,.

Det'elopm,ntal Psychology1973, Vol. 6, No. 3, 469-475

of personal time in TATstories, Joumal of Pro-

f Personality Assessmem, Some Immediate Effects of Televised Violence onChildren's Behavior1Ily 13, 1971)

ROBERTM. LIEBER'f2

State Universityof New Yorkat Stony Brook

ROBERT A. BARON

Purdue University

The hypothesis that exposure to televised violence would increase thewillingness of children to hurt another child was investigated. Boys andgirls of two age groups (5-6 and 8-9 years) first viewed excerpts fromactual television programs depicting either aggresive or nonaggressivescenes, and were then provided with an opportunity to aggress againsta peer. All subjects were subsequently placed in a free play situation andthe frequency of their aggressive responses observed. Results indicatedthat children exposed to the aggressive program engaged in longer attacksagainst an ostensible child victim than.subjects exposed to the nonaggres-sive program. The aggressive program also elicited a higher level of ag-gressive play than the nonaggressive one, particularly among the youngerboys.

~-_. - -... his review of the SOCIaland sCIentificis- .

sues surrounding the portrayal of violence inthe mass media, Larsen (1968) noted thatwe may begin with two facts: "(1) Massmedia content is heavily saturated with vio-lence, and (2) people are spending more andmore time in exposure to such content [po115]," This state of affairs has been used byboth laymen and professionals as the basisfor appeals to modify the enLertainment fareto which viewers, particularly children andadolescents, are exposed (Merripm, 1964;Walters, 1966; Walters & Thomas, 1963;Wertham, 1966). Other writers, however,have argued that the kind of violence foundon television or in movies does not necessar-ily influence observers' "real-life" social be-havior (Halloran, 1964; Klapper, 1968). Afew have even characterized the portrayal of

1 This study was supported, in part, by NationalInstitute of Mental Health Contract HSM-42-70-38 and Was conducted while both authors were af-filiatedwith the Fels Research Institute, Gratefulacknowledgementis due to Robert Devine, JoanKleban, Diane E. Liebert, Carol Lyons, CherylRussell,and Sharon Swenson for their many con-tributions. The contributions of the co-authors totheproject were approximately equal.

2 Reprints may be obtained from Robert M.Liebert, Department of Psychology, State Univer-Sityof New York, Stony Brook, New York 11790.

violence as potentially preventing the overtexpression of aggression, at least under somecircumstances (Feshbach, 1961; Feshbach &Singer, 1971).

In view of the controversy, it is hardlysurprising that recent years have seen a sub-stantial increase in the number of experi-mental studies directed to this issue. An ef-fort has been made to determine whetherchildren will learn and/or be disinhibited intheir performance of aggressive acts as afunction of expospre to symbolic aggressivemodels (e.g., in cartoons, movies, stories,and simulated television programs). This re-search has indicated consistently that chil-dren may indeed acquire, from even a verybrief period of observation, certain motoricand verbal behaviors which are associatedwith aggression in life situations. More spe-cifically, it has been repeatedly shown thatafter viewing a film which depicts novelforms of hitting, kicking, and verbal abuse,children can, when asked to do so, demon-strate this learning by reproducing these pre-viously unfamiliar behaviors with a remark-able degree of fidelity (Bandura, 1965;Hicks, 1965). Taken together with the largebody of research on the observational learn-ing of other behaviors (Flanders, 1968), theavailable evidence appears to leave little

469

i

: 'i: Ii'

.' I,

Ii

{j "

ij,'Hlli'lh .,!FC'i[,:,";'tr/t~:i

J\t:

:'.

:

.r';

~

"

.;,

i:, I,.; .:1"."

. ,r

i!

:,

'

11'1

:1'HI

iiIi

" !,I;

470 ROBERT M. LIEBERT AND ROBERT A. BARON

("'

doubt that the learning of at least someaggressive responses can and does resultfrom television or movie viewing.

Equally important, however, is the ques-tion of whether the observation of violencewill influence children's performance of ag-gressive acts when they have not been spe-cifically asked to show what they have seenor learned. Several experiments appear toprovide evidence relating to this issue (Ban-dura, Ross, & Ross, 1961, 1963a, 1963b;Rosekrans & Hartup, 1967). In these stud-ies, subjects have typically been exposed tolive or filmed aggressive scenes, then placedin a free play situation with a variety of toysor other play materials. Results obtainedwith these procedures have shown repeatedlythat the exposure of young children to ag-gression produces increments in such playactivities as punching inflated plastic clowns,popping balloons, striking stuffed animals,and operating mechanized "hitting dolls."

It has been argued by critics (Klapper,1968) that findings such as those reviewedabove are not directly relevant to the ques-tion of whether exposure to televised aggres-sion will increase children's willingness toengage in behavior which might actuallyharm another person. Since this criticismwas advanced, a human victim has replacedthe inanimate target in at least four more re-cent investigations (Hanratty, 1969; Han-ratty, Liebert, Morris, & Fernandez, 1969;Hanratty, O'Neal, & Sulzer, 1972; Savitsky,Rogers, Izard, & Liebert, 1971). '])lese laterstudies have demonstrated clearly that expo-sure to the behavior of filmed aggressivemodels may lead young children to directlyimitate aggression against a human, as wellas a "toy," victim.

Despite the newer evidence, critics maystilI question whether exposure to the type ofviolence generally depicted on regularlybroadcast television shows will produce simi-lar effects. Likewise, it is important to con-sider the possible disinhibitory effects (cf.Lovaas, 1961; Siegel, 1956) rather thanonly the direct imitative effects of observingaggressive models. Although such effectshave previously been observed with adultsubjects and violent scenes taken from mo-tion pictures (e.g., Berkowitz, 1965; Ber-kowitz & Rawlings, 1963; Walters &

,j! ;~

1'.,.,1"to!:

~;':t!

,

,

"'

,

' ' !' ,., !c, , '

~'4nli'j

H

i

li. i

~-! ! I

~.4 I; ~ .;

r,::,

',

.,j,

'

,"

!

,

'

,,

';t ,'!I'i "':,! I

: ;"1,i! Ii I"!; ~iI<j. 'Ii

Ii.,.,Ii! '

,_!

~

;

,

:,' ~",

U f:,; ,j'

1

'

,,

'~' . " ,

i( :,

:",

:1

1

:I::I:t ' iI iI '

"'"

, "

.. ", <"

,,"

~ " 'I" ";' t

YhiH

I'

,

>

",

'F,

:

,j'""""~ i.."

,n,H'%~'!

~il

. .~ ~

Thomas, 1963), in no previous investigationknown to the authors has the influence oftelevised violence on interpersonal aggres-sion been examined for young children. Itwas with these latter questions that the pres-ent research was primarily concerned. Wesought to determine whether exposure to vi-olent scenes taken directly from nationallytelecast programs increases the willingness ofyoung children to engage in aggressive actsdirected toward another child.

Method

Participants

Population sampled. The sample was drawnboth from Yellow Springs, Ohio, a small collegetown, and from a larger and more conservativeneighboring community, Xenia. The participantswere brought to Fels Research Institute in Yellow!Springs by one of their parents, in response to a .,.newspaper advertisement and/or a letter distributed ~.

in local public elementary schools asking for ~volunteers to participate in a study of the effects '~of television on children. To assure that no poten- ;tial participants were turned away because of ~scheduling inconveniences, parents were invited to~:select their own appointment times (including eve- '*nings or weekends), and transportation was of.

J

:

fered to

,

those who could not provide it for them-

,',

'"

,.

'

,

selves. -.' ,Subjects. The subjects were 136 children, 68 '

boys and 68 girls. Sixty-five of the participantswere 5 or 6 "ears of age at the time of the study; ,

the remaining 71 subjects were 8 or 9 years of ,~age. Within each age group and sex the children ...were assigned randomly to the treatment condi- , '

tions. Approximately 20% of the children in thisstudy were black; virtually all of the remainder ~were white. The economic backgrounds from'which these participants came was widely varied. .Although economic characteristics were not used .

as a basisfor assignmentto treatments,inspection~

suggested that the procedure of random assign.;ment had adequately distributed them among the ':experimental groups. ;j

Experimental personnel. One of the investiga- 'Jtors greeted the parent and child at the outset, iserved as the interviewer, and obtained informed ~parental consent for the child's participation. A ~

,

".,28-year-old white female served as experimenter $'for all the children, and two other adult femal~s ~served as unseen observers throughout the expen- ~

ment. ~r.(

D . .eSlgn ".,.""~'

A 2 X 2 X 2 factorial design was employed. ;~The three factors were sex, age (5-6 or 8-9 years f:

old) ,and treatment (observation of aggressive or 1,

1

nonaggressive television sequences). i I

~

rrocediInfroa

llf parerc~corted,ive malrarent v.the interJi~closedan~werellicipatior

Expeninterviev.the chilleach subwntaininvideo-tapon by t~child watfor him,televisionthe subj~throughFor all !consistcdduring e~effectiven~econd al'nlc comand atten

Thereaunserveda popularThc sequ(containedshootings,thc contrminute vi,letes comthc like. ]thc progr:bile tires,thc expernounced tl

Assess"The subjecfrom theseated at :the one (( 1966). 1measured

3 Sincevided in ~phone, it i

company tassure thaformed cothe costs 0who remainate econoappeared ithe parentpeared forher child tc

::>N

10 previous investigationJrS has the influence of10 interpersonal aggres-I for young children. Itr questions that the pres-rimarily concerned. Wewhether exposure to vi-directly from nationallycreases the willingness ofngage in aggressive actscher child.

\ethod

The sample was drawndngs, Ohio, a small college,'ger and more conservativety, Xenia. The participantsResearch Institute in Yellow~ir parents, in response to aent and/or a letter distributed.entary schools asking forlte in a study of the effectsen. To assure that no poten-~ turned away because oflces, parents were invited tontment times (including eve-and transportation was of-

mid not provide it for them-

:cts were 136 children, 68ixty-five of the participantsage at the time of the study;jects were 8 or 9 years ofgroup and sex the childrenlly to the treatment condi-10% of the children in this

"tually all of the remainder:onomic backgrounds fromIts carne was widely varied.haracteristics were not usedlent to treatments, inspection'ocedure of random assign-distributed them among the

'me!. One of the investiga-ot and child at the outset,.ver, and obtained informedhe child's participation. Aale served as experimenterod two other adult females'vers throughout the experi-

:orial design was employed.e sex, age (5-6 or 8-9 years

,observation of aggressive or1 sequences).

I,lI

I

I,II

EFFECTS OF TELEVISED VIOLENCE ON CHILDREN'S BEHAVIOR

Procedure

Introduction to the situation. Upon the arrivalof parent and child at the institute, the child wasescorted to a waiting room containing nonaggres-sive magazines and other play materials while theparent was interviewed in a separate room. Duringthe interview, the nature of the experiment wasdisclosed to the parent, questions were invited andanswered, and a written consent to the child's par-ticipation was obtained.3

Experimental and control treatment. After theinterview, but without permitting the parent andthe child to interact, the experimenter escortedeach subject individually to a second waiting roomcontaining children's furniture and a televisionvideo-tape monitor. The television was then turnedon by the experimenter, who suggested that thechild watch for a few minutes until she was readyfor him. The experimenter left the child to watchtelevision alone for approximately 6~ minutes;the subjects were in fact continuously observedthrough a concealed camera and video monitor.For all groups, the first 120 seconds of viewingconsisted of two I-minute commercials video-tapedduring early 1970. The first of these depicted theeffectiveness of a certain paper towel, and thesecond advertised a humorous movie (rated G).The commercials were selected for their humorand attention-getting characteristics.

Thereafter, children in the experimental groupobserved the first 3~ minutes of a program froma popular television series, "The Untouchables."The sequence, which preserved a simple story line,contained a chase, two fist-fighting scenes, twoshootings, and a knifing. In contrast, children inthe control group viewed a highly active 3~.minute video-taped sports sequence in which ath-letes competed in hurdle races, high jumps, andthe like. For all subjects, the final 60 seconds ofthe program contained a commercial for automo-bile tires. Before the end of this last commercial,the experimenter reentered the. room and an.nounced that she was ready to begin.

Assessment of willingness to hurt another child.The subject was next escorted by the experimenterfrom the television room to a second room andseated at a response box apparatus modeled afterthe one employed by Mallick and McCandless(1966). The gray metal response box, whichmeasured approximately 17 X 6 inches, displayed

471

a red button on the left, a green button on theright, and a white light centered above these twomanipulanda. The word "hurt" appeared beneaththe red button, while the word "help" appearedbeneath the green button. Several plastic wires ledfrom the response box to a vent in the wall. Theexperimenter explained to the subject that thesewires were connected to a game in an adjacentroom and that "one of the other children is in thenext room right now and will start to play thegame in just a minute." She further explained thatthe game required the player in the other room toturn a handle and that the white light would comeon each time the other child in the next roomstarted to turn the handle, thus activating the redand green buttons.

The experimenter continued:

When this white light comes on, you haveto push one of these two buttons. If youpush this green button, that will make thehandle next door easier to turn and willhelp the child to win the game. If you pushthis red button, that will make the handlenext door feel hot. That will hurt the child,and he will have to let go of the handle.Remember, this is the help button, and thisis the hurt button [indicating]. See, it sayshelp and hurt. . . . You have to push one ofthese two buttons each time the light goeson, but you can push whichever one youwant to. You can always push the same'button or you can change from one buttonto the other whenever you want to, but justremember, each time the light goes on, you.can push only one. So if you push this

." green button then you help the other childand if you push this red button then you hurtthe other child. Now if you push this greenbutton for just a second, then you help theother child just a little, and if you push thisred button down for just a second, then youhurt the other child just a little. But if youpush this green button down a little longer,then you help the other child a little more,and if you push this red button down a littlelonger, then you hurt the other child a littlemore. The longer you push the green but-ton, the more you help the other child andthe longer you push the red button, the moreyou hurt the other child.

This explanation, with slightly varied wording,was repeated a second time if the child did not in-dicate comprehension of the instructions. Afterbeing assured that the subject understood the task,the experimenter left the room.4

4Nine children, all in the 5-6-year-old agegroup, were terminated prior to the collection ofdata because they refused to remain alone, cried,or left the experimental situation. Twenty-threeother children participated in the entire experi-ment but were not included in the sample. Ofthese, 14 (5 in the younger age group and 9 in the

,I

f

I

I

I

{

I

II

3 Since no specific information could be pro-vided in public announcements or over the tele-phone, it appeared necessary to have parents ac-company their children to the institute in order toaSsure that no child participated without the in-formed consent of his parents. In order to defraythe costsof transportation,babysitters for siblingswho remained at home, and the like, and to elimi-nate economic biases which might otherwise have.appeared in the sample, a $10 stipend was giventhe parent of each participant. No parent who ap-peared for the interview declined to allow his orher child to participate.

~ ~

i'\

472 ROBERT M. LIEBERT AND ROBERT A. BARON

Results

'i

Although all the subjects were led to believethat other children were participating, there was,in fact, no other child: the entire procedure wascontrolled in the next room so as to produce 20trials, with an intertrial interval of approximately15 seconds. Each child's response to each trial(appearance of the white light) and the durationof the response, recorded to the hundredth of asecond, was automatically registered. When thesubject had completed 20 trials, the experimenterreentered the room and announced that the gamewas over.

Assessment of aggressive play. The influence oftelevised violence on the children's subsequentplay activities was also explored, although thisissue was of secondary interest in the present re-search (the study being primarily concerned withinterpersonal aggression rather than aggressionaimed at inanimate objects). After completing thebutton-pushing task, the child was escorted to athird room (designated the "play roo,m") acrossthe hallway. The room contained two large tables,on each of which appeared three attractive non-aggressive toys (e.g., a slinky, a cookset, a space-station) and one aggressive toy (a gun or a knife).Two inflated plastic dolls, 36 inches and 42 inchesin height, also stood in the room. The child wastold that he would be left alone for a few minutesand that he could play freely with any of the toys.

All the children were observed through a on'e-way vision mirror, and their aggressive behaviorwas recorded using a time-sampling procedure.One point was scored for the occurrence of eachof three predetermined categories of aggressiveplay (playing with the knife, playing with the gun,assaulting either of the dolls) during the first 10seconds of each of ten ~-minute periods. In orderto assess interobserver reliability for this measure,10 subjects were observed independently by thetwo observers. Their agreement using the scoringprocedures was virtually perfect (r = .99).

At the end of the play period, the experimenterreentered the rootn and asked the child to recallboth the television program which he had seenand the nature of the game he had played. (Allchildren included in the analyses were able to re-call correctly the operation of the red and greenbuttons and the essential content of the televisionprograms to which they had been exposed.) Thechild was then escorted to the lounge where theparent was waiting, thanked for his or her p.1lrtici-pation, rewarded with a small prize, and asked notto discuss the experiment with his or her friends.

Ii,;

fii.~f'

jl;

I

~

li; ;

I, ! ;I' ,II '.,

i i .~ ~ .i

t

': i, I

" , ,

" "I

.. '" ,

I , \ I',: I;":, ii

iii.

'

~

"tr'""'; :;..

r i', .<t ;;'- '

I

J

'.. ,1 ~:.-, "" "

r, I

~

~'i I!I.' j If; , i,~~ ' .t

.,

'.

',

' i i

J.~J.I '

U." I,ftIf,f

H'

'[

older group) did not understand or follow instruc-tions for the response box; 7 (3 younger and 4older children) played or explored the room in-stead of watching television. The data for the re-maining 2 children were not recorded properlydue to the technical difficulties. All potential par-ticipants brought to the institute by their parentswho were not eliminated for the reasons listedabove were included in the experimental sample.

~

Willingness to Hurt Another Child

The single overall measure which appearsto capture the greatest amount of informa-tion in this situation is the total duration inseconds of each subject's aggressive reosponses during the 20 trials. Since markedheterogeneity of variance was apparentamong the groups on this measure, the over-all 2 x 2 x 2 analysisof variance was per.formed on square-root transformed scores(Le., x = Vi + vx+T; Winer,1962).The means for all groups on this measureare presented in Table 1. The analysis itselfreveals only one significant effect: that fortreatment conditions (F = 4.16, p < .05).Children who had observed the aggressiveprogram later showed reliably more willing-ness to engage in interpersonal aggressionthan those who had observed the neutralprogram. ,

Several supplementary analyses, whichmay serve to clarify the nature of this overalleffect, were also computed. For example, a""subject's total duration score may be viewed -as the product of the number of times he ag- '~

gresses and t?e average duration of each o!."these aggressIve responses. Moreover, these..- ' .~two measures are only moderately, althougb .reliably,relatedin the ?veralls.ample(r ~t+ .30, p < .05). AnalysIsof vanance for the 5.average duration of the hurt responses r;:veals only a significantprogram effect that.directly parallelsthe effectsfor total duration~(F = 3.95, P. < .05). The means for ~.

groups on tillS measure are preseuted ,~. TABLE 1 . :,.

MEAN TOTAL DURATION (TRANSFORMED) OF Aa- 21

GRESSIVE RESPONSES IN Au. GROUPS 'i~f

';1"

~j.~f~

Table 2quency Jeffects,younger

Helping

One]total aggsive prowere simgressivethis interiance wathe helJ:square-r<Presumalthe effec:sive proghelp scmgroups. '

arousal Iments onoverall Fsubjects'the non aonly 1.1inificancewas a Pre=3.91, J3, in whi<teractionresponsesaggressiv(helping r.who saw t

As a S(

MEAN AVENUMBER OF

Program she

~~..

AggressiveNonaggressi'

4 Nole.-The rthe same as tha

.-,.

5-6-year-olds 8-9-year-olds

Program shown

Boys Girls Boys Girls

Aggressive 9.65 8.98 12.50 8.53N 15 18 20 17

Nonaggressive 6.86 6.50 8.50 6.27N 15 17 18 16

'N

suits

4notlzer Child

measure which appears'st amount of informa-is the total duration inlbject's aggressive re-O trials. Since markedIriance was apparentthis measure, the over-.is of variance was per-'ot transformed scores

'x + I, Winer, 1962).roups on this measuree I. The analysis itselflificant effect: that for(F = 4.16, p < .05).bserved the aggressive! reliably more willing-1terpersonal aggression

observed the neutral

ntary analyses, which'le nature of this overallputed. For example, an score may be viewednumber of times he ag-

ge duration of each of {onses. Moreover, these.v moderately, although

I

;: overall sample (r = '

ysis of variance for thethe hurt responses ,re- " .

It program effect that}fects for total duration I). The means for all "

,ure are presented in

\LE 1

(TRANSFORMED) OF AG-i IN ALL GROUPS

EFFECTS OF TELEVISED VIOLENCE ON CHILDREN'S BEHAVIOR

I

I,I!

Table 2. In contrast, analysis of the fre-quency measure fails to show any significanteffects, although the tendency for theyounger children is in the same direction.

Helping Responses

One possible explanation of the highertotal aggression scores shown by the aggres-sive program group is that these childrenwere simply more aroused than their nonag-gressive treatment counterparts. To check onthis interpretation, an overall analysis of var-iance was performed on the total duration ofthe help responses, employing the samesquare-root transformation described above.Presumably, if general arousal accounted forthe effects of the hurt measure, the aggres-sive program groups should also show largerhelp scores than the nonaggressive programgroups. However, contrary to the generalarousal hypothesis, the effect of the treat-ments on this measure was not significant; theoverall F comparing the aggressive programsubjects' prosocial responses with those ofthe nonaggressive program observers wasonly 1.17. The one effect of borderline sig-nificance which did appear in this analysiswas a Program X Sex X Age interaction CF= 3.91,p == .05). As can be seen in Table3, in which these data are presented, the in-teraction results from the very large helpingresponses shown by older girls who saw theaggressive program and the relatively largehelping responses shown by younger girlswho saw the nonaggressive one.

As a second check on the possibility that

473

. TABLE 3

MEAN TOTAL DURATION (TRANSFORMED) OF HELPINGRESPONSES IN ALL GROUPS

Note.- The number of subjects for each cell in this analysis isthe same as that shown in Table 1.

the longer durations in the aggressive pro-gram groups simply reflected a generalarousal, a similar analysis was performed onthe average duration scores of the help re-sponses. In contrast to the comparable meas-ure for aggressive responses, no significantdifferences for any of the main effects or in-teractions appeared on this measure C maineffect for treatments, F = 1.24) although

r paralleling the total duration measure, theolder girls who saw the aggressive programshowed particularly long average durations.Finally, to show from a correlational ap-proach that the overall help and hurt scoreswere 'riot merely alternate measures of thesame phenomenon, the product-moment cor-relation betwi:en the two sets of scores wascomputed. The resulting r of - .24 reflects aweak but significant CP < .05, two-tailed)negative relationship. Thus, overall, it ap-pears clear that a specific disinhibition re-garding aggressive behavior was produced byobserving the televised aggression. This can-not be explained as a general arousal effect.

Aggression in the Play Situation

The mean aggressive play scores for allsubjects are presented in Table 4. A 2 X 2x 2 analysis of variance of these data re-vealed significant main effects for treatment(F = 8.01, df = 1/128, p < .01) andsex CF = 37.87, df = 1/128, p < .001).In addition, the Treatment X Sex CF =4.11, df = 1/128, p < .05), Treatment XAge CF= 4.28, df = 1/128, p < .05), andTreatment X Sex x Age CF= 4.68,df =1/128, p < .05) interactions were all sig-

: ,iI

, I: I

i I

, I

,

),

:/'

~::-.

1",

Ii

,

i

I!~

:i

i ~: i

j:hI-!i

1

./"

.,

,

'

,;

I

;:!i

II

i,

:')'n

I:;!'i Ii

, ,!

I

I

. .

I

I

TABLE 2

MEAN AVERAGE DURATIONS (TOTAL DURATIONjNUMBER OF HURT RESPONSES) OF AGGRESSIVE RE-

SPONSESIN ALL GROUPS

NOle.-The number of subjects for each cell in this analysis isthe same as that shown in Table 1.

'-aIds 8-9-year-olds

Girls Boys I Girls I8.98 12.50 8.53 I

18 20 17

I6.50 8.50 6.2717 18 16 I

5-6-year-olds 8-9-year-olds

Program shown

Boys Girls Boys Girls

Aggressive 10.81 11.66 11.32 19.97Nonaggressive 10.76 14.12 11.59 10.69

5-6-year-olds 9--8-year-olds

Program shown

Boys Girls Boys Girls

Aggressive 3.42 2.64 5.18 3.07Nonaggressive 2.55 2.09 2.07 1.57

474 ROBERT M. LIEBERT AND ROBERT A. BARON

TABLE 4

MEAN NUMBER OF Tl~IE-SAMPLED AGGRESSIVE PLAYRESPO:>:SESIN ALL GROUPS

Note.- The number of subjects for each cell in this analysis isthe same as that shown in Table 1.

, :: i!< \f . nificant. As is apparent from inspection of

Table 4, these interactions arose from thefact that, although children exposed to theaggressive program tended to show a higherlevel of aggressive play than children ex-posed to the nonaggressive one in all simplecomparisons, the effect was much greater forthe younger boys than for any of the remain-ing groups.

.. i""

. ""'"~;j~fi

f',1ttjl1.

,

:,

"'~

, ,,

"

,,

' ' I..,",~ ~,

d,tt; ,Hi, ,

!

'"

Discussion

The overall results of the present experi-ment provide relatively consistent evidencefor the view that certain aspects of a child's ..

willingness to aggress may be at least tempo-rarily increased by merely witnessing ag-gressive television episodes. These findingsconfirm and extend many earlier reports re-garding the effects of symbolically modeledaggression on the subsecp.1entimitative ag-gressive behavior of young observers towardinanimate objects (e.g., Bandura, Ross, &Ross, 1963a; Hicks, 1965; Rosekrans &Hartup, 1967). Likewise, the present dataare in accord with other studies which haveshown disinhibition of both young children'saggressive play and older viewers' willing-ness to shock another person after observingfilmed aggressive modeling. As in many earl-ier studies, subjects exposed to symbolic ag-gressive models regularly tended to behavemore aggressively than control group sub-jects tested under identical circumstances.Further, the present results emerged despitethe brevity of the aggressive sequences (lessthan 4 minutes), the absence of a strongprior instigation to aggression, the clear

. .

availability of an alternative helping re.-sponse, and the use of nationally broadcastmaterials rather than specially prepared lab-oratory films.

The various measures employed, consid.ered together, provide some clarification ofthe nature of the effects obtained in the over.all analysis. The significant effect for thetotal duration measure appears to stem pre.dominantly from the average duration of thesubjects' aggressive responses. In fact, asseen in Table 2, the group means on thismeasure did not overlap; the lowest individ.ual cell mean among those who observed theaggressive program was higher than thehighest mean among those groups who ob-served the nonaggressive program. .

It should also be recalled that the instruc.tions given to all children emphasized that abrief depression of the hurt button wouldcause only minimal distress to the otherchild, while longer depressions would causeincreasingly greater discomfort. This fac~coupled with the finding that the overall av-erage duration of such responses was morethan 75% longer in the aggressive programgroup than in the control group, suggestsclearly that the primary effect of exposure to>:the aggressive program was that of reducing"';'

.. .subjects' restraints against inflicting severe~~

discomfort on the ostensible peer victim.'1that is, of increasing the magnitude of the~!hurting response. With the exception of the~.""'1older girls, this effect was not paralleled by:,an increment in the corresponding measures,r.:

of helping; thus it cannot be attributed t~J"

.

simple arousal effects. .iIIt should be noted that the measureof.:J

aggressive play responses was obtained,~after all the subjects had been given an opo4portunity to help or hurt another child. Thus .,ithe observed effects might reflect an inter-.,.'action between the programs and some as-,lpect of the hurting/helping opportunity}rather than the simple influenceof the pro- "

grams themselves. While the present datado not permit us to address the possibility ofsuch interactions directly, it is clear that theobtained results are consistent with earlierstudies in which other types of aggressivescenes were used and where there were nosuch intervening measures.

The present experiment was designed pri- .,

5-6-year-olds 8-9-year-olds

Program shown

Boys Girls Boys Girls

Aggressive 7.13 2.94 5.65 3.00Nonaggressive 3.33 2.65 5.39 2.63

IARON

an alternative helping reouse of nationally broadcastthan specially prepared lab-

neasures employed, consid.,rovide some clarification ofeffects obtained in the over-

e significant effect for theeasure appears to stem pre-1the average duration of the,ive responses. In fact, as:, the group means on thisoverlap; the lowest individ.

long those who observed theram was higher than the110ngthose groups who ob-gressive program.be recalled that the instruc-

I children emphasized that aof the hurt button would

limal distress to the otherger depressions would cause':tter discomfort. This fact. finding. drat the overall av~If such responses was morer in the aggressive program:he control group, suggestsprimary effect of exposure torogram was that of reducingIts against inflicting severehe ostensible peer victim,

asing the magnitude of the" With the exception of the'ffect was not paralleled by:he corresponding measures

. it cannot be attributed tofects.noted that the measure of

responses was obtained"ects had been given an op-or hurt another child. Thus

:cts might reflect an inter-he programs and some as-frting/helping opportunity:mple influence of the pro-;. While the present data,0 address the possibility ofdirectly, it is clear that theJre consistent with earlierother types of aggressiveand where there were no

measures.

xperiment was designed pri-

EFFECTS OF TELEVISED VIOLENCE ON CHILDREN'S BEHAVIOR 475

HANRATIY,M. A., LIEBERT.R. M., MORRIS,L. W.,& FERNANDEZ,L. E. Imitation of film-mediatedaggression against live and inanimate victims.Proceedings of the 77th Annual Conventionof the American Psychological Association,1969, 4, 457-458. (Summary)

HANRATIY,M. A., O'NE.~L,E., & SULZER,J. L.The effect of frustration upon imitation of ag-gression. Journal of Personality and Social Psy-chology, 1972, 21, 30-34.

HICKS, D. J. Imitation and retention of film-mediated aggressive peer and adult models.Journal of Personality and Social Psychology,1965, 2, 97-100.

KLAPPER,J. T. The impact of viewing "aggres-sion": Studies and problems of extrapolation.In O. N. Larsen (Ed.), Violence and the massmedia. New York: Harper & Row, 1968.

LARSEN,O. N. Violence and the mass media. NewYork: Harper & Row, 1968.

LESSER,G. S. Designing a program for broadcasttelevision. In F. F. Korten, S. W. Cook, & J. I.Lacey (Eds.), Psychology and the problems of

REFERENCES society. Washington, D.C.: American Psycho-

BANDURA,A. Influence of models' reinforcement logical Association, 1970.contingencies on the acquisition of imitative re- '. .'LovAAs, ~. I. Effect ~f expos~re to. symbolicsponses. Journal of Personality and Social Psy- aggressIOn on aggressIve behaVIor. ChIld Devel-cIlOlogy,1965, 1, 589-595. opment, 1961, 32, 37-44.

BANDURA,A., Ross, D., & Ross, S. A. Transmis- MALUCK,.S. K., & M~ANDLESS,B. R. A study ofsion of aggression through imitation of aggres- catharsIs. of aggressIon. Journal of Personalitysive models. Journal of Abnormal and Social and SocIal Psychology, 1966, 4, 591-596.Psychology, 1961, 63, 575-582. ME~M, ~. We're teaching our children that

BANDURA,A., Ross, D., & Ross, S. A. Imitation violence IS fun. The Ladies' Home Journal,of film-mediated aggressive models. Journal of 1964, 52, 44, 49, 52.Abnormal and Social Psychology, 1963, 66, 3- R<?SEKRANS,M. A., ~ HARTUP, V(. W'. Imitative11. (a)lpfiuences of consIstent and mconslstent re-

BANDURA,A., Ross, D., & Ross, S. A. Vicarious . sponsc:s c?nse~uences to a model on aggressivereinforcement and imitative learning. Journal of beh~vlor In children. Journal of Personality andAbnormal and Social PsychO/ogy, 1963, 67, SocIal Psychology, 1967, 7, 429-434.601-607. (b)" SAVITSKY,J. C., ROGERS,R. W., IzARD,C. E., &

BE,?,OWlTZ,L. Some aspects of observed aggres- LreBER~, R. ~. !h~ role of frustration ~dsion. Journal of Personality and Social Psychol- ang~r In the IIDI~a~on of filmed aggressIonogy, 1965,2, 359-369.". agamst a human victim.PsychologicalReports,

BERKOWITZ,L., & RAWLINGS;It. Effects of film 1971, 29, 807-810.violence on inhibitions against subsequent ag- SIEGEL,A. E. Film-mediated fantasy aggressiongression. Journal of Abnormal and Social Psy- and strength of aggressive drive. Child Develop-chology, 1963, 66, 405-412. ment, 1956, 27, 365-378.

FESHBACH,S. The stiumulating versus cathartic WALTERS,R. H. Implications of laboratory studieseffects of a vicarious aggressive activity. Jour- for the control and regulation of violence. Thenal of Abnormal and Social Psychology, 1961, Annals of the American Academy of Political63, 381-385. and Social Science, 1966,364, 60-72.

FESHBACH,S., & SINGER,R. D. Television and ag- WALTERS,R. H., & THOMAS,E. L. Enhancementgression. San Francisco: Jossey-Bass, 1971. of punitiveness by visual and audiovisual dis-

FLANDE~,J. P. A review of research on imitative plays. Canadian Journal of Psychology, 1963,behavior. Psychological Bulletin, 1968, 69, 316- 16, 244-255.337. WERTHAM,F. Is T.V. Hardening us to the war

HALLORAN,J. D. Television and violence. The in Vietnam? New York Times December 4Twentieth Century, 1964, 174, 61-72. 1966. .,

HANRA~, M.. A. !mitati°I?-of. film-m~d~atedag- WINER,B. J. Statistical principles in experimentalgreS~IO11_"ag~lDstbve and mammate VlctlIDS.Un- design. New York: McGraw-Hili 1962.pubhshed master's thesis Vanderbilt University ,196~ ' I

marily to determine whether children's will-incness to engage in interpersonal aggressionw;uld be affected by the viewing of violenttelevised material. Within the context of theexperimental situation and dependent meas-ures employed, it appeared that this was in-deed the case. However, it is clear that theoccurrence and magnitude of such effectswiIIbe influenced by a number of situationaland personality variables. It is thus impor-tant to examine the antecedents and corre-lates of such reactions to violence in greaterdetail. In view of the fact that a child borntoday will, by the age of 18, have spentmoreof his life watching television than inany other single activity except sleep(Lesser, 1970), few problems seem moredeserving of attention.

(Received August 19, 1971)

'\