7/30/2019 Janis & Feshback (1953)

1/15

TH B J O U R N A L O F A B N O R M A L A N D S O C I A L P S Y C H O L O G YVol. 48, No. i, 1053

EFFECTS OF FEAR-AROUSING COMMUNICATIONSIRVING L. JANIS AND SEYMOUR FESHBACH

Yale UniversityIT is generally recognized that when beliefsand attitudes are modified, learning proc-esses are involved in which motivationalfactors play a prim ary role. Symb ols in masscommunications can be manipulated in avariety of ways so as to arouse sociallyacquired motives such as need for achieve-ment, group conformity, power-seeking, andthe more emotion-laden drives arising fromaggression, sympathy, guilt, and anxiety.The present experiment was designed tostudy the effects of one particular type ofmotive-incentive variable in persuasive com-munications, namely, the arousal of fear oranxiety by depicting potential dangers towhich the audience might be exposed.1 Fearappeals of this sort are frequently used toinfluence attitudes and behavior. For ex-ample, medical authorities sometimes try topersuade people to visit cancer detectionclinics by pointing to the dangerous conse-quences of failing to detect the early symp-toms of cancer; various political groups playup the threat of war or totalitarianism in anattempt to motivate adherence to their politi-cal program. Our interest in such attemptsis primarily that of determining the condi-tions under which the arousal of fear is effec-tive or ineffective in eliciting changes inbeliefs, practices, and attitudes.Implicit in the use of fear appeals is theassumption that when emotional tension isaroused, the audience will become morehighly motivated to accept the reassuringbeliefs or recommendations advocated by the1This study was conducted at Yale University aspart of a coordinated program of research on attitudean d opinion change, financed by a grant from theRockefeller Foundation. The attitude change researchproject is under the general direction of Professor CarlI. Hovland, to whom the authors wish to express theirappreciation for many valuable suggestions concerningthe design of the experiment. Special thanks are dueto Dr. Isador Hirschfeld of New York City and Dr.

Bert G. Anderson of the Yale Medical School for theirhelpful advice in connection with the preparation of theillustrated talks on dental hygiene. The authors alsowish to thank Dr. S. Willard Price, Superintendent ofSchools at Greenwich, Connecticut, and Mr. AndrewBella, Principal of the Greenwich High School, for theirgenerous cooperation.78

communicator. But the tendency to acceptreassuring ideas about ways and means ofwarding off anticipated danger may notalways be the dominant reaction to a fear-arousing com m unication. Un der certain con-ditions, other types of defensive reactions m ayoccur which could give rise to highly unde-sirable effects from the standpoint of thecommunicator.Clinical studies based on patients' reactionsto psychiatric treatment call attention to threemain types of emotional interference whichcan prevent a person from being influencedby verbal communications which deal withanxiety-arousing topics.1. When a communication touches offintense feelings of anxiety, communicateeswill sometimes fail to pay attention to whatis being said. Inattentiveness may be amotivated effort to avoid thoughts which

evoke incipient feelings of anxiety. Thisdefensive tendency may be manifested byovert attempts to change the subject of con-versation to a less disturbing topic. W hensuch attempts fail and anxiety mounts to avery high level, attention disturbances maybecome much more severe, e.g., "inabilityto concentrate," "distractibility," or othersymptoms of the cognitive disorganizationtemporarily produced by high emotionaltension (4).2. When exposed to an anxiety-arousingcommunication, communicatees will occa-sionally react to the unpleasant ("punishing")experience by becoming aggressive towardthe com mu nicator. If the com mu nicator isperceived as being responsible fo r producingpainful feelings, aggression is likely to takethe form of rejecting his statements.3. If a communication succeeds in arousingintense anxiety and if the communicatee'semotional tension is not readily reduced

either by the reassurances contained in thecommunication or by self-delivered reassur-ances, the residual emotional tension maymotivate defensive avoidances, i.e., attemptsto ward off subsequent exposures to the

7/30/2019 Janis & Feshback (1953)

2/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N S 79anxiety-arousing content. The experience ofbeing temporarily unable to terminate thedisturbing affective state elicited by a discus-sion of a potential threat can give rise to apowerful incentive to avoid thinking or hear-ing about it again; this may ultimately resultin failing to recall what the communicatorsaid, losing interest in the topic, denying orminimizing the importance of the threat.The above reaction tendencies, whileformulated in general terms, take accountof three specific types of behavior observedduring psychoanalytic or psychotherapeuticsessions (i, 2, 3). The first two refer toimmediate reactions that often occur whena therapist gives an interpretation whichbrings anxiety-laden thoughts or motives intothe patient's focus of awareness: (a ) atten-tion disturbances, blocking of associations,mishearing, evasiveness, and similar formsof "resistance"; and () argumentativeness,defiance, contempt, and other manifestationsof reactive hostility directed toward thetherapist. The third refers to certain typesof subsequent "resistance," displayed duringthe later course of treatment, as a carry-overeffect of the therapist's disturbing commentsor interpretations.Although the three types of defensive be-havior have been observed primarily in clini-cal studies of psychoneurotic patients (whoseanxiety reactions are generally linked withunconscious conflicts), it seems probable thatsimilar reactions may occur among normalpersons during or after exposure to communi-cations which make them acutely aware ofsevere threats of externa l danger. Neverthe-less, it remains an open question whethersuch sources of emotional interference playan y significant role in determining the neteffectiveness of fear-arousing material in masscommunications, especially when the com-munications are presented in an impersonalsocial setting where emotional responses ofth e audience are likely to be greatlyattenuated.

The present experiment was designed toinvestigate the consequences of using fearappeals in persuasive communications thatare presented in an impersonal group situ-ation. One of the main purposes was toexplore the potentially adverse effects whichmight result from defensive reactions of the

sort previously noted in the more restrictedsituation of psychotherapy.M E T H O D

The experiment was designed so as to providemeasures of the effects of three different intensitiesof "fear appeal" in a standard communication ondental hygiene, presented to high school students.The influence of the fear-arousing material wasinvestigated by means of a series of questionnaireswhich provided data on emotional reactions to thecommunication and on changes in dental hygienebeliefs, practices, an d attitudes.The Three Forms of Communication

A 15-minute illustrated lecture was prepared inthree different forms, all of which contained th esame essential information about causes of toothdecay and the same series of recommendationsconcerning oral hygiene practices. T he three (re-corded) lectures were of approximately equal lengthan d were delivered in a standard manner by thesame speaker. Each recording was supplementedby about 20 slides, which were shown on the screenin a prearranged sequence, to illustrate variouspoints made by the speaker.The three forms of the illustrated talk differedonly with respect to the amoun t of fear-arousingmaterial presented. Form i contained a strong fearappeal, emphasizing the painful consequences oftooth decay, diseased gums, and other dangers thatca n result from improper dental hygiene. Form 2presented a moderate appeal in which th e dangerswere described in a milder and more factual man-ner. Form 3 presented a minimal appeal whichrarely alluded to the consequences of tooth neglect.In Form 3, most of the fear-arousing material wasreplaced by relatively neu tral informatio n dealingwith the growth an d functions of the teeth. In allother respects, however, Form 3 was identical withForms i and 2.The fear appeals were designed to represent typi-ca l characteristics of mass communications whichattempt to stimulate emotional reactions in order tomotivate the audience to conform to a set ofrecommendations. The main technique was thatof calling attention to the potential dangers thatca n ensue from nonconformity. For example, theStrong appeal contained such statements as thefollowing:

If yo u ever develop an infection of this kind fromimproper care of your teeth, it will be an extremelyserious matter because these infections ar e reallydangerous. They can spread to your eyes, or yourheart, or your joints an d cause secondary infectionswhich may lead to diseases such as arthritic paralysis,kidney damage, or total blindness.One of the main characteristics of the Strong

appeal was the use of personalized threat-referencesexplicitly directed to the audience, i.e., statements tothe effect that "this ca n happen to you." TheModerate appeal, on the other hand, described thedangerous consequences of improper oral hygienein a more factual way, using impersonal language.

7/30/2019 Janis & Feshback (1953)

3/15

8 o I R V I N G L . J A N I S A N D S E Y M O U R F E S H B A C HIn the Minimal appeal, the limited discussion ofunfavorable consequences also used a purely factualstyle.The major differences in content are summarizedin Table i, which is based on a systematic contentanalysis of the three recorded lectures. The datain this table show ho w often each type of "threat"w as mentioned. It is apparent that th e main differ-ence between the Strong appeal and the Moderateappeal was not so much in the total frequency ofthreat references as in the variety an d types ofthreats that were emphasized. The Minimalappeal, however, differed markedly from the othertwo in that it contained relatively few threat refer-ences, almost all of which were restricted to"cavities" or "tooth decay."

TABLE 1CONTENT A N A L Y S I S O F T H E THREE F O R M S O F T H E

C O M M U N I C A T I O N : R E F E R E N C E S TO CONSE-Q U E N C E S O F I M P R O P E R C A R E O FTHE TEETHTYPE F O R M i F O R M 2O F ( S T R O N G ( M O D E R A T E

R E F E R E N C E A P P E A L ) A P P E A L )Pain from toothachesCancer, paralysis, blindnessor other secondary dis-easesHaving teeth pulled, cavitiesdrilled, or other painfuldental workHaving cavities filled orhaving to go to thedentistMouth infections: sore,swollen, inflamed gumsUgly or discolored teeth"Decayed" teeth"Cavities"Total references to unfavor-able consequences

ii69o

184M97 i

ioi5

1621212

49

F O R M S( M I N I M A LA P P E A L )o

0

0

Ia069

18

One of the reasons for selecting dental hygieneas a suitable topic fo r investigating th e influence offear appeals was precisely because discussions of thistopic readily lend themselves to quantitative andquali tative variations of the sort shown in Table I.Moreover, because of the nature of the potentialdangers that are referred to, one could reasonablyexpect the audience to be fairly responsive to suchvariations in contentthe teeth and gums probablyrepresent an important component in the averageperson's body image, and, according to psycho-analytic observations, the threat of damage to theteeth and gums can sometimes evoke deep-seatedanxieties concerning body integrity. In any case,by playing up the threat of pain, disease, and bodydamage, the material introduced in Form I is prob-ably representative of the more extreme forms offear appeals currently to be found in persuasivecommunications presented via the press, radio, tele-vision, and other mass media.

The fear appeals did not rely exclusively uponverbal material to convey the threatening conse-quences of nonconformity. In Form i, the slidesused to illustrate the lecture included a series ofeleven highly realistic photographs which vividlypor t rayed tooth decay and mouth infections.Form a, the Moderate appeal, included nine photo-graphs which were milder examples of oral pathol-ogy than those used in Form i. In Form 3,however, no realistic photographs of this kind werepresented: X-ray pictures, diagrams of cavities, an dphotographs of completely healthy teeth were sub-stituted for the photographs of oral pathology.Subjects

The entire freshman class of a large Connecticuthigh school was divided into four groups on arandom basis. Each of the three forms of the com-munication was given to a separate experimentalgroup; the fourth group was used as a control groupand was exposed to a similar communication on acompletely different topic (the structure an d func-tioning of the human eye). Altogether there were200 students in the experiment, with 50 in eachgroup.

The four groups were well equated with respectto age, sex, educational level, and IQ. The meanage for each group was approximately 15 years andthere were roughly equal numbers of boys and girlsin each group. The mean and standard deviationof IQ scores, as measured by the Otis group test,were almost identical in all four groups.Administration of the Questionnaires

The first questionnaire, given one week beforethe communication, was represented to the studentsas a general health survey of high school students.The key questions dealing with dental hygiene wereinterspersed among questions dealing with manyother aspects of health and hygiene.One week later the illustrated talks were given aspart of the school's hygiene program. Immediatelyafter the end of the communication, the studentsin each group were asked to fill out a short ques-tionnaire designed to provide data on immediateeffects of the communication, such as the amountof information acquired, attitudes toward the com-munication, and emotional reactions. A follow-upquestionnaire was given one week later in order toascertain the carry-over effects of the different formsof the communication.

R E S U L T SAffective Reactions

Evidence that the three forms of the illus-trated talk differed with respect to theamount of emotional tension evoked duringthe communication is presented in Table 2.Immediately after exposure to the communi-cation, the students were asked three ques-tions concerning the feelings they had just

7/30/2019 Janis & Feshback (1953)

4/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N STABLE 2

FEELINGS OF WORRY OR CONCERN EVOKED DURING THE COMMUNICATION

81

Q U E S T I O N N A I R E R E S P O N S E SS T R O N G M O D E R A T E M I N I M A LG R O U P GROUP GROUP

(AT=50) (JV=5o) (2V=50>Felt worried-a "few times" or "ma ny t imes"about own mouth conditionFelt "somewhat" or "very" worried about improper care o own teethThought about condi t ion of ow n teeth "most of the time"

74% 60% 48%66% 36% 34%42% 34% 22%

experienced "while the illustrated talk wasbeing given." Their responses indicate thatthe fear stimuli were successful in arousingaffective reactions. On each of the threequestionnaire items shown in the table, thedifference between th e Strong group and theMinimal group is reliable at beyond the .05confidence level.2 The Moderate group con-sistently falls in an intermediate position butdoes not,in most instances, differ significantlyfrom the other two groups.Further evidence of the effectiveness of thefear-arousing material w as obtained fromresponses to the following two questions,each of which had a checklist of five answercategories ranging from "Very worried" to"Not at all worried":i. When you think about the possibility8All probability values reported in this paper arebased on one tail of the theoretical distribution, sincethe results were used to test specific hypotheses whichpredict the direction of the differences.

that you might develop diseased gums, howconcerned or worried do you feel about it?2. When you think about the possibilitythat you m ight developed decayed teeth, howconcerned or worried do you feel about it?Since these questions made no reference toth e illustrated talk, it was feasible to includethem in the pre- and postcommunicationquestionnaires given to all four groups.Systematic comparisons were made interms of the percentage in each group whoreported relatively high disturbance (i.e.,"somewhat" or "very worried") in responseto both questions. The results, presented inTable 3, show a marked increase in affectivedisturbance among each of the three experi-mental groups, as compared with the controlgroup. Paralleling the results in Table's, thegreatest increase is found in the Stronggroup. The difference between the Moderateand the Minimal groups, however, isinsignificant.

TABLE 3P E R C E N T A G E O P E A C H G R O U P W H O R E P O R T E D F E E L I N G S O M E W H A T O R V E R Y W O R R I E D A B O U T D E C A Y E D T E E T HA N D D I S E A S E D G U M S

One week before the communicationImmediately after the communicationChange

S T R O N GGROUP(W=5o)3476+42%

M O D E R A T EGROUP(N=5o)245+26%

M I N I M A LGROUP(N=5b)2246+24%

C O N T R O LGROUP< JV = 5o)3o38+8%

GROUP R E L I A B I L I T Y OF DIFFERENCE *CR pStrong vs. ControlStrong vs. M inimalStrong vs. ModerateModerate vs . ControlModerate vs . MinimalMinimal vs . Control

3.061.591-371-540.171.43

7/30/2019 Janis & Feshback (1953)

5/15

82 I R V I N G L . J A N I S A N D S E Y M O U R F E S H B A C HIn order to obtain an over-all estimate ofthe relative degree of emotional arousalevoked by the three form s of the communi-cation, a total score was computed for eachindividual in each experimental group, based

on answers to all five questions: tw o pointscredit was given to each response specifiedin Tables 2 and 3 as indicative of high dis-turbance; one point credit was given tointermediate responses on the checklist; zerocredit was given for the last two responsecategories in each check list, which uniformlydesignated a relative absence of worry orconcern. Hence individual scores rangedfrom zero to ten. The mean scores for theStrong, Moderate and Minimal groups were7.8, 6.6, and 5.9 respectively. The Stronggroup differs reliably at the one per centconfidence level from each of the other tw ogroups (t=2.T, an d 3.6). The difference be-tween the Moderate and Minimal groupsapproaches reliability at the .08 confidencelevel (t=i.4).In general, the foregoing evidence indicatesthat after exposure to the communications,the Strong group felt more worried about thecondition of their teeth than did the othertw o groups; th e Moderate group, in turn,tended to feel more worried than the Mini-mal group.Information Acquired

Immediately after exposure to the illus-trated talk, each experimental group wasgiven an information test consisting of 23separate items. The test was based on thefactual assertions comm on to all three formsof the communication, including topics suchas the anatomical structure of the teeth, thecauses of cavities and of gum disease, the"correct" technique of toothbrushing, an dthe type of toothbrush recommended bydental authorities. No significant differenceswere found among the three experimentalgroups with respect to information testscores. Comparisons with the Control groupshow that the three forms of the dentalhygiene communication were equally effec-tive in teaching the factual material.Attitude Toward the CommunicationThe questionnaire given immediately afterexposure to the illustrated talk included aseries of seven items concerning the students'

appraisals of the communication. From theresults shown in Table 4, it is apparent thatthe Strong group responded more favorablythan the other tw o groups.8These findings imply that interest in thecommunication and acceptance of its educa-tional value were heightened by the Strongappeal. But this conclusion applies only torelatively impersonal, objective ratings of thecomm unication. Ad ditional evidence pre-sented in Table 5, based on questions whichelicited evaluations of a more subjective char-acter, reveals a markedly different attitudetoward the communication among those ex-posed to the Strong appeal.One of the additional questions was thefollowing: "Was there anything in the illus-trated talk on dental hygiene that you dis-liked ?" Un favora ble ("dislike") answerswere given by a reliably higher percentage ofstudents in the Strong group than in theModerate or Minimal groups (first row ofTable 5). A tabulation was also made ofthe total number of students in each groupwho gave complaints in their answers toeither of two open-end questions which askedfor criticisms of the illustrated talk. Theresults on complaints about the unpleasantcharacter of the slides are shown in row twoof Table 5; the difference between the Stronggroup and each of the other tw o groups isreliable at the .01 confidence level. Similarly,a reliably higher percentage of the Stronggroup complained about insufficient materialon ways and means of preventing tooth andgu m disease (row three of Table 5).* Thelatter type of criticism often was accompaniedby the suggestion that some of the disturbingmaterial should be eliminated, as is illus-trated by the following comments from tw o8The Strong group differs significantlyfrom th e Minimal group on five of the seven itemsan d from the Moderate group on three items; theModerate group does not differ reliably from theMinimal group on any of the items.4 In row three of Table 5, the difference between theStrong and Moderate groups is reliable at the .01confidence level, and the difference between th e Strongan d Minimal groups is significant at the .08 level.Other types of criticisms, in addition to those shownin Table 5, were also tabulated . Most of these inyolved

minor aspects of the presentation (e.g., "a movie wouldhave been better than slides") an d were given byapproximately equal percentages of the three groups.The vast majority of students in the Moderate an dMinimal groups expressed approval of the illustratedtalk or stated that they had no criticisms.

7/30/2019 Janis & Feshback (1953)

6/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N S 8 3TABLE 4

PERCENTAGE OF EACH GROUP WH O EXPRESSED STRONGLY FAVORABLE APPRAISALS os THE COMMUNICATION

APPRAISAL RESPONSE

The illustrated talk does a very good teaching job.Most or all of it was interesting.It was very easy to pay attention to what the speaker was saying.My min d p ractically never wandered.The slides do a very good job.The speaker's voice was very good.The illustrated talk definitely should be given to all Connecticut high schools.

S T R O N GGROUP

62807458526674

M O D E R A T EGROUP

50683646205658

MINIMALG R O U P

(N=5>406442225870

students in the Strong group: "Leave out theslides that show the rottiness of the teeth andhave more in about how to brush your teeth";"I don't think you should have shown somany gory pictures without showing moreto prevent it." Com ments of this sort, to-gether with the data presented in Table 5,provide additional evidence of residual emo-tional tension. They imply that th e Strongappeal created a need for reassurance whichpersisted after the communication was over,despite the fact that the communicationcontained a large number of reassuringrecommendations.The apparent inconsistency between theresults in Tables 4 and 5 suggests that theStrong appeal evoked a more mixed orambivalent attitude toward the communi-cation than did the Moderate or Minimalappeals. Some of the comments, particularlyabout the slides, help to illuminate the differ-entiation between the individual's objectiveevaluation of the communication and hissubjective response to it. The followingillustrative excerpts from the Strong groupwere selected from the answers given to theopen-end question which asked for criticismsand suggestions:

I did not care for the "gory" illustrations ofdecayed teeth and diseased mouths but I really thinkthat it did make me feel sure that I did not wantthis to happen to me.

Some of the pictures went to the extremes butthey probably had an effect on most of the peoplewho wouldn't want their teeth to look like that.

I think it is good because it scares people whenthey see the awful things that can happen.Such comments not only attest to the moti-vational impact of the Strong appeal, but alsosuggest one of the ways in which the dis-crepancy between subjective and objectiveevaluations m ay have been reconciled. Insuch cases, the ambivalence seems to havebeen resolved by adopting an attitude to theeffect that "this is disagreeable medicine, butit is good for us."Conformity to Dented Hygiene Recommen -dations

The immediate effects of the illustratedtalks described above show the type of affec-tive reactions evoked by the fear-arousing m a-terial but provide little information bearingdirectly on attitude changes. The question-naire administered one week later, however,was designed to measure some of the majorcarry-over effects of fear appeals, particularly

TABLE 5PERCENTAGE OF EACH GROUP WHO EXPRESSED COMPLAINTS ABOUT THE COMMUNICATION

TYPE OP COMPLAINT STRONG MODERATE MINIMALGROUP GROUP GROUP(JV=5o) (2V=5o) (W=5o)Disliked something in the illustrated talk.The slides were too unpleasant ("horrible," "gory," "disgusting," etc.).There was not enough material on prevention.

2834

7/30/2019 Janis & Feshback (1953)

7/15

84 I R V I N G L . J A N I S A N D S E Y M O U R F E S H B A C Hwith respect to changes in dental hygienepractices, beliefs, and preferences. The re-sults provide an empirical basis fo r estimatingthe degree to which such communicationssucceed in mod ifying a t t i tudes.

Personal practices were investigated byasking th e students to describe the way theywere currently brushing their teeth: the typeof stroke used, the a moun t of surface areacleansed, the a moun t of force applied, thelength of t ime spent on brushing the teeth,and the t ime of day that th e teeth werebrushed. The same five questions wereasked one week before the communication

four groups had very lo w scores and thegroup differences were insignificant. Bycomparing th e score that each individualattained one week after the communicationwith that attained tw o weeks earlier, it waspossible to determine fo r each group thepercentage who changed in the direction ofincreased or decreased conform ity.The results, shown in Table 6 , reveal thatthe greatest amount of conformity was pro-duced by the communication which con-tained the least amount of fear-arousingmaterial. The Strong group showed reliablyless change than the Minimal group; in fact,

TABLE 6EFFECT OF THE I L L U S T R A T E D TALK ON CONFORMITY TO DENTAL HYGIENE R E C O M M E N D A T I O N S

TYPE OF C H A N G E

Increased conformityDecreased conformityNo changeNet change in conformity

S T R O N GG R O U P

(JV=5)28 %20%52%+8%

M O D E R A T EG R O U P(W=5o)

4 4 %22%34%+22%

M I N I M A LG R O U P(JV=5o)50 %14%36%+36%

C O N T R O LG R O U P(W=5o)22%22%56 %o%

G R O U P CR R E L I A B I L I T Y O F DIFFERENCE

Control vs. MinimalControl vs . ModerateControl vs. StrongStrong vs. ModerateStrong vs . MinimalModerate vs. Minimal

2 - 5 41.500.590 . 951.960.93

7/30/2019 Janis & Feshback (1953)

8/15

E F F E C T S O F F E A R - A R O U S IN G C O M M U N I C A T I O N Sfindings represent changes in overt behavioralconformity, since the observations are basedon the 5s' ow n verbal reports. What remainsproblematical, how ever, is whether the verbalresponses reflect only "lip-service" to therecommendations or whether they also reflectinternalized attitudes that were actually car-ried out in action. The results, neverthe-less, demonstrate that the Strong appeal wasmarkedly less effective than the Minimalappeal, at least with respect to eliciting verbalconformity.Further evidence in support of the sameconclusion comes from responses pertinent toa different type of dental hygiene behaviorwhich had also been recommended in theillustrated talk.8 The students were asked togive the approximate date on which they hadlast gone to a dentist. The percentage ineach group whose answers indicated that theyhad gone to the dentist during the week fol-lowing exposure to the illustrated talk wereas follows: 10 per cent of the Strong group,14 per cent of the Moderate group, 18 per centof the Minimal group, and 4 per cent of theControl group. The percentage differencebetween the Minimal group and the Controlgroup was found to be statistically reliableat the .04 confidence level; none of the othercomparisons yielded reliable differences.Although not conclusive evidence, these find-ings are in line with those in Table 6: theMinimal appeal again appears to have beensuperior with respect to eliciting conformityto a recommended practice.Beliefs Concerning the "Proper" Type ofToothbrush

The illustrated talk presented an extensivediscussion of the "proper" type of toothbrushrecommended by dental authorities. Fourmain characteristics were emp hasiz ed: (a )the bristles should be of medium hardness,(b) the brush should have three rows ofbristles, (c) the handle should be completelystraight, an d (d) the brushing surface shouldbe completely straight. Personal beliefs con-cerning the desirability of these four charac-5 In all three forms of the illustrated talk, an explicitrecommendation was made concerning the desirabilityof obtaining advice from a dentist about one's owntoothbrushing technique. In addition, several referenceswere made to the importance of going to a dentist forprompt treatment of cavities, before the decay spreadsto the inner layers of the tooth.

85teristics were measured by four questionswhich were included in the precommuni-cation questionnaire as well as in thequestionnaire given one week after the com-munication. The main finding was that allthree experimental groups, as compared withthe Control group, showed a significantchange in the direction of accepting the con-clusions presented in the communication.Among th e three experimental groups, therewere no significant differences with respectto net changes. Nevertheless, as will be seenin the next section, the fear-arousing materialappears to have had a considerable effect onthe degree to which the students adhered tosuch beliefs in the face of counteractingpropaganda.Resistance to Counteracting Propaganda

In addition to describing the four essentialcharacteristics of the "proper" toothbrush,the illustrated talk contained numerous com-ments and illustrations to explain the needfor avoiding the "wrong" kind of toothbrush.Much of the material on cavities and otherunpleasant consequences of tooth neglect waspresented in this context. The importanceo f using the proper fond of toothbrush wasthe theme that was most heavily emphasizedthroughout th e entire communication.The key questionnaire item, designed todetermine initial attitudes before exposure tothe communication, was the following:

Please read the following statement carefully anddecide whether you believe it is true or false.It does not matter what kind of toothbrush aperson uses. Any sor t of toothbrush that is sold ina drugstore will keep your teeth clean and healthy

if you use it regularly.Do you think that this statement is true or false?(Check one.)One week after exposure to the communi-cations, the question was asked again, inessentially the same form, with the samechecklist of five answer categories (rangingfrom "Feel certain that it is true" to "Feelcertain that it is false"). But in the post-communication questionnaire, the questionwas preceded by the following propagandamaterial which contradicted the dominanttheme of the illustrated talk:A well-known dentist recently made the followingstatement:

7/30/2019 Janis & Feshback (1953)

9/15

86 I R V I N G L . J A N I S AND S E Y M O U R F E S H B A C HSome dentists, including a number of so-called"experts" on dental hygiene, claim it is importantto use a special type of toothbrush in order to clean

the teeth properly. But from my own experience,I believe that there is no sound basis for that idea.My honest opinion, as a dentist, is that it does notmatter what kind of toothbrush a person uses. Anysort of toothbrush that is sold in a drugstore willkeep your teeth clean and healthyif you use itregularly.That this propaganda exposure had a pro-nounced effect is revealed by the atti tudechanges shown by the Control group. Astatistically reliable change in the direction

definite answer emerges from the results inTable 7 , which shows the percentage of eachgroup who changed in the direction of agree-ment or disagreement with the counterpropa-ganda statement.Before exposure to the illustrated talk, thegroup differences were negligible: approxi-mately 50 per cent of the students in each of

the four groups agreed with the statementthat "it does not matter what kind of tooth-brush a person uses." But two weeks later(immediately after exposure to the counter-TABLE 7

EFFECT OP THE ILLUSTRATED TALK ON REACTIONS TO SUBSEQUENT COUNTERPROPAGANDA: NET PERCENTACIOF EACH GROUP WHO CHANGED IN THE DIRECTION OF AGREEING WITH THE STATEMENT THAT"!T DOES NOT MATTER WHAT KIND OF TOOTHBRUSH A PERSON USES"

TYPE OF C H A N G E

More agreementLess agreementNo changeNet changeNet effect of exposure to the illustrated talk

S T R O N GGROUP(JV=5o)303832828

M O D E R A T EG R O U P(JV=5o)2842301434

MINIMALGROUP(JV=5o)1454324 060

CONTROLGROUP(W=5o)442432+20

GROUPRELIABILITY OF THE DIFFERENCES IN NET CHANGECR p

Control vs. MinimalControl vs. ModerateControl vs. StrongStrong vs. ModerateStrong vs. Min imalModerate vs. M i n i m a l

3-662 .051.710.362 . 0 31.66

7/30/2019 Janis & Feshback (1953)

10/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N S 87the write-in answers given by those studentswho had disagreed with the counterpropa-ganda. In their refutations, some of thestudents made use of material that had beenpresented one week earlier, either by referringto the illustrated talk as an authoritativesource or by citing one of the main argumentspresented in the illustrated talk. From theresults presented in the first two rows ofTable 8, it is apparent that such refutationswere given more frequently by the Minimalgroup than by the other experimental groups.The comparatively low frequency of suchanswers in the Strong an d Moderate groupswas not compensated for by an increase inany other type of specific reasons, as indicatedby the results in the last row of the table.7

were inclined to avoid recalling the contentof the fear-arousing communication.D I S C U S S I O N

The results in the preceding sections indi-cate that the Minimal appeal was the mosteffective form of the communication in thatit elicited (a) more resistance to subsequentcounterpropaganda and (b) a higher inci-dence of verbal adherence, and perhaps agreater degree of behavioral conformity, to aset of recommended practices. The absenceof any significant differences on other indi-cators of preferences an d beliefs implies thatthe Moderate and Strong appeals had nounique positive effects that would compen-sate for the observed detrimental effects.

TABLE 8TYPES OF REFUTATION GIVEN BY STUDENTS WHO D I S A G R E E D WITH THE C O U N T E R P R O P A G A N D A

S T R O N G M O D E R A T E M I N I M A L C O N T R O LTYPE OF R E F U T A T I O N G R O U P G R O U P G R O U P G R O U P

Explicit reference to the illustrated talk as an authori-tative source for the opposite conclusionOne or more arguments cited that had been presentedin the illustrated talkOne or more arguments cited that contradicted thecontent of the illustrated talkNo answer or no specific reason given

7%43%o%

50%

14%38%0%5 2 %

18%59 %o%36%

0%

28%22%

Although the group differences are notuniformly reliable, they reveal a consistenttrend which suggests an "avoidance" tend-ency among th e students who had beenexposed to the fear appeals. Ap paren tly, eventhose who resisted the counterpropagandaTOn the first type of reason (reference to the illus-trated talk), the only difference large enough to approachstatistical reliability was that between th e Minimalgroup and the Control group (p=.o8). On the secondtype of reason (arguments cited from the illustratedtalk), th e difference between the Minimal group andth e Control group was found to be highly reliable(/>=:. 03) while the difference between the Minimalan d Mo derate groups app roached statistical reliab ility(/>=.o8). The Control group differed reliably fromeach of the experimental groups (a t beyond the .10confidence level) with respect to giving argumentswhich contradicted those contained in the illustratedtalk (row three of the table). None of the other

percentage differences in Table 8 were large enoughto be significant at the .10 confidence level. (In somecolumns, the percentages add up to more than 100 percent because a few students gave more than on e typeof refutation.)

Thus, the findings consistently indicate thatinclusion of the fear-arou sing ma terial notonly failed to increase the effectiveness of thecommunication, but actually interfered withits over-all success.The outcome of the present experiment byno means precludes th e possibility that, undercertain conditions, fear appeals may prove tobe highly successful. For instance, the Strongappeal was found to be maximally effectivein arousing interest and in eliciting a highdegree of emotional tension. The evocationof such reactions might augment the effective-ness of mass communications which aredesigned to instigate prompt audience action,such as donating money or volunteering toperform a group task. But if the communi-cation is intended to create more sustainedpreferences or attitudes, the achievement ofpositive effects probably depends upon anumber of different factors. Our experi-

7/30/2019 Janis & Feshback (1953)

11/15

I R V I N G L. J A N I S AND S E Y M O U R F E S H B A C Hmental results suggest that in the latter case,a relatively low degree of fear arousal islikely to be the optimal level, that an appealwhich is too strong will tend to evoke someform of interference which reduces the effec-tiveness of the communication. The findingsdefinitely contradict th e assumption that asthe dosage of fear-arousing stimuli (in a masscommunication) is increased, the audiencewill become more highly motivated to acceptthe reasssuring recommendations containedin the comm unication. Beneficial mo tivatingeffects probably occur whe n a relatively slightamount of fear-arousing material is inserted;L ut for communications of the sort used inth e present experiment, th e optimal dosageappears to be far below the level of thestrongest fear appeals that a communicatorcould use if he chose to do so.Before examining the implications of thefindings in more detail, it is necessary to takeaccount of the problems of generalizing fromthe findings of the present study. Thepresent experiment shows the effects of onlyone type of comm unication, presented in aneducational setting to a student audience.Until replications are carried outusing othermedia, topics, an d fear-eliciting stimuli, in avariety of communication settings, with dif-ferent audiences, etc.one cannot be certainthat the conclusions hold true for o ther situ-ations. The results from a single experimentare obviously not sufficient fo r drawing broadgeneralizations concerning the entire rangeof fear-arousing communications which arecurrently being brought to the focus of publicattention. Nor can unreplicated results berelied upon fo r extracting dependable rubricsthat could be applied by educators, editors,public relations experts, propagandists, orother communication specialists who face th epractical problems of selecting appropriateappeals for motivating mass audiences.Nevertheless, the present experiment helpsto elucidate the potentially unfavorable effectsthat may result from mass communicationswhich play up ominous threats, alarmingcontingencies, or signs of impending danger.For instance, the findings tend to bear outsome of the points raised concerning the needfor careful pretesting and for other cautionswhen warnings about the dangers of atomicbombing are presented in civilian defense

communications that are intended to preparethe public for coping with wartime emer-gencies (6). Moreover, despite our inabilityto specify the range of communications towhich our conclusions would apply, we canderive tentative inferences tha t m ay have im -portant theoretical implications with respectto the dynamics of "normal" fear reactions.W e turn now to a central question posedby th e experimental findings: Why is it thatthe fear-arou sing stimuli resulted in lessadherence to recommended practices and lessresistance to counterpropaganda ? Althoughour experiment cannot give a definitiveanswer, it provides some suggestive leadsconcerning potential sources of emotionalinterference.In the introduction, we have describedthree forms of "resistance" frequently ob-served in psyc hotherapy that m ight alsooccur among normal personalities exposed tomass communications which evoke strongfear or anxiety: (a) inattentiveness duringthe communication session, () rejection ofth e communicator 's statements motivated byreactive aggression, and (c ) subsequent de-fensive avoidance motivated by residual emo-tional tension. W e shall discuss briefly th epertinent findings from the present experi-ment with a view to making a preliminaryassessment of the importance of each of thethree types of interfering reactions.i. Our results provide no evidence that astrong fear appeal produces inattentivene ss oran y form of distraction that would interferewith learning efficiency during the communi-cation session. The three form s of the com-munication were found to be equally effective

in teaching the factual material on dentalhygiene, as measured by a comprehensiveinformation test given immediately afterexposure to the com mu nication. Beliefs con-cerning the desirable characteristics of the"proper" type of toothbrush were alsoacquired equally well. One might even sur-mise (from th e results in Table 4 ) that th eStrong appeal may have had a beneficial effecton attention, because a significantly higherpercentage of the Strong group reported that(a) it was very easy to pay attention to whatthe speaker was saying and (b) they experi-enced very little "mind-wandering."The absence of any observable reduction of

7/30/2019 Janis & Feshback (1953)

12/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N S 89learning efficiency is consistent with numer-ous clinical observations which imply thatnormal personalities can ordinarily tolerateunpleasant information concerning potentialthreats to the self without manifesting anymarked impairment of "ego" functions. Ourfindings definitely suggest that the use offear-arousing material of the sort presentedin the illustrated talks would rarely give riseto any interference with the audience's abilityto learn the content of the communication.

It is necessary to bear in mind, however,that in the present experiment the communi-cation was given to a "captive" classroomaudience. When people are at home listeningto the radio, or in any situation where theyfeel free to choose whether or not to termi-nate the communication exposure, the use ofstrong emotional appeals might often havedrastic effects on sustained attention. Con-sequently, the tentative generalization con-cerning the low probability of inattentivenesswould be expected to apply primarily to thosefear-arousing communications which are pre-sented under conditions where social normsor situational constraints prevent the audiencefrom directing attention elsewhere.Even with a "captive" audience, it is quitepossible that under certain extreme conditionsa strong fear appeal might interfere withlearning efficiency. For instance, the samesort of temporary cognitive impairment thatis sometimes observed when verbal stimulihappen to touch off unconscious personalconflicts or emotional "complexes" might alsooccur when a mass communication elicitssharp awareness of unexpected danger, par-ticularly when the audience immediatelyperceives the threat to be imminent andinescapable. Hence, the inferences from ourexperimental findings probably should berestricted to fear appeals which deal withremote threats or with relatively familiardangers that are perceived to be avoidable.

2. The fact that the Strong group expressedthe greatest amount of subjective dislike ofthe illustrated talk and made the most com-plaints about its content could be construedas suggesting a potentially aggressive attitude.But if the aggressive reactions aroused by theuse of the Strong fear appeal were intenseenough to motivate rejection of the conclu-sions, one would not expect to find this group

giving the most favorable appraisals of theinterest value of the illustrated talk, of thequality of its presentation, and of its over-alleducational success. Thus, although the pos-sibility of suppressed aggression cannot beprecluded, it seems unlikely that this factorwas a major source of emotional interference.In drawing this tentative conclusion, how-ever, we do not intend to minimize theimportance of aggression as a potential sourceof interference. In the present experiment,the communication was administered as anofficial part of the school's hygiene programand contained recommendations that wereobviously intended to be beneficial to theaudience. Under markedly different condi-tions, where th e auspices and intent of thecommunication are perceived to be less be-nign, the audience would probably be lessdisposed to suppress or control aggressive re-actions. The low level of verbalized aggres-sion observed in the present study, however,,suggests that in the absence of cues whicharouse the audience's suspicions, some factorother than reactive hostility may be a muchmore important source of interference.3. Subsequent defensive avoidance arisingfrom residual emotional tension seems to bethe most likely explanation of the outcome ofthe present study. We have seen, from the dataon immediate affective reactions, that th e dis-turbing feelings which had been arousedduring the illustrated talk tended to persistafter th e communication had ended, despite-the reassuring recommendations which hadbeen presented. The analysis of complaints,made by the three experimental groups(Table 5) provides additional evidence thatth e need fo r reassurance persisted primarilyamong the students who had been exposedto the Strong appeal. Such findings supportth e following hypothesis: When a mass com-munication is designed to influence a n audi-ence to adopt specific ways an d means ofaverting a threat, the use of a strong fearappeal, as against a milder one, increases thelivelihood that the audience will be left in astate of emotional tension which is not fullyrelieved by rehearsing the reassuring recom-mendations contained in the communication.This hypothesis is compatible with the gen-eral assumption that when a person is exposedto signs of "threat," the greater the intensity

7/30/2019 Janis & Feshback (1953)

13/15

90of the fear reaction evoked, the greater thelikelihood that his emotional tension willpersist after the external stimulus hasterminated.Whether or not the above hypothesis iscorrect, the fact remains that "unreduced"emotional tension was manifested immedi-ately after th e communication predominantlyby the group exposed to the Strong appeal.Our findings on subsequent reactions providesome suggestive evidence concerning theconsequences of experiencing this type ofresidual tension. In general, th e evidenceappears to be consistent with the followinghypothesis: When fear is s trongly arousedbut is not fully relieved by the reassurancescontained in a ma ss communication, theaudience will become motivated to ignore orto minimize the importance of the threat .This hypothesis could be regarded as aspecial case of the following general propo-sition which pertains to the effects of huma nexposure to any fear-producing stimulus:other things being equal, th e more persistentthe fear reaction, the greater will be the(acquired) motivation to avoid subsequentexposures to internal and external cues whichwere present at the time the fear reaction wasaroused. This proposition is based on thepostulate that fear is a stimulus-producingresponse which has the functional propertiesof a drive (2 , 7).88 In the sphere of human communication, the keytheoretical assumption could be formulated as follows:If rehearsal of the reassuring statements contained in acommunication fails to alleviate th e emotional tensionelicited by the use of a fear appeal, the audience willbe motivated to continue trying out other (symbolic orovert) responses until one occurs which succeeds inreducing fear to a tolerable level. Thus, a strong fearappeal which is intended to motivate th e audience totake account of a realistic threat of danger could havethe paradoxical effect of motivating the audience toignore the threat or to adopt "magical," "wishful" orother types of reassuring beliefs that are antithetical toth e comm unicator's intentions. Moreover, according toth e same theoretical assumption, when a communicationproduces a high degree of persistent fear, the audiencewill be motivated to engage in overt escape activities,some of which m ay prove to be incompatible with theprotective actions recommended by the communicator.Unintended effects of this kind can be regarded asspontaneous "defensive" reactions which are motivatedby residual emotional tension. In the present experi-ment , it would be expected that, in addition to thetendency to avoid thinking about th e threat, otherdefensive reactions would also occur. For example,fol lowing exposure to the Strong appeal, some of thestudents may have succeeded in alleviating their residualemotional tension through spontaneous interpersonalcommunication with fellow students.

I R V I N G L . J A N I S A N D S E Y M O U R F E S H B A C HIn the context of the present experiment,one would predict that th e group displayingthe greatest degree of residual fear would bemost strongly motivated to ward off thoseinternal symbolic cuessuch as anticipations

of the threatening consequences of improperdental hygienewhich were salient duringand immediately after the communication.This prediction seems to be fairly well borneout by the evidence on carry-over effects, par-ticularly by the finding that the greatestdegree of resistance to the subsequent coun-terpropaganda was shown by the groupwhich had been least motivated by fear. Theuse of the Strong appeal, as against theMinimal one, evidently resulted in less rejec-tion of a subsequent communication whichdiscounted and contradicted what was saidin the original communication. In effect, th esecond communication asserted that onecould ignore the alleged consequences ofusing th e wrong type of toothbrush, and, inthat sense, minimized the dangers which pre-viously had been heavily emphasized by thefear-arousing communication.T he results obtained from th e students'reports on their dental hygiene practices couldbe interpreted as supporting another predic-tion from the same hypothesis. It wou ld beexpected that those students who changedtheir practices, after having heard and seenone of the three forms of the illustrated talk,were motivated to do so because they recalledsome of the verbal material which had beengiven in support of the recommendations,most of which referred to the unfavorableconsequences of continuing to do the"wrong" thing. In theoretical terms, onemight say that their conformity to the recom-mendations was mediated by symbolic re-sponses which had been learned during th ecommunication. The mediating responses(anticipations, thoughts, or images) acquiredfrom any one of the three forms of the illus-trated talk would frequently have, as theircontent, some reference to unpleasant conse-quences for the self, an d consequently wouldcue off a resolution or an overt action thatwould be accompanied by anticipated successin warding off the threat. But defensiveavoidance of the mediating responses wouldreduce the amount of conformity to whateverprotective action is recommended by the

7/30/2019 Janis & Feshback (1953)

14/15

E F F E C T S O F F E A R - A R O U S I N G C O M M U N I C A T I O N S 91fear-arousing communication. Hence theprediction would be that when rehearsal ofstatements concerning potential danger isaccompanied by strong emotional tensionduring and after the communication, theaudience will become motivated to avoid re -calling those statements on later occasionswhen appropriate action could ordinarily becarried out. An inhibiting motivation ofthis kind acquired from th e illustrated talkwould tend to prevent th e students fromadopting the recommended changes in theirtoothbrushing habits because they would failto think about th e unpleasant consequencesof improper dental hygiene at times whenthey subsequently perform the act of brush-ing their teeth.Much more direct evidence in support ofthe "defensive avoidance" hypothesis comesfrom the analysis of spontaneous write-inanswers in which the students explained whythey disagreed with th e counterpropaganda(Table 8). Those who had been exposed tothe least amount of fear-arousing materialwere the ones who were most likely to referto the illustrated talk as an authoritativesource and to make use of its arguments.The relative absence of such references inthe spontaneous answers given by those whohad been exposed to the Moderate andStrong appeals implies a tendency to avoidrecalling the content of the fear-arousingcommunication.Although the various pieces of evidencediscussed above seem to fit together, theycannot be regarded as a conclusive demon-stration of the defensive avoidance hypothe-sis. What our findings clearly show is thata strong fear appeal can be markedly lesseffective than a minim al appeal, at least unde rthe limited conditions represented in ourexperiment. Exactly which conditions andwhich mediating mechanisms are responsiblefor this outcome will remain problematicaluntil further investigations are carried out.Nevertheless, so far as the present findingsgo, they consistently support the conclusionthat the use of a strong fear appeal will tendto reduce the over-all success of a persuasivecommunication, if it evokes a high degree ofemotional tension without adequately satisfy-ing the need fo r reassurance.

S U M M A R Y A N D C O N C L U S I O N SThe experiment was designed to investigatethe effects of persuasive communicationswhich attempt to motivate people to conformwith a set of recommendations by stimulating

fear reactions. An illustra ted lecture ondental hygiene was prepared in three differ-ent forms, representing three different inten-sities of fear appeal: the Strong appealemphasized and graphically illustrated th ethreat of pain, disease, an d body damage; theModerate appeal described the same dangersin a milder and more factual manner; theMinimal appeal rarely referred to the un-pleasant consequences of improper dentalhygiene. Although differing in the amountof fear-arousing material presented, the threeforms of the communication contained thesame essential information and the same setof recommendations.Equivalent groups of high school studentswere exposed to the three different forms ofthe communication as part of the school'shygiene program. In addition, th e experi-ment included an equated control groupwhich was not exposed to the dental hygienecommunication but wa s given a similarcommunication on an irrelevant topic. Alto-gether there w ere 200 stud ents in the experi-ment, with 50 in each group. A questionnairecontaining a series of i tems on dental hygienebeliefs, practices, an d atti tudes was adminis-tered to all four groups one week before th ecommunications were presented. In order toobserve the changes produced by the illus-trated talk, postcommunication question-naires were given immediately after exposurean d again one week later.1. The fear appeals were successful inarousing affective reactions. Immediatelyafter the communication, the group exposedto the Strong appeal reported feeling moreworried about the condition of their teeththan did the other groups. The M oderateappeal, in turn, evoked a higher incidence of"worry" reactions than did the Minimalappeal.2. The three forms of the illustrated talkwere equally effective with respect to (a)teaching the factual content of the com-munication, as assessed by an informationtest, an d (b) modifying beliefs concerningfour specific characteristics of the "proper"

7/30/2019 Janis & Feshback (1953)

15/15

I R V I N G L . J A N I S A N D S E Y M O U R F E S H B A C Htype of toothbrush. The evidence indicatesthat th e emotional reactions aroused by theStrong appeal did not produce inattentivenessor reduce learning efficiency.3. As compared with the other two formsof the communication, the Strong appealevoked a more mixed or ambivalent attitudetoward th e communication. T he studentsexposed to the Strong appeal were morelikely than the others to give favorable ap-praisals concerning th e interest value and thequality of the presentation. Nevertheless,they showed the greatest amount of subjectivedislike of the communication and made morecomplaints about the content.4. From an analysis of the changes in eachindividual's reports about his current tooth-brushing practices, it was found that th egreatest amount of conformity to the com-municator's recommendations was producedby the Minimal appeal. The Strong appealfailed to produce any significant change indental hygiene practices, whereas the Mini-mal appeal resulted in a reliable increase inconformity, as compared with the Controlgroup. Similar findings also emerged froman analysis of responses which indicatedwhether th e students had gone to a dentistduring th e week following exposure to theillustrated talk, reflecting conformity toanother recommendation made by the com-municator. The evidence strongly suggeststhat as the amount of fear-arousing materialis increased, conformity to recommended(protective) actions tends to decrease.5. One week after the illustrated talk hadbeen presented, exposure to counterpropa-ganda (which contradicted th e main themeof the original communication) produced agreater effect on attitud es in the Controlgroup than in the three experimental groups.The Minimal appeal, however, proved to bethe most effective form of the illustrated talkwith respect to producing resistance to thecounterpropaganda. The results tend to sup-port th e conclusion that under conditions

where people are exposed to competing com-munications dealing with the same issues, theuse of a strong fear appeal is less successfulthan a minimal appeal in producing stableand persistent attitude changes.6. The main conclusion which emergesfrom the entire set of findings is that theover-all effectiveness of a p ersuasive com-munication will tend to be reduced by theuse of a strong fear appeal, if it evokes ahigh degree of emotional tension without ade-quately satisfying th e need fo r reassurance.T he evidence from th e present experimentappears to be consistent with the followingtwo explanatory hypotheses:a. When a mass communication is de-

signed to influence an audience to adoptspecific ways an d means of averting a threat,the use of a strong fear appeal, as against amilder one, increases the likelihood that th eaudience will be left in a state of emotionaltension which is not fully relieved by rehears-ing the reassuring recommendations con-tained in the communication.b. When fear is strongly aroused but is notfully relieved by the reassurances containedin a mass communication, the audience willbecome motivated to ignore or to minimizethe importance of the threat.

REFERENCES1. A L E X A N D E R , F., & FRENCH, T. M. Psychoanalytictherapy. New York: Ronald, 1946.2. D O L L A R D , J. , & MILLER, N. E. Personal i ty an d psy-chotherapy. New York: McGraw-Hill , 1950.3. FENICHEL, O. Problems of psychoanalyt ic technique.New York : Psychoanalytic Q uarterly, 1941.4. H A N F M A N N , E U G E N I A . Psychological approaches tothe study of anxiety. In P. H. Hoch and J.Zubin (Eds.), Anxiety. New York: Grune &Stratton, 1950. Pp. 51-69.5. HOVLAND, C. I., LUMSDAINE , A. A., & SHEFFIELD ,F. D. Experiments on mass communication.Princeton: Princeton Univer. Press, 1949.6. J A N I S , I . L. Air war and emotional s tress . NewYork: McGraw-Hill , 1951.7 . M O W R E R , O. H. Learning theory an d personal i tydynamics: Selected papers . New York: Ronald,1950.

Received March 14, 1952.