The Influence of Positive Affect on the Unusualness of ......The Influence of Positive Affect on the...

I of Personality and Social PsychologyI'D!. 48, No. 6, 141.1-1426

Copyright 1985 by the Am rican Psychological Association, Inc.0()22-35I4/H5/$0<>.75

The Influence of Positive Affect on the Unusualnessof Word Associations

Alice M. IsenUniversity of Maryland

Mitzi M. S. Johnson, Elizabeth Mertz, and Gregory F. RobinsonOhio State University

A pilot study and two experiments investigated the influence of positive affect,induced in three differing ways, on the uniqueness of word associations. Personsin the positive-affect conditions gave more unusual first-associates to neutralwords, according to the Palermo & Jenkins (1964) norms, than did subjects inthe control conditions. In Study 3, where word type (positive, neutral, negative)was a second factor along with affect, in a between-subjects design, associates topositive words were also more unusual and diverse than were those to otherwords. These results were related to those of studies suggesting that positive affectmay facilitate creative problem solving and to other work suggesting an impactof positive feelings on cognitive organization.

A growing body of research suggests thatpositive affect may have a wide range ofeffects on cognitive processes. For example,positive feelings have been shown to serve asa retrieval cue for positive material in mem-ory, influencing such measures as the reactiontime for recall of positive words and thesubset of words likely to be recalled from amemorized list (e.g., Isen, Shalker, Clark, &Karp, 1978; Laird, Wagoner, Halal, & Szegda,1982; Nasby & Yando, 1982; Teasdale &Fogarty, 1979). Moreover, a mild positiveaffective state has also been shown to influencejudgments of various kinds, decision-makingstrategies, and willingness to take risks (e.g.,Isen & Means, 1983; Isen, Means, Patrick, &Nowicki, 1982; Isen & Patrick, 1983; Isen,Pratkanis, Slovic, & Slovic, 1984; Isen &Shalker, 1982; Isen et al., 1978; Johnson &Tversky, 1983; Schiffenbauer, 1974).

Recently, it has been suggested that positiveaffect may influence the way in which cog-nitive material is organized or interpreted,and that material may have a different mean-ing or function when one is feeling happyfrom that which it has at other times (Isen,

This research was supported in part by NIMH ResearchGrant 37452 to the first author.

Requests for reprints may be sent to Alice M. Isen,Department of Psychology, University of Maryland, Ca-tonsville, Maryland 21228.

1983; Isen, 1984; Isen & Daubman, 1984).This suggestion is compatible with the con-textualist position in cognitive psychology,which holds that the cognitive context (asdistinct from the simple stimulus context) inwhich material appears may influence themeaning or interpretation given to that ma-terial (e.g., Bartlett, 1932; Bransford, 1979).That is, as noted, positive affect has beenfound to cue positive material in memoryand thus influence the contents of workingmemory; and the contextualist or construc-tivist position suggests that the cognitive con-text, or contents and complexity of workingmemory, can influence the interpretationgiven to other cognitive material. Thus, ac-cording to the contextualist position, in com-parison with someone in a neutral state, aperson who is feeling happy might also beexpected to give a different meaning to cog-nitive material (not just a more positivemeaning), because he or she might be ex-pected to think about things and performcognitive tasks with a different and morecomplex cognitive context present.

Moreover, it makes sense that cognitivecontext (the specific items present in workingmemory at any given time and the interpre-tation given to them) might influence the wayin which the entire set of material would beorganized as well. Recent studies have dem-onstrated this very point. Tversky & Gati

1413

1414 ISEN, JOHNSON, MERTZ, AND ROBINSON

(1978), for example, found that the perceivedsimilarity, or categorization, of a given pairof items varied as a function of the otheritems (the cognitive context) that were presentat the same time that the judgment was beingmade. Thus, because positive affect cues abroad set or range of material (positive ma-terial has been found to be more extensivethan other material: e.g., Boucher & Osgood,1969; Bousfield, 1944), the affective state mayalso influence the organization of workingmemory; and the happy person, besidesthinking of or interpreting material differently,may see a different set of relations amongitems under consideration.

Experimental results compatible with thissuggestion regarding the influence of affecton cognitive organization, or the noting ofpotential relatedness, have recently been re-ported (Isen, 1983; Isen, 1984; Isen & Daub-man, 1984). One series of studies found thatpersons in whom positive affect had beeninduced in any of three ways tended tocategorize more stimuli together, as reflectedby either a sorting task or a rating task.Moreover, two additional studies found thatpositive affect facilitated performance on atask requiring creative problem-solving ability(i.e., requiring that subjects see potential re-lations not usually noticed between elementsof the stimulus materials). Another experi-ment found that positive affect facilitatedrecall of 10 relatable, but not obviously re-lated, words within a list of 30 to be recalled.All of these results converge to suggest thatmaterial may be organized differently whenone is happy, or at least that the cognitivecontext in which the task is performed maybe different.

The present article extends these findingsby examining in more detail the influence ofaffect on the unusuamess of word associations.Word association has a long history as bothan independent variable and a dependentvariable in psychology (see, e.g., Cramer,1968, for a review of the work from 1952-1964 and reference to earlier reviews of stillearlier literature). The literature on wordassociation constitutes a rich source of infor-mation regarding the impact of context effectsand of the affective content of the stimuluswords on this response. A complete review

of the literature on word association is beyondthe scope of this article; however, three pointsfrom this literature should be noted.

First, it has been suggested that word-association responses may reflect cognitiveorganization or at least processes that mayinfluence cognitive organization (e.g., Cramer,1968, esp. chap. 10). Cramer (1968), forexample, attributes word association to aprocess of activation of associative domains,or associative response hierarchies (p. 277),which is compatible with current process-oriented views of memory (both network andconstrucdvist views).

Second, the following specific findings bearrelevance to the impact of affect on cognitiveorganization and unusual responding: (a)pleasant stimuli elicit a greater number ofassociates, even when the word-lists are bal-anced for frequency of occurrence in thelanguage (e.g., R. C. Johnson & Lim, 1964;Matlin & Stang, 1978); and (b) a more com-plex cognitive context (e.g., a compoundstimulus word) results in more diverse re-sponses to the stimulus word (e.g., Cramer,1968). These findings suggest that, on thewhole, word associations to positive wordsmight be expected to be more unusual orunique than those to neutral words, becausethe positive words create a more complexcognitive context by cueing more associates.'

Moreover, there is also reason to expect apositive affect state to result in more uniqueor diverse associations than would a neutralaffective state. This is because positive affectcues positive material in memory (e.g., Isenet al., 1978), and positive material in memoryis more extensive than other material (e.g.,Boucher & Osgood, 1969; Bousfield, 1944,1950; Matlin & Stang, 1978; White, 1936;

1 It should not he assumed, however, that negativewords would therefore produce fewer unique responsesthan do neutral words. Although the area is marked bycontradictory findings, a number of studies suggest thatnegative or "emotional" words elicit more idiosyncraticresponses (e.g., Cramer, 1968). This phenomenon istheorized to occur for a different reason (implicit avoid-ance responses) from the one proposed here as underlyingthe similar-appearing effect of positive feelings, and otherfindings suggest that the pool of associates to negativewords is smaller than that of positive words (Cramer,1968).

INFLUENCE OF POSITIVE AFFECT ON WORD ASSOCIATIONS 1415

White & Powell, 1936), thereby creating amore complex cognitive context for the word-association task.

The third point from the word-associationliterature that is relevant to the present for-mulation is that word association, as both anindependent variable (e.g., Freedman, 1965;Maltzman, Simon, Raskin, & Licht, I960)and a dependent measure, has been associatedwith creativity (e.g., M. T. Mednick, S. A.Mednick, & E. V. Mednick, 1964; S. A.Mednick, 1962), which may itself be relatedto the organization of cognitive material andthe ability to see interrelations among stim-ulus elements.

Thus, the word-association literature con-tains a body of information supportive of thesuggestions, (a) that positive affect may influ-ence word association, and (b) that such afinding, together with others converging onthe concept of cognitive organization, mightbe indicative of an impact of affect on cog-nitive organization. The present studies at-tempt to examine the influence of positiveaffect on the uniqueness of word association,in the interest of contributing eventually toour understanding of the influence of goodfeelings on cognitive organization and cre-ativity.

Pilot Study

A pilot study has shown that persons inwhom positive affect had been induced, bymeans of refreshments at the experimentalsession, gave more unusual first associates toa randomly selected group of words from thePalermo & Jenkins (1964) list than did mem-bers of a control group that had not beenoffered refreshments (Mp = 6.60, Me = 5.28;r(18) = 1.88, p < .05; Isen, 1983). Onepossible alternative interpretation of thesefindings is that the affect induction used(refreshments) may have constituted an un-usual stimulus situation and that it was theexperience of something unusual, rather thanof something pleasant, that resulted in agreater number of unusual first associationsamong that group of subjects. Other researchhas shown that encouraging unusual responsesor giving subjects experience with differentresponses does appear to increase the occur-

rence of originality on word-association tasks(e.g., Cramer, 1968), and that exposure tosolution of a task requiring creative problemsolving can facilitate performance on a sub-sequent, different creative problem-solvingtask (Isen, 1983). Thus, Study 1 attemptedto replicate the earlier affect-and-word-asso-ciation finding using, however, a differenttechnique of affect induction, one that wouldnot constitute an unusual stimulus. The tech-nique employed was word association to af-fectively valenced words.2

Study 1

Method

Subjects. Forty university students participated in thestudy, which was conducted in classroom-administeredsessions. The subjects were students in one of two sopho-more-level courses.

Procedure. As noted earlier, affect was induced inthis study by means of word-association to affectivelyvalenced words. The experimenter distributed word-as-sociation booklets (composed of 20 words, each on aseparate page, followed by a rating task at the end of thebooklet) to the subjects. The booklets were distributedrandomly, so that some subjects in each class wereassigned to each condition. Subjects were asked to readthe words, one at a time, to write down their firstassociation to each word, and then to fill out the scale atthe end of the booklet. They were encouraged to workquickly, giving the first response that came to mind. Atthe end of each session, the experimenter discussed thepoint of the study and certain methodological issues withthe class. This constituted a debriefing.

Stimuli. The booklets were constructed so that inone third of the cases the first 10 words were positive, inone third they were negative, and in one third they wereneutral, while the second 10 words were neutral in allcases. Thus, the first 10 words constituted the independentvariable manipulation, and the second 10 words thestimuli for the dependent measure. Words were randomlyordered within list types.

All words were nouns and adjectives from the Palermoand Jenkins (1964) list of word-association norms (whichthemselves had been selected from the Kent-Rosanoff listand A-AA frequency words of the Thorndike-Lorge list).A group of 80 nouns and adjectives from the Palermoand Jenkins list had been rated for pleasantness, on ascale from 1-7, in a pretest using a different group of 36subjects. From these, a list of 10 positive (mean rating =6.08) and negative (mean rating = 2.02) words wereselected, as were two groups of 10 neutral words each(mean ratings = 4.33, 4.29).

The rating task that followed the word associationsrequired that a list of 10 words be rated "good-bad" on

2 The authors wish to thank Paul Slovic for suggestingthis means of affect induction.

1416 ISBN, JOHNSON, MERTZ, AND ROBINSON

a 7-point scale (with "1" the positive pole of the scale).It was used to check on the manipulation of affect. Thislist was composed of 5 familiar and 5 unfamiliar wordsused by Matlin (1970), who found that unfamiliar wordstended to be rated more negatively than familiar, and byIsen (1983). who reported that, relative to a controlgroup, the rating of unfamiliar words was significantlymore positive (and that of familiar words a bit morenegative) among subjects who had been made to feelgood.

Results

Manipulation checks. As a check on themanipulation of affect, subjects were askedto rate the pleasantness of unfamiliar words(the final page of the booklet described ear-lier), a task used in other research to checkon the manipulation of affect. In previousresearch, four studies, using four differentpositive-affect inductions and two negative,found positive affect to be associated withsignificantly more positive ratings of unfa-miliar words and a nonsignificant tendencytoward more negative ratings of familiaritems. These studies have also reported, incontrast, that negative affect tends to resultin lower ratings of familiar items but notunfamiliar ones. As in the previous research,in the present study persons in the positiveaffect condition (those who had given wordassociations to 10 positive words before as-sociating to the neutral words) rated unfa-miliar words significantly more positively thandid the control group, t(21) = 2.05, p < .05,but rated familiar words more toward theopposite pole (more negatively), though notsignificantly so, t(27) = 1.40, .05 < p. Thissuggests that the positive-affect induction wassuccessful in producing the desired state.

The negative-affect induction had no effecton the rating of unfamiliar words (t < 1), justas had been found in earlier studies; butunlike the results obtained previously, it hadno effect either on the rating of familiarwords (/ = 1.30). These findings suggest that,in contrast with the positive-affect induction,the negative induction may not have beensuccessful in producing the desired state.

In order to examine this question further,subjects' word associations (each subject'sfirst associate to each word) to words 11-20(the 10 neutral words that constituted thestimuli for the dependent measure) were ratedfor pleasantness on a scale from 1-7 by a

rater who was unaware of the subject's ex-perimental condition at the time of rating.These results revealed that the positive-affectsubjects' associates to subsequent neutralwords were more positive than the controlgroup's (M = 4.54, 4.29, respectively, t = 2.33,p < .025); however, association to the 10negative words did not influence the affectivetone of those subjects' associations to theneutral words that followed (M = 4.35, t <1). These results parallel those observed beforein the word-association literature, that peo-ple's associations to positive words tend tobe positive, but that their associates to negativewords do not tend to be negative (e.g., Cramer,1968; Pollio, 1964). However, these resultsextend those findings to include, now, subjects'responses to neutral words following associa-tion to 10 positive or negative words. Thismeasure might be seen as a second form ofmanipulation check, the results of which thensuggest, consistently with those of the pleas-antness-rating task just reported, that positiveaffect was successfully induced but that neg-ative affect was not.

Measures ofunusualness. Table 1 presentsthree different measures of the unusualnessof the word associations of subjects in thethree conditions. Subjects' responses werescored for uniqueness by referring to theword-association norms (Palermo & Jenkins,1964). For the first measure, a response wasconsidered "unusual" if it had a score thatwas 5% or less of the total score of the mostfrequent associate. For example, the wordhome is listed as having been given in responseto the word house by 106 of the male, and124 of the female, (230 in total) college-levelrespondents. Thus, any word given by 11 orfewer of respondents was scored an unusualresponse to house. Using this measure, thecriterion for the definition of unusual wasdifferent for each word. A word with a single,very strong (frequent) associate had a highercutoff point; one with several weaker associ-ates would require the subject's response tobe more rare before it would be called un-usual.

Second, disregarding the strength of themost popular associate, a second measure ofuniqueness was constructed: Associationswere rated as unusual if they were given by2.5% or fewer of the respondents. Because


Table 1

Study 1: Unusualness of Word Association in Each of Three Affect Conditions

M

Mean number of unusual associates"Ms2

Mean number of unusual associates11

Ms3

Mean number of respondents in the normative sample(N = 1,000) giving the same response

Ms2

Positive affect(n = 15)

5.404.11

4.934.07

47.86880.67

Neutral affect(n = 14)

3.863.21

3.363.94

78.12734.07

Negative affect(H = 11)

4.094.69

3.363.25

63.01915.99

• Denned in terms of the frequency of the most frequent associate (see text for details).* Defined independently of the popularity of the most frequent associate (see text for details).

there were 500 college-age respondents ofeach gender, for this measure a score of 12or less was always the criterion for the des-ignation unusual. A third measure of theunusualness of subjects' associations was ob-tained by averaging the number of respon-dents in the normative sample who had giventhe same response as the subject. As can beseen from Table 1, all three measures indicatethat subjects in whom positive affect hadbeen induced by free association to positivewords had more unusual associates to theneutral words that followed the initial, affect-inducing (or control) set.

Planned comparisons revealed the pre-dicted effect of positive feelings on uniquenessof associations, for each of the three measures,r(27) = 2.17, p < .025; t = 2.12, p < .025;t = 2.87, p < .005. There was no effect ofnegative affect as induced in this study.

Discussion

These results indicate that positive feelingsresult in more positive first associates toneutral words and more unusual first associ-ates to neutral words, as reflected by each ofthree indices of uniqueness. They are inagreement with the findings reported previ-ously that positive affect induced by thepresence of refreshments and a relaxed at-mosphere led to an increased number ofunusual associates to a randomly selected listof words (Pilot Study; Isen, 1983). Theyextend the earlier work by having used a

different means of affect induction (wordassociation), by having used a written-presen-tation procedure (whereas the previous studyhad employed a technique involving oralpresentation of the words), by having used alist of pretested neutral nouns and adjectivesas the stimulus words for the dependentmeasure (whereas the earlier study had useda randomly selected list from the Palermo &Jenkins corpus of words), and by havingincluded a negative-affect induction as wellas the positive.

The attempt to induce negative feelingsand to influence the affective tone and un-usualness of associations to neutral words bythe technique of word association seems notto have been successful for negative affect,even though the negative words were ratedin pretests as significantly different from neu-tral, and relatively as negative as the positivewords were positive. At first glance, theseresults might seem surprising; yet, on recon-sideration they are quite compatible withprevious work reporting nonsymmetrical re-sults of positive and negative affect inductions.For example, Isen et al. (1978) found in-creased accessibility of positive material inmemory following success, but no parallelincrease in the accessibility of negative ma-terial following a failure. Likewise, Teasdaleand Fogarty (1979) and, recently, Nasby andYando (1982) found a similar lack of sym-metry, as have other researchers (see Isen,1984, and Isen, in press, for a fuller discussionof the data relating to this issue).


Our present results also accord with otherdata suggesting that word associations remainpositive following positive-affect induction,but quickly become neutral or positive fol-lowing negative-affect induction. Subjects re-portedly do not produce chains of negativeassociates to negative words (e.g., Cramer,1968; Pollio, 1964); neither, according to thepresent results, do they continue to havenegative associations following the activity ofgiving their first associate to each of 10negative words. These findings are compatiblewith others in the word association literaturethat suggest that the pool of associates tonegative words is small (e.g., Cramer, 1968;Pollio, 1964), even though these words havesometimes been found to produce a diversityof first-associates (e.g., Cramer, 1968).

Thus, the technique of word association asa means of affect induction may not be usefulfor inducing negative affect. It simply maynot induce affect; or it may induce the neg-ative state but also rapidly trigger, in normalsubjects, a tendency to return the negativestate to neutral or positive. Because we con-ceptualize the first 10 words as the affect-induction phase, and words 11-20 as thedependent measure, the first available manip-ulation check (ratings of the affective tone ofthe subjects' associates) did not occur untilthe 11-20th word associations. Thus, if neg-ative affect had been induced after the firstfew negative word associations of the affect-induction phase, subjects might already havebegun the mood-repair process by the timethey reached words 11-20. In either case,however, the results suggest that the techniqueof word association may not be successful formanipulation of negative affect in experi-ments, even though this technique appears tobe successful for inducing positive affect.

These results are not easily interpreted asa function of the freakishness of the indepen-dent variable manipulation—that is, that themanipulation represents an unusual stimulusor situation and thereby cues unusual asso-ciations. While this interpretation might havebeen proposed regarding the earlier studythat used refreshments to induce positiveaffect, it is not a plausible hypothesis in thecontext of the present study. The task ofword association to 10 common words hardlyconstitutes an unusual situation, and there is

no reason to believe that the positive wordsthemselves were more unusual than the neu-tral. (All words had been selected from thePalermo and Jenkins corpus, which itself hadbeen composed of words of about equalfrequency in the language.) In fact, if anything,positive words have been found to be morecommon, not more unusual (e.g., Zajonc,1968). Thus, it would be hard to argue thatthe positive words constitute a more unusualstimulus and for this reason alone primeunusual associates.

The results of this study confirm thatunusual associations are to be expected frompersons who are feeling happy. Another alter-native interpretation that does not appearviable is that the effect is due to the fact thatpeople who feel good give positive associates,but neutral words are more likely to appearin the word-association norms. The thirdindex of unusualness, scoring of the absolutenumber of respondents giving each subject'sresponse, was included to provide some evi-dence relating to this issue. Other details ofsubjects' responses that bear upon this matterwill be presented next.

Although subjects in the positive affectcondition did average significantly more 0-frequency responses (M = 2.33, M = 1.14;t(21) = 2.08, p < .05), they usually gave re-sponses that appeared in the norms, but lessfrequently than those of control subjects. Inaddition, their unusual associates were notalways stereotypically positive words. For ex-ample, subjects in the positive condition re-sponded to "house" with low frequency as-sociates (Palermo & Jenkins norms in paren-theses) such as "security" (3), "apartment"(2), and "residence" (0); and when they gaverather common associates, these were oftenwords that were less frequent than the mostusual associate, "home" (230): "family" (27),brick (26), garage (36). In this condition only3 of 15 subjects (20%) responded with themost common associate, whereas in the con-trol condition 5 out of 14 (36%) did so. Fora second example, to "carpet" people in thepositive-affect condition responded with un-usual words such as (again, Palermo andJenkins norms in parentheses) "plush" (4),"living room" (5), "fresh" (0), "green" (6),"texture" (4), and "shag" (10). As in theexample above, when subjects gave more


common associates they were more oftenmoderately popular responses such as "soft"(206) and "floor" (153) rather than the mostcommon associate, "rug" (311). "Rug" wasgiven by only 3 of 15 (20%) positive-affectsubjects, whereas 10 out of 14 (71%) controlsubjects gave that response. This pattern wasrepeated across the words, so that overallthere was a significant difference, in the ex-pected direction, between the two affect con-ditions in number of times the most commonassociate was given (A//subject = 2.27, 3.64;1(27) = 2.4 l , p < .025).

Thus, it can be seen that subjects in thepositive-affect condition had more unusualassociates than did a control group, and thatthis effect was not due simply to the positive-affect subjects having more pleasant associ-ates, which as a group happened to be lesscommon responses to the stimulus words.The associations given by the positive-affectsubjects do not appear to be organized aroundthe Affect factor (with their unusualness anartifact of this); rather the associates given bysubjects in whom positive affect had beeninduced appear to be organized around thestimulus word but more broad-ranging thanthe usual responses. These associates alsotend to be slightly more positive in tone thanthe more common associates are.

It might be argued that the word associationtechnique, even in the positive condition, didnot produce the observed effect because ofthe affect it induced, but only produced itseffects by direct priming of similar or asso-ciated words (word-to-word priming, requir-ing no mediation by affective state). That is,in our formulation, the change in rated pleas-antness of the associates generated is assumedto have been due to the type of word primedby the affective state induced by the positivewords. However, given the method of affectinduction used in this study, it could besuggested that the positive type of word mighthave been primed directly by the first set ofwords (positive words).

This interpretation of our findings is notcompelling because direct priming of positivecontent cannot account for the pattern ofresponses obtained on the rating of unfamiliarand familiar words (the manipulation check),nor for the unusualness data. On the manip-ulation check, the same response pattern

observed previously (Isen, 1983), after fourdifferent types of affect induction (pleasantfilm, funny film, free gift, and positive Veltenstatements), was observed again: improvedrating of the unusual but not the familiar (infact, a tendency in the opposite direction forthe latter). This would be difficult to explainin terms of direct priming of positive wordsby the first 10 positive words. Moreover,direct priming of positive material would notaccount for the data showing an effect ofpositive affect on unusualness. As describedabove, the unusual responses of the positive-affect subjects did not appear to be simplyan artifact of those subjects having givenmore positive responses. Thus, it seems thatgiving word associates to a group of 10positive words is sufficient to induce positiveaffect, to produce response patterns similarto those obtained with a variety of otherpositive-affect inductions, and to influencethe uniqueness of first associates given toneutral words.

Because the procedure of giving word as-sociations to negative words produced noeffect in our study, nothing definitive can besaid with regard to negative affect. However,the failure of these words, rated as significantlymore negative than neutral words, to haveproduced negative associates also arguesagainst the direct priming interpretation fornegative affect.

Study 2

As noted, it does not seem compelling thatthe results of Study 1 are attributable todirect priming of either positive or unusualresponses by the affect-induction techniqueof word association to positive words. How-ever, Study 2 addresses these issues directlyby inducing positive affect in additional ways,by means of a comedy film or a free gift.Study 2 also examines subjects' word-associ-ations to positive, negative, or neutral words,independently.

Method

Subjects. Undergraduate students (N - 190) enrolledin introductory psychology participated in this study, inpartial fulfillment of a course requirement. Three subjects(one in each of three conditions), who responded to theword-association stimuli with complete sentences, phrases,


or Ihe names of their girlfriends, were eliminated fromthe analysis.

Design. Subjects were randomly assigned to one of12 experimental conditions created by the crossing offour affect conditions (two positive and two neutral) withthree word-type conditions (positive, neutral, and negativewords). Thus, the study contained 12 conditions, with14-18 subjects in each condition, in a completely between-subjects design.

Materials. Two 5-min videotaped films were used toinduce a positive or a neutral affective state in the filmconditions. The positive film, "Gag Reel," consists of"bloopers" or humorous oultakes from several TV west-erns. The neutral film, "Area Under a Curve,1' consistsof a mathematical proof. Positive affect was induced, ina third condition by giving the subjects a small bag ofcandy (15 pieces of wrapped, hard candy in a sandwichbag decorated with cartoon characters (Glad "Funtime"Sandwich Bags) and tied with colored yam). A no-manipulation control group was included in the study asa comparison group for this treatment. These affect-induction techniques had been used in previous researchand found to induce the appropriate affective state (e.g.,

Isen, 1983; Isen & Daubman, 1984; Isen & Gorgoglione,1983; Isen et al., 1984).

Booklets for the word-association task were constructedfrom the positive, negative, and neutral word lists usedas the affect-inducing stimuli in Study 1, so as to createconditions in which subjects (in whom affect had beenmanipulated by film or candy) provided word associationsto positive, negative, or neutral words only.

Procedure

Manipulation check. Subjects who viewed one of thefilms were given a blank sheet of paper and asked to

write a few sentences about how tie film made themfeel. This task had been used previously as a means of

manipulation check (e.g., Isen, 1983) and yielded resultsconvergent with that of other means of evaluation ofaffective state (Isen & Gorgoglione, 1983).

Subjects were tested in groups ranging from 2 to 10individuals by male and female research assistants. Twoexperimenters were present for each experimental sessionin which affect was induced by means of films. One wasrandomly chosen prior to the session to act as theexperimenter; the other acted as someone who was thereonly to pretest the films, which ostensibly were to be

used for affect manipulation in a different experiment.That experimenter briefly introduced the film and, aftershowing it, asked subjects to write several sentences onhow the film made them feel. In these conditions, subjectswere assured that the session would not run overtime. Inthe conditions in which affect was induced by means ofcandy, the experimenter gave subjects a small bag ofcandy of the type described above, explaining that it wasa token of appreciation for their participation in theexperiment. (Subjects were assured that, in addition, theywould receive the expected credit for participation.)Subjects in the no-manipulation control condition receivedno affect manipulation before being introduced to theword-association task.

Word associations. The experimenter then randomlydistributed the positive, neutral, and negative booklets

Table 2

Study 2: Mean Number of Unusual FirstAssociates to 10 Positive, Neutral, or NegativeWords in Each of Four Affect Conditions

Affect condition

Valenceof word

PositiveMs2

nNeutral

Ms2

nNegative

Ms2

n

Positivefilm

(comedy)

5.79(4.16)14

5.41

(3.39)17

3.50(3.46)18

Controlfilm

(math)

4.88(6.00)17

3.80(3.03)

15

4.46(7.29)15

Freegift

5.00(2.40)16

4.40(4.97)15

3.53(2.28)15

No

manip-ulation

5.37(2.92)

16

2.93(2.79)

15

3.14(5.95)14

Note. Variances in parentheses.

containing 10 words, 1 per page. Subjects were instructedto write down their first associate to each word and to

work at a relatively quick pace. When all subjects hadcompleted their ratings, they were debriefed regardingthe nature of the experimental hypotheses and affectmanipulations.

Results

Manipulation check. Five additional in-troductory psychology students, who had notviewed the films and who were unaware ofthe nature of the experiment, rated the feel-ings conveyed by the sentences that subjectshad written describing the way the film madethem feel. A 7-point scale ranging from verypositive (1), to very negative (7) was used.Judgments of the five raters were averaged.A comparison of the mean ratings in the twoaffect conditions indicated that the affectivestate of subjects viewing the positive film wassignificantly more positive than that of sub-jects viewing the neutral film (AfPositi« Kim =2.45, ATNr!utral Fi,m = 4.76, /(94) = 10.74, p <.01). These results are consistent with thoseof other studies (Isen, 1983; Isen & Daubman,1984; Isen & Gorgoglione, 1983) in suggestingthat the positive film is a successful methodof positive-affect induction.

Word associations. Table 2 contains themean number of unusual first-associates to


negative, neutral, and positive words in eachof the four conditions of the study. Becauseall three indices constructed for Study 1yielded comparable results, only one of these(#2) was chosen for use in Study 2.

Two two-way analyses of variance (ANOVAS)intended to determine whether the two posi-tive-affect conditions (positive film and candy)differed from each other and whether the twocontrol conditions (neutral film and no-ma-nipulation) differed from each other, indicatedno main effect of induction method and nointeractions (positive-affect conditions: f\l,89) = 1.67, .20 <p; F(2, 89) = .71, .20 <p\control conditions: F(\, 86) = 1.43, .20 < p;F(2, 86) = 1.51, .20 < p, respectively). Thus,the positive-affect conditions were combinedfor further analysis, as were the control con-ditions; and all subsequent analyses treatedaffect as a variable composed of two levels(positive and neutral) rather than four.

Table 3 presents the mean number ofunusual associates to each type of word, ineach of these two affect conditions. A 2 X 3(Affect Condition X Word Type) between-subjects ANOVA revealed a significant maineffect of word type, F(2, 181) = 10.26, p <.001, and a significant interaction betweenword type and affect condition f\2, 181) =3.55, p < .05. No main effect for the affectcondition was found. Post hoc tests indicatedthat the main effect of word type was attrib-utable to the positive-word conditions, butthat the negative-word conditions, althoughtheir overall mean was lower than that of thecontrol group, did not differ from the neutral,/(123) = 2.96, p < .01; t(\22) = 1.36, .2 <p, respectively.

Paired comparisons performed in order tolocate the source of the interaction effectindicated that induced positive affect andpositive words as stimuli each produced sig-nificantly greater uniqueness scores relativeto their comparison groups. That is, givenneutral words (but not positive or negative)as the stimuli for association, persons in thepositive-affect condition had significantlymore unusual first associates than did thosein the control group, ((60) = 3.24, p < .01;at the same time, subjects in the neutral-affect condition (but not those in the positive)gave more unusual responses to positive-word

Table 3Study 2: Mean Number of Unusual First

Associates to Ten Positive. Neutral, or Negative

Words in Two Affect Conditions

Affect condition

Valenceof word

PositiveM5*

nNeutral

Ms2

nNegative

Ms2

n

Positiveaffect

5.37(3.28)30

4.94(4.24)32

3.51(2.82)33

Neutralaffect

5.12(4.41)33

3.37(2.99)30

3.83(6.86)29

Note. Positive affect includes the positive film and the freegift conditions combined. Neutral affect includes the controlfilm and the no-manipulation control conditions combined.

stimuli than to neutral or negative-word stim-uli, /(61) = 3.62, p < .01; /(60) = 2.12, p <.05, respectively. Although the mean numberof unusual associates to negative words wasslightly higher than to neutral words in theneutral-affect condition, this difference wasnot reliable, t(57) < 1. Moreover, in the pos-itive-affect condition, although the meannumber of unusual responses to negativewords was significantly lower than to neutralor positive words, J(63) = 3.06, p < .01, thismean was not significantly different from thatof the neutral-affect, negative-word condition,f(60) < 1. Additionally, just as no differencewas found between the affect groups whenthe stimuli were negative words, neither wasany observed when they were positive words,«<61) < 1.

Discussion

The results of Study 2 are compatible withthose of Study 1 and of the Pilot Study inindicating that positive affect induced by twoadditional means (candy or comedy film),each unrelated to the word stimuli, results inan increase in the number of unusual wordassociations given in response to neutralstimuli. This extends the earlier work and


further discourages the alternative hypothesisthat the effect is due to direct priming by thetask of giving associations to positive words.

These converging results lend support tothe suggestion that positive affect may influ-ence the way in which cognitive material isorganized. The findings of Studies 1 and 2are compatible with the theoretical notionthat positive affect results in a change incognitive organization, characterized by in-creased breadth of concepts and increasedperception of relatedness or integration ofcognitive material. It follows that such achange in cognitive organization might resultin a broader range of associates, because, ifmore things are seen as related, the pool ofassociates to a given word should be morewide-ranging.

Additionally, results of Study 2 show thatamong control subjects, positive words asstimuli also produced more unusual responsesthan did neutral words. This finding seemscompatible with the results of two other typesof studies on word association, viewed to-gether: Those showing an effect of complexityof the associative context on associativeuniqueness (i.e., that more unique associatesare given in a more complex cognitive con-text); and those showing that positive wordstend to have more associates (i.e., tend tohave a more complex associative context)than other words (e.g., Cramer, 1968; Pollio,1964).

It should be noted that, in contrast withother results reported in the word associationliterature (e.g., Cramer, 1965; 1968; Mandler& Fames, 1957; see also Footnote 1), subjects'responses to the negative words used in Study2 were not more unusual than their associatesto the neutral words. This may be due to themeasure of uniqueness that we used or to thefact that subjects responded to only one typeof word (positive, negative, or neutral).

In measuring uniqueness in the presentstudy, we referred to the Palermo and Jenkins(1964) norms and identified as "unusual"any response that was rare in that corpus.However, the normative data for positive andnegative words in the corpus must themselvesreflect any increase in uniqueness of responsesthat results from the affective tone of thewords. Because in our present study we de-

fined uniqueness in terms of a differencefrom the normative data, it is possible thatan effect of word valence was obscured bythe fact that such an effect was already presentin the norming corpus. That is, conceptually,respondents in the normative study can beseen as equivalent to subjects in our controlconditions; therefore, if positive or negativeaffective tone of words produces more uniqueor diverse associates, this should already berepresented in the corpus, and the effectshould be equal for the norming subjects andfor our control subjects. Therefore, even ifan effect of word valence on uniqueness doesexist, one might expect not to observe it, ifone uses a criterion based on obtaining adifference from responses present in the cor-pus of norms.

Indeed, we did observe such an effect ofpositive words, even with the measure thatwe used;3 however, it still may be true that acriterion based on obtaining differences fromthe responses given by members of the norm-ing sample may be too stringent. It may bethe use of this stringent criterion that accountsfor the failure of the negative words to pro-duce more unusual associates than did neutralwords in our study.

Thus, we constructed another index ofuniqueness, one independent of the normativecorpus and therefore perhaps more suitablefor revealing an effect of the valence of thewords. This index, the measure identified as"D" in the word association literature (e.g.,Cramer, 1968), involves simply counting,across subjects, the number of different firstassociates given by our subjects to the stimuluswords, without reference to the Palermo andJenkins norms. This index should reveal aneffect of positive or negative word valence ifone is present. That is, if negative (or positive)words produce more diverse or unique re-sponses than neutral words, and the effect

3 This result indicates that the effect of positive wordsin our study was stronger than that observed in thenorming sample used by Palermo & Jenkins (1964). Itis possible that this is due to the fact that the words towhich subjects responded in our study were of only oneaffective type, whereas in the norming study subjectsgave associates to a mixed group of words.


was merely obscured when uniqueness wascalculated using our norm-based index (be-cause it was equally present in the Palermoand Jenkins corpus), then one should find,across control subjects, more different re-sponses to negative (or positive) words thanto neutral words.

Looking only at subjects in the neutralaffect condition, we calculated these means(mean proportion of different responses perword, across the 10 words), for the threeword types. This analysis revealed a differencebetween the neutral and the positive words,but not between the neutral and the negativewords (Jl/Positive = .57 (s2 = .018), MNcutnll =.46 (s2 = .016), Motive = .51 (s2 = .023);f(18) = 1.77, p < .05; J(18) < 1, .05 < p,respectively). Thus, our data do not supportthe suggestion that negative words producemore diverse associates than neutral words.They do suggest that positive words producemore diverse associates.

General Discussion

In sum, then, results of the three studiespresented in this article indicate that peoplein whom positive affect has been induced,relative to those in control conditions, tendto give more unusual first-associates to neutralwords; additionally, associations to positivewords appear to be more diverse than toneutral. These findings are compatible withthe suggestion that positive affect influencesthe organization of cognitive material so thatinformation is more integrated and patternsof associations are changed.

One possible alternative interpretation ofthese findings is that positive affect influencesresponding—that is, willingness to give theunusual response—rather than cognitive or-ganization. Although such a factor may pos-sibly play a role in the overall impact of affecton behavior, this interpretation of the presentdata is not compelling because of the inter-action obtained between affect and word type.This interaction suggests that response biaswas not the factor producing the observedresults, because response bias presumablywould affect all responding equally. Moreover,the present findings are compatible with agrowing body of evidence from diverse studies

(not subject to the response bias interpreta-tion) that converge to suggest that cognitiveorganization is influenced by affect.

The interpretation that we wish to proposeis that positive affect influences cognitiveorganization by altering the cognitive contextin which cognitive activity takes place. Positiveaffect cues positive material (e.g., Isen et al.,1978), positive material is extensive and varied(e.g., Boucher & Osgood, 1969), and thuspositive affect cues an extensive and variedset of material. Because context is known toinfluence cognitive organization as well asother processes (e.g., Tversky & Gati, 1978),we propose that positive affect can influencecognitive organization. Indeed evidence isbeginning to confirm this possibility (e.g.,Isen & Daubman, 1984).

The studies presented here provide addi-tional evidence that cognitive organizationmay be influenced by positive affect.4 Theyshow that positive words elicit a broaderrange of associates (more unusual associates)and that a positive state results in a broaderrange of associates (more unusual associates)to neutral words. Not only does this suggestthat positive affect may influence the way inwhich cognitive material is organized (i.e.,the set of ideas that are cued by, or seen asrelated to, a given idea), but it also providesgood evidence that the cognitive context ac-companying positive affect is more complexthan that present at other times.

There are two main points here: (a) thatcognitive context is altered by positive affect—not just in the direction of positivity, but alsoin complexity; and (b) that cognitive organi-zation may be altered by positive affect, pos-sibly as a result of the complexity of thecontext—that the sets of relations perceivedin cognitive material are altered. These two

' It should be noted (a) that we still have not studiedcognitive organization directly; and (b) that the word-association paradigm currently is not in favor as a meansof studying cognitive organization (in part because of therole that response factors may play in determining theresults). However, the findings of these studies, especiallywhen considered together with those indicating changesin categorization and memory as a function of positiveaffect (e.g., Isen, 1983; Isen & Daubman, 1984), lendsupport to the view that relations among cognitive ele-ments may be altered by positive affect.


points have implications for many aspects ofdecision making and problem solving, andthus they suggest that positive affect may playan important role in these processes as well.

All of this suggests that cognitive processesmay be more flexible as a function of positiveaffect. If a more complex context is present,if more material is cued, if more diverse waysof relating ideas to one another occur topeople when they are happy, then their think-ing may reflect this increased flexibility. Ma-terial may be multiply encoded and multiplyretrievable. This may result in greater dis-tractability and the possibility of errors attimes (e.g., Isen et al., 1982); but it may alsoresult in greater creativity, as has alreadybeen suggested (e.g., Isen et al., 1982; Isen,1983). Preliminary evidence indicates thatpositive affect is associated with improvedcreative problem-solving ability; the presentresults, indicating increased unusualness ofassociations as a function of positive affect,may also be suggestive of a relation betweenpositive affect and increased creativity; andthis possibility is under further investigation.

The results that we have presented anddescribed here—unusual or flexible respond-ing as a function of positive affect—mayseem surprising if one considers affect andarousal synonymous. This is because it isunderstood that arousal facilitates the domi-nant or typical response, not the unique(Zajonc, 1965). Indeed, there is even evidencethat persons working in the presence of others(and therefore, presumably, aroused) givefewer unusual word associations than thoseworking alone (Matlin & Zajonc, 1968). Thiswould suggest specifically that arousal resultsin more common, rather than more unusual,word associations. This is exactly the oppositeof the results that we have reported here forpositive affect. Thus, perhaps positive affectand arousal should not be equated.

In this context it is important to examinethe assumptions (a) that affect is equivalentto arousal, and (b) that positive and negativeaffect are equivalent in this regard. There haslong been confusion between affect, especiallynegative affect (and most especially anxiety),and arousal. In fact, in the early part of thiscentury, anxiety and arousal were often

equated, in studies of both motivation andlearning. One early view attempted to reduceall emotion to arousal (e.g., Duffy, 1934); butother views contrasted with that one eventhen (e.g., Leeper, 1948). And this kind ofmonolithic view of emotion (or even ofarousal) has been questioned from severalperspectives (e.g., Averill, 1982; Lacey, 1967;Lacey, Kagan, Lacey, & Moss, 1963; see Isen,1984, and Isen & Hastorf, 1982, for moredetailed consideration of this issue).

More recently and specifically, it has beennoted that mild positive feeling states of thetype induced in these studies do not seem toproduce effects suggestive of global "arousal"(Isen & Daubman, 1984). Moreover, Mandler(1984), long a proponent of the role of arousalin emotion, has recently suggested that adistinction should be made between emotionsand the more pervasive feeling states (or"moods") of the type induced in these studies,with regard to arousal. In fact, he has sug-gested that arousal is not a central componentin mood states (Mandler, 1984, p. 277).

Although it is not possible to know forcertain whether arousal is an important com-ponent of a mild positive feeling state, itseems that one should not simply assumethat it is. Even less, then, would one bejustified in assuming that arousal is all thereis to positive affect of the type induced inthese studies, and that the two are synony-mous. This is especially so in view of thevagueness of the concept of "arousal" itself,the controversy that has surrounded its usein psychology, and, last but not least, theresults like those presented here and elsewhere(e.g., Isen & Daubman, 1984; Isen et al.,1982) that are now suggesting that happinessmay facilitate creative and unusual respondingrather than the more routinized pattern ofresponding usually thought typical of"aroused" persons. Hopefully, future researchwill advance our understanding of affect andrelated concepts so that we can distinguishamong them in determining their implicationsfor thought and social behavior.

References

Averill, J. (1982). Emotion: What is the proper unit of

analysis? Paper presented at the symposium, Current


issues in theories of the emotions. L. deRivera, Chair,Meeting of the American Psychological Association,Washington, DC

Bartlett, F. C. (1932). Remembering: A study in experi-

mental and social psychology. New York: CambridgeUniversity Press.

Boucher, J., & Osgood, C. E. (1969). The pollyannahypothesis. Journal of Verbal Learning and VerbalBehavior. 8. 1-8.

Bousfleld, W. A. (1944). An empirical study of theproduction of affectively toned items. Journal of GeneralPsychology. SO. 205-215.

Bousfteld, W. A. (1950). The relationship between moodand the production of affectively toned associates.

Journal of General Psychology, 42, 67-85.

Bransford, J. D. (1979). Human cognition. Belmont CA:Wadsworth.

Cramer, P. (1965). Response entropy as a function of the

affective quality of the stimulus. Psychonomic Science.3, 347-348.

Cramer, P. (1968). Word association. New York: AcademicPress.

Duffy, E. (1934). Emotion: An example of the need forreorientation in psychology. Psychological Review, 41,184-198.

Freedman, J. L. (1965). Increasing creativity by free-association training. Journal of Experimental Psychol-ogy, 69, 89-91.

Isen, A. M. (1983). The influence of positive affect oncognitive organization. Stanford conference on aptitude,learning and instruction: Affective and cognitive pro-cesses, R. E. Snow, chair.

Isen, A. M. (1984). Toward understanding the role ofaffect in cognition. In R. Wyer & T. Srull (Eds.)Handbook of social cognition (pp. 179-236). HillsdaleNJ: Erlbaum.

Isen, A. M. (in press). The asymmetry of happiness andsadness in effects on memory in normal college students:Comment on Hasher, Rose, Zacks, Sanft, and Doren.

Journal of Experimental Psychology: General.

Isen, A. M., & Daubman, K. A. (1984). The influenceof affect on categorization. Journal of Personality and

Social Psychology, 47. 1206-1217.

Isen, A. M., and Gorgoglione, J. M. (1983). Some specificeffects of four affect-inducing techniques. Personality

and Social Psychology Bulletin, 9, 136-143.

Isen, A. M., & Hastorf, A. H. (1982). Some perspectiveson cognitive social psychology. In A. H. Hastorf &A. M. Isen (Eds.) Cognitive social psychology (pp. 1 -31). New York: Elsevier.

Isen, A. M., & Means, B. (1983). The influence of positiveaffect on decision-making strategy. Social Cognition,

2. 18-31.

Isen, A. M., Means, B., Patrick, R., & Nowicki, G.(1982). Some factors influencing decision-makingstrategy and risk-taking. In M. S. Clark & S. T. Rske(Eds.), Affect and cognition: The 17th Annual CarnegieSymposium on Cognition (pp. 243-261). Hillsdale NJ:Erlbaum.

Isen, A. M., & Patrick, R. (1983). The effect of positive

feeling on risk-taking: When the chips are down.Organizational Behavior and Human Performance, 31,

194-202.Isen, A. M., Pratkanis, A. R., Slovic, P., & Slovic, L.

(1984). Positive affect promotes risk aversion. Unpub-lished manuscript, University of Maryland.

Isen, A. M., & Shalker, T. E. (1982). The influence of

mood state on evaluation of positive, neutral andnegative stimuli: When you "accentuate the positive;"do you "eliminate the negative"? Social PsychologyQuarterly, 45. 58-63.

Isen, A. M., Shalker, T., Clark, M., & Karp, L. (1978).Affect, accessibility of material in memory and behaviorA cognitive loop? Journal of Personality and SocialPsychology, 36, 1-12.

Johnson, R. C., & Lim, D. (1964). Personality variablein associative production. Journal of General Psychology,71. 349-350.

Johnson, E., & Tversky, A. (1983). Affect, generalization,and the perception of risk. Journal of Personality and

Social Psychology, 4S, 20-31.Lacey, J. I. (1967). Somatic response patterning and

stress: Some revisions of activation theory. In M. H.Appley & R. Trumbul (Eds.), Psychological stress:Issues in research (pp. 14-44). New York: Appleton-Century Crofts.

Lacey, J. I., Kagan, J., Lacey, B., & Moss, H. A. (1963).The visceral level: Situational determinants and behav-ioral correlates of autonomic response patterns. InP. H. Knapp (Ed.), Expressions of the emotions in

man. New York: International Universities Press.

Laird, J. D., Wagener, J. J., Halal, M., & Szegda, M.(1982). Remembering what you feel: The effects ofemotion on memory. Journal of Personality and Social

Psychology, 42, 646-657.

Leeper, R. W. (1948). A motivational theory of emotion

to replace "emotion as disorganized response." Psy-

chological Review, 55, 5-21.Maltzman, I., Simon, S., Raskin, D., & Licht, L. (I960).

Experimental studies in the training of originality.Psychological Monographs, 74, (6, Whole N6. 493).

Mandler, G. (1984). Mind and body: The psychology of

emotion and stress. New York: Norton.Mandler, G., & Parnes, E. W. (1957). Frequency and

idiosyncrasy of associative responses. Journal of Ab-

normal and Social Psychology, 55, 58-65.

Matlin, M. W. (1970). Response competition as a me-diating factor in the frequency-affect relationship.

Journal a/Personality and Social Psychology. 16, 536-552.

Matlin, M. W., & Stang, D. J. (1978). The Pollyannaprinciple: Selectivity in language, memory, and thought.

Cambridge MA: Schenkman.Matlin, M. W, & Zajonc, R. B. (1968). Social facilitation

of word associations. Journal of Personality and Social

Psychology, 10, 455-460.Mednick, M. T., Mednick, S. A., & Mednick, E. V.

(1964). Incubation of creative performance and specificassociative priming. Journal of Abnormal and SocialPsychology, 69, 84-88.

Mednick, S. A. (1962). The associative basis of thecreative process. Psychological Review, 69, 220-232.


Nasby, W., & Yando, R. (1982). Selective encoding andretrieval of affectively valent information. Journal ofPersonality and Social Psychology, 43, 1244-1255.

Palermo, D. S., & Jenkins, J. J. (1964). Word associationnorms: Grade school through college. Minneapolis,MN: University of Minnesota Press.

Pollio, H. R. (1964). Some semantic relations amongword-associates. American Journal of Psychology, 77,249-256.

Schitfenbauer, A. (1974). Effect of observer's emotionalstate on judgments of the emotional state of others.Journal of Personality and Social Psychology, 30, 31-35.

Teasdale, J. D., & Fogarty, S. J. (1979). Differentialeffects of induced mood on retrieval of pleasant andunpleasant events from episodic memory. Journal ofAbnormal Psychology, SS, 248-257.

Tversky, A., & Gati, I. (1978). Studies of similarity. InE. Rosen and B. B. Lloyd (Eds.) Cognition and cate-gorization. Hillsdale NJ: Erlbaum.

White, M. M. (1936). Some factors influencing recall ofpleasant and unpleasant words. American Journal ofPsychology, 48, 134-139.

White, M. M., & Powell, M. (1936). The differentialreaction-time for pleasant and unpleasant words.American Journal of Psychology, 4/t, 126-133.

Zajonc, R. B. (1965). Social facilitation. Science, 149,269-274.

Zajonc, R. B. (1968). Attitudinal effects of mere exposure.Journal of Personality and Social Psychology, 9, 1-27.

Received July 23, 1984

Revision received December 17, 1984

The Influence of Positive Affect on the Unusualness of ......The Influence of Positive Affect on the...

Documents

Transcript of The Influence of Positive Affect on the Unusualness of ......The Influence of Positive Affect on the...