Personal Responsibility and Salience of the Request for ...

Personal Responsibility and Salience of theRequest for Help: Determinants of the RelationBetween Negative Affect and Helping Behavior

Marian Rogers, Norman Miller, F. Stephan Mayer, and Shelley DuvalUniversity of Southern California

Three experiments examined the relation between negative affect and helpingbehavior. The first hypothesis (Experiment 1), that attribution of responsibilityto self for one's experimentally induced depressed mood would induce greaterinclination to offer help when subsequently asked, was tested in a 2 X 2 factorialdesign that manipulated mood (negative vs. neutral) and attribution of respon-sibility for it (internal vs. external). The obtained opposite result seemed attrib-utable to the low salience of the request for help. Experiment 2 replicated Ex-periment 1 using a highly salient request for help and confirmed the initial hy-pothesis. In Experiment 3, a negative mood was induced in all subjects, andattribution of responsibility (internal vs. external) was crossed with salience ofthe helping request in a 2 X 2 factorial design in which the confederate whorequested help was totally dissociated from the experiment and blind to theexperimental conditions. The obtained interaction confirmed the prediction thatinternal attribution of responsibility increases willingness to help (as measuredeither behaviorally or attitudinally) when the request is salient, but inhibits itwhen the request lacks salience. Self-focus, as measured by the Stroop Color-Word Interference Test, was shown to mediate these effects.

Research on the effect of negative affect onprosocial behavior has produced inconsistentresults. When compared to the level of helpoffered when in a neutral affective state, neg-ative affect often decreases a person's will-ingness to help (Moore, Underwood, & Ro-senhan, 1'973; Underwood, Fronting, &Moore, 1977; Underwood, Berenson, et al.1977). Some researchers, however, havefound the opposite relationship (Apsler, 1975;Brock & Becker, 1966; Carlsmith & Gross,1969; Cialdini, Darby, & Vincent, 1973; Fil-ter & Gross, 1975; Freedman, Wallington,

We wish to acknowledge assistance from the SocialScience Research Institute of the University of SouthernCalifornia in the preparation of this manuscript.

Requests for reprints should be sent to NormanMiller, Department of Psychology, University of South-ern California, Los Angeles, California 90007.

& Bless, 1967; Konecni, 1972; McMillen,1971; Rawlings, 1968; Regan, Williams, &Sparling, 1972; Steele, 1975; Wallace & Sad-alla, 1966). The two most recent theoreticalformulations offered to account for this con-tradiction are those of Cialdini, Darby, andVincent (1973) and Thompson, Cowan, andRosenhan (1980).

Cialdini et al. (1973) postulated that a tem-porary state of negative affect may increaseprosocial behavior. Their model argued thatone can reduce one's negative affect by en-gaging in behavior that produces positive af-fect. People typically associate helping withoutcomes that are affectively positive; con-sequently, when one experiences negativeaffect it can be reduced by behaving proso-cially. Working within this model, Cialdiniand Kenrick (1976) proposed that the incon-sistency among the findings reported above

Journal of Personality and Social Psychology, 1982, Vol. 43, No. 5, 956-970Copyright 1982 by the American Psychological Association, Inc. 0022-3514/82/4305-0956S00.75

956

NEGATIVE AFFECT AND HELPING 957

may be caused by differences in the ages andcorresponding levels of socialization of sub-jects who participated in the studies. Specif-ically, as age increases, the likelihood that theperson has been socialized to associate pos-itive outcomes with helping also increases.Consequently, older individuals should bemore likely than younger individuals to en-gage in prosocial behavior in order to reducenegative affect. To test this hypothesis, Cial-dini and Kenrick recruited subjects fromthree age groups (6-8, 10-12, and 15-18);they then induced a negative mood in onehalf of each age group and a neutral moodin the other half. As compensation for par-ticipating in the experiment, all subjects re-ceived coupons that could be exchanged forgifts with values linked directly to the num-ber of coupons. They then informed subjectsthat they could give some of their couponsto children who were unable to participatein the experiment. Although older childrendonated more coupons, there are a numberof problems with the study and its conclu-sions.

First, a small body of studies directly con-tradicts Cialdini and Kenrick's conclusion,showing instead that negative affect reduceshelping in adults (e.g., Underwood et al.,1977) and promotes it in children (e.g., Isen,Horn, & Rosenhan, 1973, Study 2), Thus,as they now acknowledge (Kenrick, Bau-mann, & Cialdini, 1979), their earlier con-clusions are only partially supported by ex-isting data. Second, Cialdini and Kenrick'sdesign implicitly assumes equal subjectivevalue of the coupons among all age groups.It may well be the case that younger childrenvalued them more and, consequent^ werepsychologically donating as much or morethan older children were. Without .informa-tion on their perceived value, the relationshipbetween age and donation rate remains am-biguous. Third, their theorizing also seemsto imply that negative affect should, at leastto some degree, reduce helping behavioramong younger subjects. The fact that youngchildren in the negative-affect condition do-nated as many coupons as did those in thecontrol group thus appears to be inconsistentwith their theoretical analysis. Finally, as oth-ers have suggested (e.g., Thompson et al.,1980) Cialdini and Kenrick's experimental

procedures did not adequately control for thepossibility that the induced mood elicited atype of cognitive activity among older chil-dren that differed from that which it elicitedamong younger children. Older subjects mayhave been more inclined to take personalresponsibility for their negative affective state.Thus, as these researchers now suggest, otherfactors must be responsible for these incon-sistent findings (Cialdini, Baumann, & Ken-rick, 1981).

In contrast to this explanation, Thompsonet al. (1980) proposed that one's focus of at-tention mediates the relationship betweennegative affect and helping. They suggestedthat in studies showing that negative affectincreases prosocial behavior, subjects focusedtheir attention on the misfortunes of others,whereas in those studies that found the op-posite effect, subjects directed their attentioninward. To test this hypothesis, Thompsonet al. asked subjects to identify with the pointof view expressed in one of three audiotapenarratives. There were two negative-affectconditions. In the one in which attention wasdirected to self, the second-person narrative(viz., you) described how the listener wouldfeel if his or her best friend was dying of can-cer, whereas when attention was directed toanother, the third-person narrative (viz., heor she) described the cancer victim's plight.In a neutral-affect control condition, subjectslistened to a second-person narrative aboutmaking a collage. As predicted, when atten-tion was directed to another, subjects helpedmore than did those in both the attention-to-self and the control conditions, which elicitedequivalent levels of help.

Although Thompson et al.'s (1980) resultsmay suggest that focusing attention on a vic-tim whose distress is clearly internal and un-controllable increases prosocial behavior, thisexplanation, too, faces problems. First, theirpredictions and findings are incompatiblewith experiments in which the manipulationof negative affect implicitly appears to in-crease both self-awareness and helping; forinstance, embarrassment (Apsler, 1975), der-ogation (Steele, 1975), and devianey (Freed-man & Doob, 1968; Filter & Gross, 1975)all increased helping. Second, their resultscontradict a pair of recent studies that ex-plicitly manipulate self-awareness and/or self-

958 ROGERS, MILLER, MAYER, AND DUVAL

focus and find that increasing it augmentsrather than reduces helping (e.g., Duval, Du-val, & Neely, 1979; Wegner & Schaefer,1978). Finally, if direction of attentional fo-cus fully mediates the impact of negative af-fect on helping, self-focused subjects mustshow decreased helping relative to the controlgroup. Their results show no difference onthis critical comparison.

A recent study by Maruyama, Fraser,.andMiller (1982) points to the importance ofpersonal responsibility as a factor affectinghelping behavior. When children were iden-tified as personally responsible for the num-ber of candies to be donated to other childrenwho were hospitalized on Halloween night,they donated more candy. This outcomealerted us to the possibility that differencesin perceived responsibility might account fordisparities among the experimental findings.A' review of the entire literature on the re-lation between negative affect and helpingappeared to implicate this variable as asource of inconsistency among studies. Spe-cifically, in studies that show increased help-ing as a function of negative affect, the ex-perimental procedures used to induce it ei-ther explicitly or implicitly seem likely tohave led subjects to feel responsible for theirnegative affect or the event that generated it,whereas those that showed decreased helpingdid not.

To illustrate, in some studies experiment-ers induced negative affect by leading sub-jects "accidentally" to break expensiveequipment (Regan, Williams, & Sparling,1972) or experimental apparatus (Brock &Becker, 1966; Wallace- & Sadalla, 1966); toupset a deck of index cards (Freedman, Wal-lington, & Bless, 1967); to cause someoneelse to drop punched computer cards (Ko-necni, 1972) or to lose experimental credit(Darlington & Macker, 1966); to delivershocks to confederates (Carlsmith & Gross,1969); or to lie about their previous knowl-edge of the experiment (McMillen, 1971).These procedures seem likely to have madesubjects feel responsible for the negativeevent and/or their negative affect. Subse-quently, in comparison to those in controlconditions, subjects were more willing tohelp either by participating in another ex-

periment, signing a petition, grading exams,donating blood, or rectifying the negativeconsequences of the "accident." This direc-tion of effect occurred regardless of whetherthe potential recipient of help was the injuredparty or a nonvictim.

In other studies, however, the experimen-tal procedures seemed likely to promote anexternal rather than an internal attributionfor the experienced negative affect. For ex-ample, in a field study showing that saddenedsubjects helped less than control subjects(Underwood, Berenson, et al., 1977), nega-tive affect was presumably induced whensubjects viewed a depressing film (e.g., LadySings the Blues). It seems likely that the view-ers saw the film as responsible for their emo-tional state. Similarly, other manipulationsthat explicitly instruct subjects to feel sad bygiving them negative mood statements toread (Aderman, 1972) or by telling them toreminisce about events that have previouslymade them sad (Moore, Underwood, & Ro-senhan, 1973; Rosenhan, Underwood, &Moore, 1974; Underwood, Froming, &Moore, 1977) may also have led to externalattributions for the negative affect (i.e., sub-jects blamed the statements or the experi-menter who made them read them). In eachof these studies negative affect either reducedor did not affect subjects' willingness to help.

In opposition to this analysis, a criticmight claim that some studies can be inter-preted as arguing against the notion that asense of responsibility affects a person's will-ingness to help. For example, Rawlings(1968), as well as Aderman and Berkowitz(1970), found that subjects who merely ob-served a negative incident (a peer receivingelectric shocks or a distressed other not re-ceiving aid from a potential helper) behavedprosocially; other researchers report similarfindings (Cialdini, Darby, & Vincent, 1973;Konecni, 1972; Regan, 1971). On the onehand, it seems reasonable that those whomerely observe might sometimes feel a weakersense of personal responsibility than that feltby those who ostensibly caused the negativeevent. On the other hand, observers mightnevertheless feel responsible for the outcomeof a negative event (e.g. a distressed personnot receiving aid from a potential helper) and


for their own bad feeling with respect to itbecause they failed to take any actions thatmight alleviate the situation.

To assess this issue properly, it is necessaryto use a control comparison condition thatstructurally precludes any possibility that theobserver can help. Unfortunately, no studiesin this literature contain such a comparison.Furthermore, none of the studies on observ-ers include among their dependent variablesan ancillary measure of felt responsibilitythat might confirm that observers do not feelresponsible whereas perpetrators do. Indeed,in support of our position, research showsthat factors that lessen an observer's sense ofpersonal responsibility (e.g. group size) alsolessen the observer's tendency to behave pro-socially (Latane & Nida, 1981). Therefore,we do not interpret the observer studies asseriously impugning a role for responsibilityas a variable that affects helping behavior.

Accordingly, we hypothesized that a feel-ing of personal responsibility for one's neg-ative affect increases prosocial behavior,whereas an external attribution of responsi-bility for one's negative affect decreases it.This view argues that making internal attri-butions of responsibility for events that havenegative consequences causes people to feelmore aware of themselves and that thisheightened self-awareness, in turn, increasesprosocial behavior rather than causing it todecrease, as argued by Thompson et al.(1980). To test this notion, we instructed sub-jects to read mood statements designed toinduce either negative or neutral affect. Sub-sequently, we made them believe either thatthey were personally responsible for theiremotional state or, instead, that an externalcause (the mood statements) induced theirfeelings. We then asked them to help a thirdparty, expecting that those made to feel re-sponsible for their negative affect would dis-play a stronger level of self-awareness andgreater willingness to help than those inducedto externalize responsibility for it. For thosein the neutral mood conditions, we did notexpect the manipulation of responsibility todifferentially affect their self-awareness ortheir inclination to help. Furthermore, weexpected these latter groups to express thesame level of self-awareness and inclination

to help as did those led to externalize re-sponsibility for their negative affect.

Experiment 1

Method

SubjectsFemale undergraduates (N = 40) participated in par-

tial fulfillment of requirements for an introductory psy-chology course. Each session lasted approximately 30minutes.

DesignThe factorial design crossed two levels of affective

state, negative and neutral, with two conditions of re-sponsibility for it, internal and external.

ProcedureSubjects were individually recruited for an experiment

entitled "Autosuggestion and Task Performance." Aseach subject arrived, she was greeted by a female ex-perimenter who briefly explained that the experimentwas designed to clarify the hitherto perplexing relation-ship between an individual's mood and her performanceon certain tasks.

Premeasure. To premeasure mood, each subject re-sponded to a word-association task during which herspontaneous associations to eight words were unobtru-sively taped. The premeasure was adapted from a mea-sure used by Velten (1967), which showed that subjectsexperiencing negative affect had a longer total associa-tion time than did subjects with neutral affect. Eightwords were selected from Velten's list of 16: paper,wood, life, music, love, parent, hate, and school.

Mood induction. To induce either a negative or neu-tral mood, subjects read 45 mood statements (Velten,1968), each typed on a 4 X 6 index card that the ex-perimenter handed to her in a slow, methodical manner.The negative statements became progressively more de-pressing (Card 1, "Today is neither better nor worse thanany other day."; Card 45, "I want to go to sleep andnever wake up."). The neutral statements consisted offactual, sentences (e.g., "The man is dressed in red.").To increase the impact of the treatment, subjects in thenegative-affect condition were instructed to reread thelast 15 statements.

Mood postmeasure. Immediately after the mood in-duction, subjects completed the Depression scale of theMultiple Affect Adjective Checklist (MAACL; Zucker-man, Lubin, & Robins, 1965).

Responsibility manipulation. To prepare the subjectsfor the manipulation of responsibility for their affectivestate, they received information concerning their re-sponses on the MAACL. In consonance with their as-signed experimental condition, they were informed thattheir responses revealed that they were feeling neitherparticularly good nor bad or that they were feeling de-pressed. To induce an internal attribution of responsi-


bility, they were then told that because their emotionswere largely a matter of their own responsibility, theyshould assume complete responsibility for their presentmood state. To induce an external attribution, they weretold that their present state was caused by the statementsthey had read and that they should not assume any re-sponsibility for the way they felt.

Measure of self awareness. We expected our proce-dure for inducing an internal attribution of responsibilityfor their negative,affect to cause subjects to focus theirattention on themselves. To assess this possibility, weadministered a modified version of the Stroop Color-Word Interference Test (Siegel, Note 1). Previous use ofa similar modification (Geller & Shaver, 1976) showedthat subjects exposed to procedures designed to induceself-awareness exhibited longer response latencies forself-relevant words than for non-self-relevant words,whereas those exposed to procedures designed to min-imize self-awareness exhibited no differences in theirresponse times to the two word sets. In the present ex-periment, subjects viewed slides of cards, measuring 30cm X 50 cm, on which five self-relevant words (me,mine, I, myself, self) and five non-self-relevant words(up, theory, a, week, tool) were printed. The words ineach set are matched for frequency of usage, length, andnumber of syllables. Each word was presented sepa-rately. Colors were randomly assigned to words withineach set of five and were counterbalanced between thetwo sets. With a concealed tape recorder, the experi-menter recorded the subjects' response latency for iden-tifying the color of each word.

Additional postmeasure of mood. Experimentallydepressed subjects write significantly fewer numbers in1 minute than do normal subjects (Velten, 1967), There-fore, a writing speed task was used as an additionalmanipulation check of mood. Subjects were given apaper and pencil and were asked to help prepare for alater task by writing the numbers from 100 to 1 in de-scending order. After unobtrusively timing the subjectfor 1 minute, the experimenter interrupted to informher that a sufficient quantity of numbers had beenwritten.

Dependent measures. Before leaving the laboratoryfor the expressed purpose of preparing materials for thesecond part of the experiment, the experimenter askedsubjects to fill out a questionnaire that had arrived toolate to be included in the survey packet that each studentin introductory psychology ,had completed during thefirst class session. It included a brief description of acampus agency that offered crisis intervention servicesto students and stated that volunteer workers wereneeded to help meet their needs. The plea for help statedthat because of cyclical escalations in the frequency ofcallers, the agency needed students to volunteer duringcertain peak weeks. An attitudinal measure of helpingasked students how willing they were to help their fellowstudents through the intervention services. A measureof behavioral intention to help asked students the num-ber of hours they were willing to volunteer during a 1-week period. Although ̂ he instructions on the question-naire asked them to provide their names and telephonenumbers, subjects were assured that only the campusagency would be aware of their responses. To increasetheir perception of confidentiality, subjects were in-

structed to place their completed questionnaires in amanila envelope partially filled with other question-naires.

Restoration of positive affect and debriefing. After thesubject responded to these written measures, the exper-imenter announced that unforeseen: scheduling prob-lems required that the experiment be prematurely con-cluded. To restore subjects to their preexperimental•state, those who had been exposed to negative-moodstatements read, both silently and aloud, 20 statementsdesigned to induce positive affect. To examine whetheror not hypothesis guessing or suspiciousness occurred,all were asked to briefly describe in a written statementany ideas they had concerning the experiment. Subjectswere thoroughly debriefed by phone during the followingsemester.

Results

Premeasure

A 2 X 2 analysis of variance (ANOVA) ofthe word association task revealed that sub-jects in the four treatment conditions exhib-ited no initial differences in their preexperi-mental mood.

Manipulation Check of Mood

To remove potential bias resulting fromnonsignificant pretreatment differences in in-dividual mood states, the scores from theword-association task were used as a covar-iate in the analysis of responses to theMAACL. The highly significant main effectof affective state, F(\, 39) = \02.Q,p < .001,confirms an effective manipulation of mood(see Table 1); the induction of negative affectcaused subjects to describe themselves as feel-ing more depressed than did those who readthe neutral statements,

Measure of Self-Awareness

The cell variances associated with the re-sponse latencies to the modified Stroop Color-Word Interference Test did not meet Har-tley's test of homogeneity (Winer, 1971). Cellmeans and variances tended to be correlated,and the response latencies were positivelyskewed. Therefore, a logarithmic transfor-mation was applied to raw latencies. Inaddition, before we analyzed the responsemeasure, we subtracted transformed re-sponse latencies associated with the five non-self-relevant words from those obtained with


the five self-relevant words. Positive scoreson this difference-score index indicate thatsubjects had longer latencies for self-relevantwords than for non-self-relevant words.

The main effect for affective state, F(l,39) = 9.5, p < .004, obtained in the analysisof the Stroop difference scores revealed thatnegative affect produced longer latencies forself-relevant words than did neutral affectand that internal responsibility inducedgreater self-focus than did external respon-sibility. The interaction between the inde-pendent variables, F(l, 39) = 4.39,p < .043,as anticipated, shows that the main effectsfor affective state and responsibility are pri-marily explained by the high level of self?awareness induced when subjects were madeto feel responsible for their negative affect(see Table 1).

Postmeasure of Mood

The main effect for affective state obtainedin the analysis of the writing-speed task, F(\,39) = 5.4, p < .05, confirms that, as expected,subjects with neutral affect wrote more num-

bers during the 1-minute period than didthose with negative affect (see Table 1).

Dependent Measures of Helping

The attitudinal and behavioral measuresof helping were highly correlated (r = .64,p < .0001), the pattern of cell means for eachmeasure was identical, and the analysis of thetwo measures resulted in a similar interactionbetween affect and locus of responsibility.Accordingly, standardized responses on thetwo indicators were summed to provide anoverall index of helping. However, becauseresponses on the behavioral measure werepositively skewed and contained a modestnumber of extreme outliers, a logarithmictransformation was performed before stan-dardizing them and combining them withresponses to the attitudinal measure (Tukey,1979). The obtained interaction, F(l,39) = 4.78, p < .035, indicates that subjectswho were made to feel responsible for theirnegative affect were least inclined to help,whereas those in the three other conditionswere equally inclined to help (see Table 1).

Table 1Experiment 1: Manipulation of Mood and Responsibility

Responsibility

Affect

Negative

M SD

Neutral

M SD

MAACLInternalExternal

StroopInternalExternal .

/Number production

InternalExternal

HelpingInternalExternal

31.1.'32.4.

1.9.6

39.0.38.3.

2.76.5.2b

4.96.6

1.3.57

5.64.1

1.71.2

13.9b14.0b

• lb~.'b

44.0b

46.0b

5.04b

4.77b

5.25.5

.70

.25

5.65.3

1.52.7

Note. Means and standard deviations are based on summed raw scores. As indicated in the text, statistical analysesof the Stroop and the helping measures were performed on transformed scores. Larger numbers indicate greaterdepression, self-awareness, number production, and willingness to help. Within each measure, means without acommon subscript differ from each other. MAACL = Multiple Affect Adjective Checklist, Stroop = Stroop Color-Word Interference Test.


Discussion

We expected that subjects who were madeto feel responsible for their negative affectwould help more than those made to exter-nalize responsibility for it. The results of Ex-periment 1 directly contradicted our predic-tions. This is particularly puzzling because,as we expected, subjects who were inducedto internalize responsibility for their negativeaffect exhibited greater self-awareness thanthose in the other three conditions, and somestudies show that raising self-awareness in-creases helping behavior.

In retrospect, it seemed likely that this un-expected result could be explained by ex-amining the influence of self-awareness onsubjects' awareness of the request or need forhelp. Latane and Darley (1970) and Bar-Tal(1976) posit that the need for help must benoticed before a person will offer it. In ad-dition, Berkowitz (1972) suggests that exces-sive self-preoccupation can interfere withone's awareness of others' needs. Indeed, arecent study shows that subjects who expe-rienced a temporary loss in self-esteem, anexperience that might focus attention in-ward, were more likely to help a confederatewhen she explicitly drew attention to herdistress than when the need for help was notsalient (McMillen, Sanders, & Solomon,1977). Thus, if a potential helper is distractedby self-concerns and the request for helplacks salience or impact, it may pass unno-ticed or may be denned as a situation thatis not serious enough to necessitate personalinvolvement.

The request for help used in Experiment1 certainly lacked salience. It was plainlytyped in black ink on standard white typingpaper and the experimenter deliberatelydownplayed its significance relative to theprimary purpose of the experiment. In ad-dition, because subjects were left alone toread and respond to it, those who were self-focused could readily shift their attentionaway from it and back to themselves. In sum,it seems possible that the self-involvement ofthose induced to feel responsible for theirdepressed state may have made them lessattentive to the request for help, and conse-quently they helped less. This line of reason-ing suggests that the salience and vividness

of the request for help is another importantvariable in predicting the effect of negativeaffect on helping behavior.

Experiment 2

In the second experiment we strongly in-creased the salience of the request for helpand reexamined the relationship betweennegative affect, attribution of responsibility,self-awareness, and helping behavior. As ar-gued above, in Experiment 1, the heightenedself-awareness of subjects in the internal neg-ative-affect condition may have disruptedtheir attention from the request for help. In-creasing its salience should serve to momen-tarily redirect self-focused attention and,thus, allow subjects to respond to it. There-fore, in Experiment 2, we again predictedthat those induced to take responsibility fortheir negative affect would be more willingto help than either those induced to exter-nalize the cause of their negative affect orthose in the neutral-affect conditions.

Method

Subjects

Subjects were 48 female undergraduates who partic-ipated in the study in partial fulfillment of a course re-quirement.

Design

Similar to Experiment 1, two levels of affective state,negative and neutral, were crossed with two conditionsof responsibility for affective state, internal and external,in a 2 X 2 factorial design.

In addition, a control condition was added to provideinformation on the comparability of subjects recruitedfor Experiments 1 and 2 and to help assess the impactof changing the format of the helping request. Eight sub-jects were randomly assigned to the affectively neutralinternal-responsibility condition and exposed to thesame low-salience request for help used in Experi-ment 1.

Procedure

The procedure exactly paralleled that used in Exper-iment 1 except that the request for help was made highlysalient for subjects in all groups other than the low-sa-lience control group. Salience was heightened by break-ing the information contained in the request into threeparagraphs, each typed in a different type, and by circlingthe request for help with asterisks and exclamationpoints and underlining it.


Results

Premeasure

As in Experiment 1, subjects in the fourtreatment conditions exhibited no initial dif-ferences in their preexperimental mood states.A comparison of the pretreatment mood ofsubjects in the helping-request control con-dition with that of subjects in the parallelexperimental condition (internal responsi-bility, neutral affect) also showed no differ-ence in mood, t(l, 16) = .16, ns.

Manipulation Check of Mood

As in Experiment 1, the scores from theword-association task were used as a covar-iate in the analysis of responses to theMAACL. As expected, the main effect foraffective state, F(l, 35) = 213.5, p < .0001,showed that affectively neutral subjects de-scribed themselves as less depressed than didthose induced to have negative affect. A com-parison between subjects in the helping-re-quest control condition and the parallel ex-perimental condition (internal responsibility,neutral affect) showed no difference in moodstate, t(l, 16) = .57, ns (see Table 2).

Measure of Self-Awareness

As in Experiment 1, a logarithmic trans-formation was applied to the raw responselatencies of the modified Stroop Color-WordInterference Test to eliminate differences inthe cell variances. The results on this mea-sure parallel those of Experiment 1. Themain effects revealed that subjects inducedto internalize responsibility for their affectwere more self-focused than those led to ex-ternalize it, F(l, 34) = 11.9, p < .001, andthat negative affect produced higher levels ofself-awareness than did neutral affect, F(l,36) = 13.07, p < .001. Again, the interactionbetween the independent variables showedthat these main effects were attributable tothe high self-focus of those made to feel re-sponsible for their negative affect, F(\,36) = 17.26, p < .001. Planned comparisonsrevealed that subjects in the control condi-tion differed only from those in this lattercondition f(l, 16) = 3.97, p < .001; they ex-hibited the same relatively low level of self-awareness as did those in the other three con-ditions—internal neutral, t(l, 16) = .28, ns;external negative, t(l, 16) = 1.07, ns; externalneutral, t(\, 16) = .34, ns (see Table 2).

Table 2Experiment 2: Manipulation of Mood and Responsibility

Responsibility

Experimental groups

Negative affect Neutral affect

M SD M SD

Low-salience controlgroup

Neutral affect

M SD


StroopInternalExternal

Number productionInternalExternal


33.1,31.1,

2.20a

-.25b

41.5,40.2a

5.79a

3.66b

4.64.7

1.4.81

3.36.3

1.61.7

11. 9b13.3b

^.20b-.03b

44.6b47.9b

3.53b4.30b

3.84.1

.77

.82

5.09.9

2.21.9

11. lb

.10b

46.2b

3.02b

3.1

.81

4.9

1.6

Note. Means and standard deviations are based on summed raw scores. As indicated in the text, statistical analysesof the Stroop and the helping measures were performed on transformed scores. Larger numbers indicate greaterdepression, self-awareness, number production, and willingness to help. Within each measure, means without acommon subscript differ from each other. MAACL = Multiple Affect Adjective Checklist, Stroop = Stroop Color-Word Interference Test.


Postmeasure of Mood

Again, on the postmeasure of mood, themain effect obtained for affective state, F(l,36) = 6.6, p < .01, showed that subjects withnegative affect wrote significantly fewer wordsduring the 1-minute interval than did thosewith neutral affect. Planned comparisonsshowed that whereas the writing speed ofthose in the control condition was faster thanthat of both groups with negative affect—in-ternal, /(I, 16) = 1.72, p < .05; external, t(l,16) = 2.17, p < .05—the writing speed didnot differ from that of those in the two neu-tral-affect conditions—internal, t(l, 16) =.17, ns; external, t ( l , 16) = 1.04, ns.

Dependent Measure of Helping

As in Experiment 1, the attitudinal andbehavioral measures of helping were stronglycorrelated (r = .70, p < .001), the pattern ofcell means for each measure was identical,and separate analyses of each showed parallelinteractions between affective state and locusof responsibility. Therefore, the two mea-sures were again combined to form a generalindicator of helping. As in Experiment 1, thetreatment means and variances of the be-havioral measure were correlated and the celldistributions were asymmetric and positivelyskewed. Therefore, before we combined thevariables, we logarithmically transformed thebehavioral measure.

Increasing the salience of the helping re-quest produced the anticipated interaction,F(\, 36) = 6.86, p<.Ql3. Those made tofeel responsible for their affect expressed agreater interest in helping than did those inthe other three experimental conditions; thelatter groups did not differ in their inclinationto help. Finally, planned comparisons showedthat subjects in the low-salience control con-dition did not differ from neutral subjects intheir willingness to help: internal, t(l,16) = .20, ns; external, t(l, 16)= 1.56, ns;nor did they differ from those in the externalnegative treatment condition, t(l, 16) = .83,ns. These subjects were, however, less in-clined to help than were those in the internalnegative treatment condition, t(l, 16) = 3.20,p < .005, (see Table 2).

Discussion

Unexpectedly, in Experiment 1, subjectsmade to feel responsible for their negativeaffect helped less than did either those led toexternalize responsibility for their negativeaffect or those with neutral affect. In explain-ing this outcome, we suspected^ that self-awareness obstructs the processing of the in-formation contained in the request for help.Therefore, in Experiment 2, the format ofthe request for help was altered to make itmore vivid and salient. Again, we predictedthat when made to feel internally responsiblefor their negative affect, subjects would bemore inclined to help. The results of Exper-iment 2 confirmed these predictions. Thosemade to feel internally responsible for theirnegative affect not only exhibited greater self-awareness than other subjects, but more im-portant and in contrast to Experiment 1, dis-played the greatest willingness to help dis-tressed students.

The results of Experiments 1 and 2 to-gether suggest that the inconsistency amongpreviously published findings reflects varia-tion both in the salience of the request forhelp and in subjects' feelings of personalresponsibility for their negative affect.Although the results clearly support this in-terpretation, the alternative theoretical ex-planations previously reviewed warrant dis-cussion. The negative-state relief interpreta-tion might imply that externalizing re-sponsibility alleviates negative affect whereasinternalizing it heightens negative feelings;thus, when exposed to a salient request forhelp, subjects who are made to feel personallyresponsible for their negative affect helpmore in order to reduce their higher level ofnegative affect. The results of Experiments1 and 2 show that this is not a viable inter-pretation. Its plausibility requires that sub-jects who are led to internalize responsibilityfor their affect exhibit greater depression(more negative feeling) than their counter-parts who externalize responsibility. Al-though the analysis of the measure of moodtaken after the responsibility manipulationrevealed strong differences in the mood levelsof subjects who received a negative or neutraltreatment, thereby confirming that it is a sen-


sitive measure, those induced either to inter-nalize or externalize responsibility for theirnegative affect did not differ. The results alsocounter the Thompson et al. (1980) thesis,which postulates that an external focus en-hances helping, whereas self-focus reduces it.Experiment 2 produced the opposite effect;when exposed to an impactful helping re-quest, subjects in the condition that pro-duced the highest level of self-awarenessshowed the greatest willingness to help.

The designs of our two experiments, how-ever, necessitate that the results be inter-preted cautiously. As described in the Methodsection, the experimenter who presented therequest for help was not blind to the subjects'condition. Although the fact that the resultsof Experiment 1 were antithetical to our ownpredictions would seem to suggest otherwise,the experimenter may have unintentionallypresented this material in a biased manner.A second problem concerns the salience ofthe request for help. It was not initially an-ticipated that this variable would affect thewillingness of self-focused subjects to help,and as is apparent, it was manipulated across,not within, the two experiments. Althoughamong the six measures common to bothexperiments, the index of helping was theonly one on which the comparable cells ofthe two experiments differed,1 it neverthelessis conceivable that salience of the request forhelp was not critical. Instead, some chancefactor may have produced the between-ex-periment differences on the helping measure.Experiment 3 was undertaken to rule outthese residual sources of interpretive ambi-guity.

Experiment 3

The third experiment explicitly tests theprediction that when a request for help isimpactful, subjects made to feel responsiblefor their negative affect will be more inclinedto help others than will those led to exter-nalize responsibility for it. To do so, it di-rectly manipulates the salience of the requestfor help. If the manipulation of salienceyields the predicted results, the possibilitythat the differences between the results ofExperiments 1 and 2 were produced by fac-

tors other than salience is virtually elimi-nated. Experiment 3 also eliminates otheralternative explanations for the results of thefirst two experiments by having a confederatewho was blind to the subjects' experimentalcondition present the helping request, therebyruling out the possibility that experimenterbias or demand might affect responses to therequest for help.

Method

Design

In Experiment 3, negative affect was induced in allsubjects. The 2 X 2 factorial design crossed two levelsof responsibility for negative affect, internal and exter-nal, with two levels of salience of the helping request,high and low.

Subjects <

Subjects were 60 female undergraduates who partic-ipated in the study in partial fulfillment of the require-ments of an introductory psychology class.

Procedure

Premeasure of mood. After the female experimenterdelivered the identical cover story used in Experiment1, subjects responded to the Depression scale of theMAACL.

Mood induction. Using the procedures described inExperiments 1 and 2, we then induced all subjects toexperience a negative mood.

Responsibility manipulation. After reading the moodstatements, subjects were handed a form entitled "Sub-ject Release Form" and instructed to read and sign it.The form consisted of a statement designed to makesubjects feel that either they or the mood-induction taskwas responsible for their affective state. Subjects in theexternal responsibility condition received the followingstatement: "You are not responsible for your presentmood state. It is entirely due to reading the mood state-ments." The statement given to subjects in the internalresponsibility condition stated, "Because a person'semotions are largely a matter of their own responsibilityyou are primarily responsible for your present moodstate."

1 Comparison between the mean responses on thehelping index by subjects in the neutral control groupof Experiment 2 and its counterpart in Experiment 1shows that the former expressed less willingness to helpthan the latter. Statements obtained during the debrief-ings in Experiment 2 suggest that this overall decreasein helping occurred because Experiment 2 was con-ducted immediately before students' final exams, whereasExperiment 1 was conducted earlier in the semesterwhen subjects felt less pressure for their time.


Postmeasure of mood. As in Experiments 1 and 2,a writing-speed task was then administered to providea postmeasure of affective state.

Manipulation check of responsibility. On a 7-pointscale subjects indicated how responsible they felt fortheir present mood state.

Termination of laboratory procedures and exposureto helping request. At this point the experimenter an-nounced that she was behind schedule and, thus, forcedto prematurely conclude the experiment. As she thankedeach subject for participating in the study, she guidedher to the building's exit. After the subject walked ap-proximately 200 feet away from the building a femaleconfederate, blind to the subject's experimental condi-tion, approached her. She introduced herself as a rep-resentative of the campus radio station and asked thesubject to take a few moments to read some literatureconcerning it. The material stated that as a consequenceof inadequate financial support, the station was forcedto discontinue broadcasting; it needed $23,999 to builda new studio and update its equipment. The helpingrequest solicited volunteers to telephone interested alumniand local businessmen and ask them for donations. Forthose subjects exposed to a highly salient request, theprinted material was presented in six paragraphs, anddifferent typefaces were used for each. In addition, thespecific sentence containing the request for help wascircled in black ink and surrounded by asterisks andexclamation points. Those exposed to the request de-signed to lack salience read a plainly typed single-spacedparagraph that contained the same information.

Helping measures. The page following the materialconcerning the radio station contained an identical setof questions for all subjects. A behavioral measure askedhow many phone calls they would agree to make. (In-formation contained in the request for help had indi-cated that the representative could supply subjects witha maximum of 10 phone numbers.) An attitudinal mea-sure asked subjects to indicate on a 7-point scale howwilling they were to help the radio station.

Additional measures. All subjects then responded totwo additional questions that measured, both generallyand specifically, the extent to which the salience of therequest for help influenced their perceptions of howmuch help the radio station needed; on two 7-pointscales, they indicated how much financial assistance thestation needed and how much assistance, in general, itneeded.

Results

Premeasure of Mood

Analysis of the responses to the MAACLshowed no differences in subjects' pretreat-ment mood level.

Postmeasure of Mood

As expected, the treatments did not dif-ferentially influence subjects' writing speed.In other words, the responsibility manipu-

lation did not differentially affect the moodlevel of subjects in the four treatment con-ditions (see Table 3).

Manipulation Check of Responsibility

The main effect for responsibility, F(\,56) = 19.3, p < .001, found in the analysisof subjects' responses to the measure de-signed to assess their perception of the locusof responsibility for their affective state con-firms an effective manipulation of responsi-bility (see Table 3).2

Helping Measures

Analyses of both the attitudinal and be-havioral measures of helping produced astrong interaction between responsibility andsalience of the request for help. Because thepattern of means for both measures wereagain similar, the two response scales werestandardized and summed to provide a gen-eral index of willingness to help. The analysisrevealed main effects for salience, F(l,56) = 6.8, p < .01, and responsibility, F(l,56) = 5.13, p < .027. The direction of dif-ferences between the means showed that asalient request and internal responsibility forone's negative affect produced more willing-ness to help than did the respective compar-ison conditions. More important, however,the interaction between responsibility andsalience, F(l, 56) = 29.4, p < .001, showedthat these main effects were primarily due tothe greater inclination to help exhibited bythose made to feel responsible for their neg-ative affect and then exposed to the salientrequest for help. Additional analysis revealed

2 Experiments 1 and 2 previously showed that ourmanipulation of responsibility affected self-awareness.However, because two experimenters were used to con-duct Experiment 3, the procedures for manipulating re-sponsibility were slightly modified in order to reducepotential experimenter variability (see Method section,Experiment 3). Given this change, we thought it im-portant to assess the effectiveness of the new procedure.Therefore 14 subjects were asked to respond to the mod-ified Stroop Color-Word Interference Test after readingstatements designed to produce perceptions of either in-ternal or external responsibility. Analysis of the trans-formed response latencies confirmed that those led tofeel personally responsible for their negative affect hadlonger response latencies than those led to externalizeresponsibility.


Table 3Experiment 3: Manipulation of Salience of the Helping Request and Responsibility

Salience

Responsibility

High

M SD

Low

M SD


Number productionInternalExternal


Perceived need for helpInternalExternal

Perceived responsibilityInternalExternal

13.1.13.6a

45.1.42.1,

11.67.4.20b

4.92,4.73,

5.2a

2.1b

6.15.7

4.95.3

4.12.9

2.22.3

1.11.0

14.2a

13.9.

44.8a46.1 a

3.87b6.93C

4.65a5.01a

5.6a

2.5b

5.46.2

5.94.7

2.83.4

1.71.9

.91.4

Note. Means and standard deviations are based on summed raw scores. Larger numbers indicate greater depression,number production, willingness to help, perceived need for help, and personal responsibility. Within each measure,means without a common subscript differ from each other. MAACL = Multiple Affect Adjective Checklist, Stroop =Stroop Color-Word Interference Test.

that these subjects helped more than thosein the other treatment conditions. As shownin Table 3, when exposed to a highly salientrequest for help, those made to feel respon-sible for their negative affect were more will-ing to help than were those led to externalizeresponsibility for it. On the other hand,among subjects exposed to a request thatlacked salience, those made to feel respon-sible for their negative affect were less willingto help than were those led to externalizeresponsibility for it.

Measure of Perceived Need

No differences occurred in subjects' per-ceptions of the amount of help needed by theradio station (see Table 3).

General Discussion

The results of all three experiments con-sistently revealed that internal attributions ofresponsibility for negative affect heightenself-awareness. The results of Experiment 1

suggest, however, that when the helping re-quest lacks impact, being made to internalizeresponsibility for negative affect decreasesthe tendency to help. When, instead, the re-quest for help is made salient, as was the casein Experiment 2, internalizing responsibilityfor negative affect increases the willingnessto help. The results of Experiment 3 sub-stantiate these findings. Those made to feelresponsible for their negative affect and thenexposed to an impactful helping request weremore willing to help than those led to exter-nalize their responsibility but otherwise ex-posed to similar experimental manipula-tions. A less salient request for help producedthe opposite relationship between focus ofresponsibility and willingness to help. Theseresults confirm our explanation of the un-expected results from Experiment 1, namely,that the self-awareness caused by internaliz-ing responsibility for negative affect obstructsawareness of or responsiveness to the needfor help.

Although the results of the three experi-ments greatly clarify the relations among


negative affect, feelings of personal respon-sibility, self-awareness, salience of the requestfor help, and one's willingness to offer it, twoaspects of the results of Experiment 3 warrantfurther comment. First, among those led toexternalize their negative affect, more helpwas offered by those exposed to the requestthat lacked salience. The most likely expla-nation of this difference points to greater re-actance elicited in those who received thehighly salient request for help. Whereas theprocedures of Experiments 1 and 2 were un-likely to produce reactance in that the requestfor help was read in private by the subjects,this was not the case in Experiment 3. There,the confederate directly handed the helpingrequest to the subject and waited for her torespond to it. Presumably, when the requestwas made highly salient the confederate ap-peared to be demanding help, whereas theless salient request was not viewed in thismanner. Reactance was not elicited amongthose led to view themselves as responsiblefor their bad feelings presumably becausetheir self-focused attention precluded its elic-itation.3

A second puzzling aspect of the data fromExperiment 3 concerns the measure of per-ceived need for help. On first thought, onemight expect a more salient request for helpto produce perceptions of greater need forhelp. Clearly, as seen from the data on per-ceived need for help, this was not the case.This outcome, however, should not be in-terpreted as questioning the effectiveness ofthe salience manipulation; other effects con-firm its validity. Instead, it argues that thesalience manipulation did not interfere withthe encoding of information. When explicitlyasked, those in the high-salience and low-sa-lience conditions could retrieve the relevantinformation equally well, and consequently,their judgments concerning the need for helpwere also equal. However, when subjectswere led to attribute responsibility for theirnegative affect to themselves, and thus, wereself-focused, they ignored this informationeven though it had been encoded. The factthat the measures of perceived need for helpwere administered after those that assessedwillingness to help adds further confirmationto this interpretation.

Whereas we initially viewed the theoretical

formulations of both Cialdini et al. (1973)and Thompson et al. (1980) as lacking, ourresults suggest that both are at least partiallycorrect. The salience of the request for helpdetermines when the Thompson et al. viewthat self-focused attention decreases helpingwill indeed be correct. On the other hand,the fact that a highly salient request for helpaugmented a willingness to help among thoseled to feel responsible for their negative af-fect, whereas it produced reactance and adisinclination to help among those led to ex-ternalize responsibility for their negative af-fect, seems to fit comfortably with Cialdiniet al.'s (1973) negative-state relief formula-tion. Apparently, when one feels personallyresponsible for one's negative affective state,a forceful, demanding request for help doesnot produce reactance but instead readilyelicits responsiveness to the request, presum-ably because it reduces or eliminates the neg-ative affect state.

The final point that bears consideration isour reinterpretation of the literature on therelation between negative affect and helping.We initially interpreted its inconsistency asdue to unintended between-experiment vari-ation in the extent to which subjects were ledto view themselves, as opposed to externalfactors, as responsible for their bad feelings.Our results, however, imply that in previousresearch the salience of the request for helpmust also have varied between experiments,and in addition, that it covaried with attri-butions of responsibility; procedures that in-duced internal attributions must have alsocontained highly salient requests for help.Although on first thought such natural co-variation seems unlikely, inspection of theexperimental procedures suggests that it oc-curs. For example, in studies employing in-duction techniques that we believe led sub-jects to externalize responsibility, the exper-imenter or a confederate typically deliveredthe request for help and then left the subject

3 Although the comparison of the respective condi-tions of Experiments 1 and 2, in which an external at-tribution of negative affect was induced, appears to pro-vide a comparable comparison in that the salience of therequest for help varied between the two experiments, itsusefulness is only illusory. The fact that the overall levelsof help differed in the two studies precludes any mean-ingful comparison.


alone to respond to it (e.g., Moore, Under-wood, & Rosenhan, 1973; Underwood,Froming, & Moore, 1977), whereas in thesecond set of studies, experimenters or con-federates generally delivered the request forhelp and attentively waited for subjects torespond to it (e.g., Filter & Gross, 1975; Re-gan, Williams, & Sparling, 1972). Whatmight produce such consistent covariation?Differences in experimenters' style is onepossibility: Some prefer a more active ap-proach, others a more passive one. Alterna-tively (or in combination with stylistic dif-ferences), some procedures may naturally fittogether better than others.

Reference Note

1. Siegel, K. N. Focus of attention as a function of thequantification of social dimensions,' Unpublishedmasters thesis, University of Southern California,1977.

ReferencesAderman, D. Elation, depression, and helping behavior.

Journal of Personality and Social Psychology, 1972,24, 91-101.

Aderman, D., & Berkowitz, L. Observational set, em-pathy, and helping. Journal of Personality and SocialPsychology, 1970, 14, 141-148.

Apsler, R. Effects of embarrassment on behavior towardothers. Journal of Personality and Social Psychology,1975, 32, 145-153.

Bar-Tal, D. Prosocial behavior: Theory and research.New York: Halsted, 1976.

Berkowitz, L. Social norms, feelings, and other factorsaffecting helping and altruism. In L. Berkowitz (Ed.),Advances in experimental social psychology (Vol. 6).New York: Academic Press, 1972.

Brock, T. C., & Becker, L. A. "Debriefing" and suscep-tibility to subsequent experimental manipulations.Journal of Experimental Social Psychology, 1966, 2,314-323.

Carlsmith, J., & Gross, A. Some effects of guilt on com-pliance. Journal of Personality and Social Psychology,1969, 77, 232-239,

Cialdini, R. B., Baumann, D. J., & Kenrick, D. T. In-sights from sadness: A three step model of the devel-opment of altruism as hedonism. Developmental Re-view, 1981,1, 207-223.

Cialdini, R. B., Darby, B. L., & Vincent, J. E. Transgres-sion and altruism: A case for hedonism. Journal ofExperimental Social Psychology, 1973, 9, 502-516.

Cialdini, R. B., & Kenrick, D. Altruism as hedonism:A social development perspective on the relationshipof negative mood state and helping. Journal of Per-sonality and Social Psychology, 1976, 34, 907-914.

Darlington, R. B., & Macker, C. E. Displacement ofguilt-produced altruistic behavior. Journal of Person-ality and Social Psychology, 1966, 4, 442-443.

Duval, S., Duval, V., & Neely, R. Self focus, felt re-sponsibility, and helping behavior. Journal of Person-ality and Social Psychology, 1919,37, 1769-1778.

Filter, T. A., & Gross,, A. E. Effects of public and privatedeviancy on compliance with a request. Journal ofExperimental Social Psychology, 1975, //, 553-559.

Freedman, J. L., & Doob, A. N. Deviance: The psy-chology of being different. New York: Academic Press,1968.

Freedman, J. L., Wellington, S. A., & Bless, E. Com-pliance without pressure: The effect of guilt. Journalof Personality and Social Psychology, 1967, 7, 117-124.

Geller, V., & Shaver, P. Cognitive consequences of self-awareness. Journal of Experimental Social Psychol-ogy, 1976, 12, 99-108.

Isen, A. M., Horn, N., & Rosenhan, D. L. Effects ofsuccess and failure on childrens' generosity. Journalof Personality and Social Psychology, 1973,27, 239-247.

Kenrick, D. T., Baumann, D. J., & Cialdini, R. B. Astep in the socialization of altruism as hedonism: Ef-fects of negative mood on childrens' generosity underpublic and private conditions. Journal of Personalityand^ Social Psychology, 1979, 37, 747-755.

Konecni, V. J. Some effects of guilt on compliance: Afield replication. Journal of Personality and SocialPsychology, 1972, 23, 30-32.

Latanfe, B., & Darley, J. M. The unresponsive bystander:Why doesn't he help? New York: Appleton, 1970.

Latan6, B., & Nida, S. Ten years of research on groupsize and helping. Psychological Bulletin, 1981, 89,308-324.

Maruyama, G., Fraser, S. C., & Miller, N. Personal re-sponsibility and altruism in children. Journal of Per-sonality and Social Psychology, 1982, 42, 658-664.

McMillen, D. Transgression, self image, and compliantbehavior. Journal of Personality and Social Psychol-ogy, 1971,20, 176-179.

McMillen, D. L., Sanders, D. Y., & Solomon, G. S. Self-esteem, attentiveness, and helping behavior. Person-ality and Social Psychology Bulletin, 1977, 3, 257-261.

Moore, B. S., Underwood, B., & Rosenhan, D. L. Affectand altruism. Developmental Psychology, 1973,8, 99-104.

Rawlings, E. I. Witnessing harm to others: A reassess-ment of the role of guilt in altruistic behavior. Journalof Personality and Social Psychology, 1968, JO, 377-380.

Regan, D., Williams, M., & Sparling, S. Voluntary ex-piation of guilt: A field experiment. Journal of Per-sonality and Social Psychology, 1972, 24, 42-45.

Regan, J. W. Guilt, perceived injustice, and altruisticbehavior. Journal of Personality and Social Psychol-ogy, 1971, 18, 124-132.

Rosenhan, D. L., Underwood, R., & Moore, B. S. Affectmoderates self-gratification and altruism. Journal ofPersonality and Social Psychology, 1974, 30, 546-552.

Steele, C. M. Name calling and compliance. Journal ofPersonality and Social Psychology, 1975, 3, 361-369.

Thompson, W., Cowan, C., & Rosenhan, D. Focus ofattention mediates the impact of negative affect on


altruism. Journal of Personality and Social Psychol-ogy, 1980, 38, 291-300.

Tukey, J. Exploratory data analysis. Reading, Mass.:Addison-Wesley, 1979.

Underwood, B., Berenson, J. F., et al. Attention, nega-tive affect and altruism: An ecological validation. Per-sonality and Social Psychology Bulletin, 1917,3, 54-58.

Underwood, B., Froming, W. J., & Moore, B. S. Mood,attention, and altruism: A search for mediating vari-ables. Developmental Psychology, 1977,13, 541-542.

Velten, E. A. The induction of elation and depressionthrough the reading of structured sets of mood-state-ments. Unpublished doctoral dissertation, Universityof Southern California, 1967.

Velten, E. A. A laboratory task for induction of moodstates. Behavior Research and Therapy, 1968,6, 473-482.

Wallace, J., & Sadalla, E. Behavioral consequences oftransgression: The effects of social recognition. Jour-nal of Experimental Research in Personality, 1966,1,94-197.

Wegner, D. M., & Schaefer, D. The concentration ofresponsibility: An objective self-awareness of groupsize effects in helping situations. Journal of Personalityand Social Psychology, 1978, 36, 147-155.

Winer, B. J. Statistical principles in experimental design.New York: McGraw-Hill, 1971.

Zuckerman, M., Lubin, B., & Robins, S. Validation ofthe Multiple Affect Check List in clinical situations.Journal of Consulting Psychology, 1965, 29, 594.

Received October 6, 1981Revision received March 15, 1982

Helmreich Appointed Section Editor, 1984-1985

The Publications and Communications Board of the American Psychological As-sociation announces the appointment of Robert L. Helmreich, University of Texasat Austin, as Editor of the Interpersonal Relations and Group Processes section ofthe Journal of Personality and Social Psychology for a 2-year term beginning in1984. Helmreich will complete outgoing section editor Ivan Steiner's term. As ofJanuary 1, 1983, manuscripts should be directed to:

Robert L. HelmreichDepartment of PsychologyUniversity of Texas at AustinAustin, Texas 78712.

Manuscript submission patterns for JPSP sections make the precise date of com-pletion of the 1983 volume uncertain. Therefore, authors should note that althoughthe current editor, Ivan Steiner, will receive and consider manuscripts until Decem-ber 31, 1982, should the 1^83 volume be completed before that date, Steiner willredirect manuscripts to Helmreich for consideration in the 1984 volume.

Personal Responsibility and Salience of the Request for ...

Documents

Transcript of Personal Responsibility and Salience of the Request for ...