following each causal statement the subjects ratedthe likelihood of lending their notes to the student.Judgments were made on a 10-point scale anchoredat the ends with "extremely likely to help" and"extremely unlikely to help."

Results and Discussion

An analysis of variance (ANOVA) indicatedthat sex, theme, and stability did not appearas main effects or in interaction with any ofthe other variables. These factors thereforeare omitted from the remainder of theanalyses.

The mean helping judgments in the fourconditions (two levels of locus by two levelsof control) are shown in Table 1. There is amain effect for locus, F(l, 28) = 118.24, p< .0001, and for controllability, F(l, 28) =79.31, p < .0001. More important, there is ahighly significant Locus X Controllabilityinteraction, F(l, 28) = 198.47, p < .0001.Examination of Table 1 shows that helping isreported to be relatively equal and reasonablyhigh in all conditions except when the causeis internal and controllable, in which caseaid is reported as being withheld. Thus if thestudent did not try to take notes or couldhave avoided the lateness, then help appar-ently would not be given. Rated help is sig-nificantly less likely (p < .0001) in thiscondition than in the remaining three LocusX Controllability conditions.

In sum, these data tell a simple but mean-ingful story and permit an unambiguousinterpretation of the prior findings in thearea. The mismanaging subjects in the Berko-witz (1969) experiment were not aided be-cause they were perceived as personallyresponsible (able to respond; Ickes & Kidd,1976) for their plight. The interaction of twodimensions of causality—locus and control—influences judgments of help and, presumably,help-giving.

Table 1Mean Likelihood of Helping as a Function, ofLocus and Control of Cause

Locus Controllable Uncontrollable

InternalExternal

3.137.35

6.746.98

Note, The higher the number, the greater the likeli-hood of help-giving.

The above argument assumes that causescan be labeled as controllable although theycannot be subject to the personal influence ofthe actor. For example, if the teacher, by nottrying, is responsible for the student's prob-lem, then that cause may be labeled as con-trollable, even though from the perspectiveof the student the cause is beyond personalcontrol. Of course, if the disadvantages ofthis forced orthogonality between the causaldimensions are proven to exceed the advan-tages, then a different conceptual analysis iswarranted in which external causes are notdistinguished according to their control-lability.

There is additional evidence that perceivedpersonal control relates to a variety of inter-personal judgments. For example, in achieve-ment-related contexts effort is a much morepotent determinant of teacher evaluation thanis ability. High effort is rewarded more thanhigh ability, given success, and lack ofeffort is punished more than a lack of ability,given failure (see Weiner & Kukla, 1970). Itis suggested that because effort is perceived asinternal and controllable, reward and punish-ment for achievement performance are height-ened.

Furthermore, sentiments are also influencedby perceived internal controllability. Peplau,Russell, and Heim (1979) summarize anumber of investigations indicating that per-sons lonely for reasons that are perceived aspersonally controllable (e.g.; do not try tomake friends) are less liked and are givenless sympathy than are individuals lonely foruncontrollable reasons (e.g., no opportunityto meet people). It seems reasonable to sur-mise that such sentiments also would applyin achievement-related contexts: A personfailing because of a lack of effort probably

190 BERNARD WEINER

will be disliked, in addition to being judgedresponsible for the failure. Furthermore, anindividual's obesity that is attributed to self-indulgence—a deficiency in "try"—gives riseto disliking of that person, whereas this is notthe case if the obesity is ascribed to a physio-logical dysfunction—a deficiency in "can"(Dejong, 1980).

In sum, there is vast evidence demonstrat-ing the influence of perceived personal controlin interpersonal judgments, including evalua-tion, sentiments, and as documented here,thoughts about help-giving. The questionsguiding the five investigations about to bepresented, which are set within the context ofthe experiment created by Piliavin et al.(1969), are whether thoughts about helpingare directly mediated by perceptions of causal-ity or whether causal thoughts are linkedwith particular affects that in turn influencejudgments of help. The general hypothesesexamined are that perceived personal controlfor a need generates disgust or anger, whereaslack of perceived personal control gives riseto sympathy. These affects then beget ap-proach (help) or avoidance (neglect) judg-ments.

Experiment 2

Method

In Experiment 2, both thoughts about causalityand affects are examined in a help-giving context.The subjects were 40 male and female studentsenrolled in introductory psychology at the Univer-sity of California, Los Angeles, participating to ful-fill a course requirement. They were given thefollowing scenarios (based upon Piliavin et al., 1969,p. 291) in counterbalanced order:

At about 1:00 in the afternoon you are riding asubway car. There are a number of other indi-viduals in the car and one person is standing,holding on to the center pole. Suddenly, thisperson staggers forward and collapses. The per-son is carrying a black cane and apparently isill. (Alternate form: The person apparently isdrunk. He is carrying a liquor bottle wrapped ina brown paper bag and smells of liquor.)

Subjects then read: "Try to assume that youactually are on the subway and try to imagine thisscene. Describe your feelings in this situation." Threespaces were provided for affective descriptions.

The subjects then rated the causes of falling onthe three dimensions of locus, stability, and con-

trollability. The meaning of each dimension waselaborated with specific examples. The rating scaleswere anchored at the extremes with the poles of thedimensions (internal-external; permanent-temporary;controllable-uncontrollable), and responses to themwere given immediately following each scenario andthe emotional reports. For scoring purposes thescales were divided into nine equal intervals.


The 240 affects listed by the subjects (40subjects by two causes by three affects) wereclassified into 13 categories, with a 94% inter-rater agreement. Table 2 shows the categories,the percentage of responses in each category,and the number of subjects giving a categoryresponse as their first reaction. The categoriesare labeled sympathy (e.g., pity, feelingsorry); concern (concern and worry); nega-tive affect toward the person (e.g., anger, dis-gust) ; general discomfort (e.g., embarrassed,upset); fear; caution; surprise, positive ac-tion (e.g., seek help); apathy; personal short-comings (e.g., helpless, inadequate); informa-tion seeking (e.g., curious); description (e.g.,a wino); and unclassified.

The largest difference between the drunkand the ill conditions involved negative af-fects toward the person. Of the responseselicited by the description of the drunk per-son, 2T°/o were negative affects directedtoward him, whereas such negativity char-acterized only 3% of the responses towardthe ill individual. In addition, nearly onethird (13 of 40) of the subjects listed anegative emotion toward the drunk as theirinitial feeling, but this was not true in anyinstance for the case of the ill person. Con-cerning the more positive outward-directedaffective categories of sympathy and concern,there was a trend in the reverse direction,with 46% of the responses toward the sickindividual being positive emotions (19 sub-jects indicated this as their initial feeling),whereas 30% of the emotional responsestoward the drunk were positive emotions (14persons stated that this would be their firstreaction). The difference between the positiveand negative affective reactions in the twoconditions was highly significant (p < .001by the Fisher exact test; only first responseswere tested to meet the independence cri-terion). In addition, some of the responses


were action oriented rather than representingwhat is usually meant by feelings and emo-tions. Of the responses toward the ill person,10% were help related, whereas \% impliedapathy. Conversely, 2% of the responsestoward the drunk were help related, whereas7% conveyed a lack of concern.

Concerning the ratings of the three causaldimensions, drunkenness was perceived asmore controllable, t(39) = 7.39, p < .0001,and more internal, <(39) = 3.46, p < .0001,than illness, whereas the two causes did notdiffer in perceived stability (t < I ) .

In sum, the results of Experiment 2 en-couraged further pursuit of the belief thataffect, as well as attributions, mediates judg-ments of help-giving. As anticipated, therewas evidence that negative emotions such asdisgust and anger were more likely to beexperienced when people were exposed to thedrunk individual, whereas positive emotionssuch as pity and sympathy were more likelyto be experienced in connection with the illperson. These other-directed emotions werecited as the initial feeling by nearly 60% ofthe subjects. Furthermore, the causes differedin their degree of perceived personal control-lability or how responsible the person wasperceived as being for the cause of falling:One is judged as personally responsible forbeing drunk, but not for being ill. Experiment

3 pursues further the findings reported inExperiment 2, using a methodology that per-mits more complex statistical analyses.

MethodExperiment 3

The subjects were 28 male and female studentsenrolled in introductory psychology at the Univer-sity of California, Los Angeles, participating to ful-fill a course requirement. They again were giventhe drunk and the ill scenarios in counterbalancedorder. Following each scenario, the subjects ratedthe degree to which the cause was perceived aspersonally controllable (under personal control-notunder personal control), their feelings of pity andsympathy (a great deal-none), their feelings ofdisgust and distaste (a great deal-none), and theirlikelihood of helping (definitely would aid - definitelywould not aid). The order of the attribution andaffect scales was counterbalanced so that neitherwould always immediately precede the helping judg-ments. For scoring purposes the scales were dividedinto nine equal intervals. It was thus possible todetermine the relations between personal control,affect, and ratings of help, as well as to comparethe differences in these variables between the drunkand the ill conditions.


Table 3 shows the correlations across boththe drunk and the ill scenarios between per-ceptions of personal control, positive and neg-ative outward-directed affects, the resultant of

Table 2Emotional Responses, Including All Reactions and Only the First Reaction for Each Subject

III Drunk

Affective category

SympathyConcernNegative affectDiscomfortFearCautionSurprise

Positive actionApathyPersonal shortcomingsInformation seekingDescriptionUnclassified

All reac-tions (%)

331336

1067

1015510

Initialresponse

(no.)

14500437

210310

All reac-tions (%)

255

279334

271464

Initialresponse

(no.)

131

133023

010211

192 BERNARD WEINER

Table 3Correlations Between the Judgments of Control, Affect, and Help Including Both the Drunk andIII Conditions

Variable 1

1. Control — -.77***2. Sympathy (S) —3. Disgust (D)4. S-D5. Help

.55**-.64***

—

-.73***.90

-.91—

-.37*.46**

-.71***.65***—

V < .05. **p<.01. ***£<.001.

the positive minus the negative emotions, andjudgments of help-giving. Because each sub-ject made ratings in both the drunk and theill conditions, Table 3 includes only the re-sponses in the drunk condition of the 14subjects judging the drunk individual firstand the responses in the ill condition of theremaining 14 subjects judging the ill personfirst. When the other one-half of the data forthe 28 subjects are used in a separate (al-beit not independent) analysis, there is aclear replication of the findings reported inTable 3.

Examination of Table 3 reveals that per-ceptions of personal control were negativelyrelated to feelings of sympathy, positivelyrelated to feelings of disgust, and negativelyrelated to judgments of help. That is, personalresponsibility for falling was accompanied bythe absence of positive affect, the presence ofnegative affect, and neglect. Furthermore,sympathy was positively related to judged

Table 4Correlation of Variables With Judgments ofHelping, Including Both the Drunk and IIIConditions, With Individual VariablesStatistically Partialed From the Analysis

1.2.3.4.5.

Partialedvariable

NoneControlSympathy (S)Disgust (D)S-D

2

-.37*

—-.02".04.20

3

.46**

.30

—.01—

4

-.71***-.66***-.61***

——

5

.65***

.60***———

a Indicates the correlation between perceptions ofcontrol and helping ratings, with sympathy ratingspartialed out.* p < .05. ** p < .01. *** p < .001.

help-giving, whereas disgust was negativelyassociated with judgments of help. In sum,Table 3 strongly supports the hypothesizedlinkages of lack of personal control, sympa-thy, and help, and between perceived personalcontrol, disgust, and neglect.

Further analyses were conducted in whichattributions were related to ratings of help-giving with affects statistically held constant,and affects were related to helping judgmentswith personal control partialed out (see Table4). Column 1 of Table 4 reveals that whensympathy or disgust are held constant, thecorrelation between personal control andjudgments of help is no longer significant. Onthe other hand, Row 2 of Table 4 shows thatthe correlations between disgust and helpratings and between resultant affect (sympa-thy minus disgust) and help are scarcelyreduced when personal control is statisticallyheld constant. These findings are furtherelaborated with stepwise multiple regressionanalyses. These analyses indicate that re-sultant affect by itself contributes 42 % toR'3, whereas the addition of personal controlcontributes only an additional 2% to thevariance accounted for in the helping judg-ments. When personal control is entered as thefirst predictor, then it contributes 13% to R2,whereas the addition of resultant affect ac-counts for another 31% of the variance. Insum, the affect variable independently con-tributes 31%, and attribution independentlycontributes 2%, to R2, with 11% of thevariance accounted for jointly by these twovariables.

Table 4 also suggests that disgust is themore influential of the two affective ratings.However, this was the only effect not repli-


Table 5Correlations Between Judgments of Control, Pity, Disgust, and Help

Variable 1

1. Control2. Sympathy (S)3. Disgust (D)4. S-D5. Help

-.53**.15

-.42*-.12

-.55**—

-.17.83.17

.46*-.68***

—-.69-.34

-.55**.92

-.91—.32

-.36.61***

-.67***.69***—

Note, Correlations within the drunk condition are shown in the upper right half; correlations in the illcondition are shown in the lower left half.V < .05. ** />< .01 . *** p < .001.

cated using the second half of the data set,where sympathy and disgust emerged asequally strong associates of the helping judg-ments. Thus, this finding is not consideredfurther.

Correlations were then computed within thedrunk and the ill conditions. Because of se-vere restrictions in range, it was anticipatedthat the correlations reported in Tables 3 and4 would be reduced in magnitude. The upperhalf of Table S shows the correlations withinthe drunk condition, whereas the lower half ofTable 5 reports the data within the ill condi-tion. Examination of the judgments within thedrunk condition reveals the identical patternas that shown in Table 3, without the antici-pated reduction in the magnitudes of thecorrelations. Personal control correlates nega-tively with sympathy, positively with disgust,and negatively with helping judgments. Inaddition, sympathy correlates positively, anddisgust negatively, with helping ratings.

The partial correlations within the drunkcondition are also consistent with the overallanalyses and do not differ in magnitude fromthe data in Table 4 (see Table 6). With af-fective ratings held constant, causal attribu-tions are only marginally related to judgmentsof help (see Column 1, Table 6). On the otherhand, with personal control statistically heldconstant (Row 2, Table 6), the correlationsbetween the affects and the helping ratingsare only slightly reduced. Stepwise multipleregression analyses reveal that resultant affectby itself contributes 48% to jR2, whereas theaddition of personal control accounts for nofurther variance. When control is entered asthe first predictor, it then contributes 13%to R-, whereas the addition of affect accounts

for another 35% of the rating variance. Insum, affect independently contributes 35% to/?2, and personal control independently con-tributes nothing, while these two variablesshare 13% of the variance in the helping rat-ings.

The lower half of Table 5 shows the correla-tions within the ill condition. It is evidentthat these correlations are greatly reduced,although the directions of the findings areconsistent with what has already been re-ported. That is, helping judgments relatenegatively to personal control, positively withsympathy, and negatively to disgust. Further-more, resultant affect uniquely contributes 9%to R2, whereas personal control independentlyaccounts for no additional variance, and thetwo variables jointly share 1 % of the variance.

One probable reason for the stronger rela-tions in the drunk than in the ill condition isthat the variances in the former condition are

Table 6Correlations of Variables With Judgments ofHelping, in Only the Drunk Condition, WithIndividual Variables Statistically PartialedFrom the Analysis

1.2.3.4.5.

Partialedvariable

NoneControlSympathy (S)Disgust (D)S-D

2

-.36—

-.04'-.08

.03

3

.61***

.52**—

.27—

4

-.67***-.60***-.44*

——

5

.69***

.63***———

" Indicates the correlation between perceptions ofcontrol and helping ratings, with sympathy ratingspartialed out.*p < .05. **p < .01. ***p < .001.

194 BERNARD WEINER

greater, with significant differences betweenthe variances of the disgust judgments (p <.05), whereas all the other variance compari-sons approach statistical significance (p <.10). The mean of the judgments of disgust inthe ill condition was only .70 (on a 9-pointscale), SD = 1.2S. Judgments of disgust inthe drunk condition reveal M = 4.73, SD =2.80. There may be other factors operativethat gave rise to more pronounced effects inthe drunk than in the ill condition, but vari-ance limitations do appear to be among thecausative factors.

Analyses of the ratings over both conditionsalso revealed highly significant differencesbetween the drunk and the ill individuals,with drunkenness perceived as more person-ally controllable, less provoking of sympathy,more evocative of disgust, and less likely tolead to help than illness (all ps < .0001).Hence, further evidence is provided thatdrunkenness is associated with personal con-trol, disgust, and neglect, whereas illness isassociated with a lack of responsibility, sym-pathy, and help.

Experiments 4 and S

Although the two previous experiments werecorrelational, it has been contended that thereis a causal sequence of events in a motiva-tional episode. In Experiments 4 and 5 theattributions and the affects believed to medi-ate judged help-giving were manipulated in afactorial design and were related to judgmentsof aid, thus providing additional evidenceconcerning the issue of causal relations be-tween the variables.

Method

Subjects in Experiment 4 were 41 male and femalestudents enrolled at the University of California, LosAngeles, participating to fulfill an introductory psy-chology course requirement. The subjects were giventhe two scenarios used in Experiments 2 and 3. Inthis instance, however, the scenarios were factoriallycombined with two levels of perceived personalcontrol. The subjects considered the following foursituations:

A. The person who falls is carrying a black caneand apparently is ill.1. You know that this illness is not under con-

trol of the person.

2. You know that this illness is under control ofthe person and that he is refusing to takethe medication that he should.

B. The person who falls is drunk. He is carryinga liquor bottle wrapped in a brown bag andsmells of liquor.1. You think that drinking is not under per-

sonal control and is an addiction.2. You think that one should be able to control

the amount that one drinks.

The subjects were asked to "imagine that thesethoughts truly characterize your beliefs in the situ-ations." Subjects then rated the amount of help thatthey would be willing to give on a scale anchoredat the extremes by definitely would aid - definitelywould not aid. For scoring purposes the scale wasdivided into nine equal intervals. There were fourrandom story orders.

Experiment 5 followed the same format, varyingstated affective reactions rather than beliefs aboutcontrol. The subjects read:

The person who falls is carrying a black cane [isdrunk].1. You feel disgust and distaste.2. You feel sympathy and pity.

Again the subjects were asked to assume that thedescriptions characterized their feelings in the situa-tions, and help was rated on the scale previouslydescribed. There were also four random story orders.The subjects were 42 male and female students en-rolled at the University of California, Los Angeles,participating to fulfill a course requirement in intro-ductory psychology.


Figure 1 shows the data given the personalcontrol (left half) and affect (right half)manipulations. Analyses of the ratings giventhe personal control manipulation reveal sig-nificant main effects due to the drunk-illvariable, F(l, 39) = 58.04, p < .0001, anddue to perceived control, F(l, 39) = 27.23,p < .0001. Illness and perceived lack ofpersonal control promote the judged likelihoodof helping. The respective variances accountedfor in the judgments, computed with w2, are47% and 6%. The interaction term did notapproach significance (F < 1).

Analyses of the judgments given the affectmanipulation reveal significant main effectsdue to the drunk-ill variable, F(l, 40) =48.75, p < .0001, and due to the affectivefactor, F(l, 40) = 99.60, p < .0001. Illnessand positive affective reactions toward the


person increased the judged likelihood of help-ing. The respective variances accounted for inthe judgments, computed with w2, are 15%and 45%. The interaction term did not ap-proach significance (F < 1).

In sum, independent of perceptions of per-sonal control and the specified affective reac-tions, falling because of drunkenness versusillness gave rise to differential judgments ofhelping. This may be due to any number ofreasons, some of which could be affective(fear of the drunk), cognitive (anticipatingthat one's clothes might be soiled when help-ing the drunk), script guided (stay awayfrom drunks), and so forth. Some possibledeterminants of helping in addition to attri-butions and affects are considered in the dis-cussion section of this paper. But the mostdominant determinants of helping judgmentswere the specified affective reactions, which isconsistent with the data reported in Experi-ment 3. Perceptions of personal control alsosignificantly influenced the helping ratings.However, this effect was smaller than thedrunk-ill or sympathy-disgust manipulationsand accounted for relatively little variance.Furthermore, it may be that the effect wasdue to the fact that disparate perceptions ofpersonal control elicit differential affectivereactions. That is, the significant effect doesnot ensure that attributional beliefs are di-rectly linked to judgments and behavior. Thisquestion is pursued further in Experiment 6.

Experiment 6

Experiment 6 follows the same general for-mat as Experiments 4 and 5. However, a morecomplete design is employed, using the threecritical variables of drunk-ill, attributions,and affects.

Method

Subjects were 99 male and female students en-rolled at the University of California, Los Angeles,participating to fulfill an introductory psychologyrequirement. The subjects rated their likelihood ofhelping in eight conditions (two causes of falling bytwo levels of personal control by two types ofaffect). For example, after reading the falling sce-nario the subjects read:

The person who falls is carrying a black cane andapparently is ill. You feel disgust and distaste. You

8-1EXPERIMENT 4 EXPERIMENT 5

1 7

O-

UJ _i 6

u_OQ 5OO

D 4UJ

QUJ

2-

0-

UNCONTROLLABLE

CONTROLLABLE

PITY

DISGUST

DRUNK ILL DRUNK ILL

Figure 1. Mean helping judgments given drunk andill individuals as a function of perceived control-lability of the cause (left half) and affective reactionstoward the person (right half).

know that this illness is under volitional controlof the person and that the person is refusing totake the medication that he should.

There were four different orders of stories, and theorder of the affect and the control information wascounterbalanced.

Results

An ANOVA revealed significant main effectsfor the drunk-ill variable, F(l, 97) = 78.21,p < .0001; for perceptions of personal con-trol, /?(!, 97) = 54.31, p < .0001; and foraffect, F(l, 97) = 116.82, p < .0001. Theperson most likely to be helped, according tothe judgments, was ill, the cause was perceivedas not personally controllable, and the affec-tive reaction was sympathy. The respectivevariances accounted for by these effects, com-puted by a w", were 18%, 4%, and 16%.None of the interactions approached signifi-cance.

General Discussion

The six experiments allow some generalconclusions to be drawn, although many ques-

196 BERNARD WEINER

tions remain unanswered. First, there is nodoubt that there is a union between drunk-enness, perceived personal control, negativeoutward-directed affective reactions, andjudgments to neglect, as well as between ill-ness, perceived lack of personal control, posi-tive outward-directed affective reactions, andjudgments to help.

Regarding the determinants of helpingjudgments and, presumably, helping behavior,there is little doubt that drunks are helpedless than ill persons in the situation described.The judgment data are consistent with thebehavioral observations reported by Piliavinet al. (1969) and thus increase our confidencein the validity of the role-enactment method-ology. As indicated previously, the correlatesof the drunk and ill states that beget neglector help are doubtless numerous and rangefrom specific fears and cost-benefit calcula-tions to general moral beliefs.

It also is quite clear that affective reactionsare an important determinant of stated reportsabout helping. In the correlational investiga-tions affects were highly related to helpingjudgments even when perceptions of personalcontrol were statistically partialed from theanalysis. And the highly significant F valuesand substantial variance accounted for by theaffective manipulations in Experiments 5 and6 support the findings in the correlationalstudies.

Finally, thoughts about personal control atbest only weakly exert direct influence onhelping judgments. If perceptions of personalcontrol do directly influence these judgments,then that influence is less than that of theaffects employed in the present experiments.It appears that much of the relation betweenthe judgments of personal control and ratedhelp exists because perceptions of control arestrongly related to affective reactions.

A Model of Helping Behavior

An attributional model of helping behaviorthat is injerred from the judgment data intui-tively that seems best and most adequately in-corporates the data reported in the prior sixinvestigations is shown in the following dia-gram. This diagram indicates that when a per-son falls, there is probably some reflexive ap-

proach or avoidance behavior and immediateaffective reactions to the event, such as startleor fear (these are untested relations, shown inLinkages 1 and 2 in the diagram). Individuals

falling

initial causal analysis(drunk vs. ill)

elaborated causal analysis(causal dimensions)

affect

then engage in a causal search, seeking thereasons for the event (Linkage 3). This is anassumption of attribution theorists, but thereis evidence that unexpected events are particu-larly likely to elicit an attributional search(Lau & Russell, 1980; Wong & Weiner, inpress). The cause of falling is then deter-mined, in this case quite easily on the basis ofthe descriptions provided. In other situations,a causal conclusion might prove quite difficultto reach and would require information search,such as looking for a medical bracelet or dia-betic card. The ascription to drunkenness orillness gives rise to help or neglect independentof an elaborated causal (dimensional) analy-sis (Linkage 4). This linkage was suggestedin the judgment research reported here, al-though as previously stated, some interveningcognitive processes are likely, such as thoughtsabout the "cost" of helping. In addition, it ispossible that perceptions of drunkenness orillness give rise to particular affective reac-tions that are independent of further causalconsideration (e.g., one is saddened at thesight of the ill). This possible relation isshown in Linkage S. The perceived cause,then, is consciously or unconsciously anddeliberately or automatically examined interms of its underlying properties (Linkage6), with the key dimensions in this instancebeing locus and control (perceived personalresponsibility). The dimensions of causalascriptions have been identified in other re-


search (see Weiner, 1979), and the relationsbetween drunkenness and illness and perceivedpersonal control were demonstrated in thepresent studies. Causal ascriptions appear torelate weakly to help (Linkage 7) but clearlyto affective reactions (Linkage 8), as inferredfrom the research reported here and studiescited earlier. Finally, affects are strongly re-lated to judgments of help or neglect (Link-age 9), as has repeatedly been contended inthe prior pages, and presumably they alsorelate to actual helping behavior.

Needless to say, other linkages within thismodel are likely (such as help-giving influ-encing immediate and more reflective affects),and there are a myriad of determinants ofhelp-giving not included in the diagram, suchas the number of bystanders and the be-havior of models. What the diagram does at-tempt to capture is an attributional approachto judgments of aid and to helping behavior,stressing the relations between attribution,affect, and action, represented in the 3-6-8-9path in the diagram.

Generality oj the Model and Attributional-Change Programs

The model outlined in the diagram is ex-pected to generalize across a variety of help-giving situations. For example, Experiment 1suggested that within a classroom contextnotes will be lent to a student if the cause ofthe need is low ability or teacher shortcom-ings, whereas help will be withheld if the needis ascribed to a lack of effort. Low abilityand teacher shortcomings, like illness, shouldelicit sympathy inasmuch as they are notunder personal control. On the other hand,lack of effort, like drunkenness, should elicitdisgust or anger, for effort is perceived aspersonally controllable. It is hypothesizedthat these affects in turn are responsible forthe differential help-giving judgments in Ex-periment 1.

Further, the proposed behavioral sequenceshould generalize to other motivational do-mains. For example, in achievement-relatedcontexts success perceived as due to otherselicits gratitude (Weiner et al., 1978) and isanticipated to generate behavior appropriateto that affect, such as the giving of a gift. In

a similar manner, failure due to hindrancefrom others begets anger (Weiner et al., 1978)and is likely to give rise to aggression to"eliminate" the source of anger. That is, thereis again an attribution-affect-action ordering.

There is a growing literature in the achieve-ment area utilizing attributional retrainingtechniques to augment achievement strivings.Typically, the program participants are toldthat their failure is caused by a lack of effort,which is unstable and controllable (see An-drews & Debus, 1978; Dweck, 1975). Giventhis ascription, one can maintain hope in theface of failure and sustain a relatively highexpectancy of success. Subsequent improve-ments in performance have been believed tobe due to the change in the expectancy varia-ble. On the other hand, the present researchintimates that attributions of failure to a lackof effort generate particular affective reac-tions, and these self-directed emotions (shameand guilt; see Weiner et al., 1978) or inferredother-directed emotions (anger and disgust)are responsible for the subsequent incrementsin achievement-oriented behavior. This is notto suggest that expectancy is not a centraldeterminant of behavior. Rather, I want topoint out that the motivational effects of emo-tional reactions have been completely over-looked in the attributional change research,and these effects provide an equally plausibleexplanation of the change program results(see Weiner, in press).

Comparison and Contrast WithAlternate Theories

The empirical literature concerning helpingbehavior and altruism has grown enormouslyin the last decade and has spawned a varietyof conceptual analyses (see Macaulay & Berk-owitz, 1970; Wispe, 1978). I will now brieflyexamine these theories, confining my remarksto comparisons with the attributional approachadvocated here and to what I believe are theshortcomings of these alternate conceptionsand the strengths of the attributional ap-proach.

First, it is evident that there are as manydeterminants of helping behavior as there aresources of motivation. One can help to winapproval and friendship, to gain power, for a

198 BERNARD WEINER

monetary reward, and so on. These "benefits"surely will increase helping, just as "costs"such as time (Barley & Batson, 1973) andthe possibility of getting "bloodied" (Pilia-vin, Piliavin, & Rodin, 1975) retard helpingactivity. Such effects are beyond the range ofconvenience of the attributional analysis pre-sented here and are likely to be embracedwithin the differential helping judgments ac-corded to a drunk versus an ill individual,irrespective of perceived personal controlla-bility and the affects of disgust and sympa-thy. However, cost-benefit concerns cannotexplain the total pattern of data presented inthese investigations. Furthermore, this con-ceptual approach cannot account for why aperson who has failed to take notes becauseof laziness is judged less likely to be helpedthan one who does not have the notes becauseof low ability. Indeed, from a cost-benefitperspective the opposite pattern might bepredicted, for the low ability person will notbe "able" to reciprocate.

Cost-benefit concerns are less applicable towhat have been labeled "altruistic" acts, whichby definition involve no benefits to the helper.Three traditional motivational concepts havebeen widely used to account for altruism:hedonism, homeostasis, and arousal. The he-donistic doctrine has been responsible forthinking of altruism as a "paradox," for it isconsidered anomalous to engage in behaviorthat apparently does not maximize pleasureand minimize pain. I find this an unnecessaryconcern, for even Freud (1920/1955) notedthat many behaviors, including transference,traumatic dreams, and disappearance games,fall "beyond the pleasure principle." In addi-tion, ego psychologists have postulated thatmastery attempts (attribution theorists havestressed inclusion of information gain in thiscategory) are basic "springs of action." Insum, not all behaviors are in service of thepleasure principle, and attempts to accountfor such behaviors by finding some underlyinghedonic gain, such as reduction of an internalstate that brings the organism back into astate of equilibrium, seem forced.

Arguing within the framework of the he-donistic doctrine, it has been suggested thatobserving another in distress produces arousal

that in turn activates the person to reducethat arousal (Piliavin et al., 1975). Arousalis presumed to be diminished in a manner thatminimizes costs while activating the most ef-fective instrumental response (Hatfield, Wal-ster, & Piliavin, 1978). This approach has beensupported by a number of apparently con-firmatory investigations (e.g., Gaertner &Dovidio, 1977; Harris & Huang, 1973; Pili-avin et al., 1975). However, the conceptionseems to be beset by the following problemsand shortcomings:

1. There is adherence to a drive-drive re-duction framework that has been proven in-adequate in other motivational domains (seeAtkinson, 1964; Bolles, 1975; Weiner, 1972).

2. Without clear specification of additionalprinciples, the arousal position cannot specifythe direction of behavior. For example, dis-gust and pity may give rise to equal levels ofarousal, yet they respectively generate avoid-ance behavior and approach behavior.

3. The arousal position does not appear tobe generalizable to other behaviors. For ex-ample, it seems unlikely that the purchase of agift when one is grateful is undertaken to re-duce some sort of distress.

4. Some of the central confirmatory studiesmake use of a misattribution procedure inwhich arousal is misattributed to a pill (e.g.,Coke, Batson, & McDavis, 1978; Gaertner &Dovidio, 1977). However, alternative inter-pretations of the finding that helping is re-duced in this condition are quite plausible.For example, it has been contended that self-concern lessens the inclination to help (Berko-witz, 1970). It may be that preoccupationwith the self, elicited because of the ingestionof an "arousing" drug, reduces or inhibitssympathy, which then decreases relativehelping.

The conception to which the proposed at-tributional approach is most closely aligned isthe so-called empathy position (e.g., Aron-freed, 1968; Rosenhan, 1969). The only cleardifference between empathy theories and whathas been suggested here is that the empathytheories have stressed the importance of roletaking, while I have documented the impor-tance of attributions to causality, particularlyperceived personal control. But it is intimated


that any factor that might increase sympathy,such as role taking, the severity of the need, ora positive mood, will augment altruistic be-havior (see Isen, 1970). This is entirely con-sistent with a position voiced by Rosenhan(1978), who stated: "Cognitive factors playa role in the elicitation of behavior when andonly when they amplify empathic and sympa-thetic affects" (p. 110). Coke et al. (1978)have provided strong evidence in support ofthis statement. They found that empathy in-fluences helping only through its effects on"empathic emotions" (e.g., concern andworry). On the basis of these data they alsopropose a two-stage theory of helping in whichcognitions influence emotions and emotionsaffect behavior.

One of the strengths of the attributionalmodel proposed here, as opposed to the viewsof Coke et al. (1978) and Rosenhan (1969),is that the attributional conception can alsobe applied to "selfishness" (see Wilson, 1978).That is, both the maintenance of personal"fitness" (neglect of the needy) and decreasesin personal fitness (help-giving) can be ac-counted for with principles of perceived per-sonal control and their affective associates.

References

Andrews, G. R., & Debus, R. L. Persistence andcausal perceptions of failure: Modifying cogni-tive attributions. Journal of Educational Psychol-ogy, 1978, 70, 154-166.

Aronfreed, J. Conduct and conscience: The socializa-tion of internalized control over behavior. NewYork: Academic Press, 1968.

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

Barnes, R. D., Ickes, W. J., & Kidd, R. F. Effectsof the perceived intentionality and stability ofanother's dependency on helping behavior. Person-ality and Social Psychology Bulletin, 1979, 5, 367-372.

Berkowitz, L. Resistance to improper dependencyrelationships. Journal of Experimental Social Psy-chology, 1969, J, 283-294.

Berkowitz, L. The self, selfishness, and altruism. InJ. Macaulay & L. Berkowitz (Eds.), Altruism andhelping behavior. New York: Academic Press,1970.

Bolles, R. C. Theory of motivation (2nd ed.). NewYork: Harper & Row, 1975.

Coke, J. S., Batson, C. D., & McDavis, K. Empathicmediation of helping: A two-stage model. Journalof Personality and Social Psychology, 1978, 36,752-766.

Cooper, J. Deception and role playing: On tellingthe good guys from the bad guys. American Psy-chologist, 1976, 31, 605-610.

Darley, C. M., & Batson, C. D. From Jerusalem toJericho: A study of situational and dispositionalvariables in helping behavior. Journal of Personal-ity and Social Psychology, 1973, 27, 100-108.

Dejong, W. The stigma of obesity: The consequencesof naive assumptions concerning the causes ofphysical deviance. Journal of Health and SocialBehavior, 1980, 21, 75-87.

Dweck, C. S. The role of expectations and attribu-tions in the alleviation of learned helplessness.Journal of Personality and Social Psychology,1975, 31, 674-685.

Folkes, V. S. Causal communication in the earlystages of affiliative relationships. Unpublished doc-toral dissertation, University of California, LosAngeles, 1978.

Freud, S. Beyond the pleasure principle. In The stan-dard edition (Vol. 18). London: Hogarth Press,1955. (Originally published, 1920.)

Gaertner, S. L., & Dovidio, J. F. The subtletyof white racism, arousal, and helping behavior.Journal of Personality and Social Psychology,1977, 35, 691-707.

Harris, M. B., & Huang, L. C. Helping and the at-tribution process. Journal of Social Psychology,1973, 90, 291-297.

Hatfield, E., Walster, G. W., & Piliavin, J. A. Equitytheory and helping relationships. In L. Wispe(Ed.), Altruism, sympathy, and helping. NewYork: Academic Press, 1978.

Ickes, W. J., & Kidd, R. F. An attributional analy-sis of helping behavior. In J. H. Harvey, W. J.Ickes, & R. F. Kidd (Eds.), New directions inattribution research (Vol. 1). Hillsdale, N.J.: Erl-baum, 1976.

Ickes, W. J., Kidd, R. F., & Berkowitz, L. Attribu-tional determinants of monetary help-giving. Jour-nal of Personality, 1976, 44, 163-178.

Isen, A. M. Success, failure, attention, and reactionsto others: The warm glow of success. Journal ofPersonality and Social Psychology, 1970, IS, 294-301.

Lau, R. R., & Russell, D. Attributions in the sportspages. Journal of Personality and Social Psychol-ogy, 1980, 39, 29-38.

Lazarus, R. S. Emotions and adaptation: Conceptualand empirical relations. In W. J. Arnold (Ed.),Nebraska symposium on motivation (Vol. 16).Lincoln: University of Nebraska Press, 1968.

Macaulay, J., & Berkowitz, L. (Eds.). Altruism andhelping behavior. New York: Academic Press, 1970.

Peplau, L. A., Russell, D., & Heim, M. The experi-ence of loneliness. In I. Frieze, D. Bar-Tal, & J.Carroll (Eds.), New approaches to social prob-lems: Applications of attribution theory. SanFrancisco: Jossey-Bass, 1979.

Piliavin, I. M., Piliavin, J. A., & Rodin, J. Costs dif-fusion and the stigmatized victim. Journal ofPersonality and Social Psychology, 1975, 32, 429-438.

200 BERNARD WEINER

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

Rosenhan, D. L. Some origins of concern for others.In P. Mussen, M. Covington, & J. Langer (Eds.),Trends and issues in developmental psychology.New York: Holt, Rinehart & Winston, 1969.

Rosenhan, D. L. Toward resolving the altruism para-dox. In L. Wispe (Ed.), Altruism, sympathy, andhelping. New York: Academic Press, 1978.

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

Tomkins, S. S. Affect, imagery, consciousness: Vol. 2.The negative affects. New York: Springer, 1963.

Weiner, B. Theories of motivation: From mecha-nism to cognition. Chicago: Rand McNally, 1972.

Weiner, B. A theory of motivation for some class-room experience. Journal of Educational Psychol-ogy, 1979, 71, 3-25.

Weiner, B. The role of affect in rational (attribu-tional) approaches to human motivation. Educa-tional Researcher, in press.

Weiner, B., & Kukla, A. An attributional analysis ofachievement motivation. Journal of Personalityand Social Psychology, 1970, 15, 1-20.

Weiner, B., Russell, D., Si Lerman, D. Affectiveconsequences of causal ascriptions. In J. H. Har-vey, W. J. Ickes, & R. F. Kidd (Eds.), New di-rections in attribution research (Vol. 2). Hillsdale,N.J.: Erlbaum, 1978.

Weiner, B., Russell, D., & Lerman, D. The cognition-emotion process in achievement-related contexts.Journal of Personality and Social Psychology, 1979,37, 1211-1220.

Wilson, E. O. The genetic evolution of altruism. InL. Wispe (Ed.), Altruism, sympathy, and helping.New York: Academic Press, 1978.

Wispe, L. (Ed.). Altruism, sympathy, and helping.New York: Academic Press, 1978.

Wong, P. T. P., & Weiner, B. When people ask whyquestions and the heuristics of attributional search.Journal of Personality and Social Psychology, inpress.

Received March 26, 1979 •

A Cognitive (Attribution)-Emotion-Action Model of Motivated … · 2017-07-10 · A Cognitive...

Documents

Transcript of A Cognitive (Attribution)-Emotion-Action Model of Motivated … · 2017-07-10 · A Cognitive...