ED 064 855

AUTHORTITLE

INSTITUTION

PUB DATENOTE

EDRS PRICEDESCRIPTORS

IDENTIFIERS

ABSTRACT

DOCUMENT RESUME

EM 009 660

Liebert, Robert M.Television and Social Learning: Some RelationshipsBetween Viewing Violence and Behaving Aggressively(Overview).State Univ. of New York, Stony Brook. Dept. ofEducation.7142p.

MF-$0.65 HC-$3.29*Aggression; *Children; Programing (Broadcast);*Social Behavior; Socialization; *Television;Television Viewing; *ViolenceObservational Learning Theory; *Surgeon GeneralsReport Television Social Behavior

Observational learning requires exposure to modelingcues, acquisition of the ability to reproduce what is seen or heard,and acceptance of the model's behavior as a guide for one's ownactions, as imitation, counter-imitation, disinhibition, orinhibition. In this ''verview paper, the author considers a large bodyof research, especially that commissioned by the Television andSocial Behavior program of the National Institute of Mental Health,and concludes that children are exposed to a great variety ofaggressive models and do acquire and are able to reproduce aggressivebehavior. Evidence on children's acceptance of aggressive behaviorsis described as complex, subtle, and dependent on such factors asrewards to the model, whether the model is seen as fantasy orreality, the observer's home life, and the situations in which hefinds himself. However, the author feels that the accumulated weightof the evidence from so many studies justifies the conclusion that atleast under some circumstances, exposure to televised aggression canlead children to accept what they have seen as a partial guide fortheir own actions. (Alith(mr/RH)

VS FILMED PROM BUT AVAILABLE COPYI"CO ,

48 .4)0

1 M Television and Soci'alIlli

a

Learning: SomeRelationships Between

Viewing Violence andBohiving Aggressively

(Overvipw)Robort M. Llobert

WS. DEPARTMENT OP HEALTH.EDUCATION & WELFAREWPM OP EDUCATION

MIS DOCUMENT NMI SEEN REPRO.DUCID EXACTLY AS RECEIVED PROMTIM PERSON OR ORGANIZATION ORIG.INATING IT. POINTS OP VIEW OR OPIN.'IONS STATED DO NOT NECESSARILYREPRESENT OPPICIAL ONCE OP IDU.CATION POSITION OR POUO.

What arl the etil*Nts upon chihirz-at of obsorvinr.viol.Int television pro-

in.:,1s7 The; cr:....1:14 is one which'husi)eam contintrAy posccl sir,ce. the

.atIvent of television sets Its a common fixt.ore in the Ameriertn home

a EInc$1 (9,;:c;41,;; ApFor,!IS tO it. bast:0 bCIC1 On simple or:Hon

on rescytch iefieets cl:wees (tad

appreckaion of the co:Inky of f.i.:a; rhenumenon, halve ranvd floratl.int the itILIene.1- is r.niformly rerni

cietis to ecitmlly %,..:-..tchiriz tcrtitimroAt fait

1/43 ca:t lit,k to c!:ii...!Nn's tech] 17:.!;:lvior.

1/45 FILMED FROM BEST AVAILABLE COPY

(54'

(:)

.0

a

2 TCLEYISION AND SOCIA1 !YARNINGI

Although literally hundreds of studies have been focused directly orindirectly on tekvision nnd its effects upon youngsters since tho 1950s,

the series of investigations recently commissioned by the Television andSocial Behavior program of the National Institutes of Mental ilealthconstitutes one of the first systematikand purposefully coordinated at.

tempts to employ the efforts of a lino group of researchers with relo.vant expertise And diverse viewpoints. Fivt of these commissioned pa.persfour expaimentel reports and a liteikutre reviewappear in thesubsequent chapters of this book. The purpose of the present paper is

to provid: a relatively brief overview of the research from which thesecontributions grew, and to identify and consider both the points ofagreement mid the inconsistencies that exist among them., This task canbest be served by beginning %% ith synopses of the theoretical and meth-odological questions which rebte to the study of television and aggres-

sion.The scienalc issue most fundamentally releted to the particular ques-

tion of the effect% of television revolves around the nature of observa-tionanearninr. i.e., the way in which the behavior of children (andHurts) chtinNes (IA a function of exposure to the behavior of others.Thns, it is to basic and applied research and theory in this area that wewill first turn our ;mention.

Rega:dless of th:ir partimbr theoretical affiliations, investigators in-terested in sociAzation ha% e virtually all acknowledged that a child'svalues, knowleef;e. alld behavior may be developed and molded. at least

in part, by observptional learnbg. Specifically, research studies haveshown that the simple &nervation of others can be very potent in chang-

ing such widely varkd aspects orsocial behavior US a Child's willingness

to aid others (e.g.. Rosenhan and White, 1967). his ability to display self-

cant:)l (e.g., liandura and Mischel, 1965), and his learning of language

rules (e.g.. Hebert. Odom, nig, and Huff , 1969). Young children's ob-

servation of otheis on film has been shown to produce an impressive

levd of learning of unfamiliar behaviors (Coates and Frartup, 1969), toincrease children's sharing (Bryan and Walbek, 1970), and to markedly

reduce phobic reactions (Bandura and Menlove, 1968; Hill, Liebert, and

Mott, MS).Tins list represents only a few examples from the impressive body of

evidenc e. which sufgects that learning by observation is a critical aspect

of the social learning processes through which the child is informedatoout the world nicund hint and molded into can adult member of his so-

a. 19:0; Bryan and Schwahz, 1971; Licbert and Spiegler.

1970). It is therefore . understandable, that professionals and laymen

alike have become increasingly interested in and concerned about deter-

tt-ic pnd extcnt to mhich such social 1.....arning occurs as a

functic.n tzALvi.lion viev. ing by ehi:d:en.

6

'g

-1-

Dafinitions and distinctions

s

3

A number of different types of phenomena may, and have been, sub-

sumed under the general classification of observation& learning. Thus,

in order to avoid confusion regarding the terms used, a series of dal&

lions is provided below (cf. Liebert. and Spiegler..1970). These Mini.dons will be employed consistently in the remainder of thh paper 'Ind

may prove helpful in clarifying both the factual and the theoretical is-

sues which must be addressed periodically as we proceed.Modeling. Modeling refers to the observed behavior, of others,

whether presented through direct demonstrations or through films, tele-vision, or stories which are heard or read. When modeling cues are pre-

sented by direct exposure to other persons, the plwnomenon has typical-

ly been referred to as live modeling, while the behavior of others as ob-

served in movies. television. and other representative media k usually

considered to fail in the general caw:ory of symbolk modeling.

Obsvrvational learning. Any and all of the demonstrated

consequences of exposore to modeling will be subsumed under the term

observational learning. Observational learning can take several

different forms and may therefore be measured in a number of different

ways, depending upon the interests of a particular investiptor or thenature of a partieubr issue.

Acquisition vs acceptance. The broadest distinction among the var-

ious folms of ebx:ivational leng, introduced by Albert kndura andRichard Walters (l963) almost a docr.de rtgo. is Utween acquisition or

the ability to reproduce previously unfami)iar vets as a function of ob-

servational learning and the subsequent acceptance and spontaneousperformance of behaviors that are the snmc as. or similar to, those

which have been observed. A'ehild may observe and remember a partic-

ular adult's manner of speech, the particular expressions which he uses,

or the novel forms of helping (or hurting) others mhich the exemplar dis-

plays "without necessarily adopting any of these characteristics. None-

thekss, if the observer can rep:oduce or describe the behavior he has

witnessed (for exnmple, when asked to do so), then the most basic form

of observational learning, the acquisition of new behaviors, has oc-

curred.The possibility that behavior can be acquired observationally and re-

tained, ssithot neezssarily being performed immediately, has important

implications for our understand;ng of the effects of both television and

other observational learnin opportunities. If a chiki has learned some

new behavior, then he clearly pi;sseses the potential to produce it if (or

when) he finds Hmself in a situ:lion in which such a perfermance ap-

pears to be dIsirl.h:e. useful. or Rely to serve his own purposes. Thus,

although learning does not necessarily lead to action, it does make

6

4 TELIWISION AND SOCIAL I.1:ARNINO

possible the performance of etherwise unavailable forms of social res.

FAIMure to Hovel modeling cues therefore changes the poten-

tial wee of activities which a child may display when stressed, forexample, or provoked, or called upon to act in a situation where theseotherwi,e dormant skills appear to become potenthdly useful.

Finally, it should be noted that acceptance of another's behavior as a

guide for one's own does not necessarily imply an increase in shnilarity

between the behavior of the model and that of the observer. The child'

who saes a peer burned by a hot stove will typically Wome tesslikely totouch the dangerous appliance than previously; he accepts the exem-

plar's actions and consequences as a guide for what he should notdo.Direct indmire effects rs. inhibition.disinMitio».Many observation.

ally horned behaviors do lead to acceptance. either immediately or

when enviromental coeditions are conducive. Researchers have noted

two related but distiquishable kinds of imit.itive performance: direct

imiwire effects and inhibition or disinhibit/on effcas. The concept of

direct imitetioe refers to whether the observer endeavors to exactly re-

produce ot mimic the behavior which he has observed. A child who re-

peats exactly some of the expressions used by his father, and a child

who prceiselr matches We kicking, bitting. and other forms of attack

(Isaias: a plastic 13obo doll which he has seen displayed in a brief movie,

arc both show;111'. direct imitative caws. Likewise, the chikl in our ear-

lier example ho now ovoids the stove is engaging in direct Colintethni-

tutkm.in contrast, the inhibition-disinhibition dimension refers to the per-

formance of behaviors which fall in the same generalckss as those ob-

served, but may be different in virtually all particulars-i. Thus. for exam-

ple, a child who sees .his parent donate money to a charity and subse-

quently becomes more likely to share chocolate cake with his little sis-

ter, is showing a disinhibition (or response facilitation) regarding the

class of sharing.Inbibitkn as a function of observational learniqg represents the other

side of the coin from disinhibition. For example, the child who, on the

first day of class in the third grade, sees that the new. teacher punishes

one of his classmates sharply for adamantly announcing his refusal to

comply with a particular request, may subsequently become less likely

to turP in his first homework assignment late. Failing to turn in home-

work on time and speaking out inappropriazely in class are hardly identi-

cal behaviors in terms of the particular acts which they involve. Howev-

er, they fall into a common Class of behaviordisobcdience to the dic-

tates of the teecherand thus the behavior of the observer may be

traced to the behaviot of (and subsequent consequences to) the model.

s

s

°OP

00 a - a as. Mat. 5 4. .1 5

Integration of processes: tho three stages ofobservational learning

The processes of obsen ational learning outlined above may now be

seen to involve three stages: 1) the observer must be exposed to model .

ing cues; 2) be must acquire and bc able to reproduce what he has seen

or heard; 3) he may or may not accept the model's behavior as a pith)

fot his own actions. Thus, step 3 may be manifested In Imitative effccts

(i.e., the observer's behrwloi becomes more like that of the model than

previously), which can involve either direct imitation, or disinhibition,

or both. Alternatively, the effects of modeling may produce counterimi-

!alive erects (eirz.ct Coullterimitation or inhibition). Finally, exposure

and acquisitior. can occur without leading to acceptance, a result which

may be thought of as nonimitation. These steps, and the alternatives

involved at era of them. are shown traphically in.Figure 1.

a

etaatbas 0115. spdia

fowl.

H A

/0/1[..1 VISII141/0 a 10,1

p a° :of% ottii, r IOW 111111

a% 4.04,1 if V. . .14. = isanMb

DIRECT 1,4111A1 NA

000 no11114.1, MIP'str al It 0 th. R USIII*.l4R R Ass Owlet

III SA

1411fAlIVE IF rws

Cotaareal

bathai HI B

II

0Mtcr COUP. It-

couNna.eoirstrna Elf ICVSaramm

Cafasea a laRlela 40 r Ctall/0.644111q the 6 *la

jethwon ens ast a.,h1obanot ka I Le% ID

11011111TION

aq OW644

Mt TA 'iota

POSUR g ACQUISITiON

'HI C

tlf

apue., s bd.", a

awe le Oa SO

gm- Ar CAP r ANct

04.1011goturrftlte l.a U1551140%All el oaf ,.fro.o44. vs boo 14

shca 1.!,

Figure 1. The time straps of obtervationnl learning

. S.

6 TELEVISION AND SOCIAL LEARNING

Figure 1 aho suggests that children's exposure to aggressive televi-sion programming, a situation in which observational learUing may po-tentially occur, mny hve a number of differcnt effects, Evidence. ofeach of the possibilities shon is available. However, before we con-sider those substantive findings, it will be helpful to distinguish amongthc methods which have been employed to obtain them.

Research approaches to the study of televisionand aggression

Three somewhat different research strategies, each with strengths andweaknesses, have been employed to explore the question of the effectsof tekvision upon chilaen: survey techniques, comlationd siudks,and experimental investigations.

Surveys. Tha survey technique typically involves identifying a rela-tively large sample of persons (usually several hundred respondents ormoro, preferably representative of the population at large in terms ofage sociononomic and ethric backgrounds, and systematically obtain-.ing questionnaire information abont such things as the frequency withwhich they (or their chiklren) watch television, an indication of favoritepogroms, ?.ne so on. Surveys can aiso be used to determine the natureand mints oZ television's ofZerings. in the latter category falls the ex-tenfive and so-Aigicated surveys of Georg: Gerbner (1968, 1971),which have pi ovidcd valuable information on the amount and nature ofcurrent televkio.t.

Like all other research effects, surveys vary in sophistication and inthe degiee to which we can be confident that their samples refiect thelarger population (of children, of programs) which they are designed toprobe. A paper by LoSciuto (1971) provides some excellent commentson the critef in and dirilculties which characterize this approach. Briefly,

. it is important that the sample surveyed should be appropriately repre:sentative of the population. large enough that descriptions of the samplewill closely approximate those of the population. and tapped in such amanner as to minimize the probability that the information-gatheringpiocess will bias the information obtained (cf. Neale and Liebert, inpress). . .

Correfraioa.71 studies. Correlation, as the name implies, clods with theco- or joint.rciationship between two or more variables. The method isidenHy suited to answerino questions of the form: "Do variable X andvariabl! V go tov:ther or vary to;'..ether?" Such questkuts have oftenbeen asked about television and aggression. Are children who watch alot of violent tc!evision more aggressive than those who watch less? Isthe social class of the chikrs parents related to the amount or kind ofprograms he watches? Are boys more likely to be influenced by aggres-sive tc!evislon p:ogams than giris?3

A

.

7

The corrclational method is characterized by the fact that all subjectsarc observed under identical (or nearly identical) conditkms. The men-surements themselves involve preexisting characteristics of the partici-pants, so that. no effort is made to manipulate or control the fwents towhich they are exposed. While this last characteristic appears, at firstblush, to bc an advantage. it k, in fact. the method's most serious draw-back. Without systematically .varying the conditions to which subjects

.are exposed, it is usually not possible to draw causal inferences. A sub-stantial correlation between two variables*mcans that each can be "pre-dicted" from the other statistically, but it does not tell us whether eitheris the cause of the other. There is, for example, a high positive correla-tion between the number of churches in a city and the number of crimescommitted in that city; the more churches a city has, the more crimesare committed. Does this mean that religion fosters crime? Certainlynot. That crime fosters religion? Unlikely. The correlation is due to athird variablepopulation.which leads to an increase in botl. churchesand crime. Or corAder the ',Inanity of a positive relationship betweenthe amount of aggressive television which a child waches and the de-.gree to which hl behaves aggressively in life situations. From such acorrelation it mieht be argued 'that youngsters who observe neressivetelevision beebree more lil:ely to aegress as a result of the, television fareto which they me exposedthat is, that the television programs causethe aggressive behavior to some degree. This areument maybe true, butnot as a logien! ieferenee from the observed relationship.

It is possible that some children are both more likely to be aggressivein their own behavior and more likely to enjoy watching aggressive pro-grams than others, due to some unidentified "third variable." If thiswere the case, aggressive children would not be expected to becomelessaggressive by a change in their television diets, nor would relativelyunaggrissive children become more aggressive as a function of in-creased exposnre to such programming. It is therefore vital that correla-tional results be supplemented by research which permits logical infer-ences about causal relationships in order to determine whether a consist-ent pattern emerecs. For this purpose, the researcher and critic mustturn to experimental investieations:

Experimental Threvigations. In an experimental investigation all sub-jects are tteated alike except fo l. diiTetential exposure to one or moremanipulated events or independent variables. They are 'subsequentlylested on one or more measures (referred to es dependent variableswhich are expected to be conttolled by or to depend upon the independ-ent variable. If the subjects are assigned to groups randomly (so thatcaeh person hed en equal ebnee of beine in any treetment groep) or ifpmsible initial d:fferenees are otherwise con!rolled or cancelled out,then the investigator can coeclude that the differences in the independ-ent variable (or treatment) cetu,ed or praticed differences on the

8 IIIIENISION AND SOCIAL LEARNING

dependent variable(s). For this reason. experimental res9rch is widelyconsidered to be the most powerful research tool in the social sciences.

Exposure to aggressive modeling throughtelevision

The question of children's exposure to aggressive modeling throughtelevision may be considered in terms of two components: 1) How muchof available television fare provides such modeling? 2) Row much timedo children spend in exposure to such content? Excellent survey research conducted under the auspkes of NIMII appears to provide relalively clear answers to both these questions.

Frequency of tiolent content on the commerekl networks. Gerbner(1970 studicd the frequency of overt physical violence during primetime and Saturday morn inr. network programs dm in3 the fall of 1969 andcompared these data with similar 1967 and l.968 studies which they hadconducted for the National Commission on the Causes and Preventionof Vioknce. He found that in 1969. as in earlier years, about cii,ht inten p!ays still contained violence, End the frequency of violent episodeswas still about five per play and nearly eight per hour." The violenceindv: (Gerbnel's overall measure of the frequency with which violentacts src portrayed) actually rose for one of the networks, NBC, from1968 to 1969.

Further, since our primary concern is in programs which are likely tobe of particular interest to children. it is especially important to note thatcartoons, the most violent of programs in 1967, increased their lead in1969.

Frequency with which children are exposed to violent content. Lyleand Hoffman (1971) conducted an extensive survey of media use amongmore than one thousand children from widely varied baekgrounds. Theyconcluded that "television saturation was almost total; only two percentof the students stated that there was not a working TV set in theirhomc." Their duia also show that more than one-third of the first grad-ers are still watching television at 8:30 p.m. on weeknights, and morethan one-half of the sixth graders are doing so. Likewise, Stein andFriethich (1971, this volume) report that, in their sample, televisionwatcliini; was Nei:mica to Lc among :1;ttellildre.ns' inost frequent akifigactivities. LesNer (1970) has argued that a child born today will, by theare of 18, have spent more of his life watching televi;ion than in anyother cinrle activity except sleep.

The pervasiveresi of televbion can be seen even more clearly bymoving from rreenta!:e lip Ns to absolute numbers. Meintyle andTeevan (1971), citing the Violence Commission stuff reports of 1969,remind us that "on one Monday duting the period covered, over fivemilikn chil:lren under the aze of . . . were still watching between

. s $

10:30 and 11:00 p.m. . . ." They also point to the commission's observe.

tions that "dime I:, a great deal of violent content tvailable, at all times

of the day. for DO nmnner of intended audience." that "the prcsenwionof violunce is typically lib a MCM15 of achieving virtually any type ofgoal," and that "the use of violence. whether sanctioned or not, is likely

to be a successful means of obtaining such gods "Moreover, high-vieleace p:ogrants are among young children's favor.

hes. In the Lyle and Hoffman report, for example, first grade children

were found to prdcr programs of the sort heavily saturated with viol.

ence. Twentyfour percent of the chiklren said that cartoons were their

favorite type of program. while another 13 percent telected detective

and "hip adventure" pro:trams as their favorites. A similar pattern wasfound in Stein et AL's (1971) study of preschoolers, v hose parents re-ported that Ole). wiached Crams an average of more than seven hours

per wee:. Lyn vio'..ntt programs were watched more than a full

hour per week by these three- to tIve-ye:ir old children.thc Ilrct ste;e of the observational learning of aggression

from wlevision, exposure, does occur. At this writing, there appear s. to

be no cptebtion that violunt televisioa fare is available in overwhelmingabundance and that chil:tren slo mach these preraros both frequentlyand regular4. lt is to the second stile, acquisition, that we must turn

next.

Acquisitionit.o one would doubt thm children can learn novel forms of bchavior::

both words and actionsfrom simply watching others. It is., however,only through systematic research that we are able to see the degree towhich this form of 1earnin3 is effectively mediatcd by television and tele.

vision-like formats.As Stevenson (1971. this volume), in his review of the effects of tele-

vision .on preschool children has observed, the Ball and Bogatz (1970)evaluatioil of the instructional effectiveness or sesame Street prov ides

one of the most comprd.tenive demon5trations of youirj children'sacquisition of 1:nowk: from television. The conclusion dre.wn bythese writers is one with implications both for the le;:rning of amessionfrom television and also for the more appetizing prospect of usirg posi-tive tvic!ol fr.re: kr kc!rvetional rurpmcs: ". . . tekViSioll ha: beenshown to v..ork extremely well as 0 teaching medium. It achieved thisresult fracring to Sc.mw Street] not only in !earnings that involve sim-ply aFsociation (for evimple, naming letters) but also in learning thatinvolves complex cogniiive processes (sorting and classifying) . . ."

Studies desivned to !.how that brief exposure to novel kuressive be .haviors can lead to their acquisitiou by quite young children have uni-formly shown that this iniThence is potent indeed (liandura, 1965; Bun-dura, Ross nd Itc.hs. 1963a; flicks, 1965). In one such study, for

9 .s.

10Ma, 4.0. I wailed% *. TELEVISION AND SOCIAL LEARNING

example, 83 percent of the subjects (3-5 year old boys and girls) dipplayed imimive hggression of novel modeled acts even though they had

not been asked to do so and were free to play with attraetNe. nonaggres-sive toys such as a tea set, crayons, cars told trucks, plastic farm ani-malb aad the like.

Partieuhrly striking is the degree to which some of the subjects tip.peered to be virtually "carbon copies" of Me aggressive models whom

they had observed. Photographs taken by Bandura and his associatesillustrating these imitative ellects are shown in Figpre 2. The topmostframes show the female model's performance of four novel aggressiveresponses while the middle and bottom frames respectively depict II

male and a female subject spontaneously reproducing the behaviors

which they had seen earlier on film.

4../01

41 Source: Courtesy of Or. Albert Bindura

Figure 2: Photographs from the film Socid Learning of Aggr&slon Through Imitation ofAggressive Modals, Mustrating children's acquisition of aguressive responses

trrough obsentatinual learning.

Further.. there is evidence that behavior Required in this way may bc

recalled kr long periods of time, as evidenced by delayed retests ofacquisition. Hicks (1965) found that subjects shown a simulated televi-sion program timilar to those used by Bandura and his associatesshowed substi.atial acquisition of these bchaviors after a sinlle viewing;this ;:ecinisition was still in evidnee when they were tested again, with-

out further exposure, six months later. In a second study by the sameauthor (Hicks, 1963), about 40 percent of the aggressive responses were

11 11

found to bc retained for a period of eir.ht moniht. On the basis of these

data, it is rehttively easy to concur with Goranson's (1969) conclusion

that novel aggressive behaviors can be learned from television and, with

even limited practice, be retained for impressively long periods of time.

Since these studies typically involved inanimate rather than human

victims, and have been mistakenly criticized for employing this strategy,

it may be. helpful to note their rationale explicitly. Ihndura (1969) has

explained the strategy this way:A stteWearninil theory of riggresslon dktinpishes thc sequsition of instru-

met...tat resj onse>. that have tleaructive or pain.producik: poteathd from the

corAions gcverrint their subsequent performance. Armssive responsepatterns arc characteristkally bequired under nonfrumrating conlitions in

the o.bsence of injurious intent and often toward inanimate objects. Thus. forexample. militory remits acquire and perfect combat skills throngh munyhours of tar3et prectler end simulated skirmishes; boxas develop lwrtft.t1pummeNag ahi;ities hy usine ptmck'ng Lass und spa:ring r arum's whom the*

do not necessarily itar.d to 1:..att and huntsmen avpire thc !tisk: t :lei:runts

of hunting by shooting at inanimate targets before they to out in search of

game. Indefd. if 4resiive repertoires were tutvzht only v.hile indiviJulds

mere hostilely moused and entertained ;On huts &sips. many of the mots sand learners would probably be maimed durin 2. the acquisition phase.

Recall of pa:ticular physical acts must, however, be distinguished

from the acquisition and readl of the somewhat mote subtle plot themes

and rclationsbii..s which characterize tdevised stories. In order to ex-

plore an aspect of this inner issue. Leifer awl Roberts (1971, in this vol-

ume) stud:cc' age cadences in children's understanding of aggression

which they observed on television. Almost ?OD subjects served in their

experiment, including kindergarteners, third graders. sixth graders,ninth graders, and twelfth graders. The primary purpose of this study

was to identify age differences in children's understanding of the mo-

tives and consequences which nttended aggressive acts. Striking age

differences were obtained, shewbg that understanding of motives and

ccmsequences increases with age. Specifically, as Leifer and Roberts

note: "Kindergarteners could answer only about one-third of the ques-

tions about either motives or consequences, third guders only aboutone-half, and twdfth graders about 95 percent. Hence the younger sub-

jects, by oar measures, are not taking in, or retaining, much of the infor-

mation about motives and consequences in a television program."The Leifer and Roberts dat are consistent with the findings of Stein

(1971, in this volume), which show that preschool children were able to

remember some of the characters and details of :he programs which

they were shown in Ler experimental field study (to be discussed in de-

tail in a latcr scetion) but that th :Pk recall was very far from perfect.

These data seem to show that children will fail to recall much of the"nonaction" detail in a particular sequence whkh they have observed

only once.Stevenson (1971, in this volume.) has noted that enduring recall of

such meter:L.11i wonld probnbly be most likely for young chilken if the

4 4Wm. -

//

4

12 TELVISION! AND SOCIAL 1.1:ARNING"ma

material to which the child is exposed produces emotional responses, ifthe observed content is discussed with otht rs, or if a codmon theme isshown repeatedly, Are these conditions met when children watch ag-gressive television?

Osborn and Endsley (1971) hnve explored the relationship betweenemotional reactions and program content. They had four- and five-year-ohl chikiren observe a variety of programs, inch:ding one containinghuman violence and one containins cartoon violence, as well as cartoonand human programs with no violent episodes, Gnlvanic skin responses(a measure of emotional reactivity) revealed that the violent programs.and particularly the one containing human vioknce, produced moreemotional re. ponses than did the nonviolent programs. Moreover, (andof particular interest in terms of the acquisition issue), the childrenshowed quite good mall of the contents of the violence programs. Infact, recall of the human violence or eaytoon vioknce programs was sig-nificantly better than that of the program showing nonviolent human in-teraction.

With respect to the second point, Lyle and Holman (1971' found thattelevision programs were subordinate only to school as a topic of dicus-sion among younf:sters.

Finally, regarOing Stevenson's third point, we may ask, "To whatrepeated then:cls and lessons are viewers of violent programs exposed?"Gcrher (1971) h...s answered lucidly:

To be able to hit hard and to strike terror in the hearts of one's opponcn.tsthat mal:es one en,:nt when the chips are down. The battered hero triumphsover evil by milcluing the bad guy in the end. The last men to hit the dustconfirm his own ilaw of ehaNcter and cause. Hurting is a test of virwe. and ,

killing is the ultimete measure. of man. Loss of life, limb, or mind, any dimi-nution of the fievlom of action. me the wages of. weakness or sin in thesymbolic shorthand of ritual drama. . .The typkal plot ends by reaching areassuring and mally foregone conclusion . . ."

The data reviewed in this section suggest that children are likely toacquire, with the level of repeated exposure which takes place, a good

dol of the aggessive repertoire that they see in televised violence. Infact, by virtue of their popularity and their ability .to evoke emotionalresporises, programs containing violence appear particularly likely to be

learned and retained from tplevised observational experience. Thus it isnot surprising that what Goranson (1969) has referred to as the "re-sponse forms" of aggNssion can be reproduced by observers easily and

with a temarkable degree of fidelity.The degree to which children are attentive to or learn from the com-

plex and oceasionz.11y sophiszicated nuances of plot, inirige, cnd verbala!!!!ression in the more "adult" type of crime dramas is not yct well un-derstood. A NH In-asp of these procedures would require systematicallyvarying the chuacteristies of otherwise unfamiliar inputs and then ex-ploring their mean p.nd retention ft:UM various time period. Research of

this 4re woold be extretady valuable thcorcticay and could play an

12

importmit subsequent role in the development of television progftams

which are designed to teach substantive material and prosocial lessons,

an appetizing possibility hinted at both in Stevenson's paper and in the

findings of Stein and her associates about the facilitation of prosocial

and self-control behavior by Alistemgers Neighborhood.However, notwithstanding the preceding qualifications, it is clear that

children can and do acquire aggressive behaviors from the type of televi-

sion fare which is currently available. But the question of whether they

accept this inater0 as a guide for their own actions remains.

Acceptance

Results reviewed in the foregoing sections clearly show that childrenin our society are exposed to a substantial amount of modeled aggres-sion on television. and that repeated exposure to this fare is likely tolead to qcqnisition of both novel age;essive responses and the percep-tion that aggession is often a potent interpersonal technique for servingone's own ends. The remaining. tmd perhaps most important, question is

the degree to which children accept and utilize information which theyhave gathered in this way in a variety of periormance situations.

Direct imitotive and counterimiwire effects. Following the theoreti-cal distinctions advanced earlier, our discussion begins with studies of

direct imithtive and counterimitative effects. There are at least a fewinstances of direct imitetion in naturalistic situations which have been

truly unfortunate, such as a lad who was stabbed while he and hisfriends reenacted scenes from the movie Rebel Without a Cause yhichthey had seen on a television 'rerun (San Fmneisco chronicle, 1961), or

till youngster who doctored the family dinner with ground glass after

observing this tactic used successfully on a television crime show

(Schramin. Lyle, and Parker, 1951). A more general (if less dramatic)influence of tdevised ageression upon dhect imitation can be seen from

experimental studies conducted in.the laboratory.A study by Bandura (1965) is among the most important for a theoreti-

cal understanding of the nature of direct imitative effects. Bandura had

children watch a model perform a series of aggressive acts against a

plastic Bobo doll clown. One group observed the model rewarded for

this behavior, one observed The model receive no consequences, and

one observed the model punished. When the children were subsequently

put in a play situatiun, those who saw the model rewarded or perform

wiout conscee,mces showed a hill, level of direct imitation. Not sur-

prisingly, those who observed the punish:id model. in contrast to the

other gtoups, showed relativel few imitative aggressive responses (that

is, showed conaterimitation). Nonetheless. when children in all groups

were subsequently asiled by the cspaiwienter to repi 011.1%:.! :is many of

the modes ari-Jessive acts as they could, and weic offered attractive

13

14 TELEvistosi AND SOCIA.I. LEARNING

incentives for doing SO, the previously observed effects of vicariousconsequences were entirely eliminated and all groups showed a renutrk-ably high and uniform degree of learning. . .

Data like the foregoing suggest that when conditkms favor activating4.

observationally karned responses into performance, the likelihood ofthis performance's occurring will not be diminished by the presence ofvicarious punishment at the time. of observational learning. This evict-ence inay successfully rebut the argument, made by 'spokesmen for themass media, that the depiction of violence has no harmful effects onyoung observers as long as it is ultimately punished. In the light of re-peated demonstrations that vicarious punishment does not impede thelearning or recall of aressive acts (cf. Liebert and Fernandez, 1970), itappears that these vioknt offerings can still have profound effects on thebehavior of the viewers by teaching more aggressive responses.

Equally 'impressive are children's responses to questions regardingdirect imitation. When Lyle and Hoffman (1971) asked, first graderswhether they had ever copied what they had seen on television, morethan 60 percent said they had done so. Moreover, when they were askedto indicate the type of program which they imitated in play activities,adventure shows (such as Batman) led the list. Not surprisingly, theseimitative activhies were largdy interpersonal in character; the childrenwere much more likely to use television programs as a guide for theirplay activities when playing with friends than when playing alone.

A recent experiment by Mtn tin, Ge1fand, and Hartmann (1970) sug-gests why adults may often not be witness to the direct imitation of ag-gressive modeling. These investigators exposed children to an aggres-sive model and then pat them in a situation where aggressive play waspossible. Some children performed in the presence of adults, some inthe presence of peers, and some alone. The general pattern of resultsdisclosed that the presence of an adult reduced the amount of aggressionbut that the presence of a same-sexed peer incretised aggression, rela-tive to the control gioup. Thus, in life situations, the most powerfuleffects of auressive modeling may occur under circumstances wherethey cannot be observed by parents or other adults. Goransdn (1969),chlig erslier studio. Nacho(' cimiln conclusion: ". . . parental eval-uation or instruction regal ding the permissibility of aggression seen inmedia can be eacctive in controlling aggression imitation, but this com-munication may be irrdevv.nt when adults are not later present to moni-tor the child's behavior."

That direct imiti:tive effects after observing aggression may occur formore than just the type of play activities described above, is indicatedby the results of several recent experiments. In the first of these (Han-ratty. Liebert. Morris. and Fernandez. 1969), four- and five-year-oldboys from a Supd:..y sehoci! I.in,'..!rg-ttten served as subjeos. Half thesubjects observed a two-nr.d one-half-minute color sound film in which

14

an adult male model aggressed against a human clown. The behavior

disçkyed by this symbolic model inch:dcd sharp and unprovoked verbal

insults to the clown, shooting at the clown with a toy machine gun, and

beating the clown viorously with a plastic mallet. Half the group saw no

such film.

. Thereafter, half of the subjects in each of these groups were permit-

ted to play in a research room, where they found a human clown stand-

ing idly, as well as a mallet and a toy gun. The remaining children were

placed in a comparable situation, except (hat they found a plastic Bobo

doll rather than it human. The children were left in this situation for ten

minutes. durinr; which time their aggressive responses toward the

clown, plastic Jr hnman. were recorded. Not surprisingly, the brief film

did increase children's apgression against the inflated Bobo. However,

reprdess of v:hether the children bad seen the apgressive film or not ,

the majority of those who were placed- with the plastic toy exhibited

some aggressive riction. in contrast, of the children who had not ob-

served the movie, none engaged in any sort of -garessive behavior..to-

waid the human clown. There are, of course, strong inhibitions for ag-

gressing against a hunrn beilg, even one who is attired as a clown, and

there was no provocation for doing so. Nonethdess, observation of the

aggressive movie did elicit physical assaults against the MIman clown,

includiag at least one swat with the mallet which was hard enough that

the victim showed a red mark on her arm several hours later.

In a second experiment (Hanratty, 1969), it was again found that a film

of this type. withuut other provocation, would lead children w physical-

ly assault a human victim. Moreover, such aggression was displayed by

both boys and girls (only boys had participated in the first experiment)

and for films in which both an eight-year-old boy and an adult serVed as

modds. This fiadik: has been esentially replicated a third and a fourth

time with somewhat older boys (lianratty, O'Neal, and Sulzer, 1971;

Savitski, Ropers, 1zard. and Liebert. 1971). although in the Hanratty et

al, (1971) study, fatstration did interact with modeling for these older

and perhaps more inhibited observers.

Since direct imitative effects definitionally require a circumstance vir-

tually idemical to the one observed, it has been argued that they are less

impom tant, from a soei:ii puira of view, than inhit,itory an1 dicinhibitory

effects. However, it is precisely in this last eateeory that the greatest

controversy has raed regarding the adequacy and interpretation of re-

search results. This is also the area in which greatest research emphasis

has recently been placedas evidenced by the fact that all four original

studies appearing in the present volume, and almost half of all of the

studies sponsored by N1MH. sought evidence about the inhibitory or

disinhibitory effects of observing aggcssive television in terms of a will-

ingness to agge%s ;Tains( ollwr pzraOnS.

16TELEVISION AND SOM. LHARNING

hlhibitory aad disinhibitoryeffects: correlationalstudies. Several re-

cently completed correlational studies bear directly on the rossibility of

a relationship between the amount of violence a child obFerves and the

amount of aggressive behavior which he displays in naturalistic situa-

tions.WOrking with adolescent subjects, McIntyre and Teem (1971) found

a relatively small but consistent relationship between objective ratings

of the amount of violence on programs which youngsters reported

watching and deviant behavior as measured by a variety of indices.

Interestingly, this relationship becomes somewhat stronger if the degree

of violence of the television programs is assessed by the subject's own

perceptions than if it is assessed by objective ratings. Likewise, these

investigators found a positive relationship between the violence rating

of the subject's favorite programs and the degree to which they

expressed approval of v'oience. Further, while violence was more likely

to be approved if it had been rewarded than if it had been punished,

McIntyre and Teevan point out that "whether the characters were

viewed as behaving the way people ought to act has no.effect on the

frequency of approval." They specilicaHy note;

Those adolescents who5e favorite programc are more violent more frequent-

ly epprove of a teenage boy punching or knifing another teenage boy. If the

ft:mite proolm is de .;:risoed ns dcrictin; vio:enee ;lb a means to an end, or

violence rewarded. teen violence is approved more often than if the program

were not io describeJ,Whzth.:r or not the program "bhows the way people

ought to act" does not influence frequency of upproval.

in another NIMII-sponsored correktional study, Dominick and

Greenberg (1971) determined. through the use of a questionnaire tech-

! nique, the amount of exposure to television violence for 434 fourth-,

and sixth-grade boys enrolled in Michigan public schools during

the spring of 1970. Evaluation of the violent content of the programs

themselves was based on earlier analyses of newspaper and magazine

critics (see Greenberg and Gordon, 1971). Exposure was then related to

the boys' approval of aad willingness to use violence, as measured by

the items drawn from the Sears Antisocial Aggression scale (1961) and

the Buss-Durkee Hostility Inventory (Buss. 1957) respectively. Mea-

,. sures were also obtained of the degree to which the boys perceived viol-

. ence as effective and the degree to which they suggested violent solu-

tions to conflict situations when presented with open-ended questions:

In this investigation, specific predictions regarding. interactions be-

tween family attitudes and social class on the one hand, and exposure to

violent content or. television on the other hand, had been advanced.

Consistent with predictions, exposure to aggressive television was relat-

ed to the. boys' ..tatcd willingness to use violence and to their percep-

tions of its effectiveness when used. The direction of these influences

was concistent with hypotheses; higher exposure was associated with

greatet approval.

16

As Dominick and Greenber g. note, one of the most intriguing aspectsof their findings is the tenckney for violent exposure to be more oftenassociated with violent nttitudes for mkidle-class than for lower.classyoungsters. It is the cese, however, that when the home environment isweak in efforts to control the development of aggressive attitudes, thenthe relationship between exposure to violence on television and thechild's own attitudes become stronger. In the investivors' own words:". . . for relatively average children from average home environments,continued exposure to violence is positively related to acceptance ofaggression as a mode of behavior. When the home environment alsotends to ignore the child's development of aggressive attitudes, this relartionship is even more substantial and pet haps more critical"

Dominick end Greenberg used the same methods to relate televisionviolence and aggressive attitudes for girls. The results closely paralleledthose for boys. with exuosure to such aggressive fare making a "consist-ent independent contAution to the child's notions about violence. Thegreater the level of exposure to television violence, the more the childwas willing to use violence, to suggest it as a solution to conflict, and toperceive it as effective." . -

In yet another cerrehitionel study conducted for the Television andSocial BehtMor program, McLeod, Atkin end Chaff 2c (1971) examinedthe leletionship between viewing of televised violence and a variety ofmeasures or hg.ressivc tehavior in two relatively inge samples of ado-lcseent. on .3 in Maryland and another in Wisconsin. "f he study is note-.wortLy for its careful consideration of multiple correlations among avariety of predictor variabls and for reports of internal consistencyamong all thc measures employed. The outcome of these corielationalstudies may be summarized in the author's own words:

Our research shows that among both boys and girls at two grade levels (junior'high and senior high) the mote the child watches vioknt tele% ision fare, themore aggressive he is Rely to be as measuted by a variety of self-report mea-sures. . . .PPItialing out (tot;:l; viewing time slightly reduces the positive corre-lations of violence viewing and ngressive behavior in most eases. but the basicresult is the sanw as for t)to raw cot relations . . . .Similarly. the partiLling outof socioeconomic status and school pei formance does not alter the basic patternof raw co:relations. . . .We may conclude. then. tha adolescents viewing highlevels of viukra content on television tend to have hie: levels of rgtessive be-havior. regadkss of televkion viewing time. socioeconomic matus. or schoolpet fort:lance. Thes: pati:11%;:rq-rr to rule twt vlernatise sirn-pie televicion exposure, social shims and general competence as a student.

A particularly sophisticzaed correlational study of television andaggression was undertaken by 1 efkowitz, Eron. Walder. and Huesmann(1971). The Lefkowitz et al. report is based on a longitudinal study of theentirepopulation of children of a particular age in a mat Nev.. Yorkcounty and invo;ved apploNimately 900 youngsters. Desimted from theoutset to relate children's aggiessive behavior to various familial, social,and experimental factors v..hich might influence it, these investigators

4

18 TELEVISION AND SOCIAL LEARNING

eMployed a peer measurement technique of aggressimi (focusingexclusively upon acts which would harm or irritate other persons). Themeasurement instrument hns been amply researched (Walder. Abelson.Eron. Banta and Lanlicht, 1961) and accepted by investigators ofdifferent theoretical persuasions (e.g.. Feshbach and Singer. 1971).

The first measures obtained in this study of television and aggressionreveakd a significant rekttionship, for male subjects, between theamount of television violence which they watched in the third grade andindependently assessed peer ratings of aggression in the classroom atthat time (Eron, 1963). This correlational liading was later replicated

with a dill'erent sampleeighth grade boys apd girls in an urban city in

the South.Lersowitz et al. rezently cor*eted the longitudinal phase of their

study by obtaining data from more than 400 oi the youngsters whomthey had studied ten years earlier. The measures obtained included peerratings of aggiession at this age, self-reports of aggression in aninterview. and self-reports of various aspects of television viewing. The

results of this ten-year followup showed that (but again only for boys)amount of agslession watched in the third grade was significantly related

to pece ratings of aggression at age 19.

Additionally, using a sophiq icated approach technically referred to as

a (loss-lagged pond MO/. Lefkowitz et al. showed that their findingsprovide stronr evidence for a causal rela.sionship than is usually avail-

able from correlational stuthes. To unfrstaad the bask logic behind

this approach, 4 consider the possibility. raised in our earlier, discussionof correlation studies. that a relationship will appear between overt ag-gression and preferences for aggressh,e television simply because per-sons who ate more willing to use aggression themselves are also morelikely to enjoy seeing it used by others in television dramas. This is animportant "riv..d hypothesis" to the.notion that seeing aggressive televi-

sion causes aggressive behavior. However, if the 'rival hypothesis werecorrect, prefvr:mces for aggressive television at age 19 in the Lefkowitz

et al. study should "go togeth:h.r" with overt azgression in the third grade

as closely as preferences for aggressive programs in the third grade gowith r.;.gressiz, i. at age 19. In othcr words, the relationships. if accounted

for by a constant third variable, should go both ways in time.

In contrast. if television aggression does cause aggressive behavior

later, it would be plausible to find a link between earlier televisionwatching and kiter aggression but not vice versa. This is exactly whatwas ..liselosed by the Lefkowitz et al. data. Th ird -grade preferences for

aggressive television.predicted later aggression, but later television pref-

erences did not t.!iate to the youngsters' eadier aggressive behavior at

Foi ressnn. it k rea,onaNe It; vgree with the investigators' inter-

pretation of thdr Endin:...: that, for boys, . .on the basis of the cross-lagged correlations. the most plausible single causal hypothesis would

119

19

appear to be that watching violent television in the third grade karal to

the building of aggressive habits" and ". . .that a substantial compo-

nent of aggression at all three grade levels and a particularly large coin-. ponent at the thhteenth grade can be predicted better by the amount of

television violence which the child Watched in thc third grade than by

tiny other causal variable measured and reinforces the contentkm that

there is Li Cause and effect relation between the violence content of tele-

vision and overt at.gressive behavior."The possibility that many of the cifeets of viewing aggressive televi-

sion arc cumulative over many years as suggested by Lefkowitz et al, is

also raised by the Stein and Friedrich report. These latter investigators

did not find a relationship between interpersonal aggression during the

first few weeks of nursery school and home viewing of aggression for

their very young (3-5-year-uld) subjects. although such a rehitionshiphas been repeatedly found for older children and adolescents. (There are

of course, other possibilities, and Friedrich list several of them.)Tile correlational studies of Dominick and Greenberg (1971), of. Lef-

kowitz et al. (1971), of McIntyre and Teevan (1971) and of McLeod et al.

(1971) uniformly show an associ:;tion between exposure to aggressivetele...ision and at3gressive attitudes and/or behavior for elementaryschool-age and adolescent subjects. The fact that the Lefkowitz et al.study shows such an effect only for boys is less inconsistent than it ap-

pears at first blush, since the measure focused on the actual perform-

ance of aggressive acts. negatively sanctioned for girls in our culture,

while the other studies (mused on approval for aggression or other atti-

tudinal measures. This last study is also best able to stand in its ownright, because it uses the longitudinal cross-lagged correlational ap-

proach.All these correlational studies may be legitimately challenged as not

.showing causation; they should, however, increase our confid-

ence in the external validity (the i.pplicability beyond the laboratory) ofexperimental studies degigned to elucidate causd relationships and un-

&dying processes if a consistent pattern emerges.

Disinhibitory effects: experimental studies

A number of studies conducted during the 1960s showed that observa-

tion of filmed or televised aggression would disinhibit children's willing-

ness to aprresS on a variety of measures.In a relatively early experiment. Lovaas (1961) showed that nursery

school chien's aggressive behavior would be increased following

exposure to symbolic nivessive stimuli. One group of subjects saw

sequences from an itggressive film entitled Racsling Match, which prov-

ided an almost continual dihplay of one cato6n fir:are 41:ressing -i:nst

another by itittin;,:, bitinit, and the like. A second group of children saw a

19

20 TELEVISION AND SOCIAL LEARNINO

film dipicting three baby bears and a mother bear engaging in gentle playactivhy. rollowing the film. subjects Were presented 'with two largetoys, and their play activities were observed. Depressing the lever on

the aggressive toy mechanism automated a doll who turned and hit asecond doll on the head with a stick. Depressing the lever on the nonag-

. gressive toy apparatus triggered a wooden ball, enclosed in a cage, tojump through obstacles. The subjects were presented with the doll toyand the hall toy side by side, so that they could operate either toy or bothsimultaneously, if they wished. Children engaged in significantly moreplay with th e! hitting dolls after exposure to the aggressive film than afterthe nonaggressive film.s

In view of the fact that most television programs appear to depict ag-gresFlon as a potent technique for power and achievement, studieswhich have focused upon the inhibiting and disinhibiting effects of con-sequences accruing to a model for aggression are of particular import-

- :ince. In one such study, Bandura, Ross, and Ross, (1963) exposed onegroup of nursery SCAM), I IJoys and girls to 3 television program in which

one character, Johnny. refuses another, Rocky, the opportunity to playwith some toys. The program goes on to show a series of aggressive res-

. ponses by Rocky,.ineludii.g hitting Johnny with a rubber ball, shootingdarts at Johnny's cars. hitting Johnny with a baton, lassoing him with ahuhohoop. t.ud so on. At the end of this sequence. Rocky, the aggressor,is playing with ail of Johnny's toys. treating himself to sweet beveragesand cookies. and.finally departs with Johnny's hobby horse under hisarm and a sack of Johnny's toys over his shoulder. At this point, a com-mentator announces that Rocky was victorious. In a second group. theprogram was rearranged so that after Rocky's initial aggression, Johnnyretaliated in kind by administering a sound thrashing to the aggressor.

Two other groups served as controls; in one. a nOnaggressive buthighly expressive television program was observed, and in the secon0no television program was seen. Children's subsequent aggressive res-

Mises while PlaYing for twenty minutes in a special test room (=ginn-ed the primary dependent measule. The results clearly showed that chil-dren who observed a iewarded aegressor showed far more aggressionthemselves than children in the oiher groups. Moreover, at the conclu-sion of the experiment the children were asked to state which of thecharz.;teis, Rozl.y or Johnay, they would prefer to emulate. Sixty per-cent of those who observed Rocky rewarded for his behavior indicatedthat they would select him as a model, whereas only 10(.4. of those whosaw him punished indicated that they would choose to emulate him.Additionally, the authors noted a classic example of how socially rt.pre-!tensible but soccessfnl plodded aggressive acts may iduchee children.One of the girk, vho had ex messed mat ked disapproval of Rocky's ag-gressive behavior as it occurred, later exhibited many of his aggressiveresponses. Vinally. in an apparent effort to make her emulation of theruthless but suceesful 1hky complete. she turned to the experimenterand inquit ed , "Do you have a sack here?"

"V

21

Like research in direct imitative effects, investiptions of disinhibilo-

ry effects have nal been lim;ted to the study of play netivities. For aunt-

pk, in a series of three experiments, Walters and Llewellyn-Thomas

09(1) evaluated the influence of film-mediated aggressive models upon

hospital auendants, high school boys, and young femak adults. In these

studies subjects in experimental groups watched the knife fight scene

from the movie Rebel Withuut a Cause, while control subjects watched

a sequence which showed adokscents engaging in constructive activi-

ties. Both before and after exposure to the film all subjects were asked

to participate in what was ostensibly a "conditioning" experiment

which required them to administer electric shocks to another person for

making "errorc" on a learning task. The difference in the intensity of

shocks which the subjects administered in the two sessions served as the

primary dependent measure. In all three.experiments. subjects exposed

to the ateressive film significantly raised the shock levels vilikh theyadministered relative to the controls. It is important to note that this

heiAtened aggression was manifested in a situation entirely different

from the one depicted in the film, and by subjects drawn from three rath-

er different subcultures.A punk-Only well.desiuned study of the effects of observing aggres-

sion yeas conducted by Hartmann 09(F9). In this investigation, delin-

quent adolescent boys were either angered or treated neutrally and then

showed one of three films, two of whkh were te,:r.ressiVe in content.

Regardless of %%Whet. they .were angered or not, seciag an wzgressive

film produced more subsequent allgression (ostensible electric shods to

another person) than did the neutral film. iloreover. and of particular

importance, boys with a past histoiy of aggressive behavior were more

aggressive than other boys. This finding, and a similar finding by Wolf

and Baron (1971) [comparing college students and convicts with records

of assaultivc crimes) provide validetion for the assertion that laboratory

measures involving button pressing which (ostensibly) inflects harm are

related to interpersonal tygression in natuMistic siulations.

A study reported by Feshbach and Singer (1971a), also involving ado-

lescent boys. is directly at odds v. ith both the fi tdings of the Walters and

Thomas and Hartmann studio and with the preponderance of research

in the area. The study was conducted with approximatdy 400 boys who

were enrolled either in tesijential private schook or in boys' homes

serving you ngslers who are either mildly disturbed or whose families arc

unable to care for them. Approximately half the subjects in each institu-

tion were 1-i:witted to watch a predomiaantly "aggressive" diet of Me-

vision pro;:rams while the remaining subjects were permitted to watch a

diet con;aioinv primarily nonagpressive pro;.rams.

Among the subjects Nom the four boys care homes. significant dif-

ferences bei v. een the groups were fot:nd on measures of aggressive

1,cliavior in Once of them (. rated by institutional personnel) during the

six weeLs of the 0.peiim,:at. srweitieally. in thet.e inNtita lions, boys

exposed to the predomiaantly uggressive diet were kss aggressive than

22 TELEVISION AND SOCIMAEARNING

those exposed predominantly to nonaggressive programs. In eontrnst,the data reveal virtually no significant differences between the aggrcs-sive and the control program groups for boys in the private schools.However, olthotigh limited to lower-class institutionalized nudes or apartkulur age range, the results reported by Feshbach and Singer remainan anomalous oo tome. Unfortunately, there are tt number of funda-mental design problems in this study which cast doubt on both the inter-nal and the external validity of the outcome. For example, in two of thethree institutions in which significantlesults were obtained the control("non-aggressive diet") subjects objected very strongly because Bat-man was not available to them. The experimenters then acceded to thedemand by adding this program to the control diet. Another problem isthat the data themselves were collected by untrained institutional em-ployees rather than by trained personnel, but no adequate reliabilitycheeks appear to have been made nor was the probability of rater biasdealt with systematically. Additionally, the control group boys mayhave behaved more aggressively because of being deprived of some oftheir favorite programs. A more complete discussion regarding theFeshbach and Singer study appears in another Volume 5 in this series.

In a more recent study, Liebert and Baron (1971, in this volume)sought to investigaw.the question of whether exposure to vggression, asmodeled in actuul telcvision fare, would disinhibit you:1;4er children interms of their wihiligness to hurt another child. 11w investigators ex-posed children of 'eoth sexes and two age groups (5-6. years) to briefexcerpts tal:en directly from publicly broadcast television shows. Forchildren in one I:rovp. these excerpts depicted instances of aggression (abrutal fist fight. a shooting, and the like), while for children in a secondcondition, exciting but nonageressive sporting events were shown. Fol-lowing exposure to oat: of these two programs, children in both groupswere provided with a series of opportunities to either hurtor help anoth-er child by pushing, respectively, either a red or a green button. The .

. children wei e to!d that pushing the green button would help this child(who was not actually present during the study) to win a pr;ze, but thatpushing the red button %% ould hurt him. In addition, they were informedthat the longer they pushed either button. the more the other child wouldbe helped or hurt. Resuhs'indicated that children who had observed the

. violent scenes pushed the red button for a significantly longer period oftime than tho.w v. ho ha oberved the non4gresske s%:enes.

In anothU study concerned with disinhihitory effects, Leifer and'Roberts (1971, in this volume) obtained information on the subsequentwiringness to mress of children and adolments (kintk.gartenersthrough twelfth graders) after they watched television programs whichdiffered in the ameant of violent content dkplaycd. The prlegams Weretaken directly from the air. k ithout editing. 1 he children were first test-ed on their tecall and understanding of the motives and consequences ofthe violent acts hich they had seen and wele then tested on a spxially

ammo*

23

&signed response hierarchy of thdr willingness to aggress. The..chiklwas presented with a series of reallife situations ("You're stamfing inlin(l for 8 drink of water. A kkl comes alon g. and just pushes you out ofline. What do you do?") and asked to choose between a pair of alterna-tive responses. One of the alternatives wo typically aggressive ("Pushthem"). while the other was not ("Go awa)").

In addition to finding that children chose physical aggression on the

response hierarchy described above more often at older than at youngerages, they also found an important relationship between the amount ofviolence.in the program a child viewed and his subsequent willingness toselect physically aggressive responses after television viewing.

Specifically, the more violent programs reliably produced higher lev-els of aggressive responding than the less violent ones. It is of furtherinterest that. in this experiment, understanding the motivations for, andconseqtrences of, violence in a programdid not account to a significantdegree for the aggression scores. It appears that the instigating effect ofviewing violence is not reduced by an increased understanding of* the

motivations and consequences which surround it, at least tor this mea-sure and these age groups. It k also of interest to note that the differen-tial effects of program content upon willingness to aggress tended to

vary with age. ahhounh these eitferences are not significant. The natureof the tendency is that the relationship between amount of violence andsubsequent physical aggression on the response hierarchy measure tend-ed to be greater for kindergarteners and third grade children than forchildren la the skill. ninth, or twelfth grades.

In interpreting their results and correctly identifying the study's majormethodological weaknesses, Leifer and Roberts note:

Whatever analysis was performed. the amount of vioknee in the program af-fected the amount of agl:ression snbsequently chosen. Nothing else about theprogramthe czintet within v. hia vioknee %vas presentedseemed to influ-ence subsequent argresskm. Furtho more. w.ir measuies of understanding ofthe cues hypothesized to control e m t iv at ions and consequencesdid not relate at all to aggression choices. These results are not encouraging intheir implkations: however. they should he interpreted uith some caution. Allchikhea were te,ted on their understanding of thc motivations and consequenc-es in the prov rams before they were tested on the response his:la:ay.

Leifer and Roberts al..o report six other studies, varying in their majorput poses. in which the effects of televised aggression upon children'saggressive choices could be assessed. Three of these provide furtherevidence for a disinhibitory effect while none suggests a decrease in ag-

gressiveness after exposure to aggression. This latter neizative finding is

of some interest. Even observing aggression whkh had both bad mo-tives and bad conseqnences (in programs produced by special editing)did not reduce :;;;;:ression Mative to a nonaggressive program.

While television and film programs, a typically observed in both na-turalistic and laboratory situations, do not formally state whether theyare real or liciional in character, the provision of such specific introduc-tions and descriptions may potentially pla; an important rcle in their

eip

..;

24 TELEVISION AND SOCIAL LEARNING

effects. For example, Orson Welles's radki play version or II. G.Wells's War of the Worlds was apparently perceived as reality ratherthan fiction by many adult Ikteners, some of whom went into the streets

armed in order to defend themselves against invading Martians. Parentswho obseive that their children are becoming upset by stories, movies,and television plays often provide reassurance that what is being seen or

heard is "just a story" or "not real." Do such reminders, explanations,

or "identifying statements" moderate the influence of observing aggres-sion on television?

A series of experiments conducted by Feshbach (1971, this volume)was designed to determine if the effects of film-mediated violence uponchildren varies as a function of whether the material was said to be taken

from "real life" or was specifically identified and labeled at the outset as

fictional or fantasy material.Feshbach advanced the hypothesis that ". . .aggressive tendencies

should be kssened or unaffected to the extent that dramatic contentfunctions as fontasy in the larger, cognitive sense and is perceived asfantasy in the narrower, fictional sense. If the dramatic content is per-ceived as 'real,' the possiNlity of facilita!ing aggression through such

processes as imitojion, instruction, and disinhibition should be consider-ably enhanced." He reported three experiments related to this hypothe-sis. In the first of these, in which the subjects were 9-11-year-old boysfrom either lower- or middle-class families, three experimental condi-tions were used: real aggression, fantasy aggression, and control.Among those in the aggression conditions, half the subjects in eachgroup witnessed a war sequence and the other half saw a police actionsequence. Control children were shown either a circus film or no televi-sion whatsoever. The dependent measure was a response box designed

as an "aggression machine" and was similar to the widely accepted typeof apparatus used by Mallick and McCandless (1966) and by Hebert and13aron (1971) in their report in this volume.

In addition to the fact that actual newsreels or movie scenes from Hol-lywood were used, subjects in each of the groups were specifically"set" by the experimenter in terms of what they were about to see.Thus, for example, boys exposed to the reality-army film were told that

they were ". . .going to sec a war film made by a Holly wood studio."Following observation of one of the films, the subjects were adminis-tcred air adjective check list to assess their moods. Subsequently, theypet formed on the aggression machine. While some intetesting changesboth in mood and in progrem ev:duations were noted. no significantoverall differences were obtained on the measure of aggression against

another person.In his second experiment, Feshbach employed the same violent film

for both fantasy and reality groups, but, as he notes, ". . .under clearlydifferent set conditions. In one experimental treatment, the subjeet be-lieved he was seeing a Min of a real event; in another treatment. the sub-ject was shown the same film but was led to believe that it was

. ip m.o. 0.. I. . 1.

25

fictional." In the fantasy set condition, the subjects were told: "We are

going to show you a film that was made in a Hollywood studio. 'l'he story

is about a student riot. You might have seen some of the actors on

television before." The aggressive film was a combination of the

campus riot reality and fantasy films of his first experiment.

The adjective mood list %vas again presented prior to the opportunity

to aggress, so the study shares the methodological probkm of.an "inter-

vening test" thnt also weakens the Leifer and Roberts investigation.

Results clearly indicated that children exposed to the aggr'eSsive reality

film arc more aggressive than those exposed either to no film or to the

fantasy one. Likewise. as Feshbach had expected. children in the fanta-

sy set condition showed less aggiession than children who had not ob-

served a film at all. No important changes in mood. as measured by a

questionnaire. m ere detected. It is unfortunate that this experiment did

not include a control group in which no "set" at all was given. Thus,

especially since many young childien may perceise fictional teleqsion

drama as presenting real-life situations, it would be difficult to extnipo-

late from the data as they stand to naturalistic television viewing by chil-

dren. Nonethe!ess, the finding is potentially quite important, and the

hyposhesis Clearly merits further research.

In a third experiment, Feshbach endeavored to determine the differ-

ential ell ects of reality vs fantasy set on aggression machine belfavior

when subjects were told that the aggre.ssien machine was to be discon-

nected so that they would only be "imagining" the consequences of

pushing various buttons. No statistically significant differences were

obtained in that small, exploratory study.Labonitory studies of the sort described above provide the best

source of information about basic processes and causal relationships.

However, to assure generality of such findings to the more complex nat-

ural environment, such investigation must be supplemented by correla-

tional studies (such as those considered in the preceding section) and, if

possible, also by experimental field studies.

The investigation conducted by Stein and Friedrich (1971) employed

the experimental method in a relatively naturalistic situation in order to

determine some or the cumulative or longer-range effects of observing

television upon ekikiren. The subjects in the Stein ex rcriment were 97

children (52 boys and 45 girls) between three and one-half and five and

one-half years of ;:ge. who were systematically exposed to televkion

ptograms of differing content during the course of their participation in a

summer nursery school.This carefully designed experiment involved an initial measurement

period in which the free play of children in the nursery school was ob-

served and rated according to a variety of categories; a four-week exper-

imental period in wIlich children were systematically exposed either to

aggressive cartoons (lialdtan and Snrounanl. neulrai television pro-

gramming (children wotking on a fat m and the like), and prosocial pro-

gramming (episodes from the prognun Misterot;ers Veighborhood); and

26TELEVISION AND MAL LEARNING

a two-week postvicwing period in which effects could he observed and

assessed.The children were exposed to the programs or films for approximately

20 minutes per day, three times a week during a four-week period.

During this time, and during the two-week postviewing period, the chil-

dren's behavior was again systematically observed in the naturalistic

preschool situation. Behavior ratings included measures of aggression,

proemial behavior, and self-contro1.6 They were checked carefully for

enability and collected by raters who were "blind" to the children's

treatment. Analyses of data were based on four observation periods,

becuse the first and second two weeks of the experimental period were

separated so as to identify changes during this time in the effects of the

programs.Stein and Friedrich found that child, en who were initially in the upper

half of the sample in interpersonal aggression subsequently showed

greater interpersonal aggression if they were exposed to the aggressive

programming than if they were exposed to either the neutral or prosocial

programming: but children who were initially low in aggression did not

respond differentially to these treatments. Children exposed to the pro-

socH and neutral television programs did not differ from each other on

these measures of interpersonal aggression. The investigators up pro-

piltely described their findings on thismeasure as follows:

. s

These results suggest tlmt children who initlay arc high in per ressiun respond

to gotressive plogams with 1.4:.:hcr eels of anrescion titan they

would under neutral conditions. These cflect o. ' et; in nal uralistie ham ior

that was removed ht..th in 601 and in environmea . Ntuing from fire viewing

ex n:rience. They occurred with a small amount of expo.ure. particularly in rda-

don to the amount the dildren received at hume, and they endured during the

postvic win period.

In contrast to the effects detected for interpersonal aggression, the

television programs did not have :i systematic effect upon either fantasy

aggression or ituression toward inanimate objeas.

A second measure.on which the programs were shown to have differ-

ential effects upon children was prosocial self-control. This milt refers

to measures of rule obedience, tolerance of delay, and task persistance

which the children showed in a variety of nursery school situations.

Results showed clearly that children exposed to the prosocial television

programs subsequently displayed higher levels of self-control on each of

these measures than did children exposed to ihe aggressive programs,

while those observing the neutral programs generally showed self-con-

trol which fell lietween the other tw o groups. These findings were partic-

ularly true for children with relatively hiyh lQs. Moreover, the direction

of the changes ia the two groups appeared to bye been antithetical; that

is, children who observed the agressive programs decreased on these

me,isures of sclf-control relative to the baseline, while those who ob-

served prosocial programs generally increased. Thus, as Stein and .

11.0., . *Oa.. s . . ...I . .27

Friedrich notc: "following the procedure of using the Neutral group as

the comparison point for evaluating the effects of the experimental treat-

ments, it appears that the aggressive programs have a deleterious af-

fect on children's willingness to tolerate delay, and to a lesser extent, on

rule obedience."The effects of the programs on children's prosocial behavior were

somewhat more complicated, because they interacted with the socio-

economic background of the children. Specifically, exposure to the pro-

social television programs produced an increase in prosocial interper-

sonal behavior among the lower-class children, while exposure to the

anressive programs resulted in a similar increase moat; the children

from higher socioeconomic bacl:grounds. Thus, the investigators note

that for the latter group, the reduction in self-control ptodueed.by ag-

gressive programs was accompanied by:Increased social interaction that was primarily cooperative. It appears. there-

fore, that the af...uessive programs had t gener:d mimulating viTezt for the higher

SES chiMren that led to hher interaetis,n and liAcr !evils of personal.'

eontrol. For those who were already aggreisive. it led to aggression as wdl.

EVALUATION AND SUMMARY OF RESULTS

Recent research on the relationship between children's viewing of tel-

evision violence and their aagressive behavior was reviewed in the light

of theoretical and methoddogical issues. The data suggest comhistentlythat children are exposed to a heavy dose of violence on tekvision. lt is

also clear that they Can and ep retain some.of the aggressive behaviors

which they see, and ate often abk.to reproduce them. Differences in

recall as a function of age are in the expected direction (better recall

with increasing age). Differences in recall as a function of content are

less clearly understood, but violent content appears to be learned and

remembered at least as well as nonviolent fare. .

Children often accept and directly imitate observed aggression in their

play activities if the observed performances have been rewarded or have

reaped no consequences. They also-report copyina actual televised se-

quences in phy. Punkhment to an aggressive model leads children to

avoid reproduction of the exemplaty behavior, but does not prevent

learning or subsequent performance under more favorable circum-

stances.Correbtionat studies show a regular .association between aggressive

te!evkion viewing z.nd variety of measures of aggression. For

meawiei of atiitudes and approl, al of aggression, such a result often

appears for both Sexes. When measures of aggressive behavior arc con-

sidered. the relationship may Ue limited to boys. These correlational

studies have emrloyed impresskely broad samples in terms of range of

'economic bael:eiounds and family chat acterist ies.

28 TELEVISION AND SOCIAL LEARNING

It is.important to note that the correlational resuks, while generallyconsistent, point to a moderate (rather than a strong) relationship be-tween watching television violence and subsequent aggressive attitudesand behavior. Further the relationship is attenuated by the presence ofcertain family and cultural characteristics. Just as a child's food diet orexercise are related to his health but do not predict it exclusively or evenpredominantly, children's exposure to television violence is related tosome measures of aggression. Correlational studies support, but du notprove, the hypothesis of a causal relationship between exposure to tele-vision violence and aggression.

Experimental studies of disinhibitory and inhibitory effects prepon-'derantly show that observing violence can lead to an increase in a childobserver's willingness to agoess. These findings are also consistent withthe correlational data. But, as Berkowitz has noted, it is important todistinguish between the statement that observudion of violence ear/ havesuch effects and that it will have them for any particular child or pro-gram, since ". . .a good many situational and personal factors influencethe relationship between witnessed violence and the likelihood of ag-gressive actions by observers, including the observers' attitudes towardthe violent event, the extent to which they are sct to act aggressively,their aggressive habits, etc."

Recent studies provide suppoa for each of the qualifiers mentionedabove. For example. Stein and Friedrich found that naturalistic aggres-sion was increased by watching aggressive cartoons only for childrenwho were relativdy aggressive initially; Feshbach found that a.specifiefantasy set actu:dly reversed the usual impact of having aggression; andEl:man. Liebert, Friesen, Harrison, Zlatchin, Malmstrom.. and Baron(1971) found that the 5.6-year-old boys in the Liebert and Baron studywere instigated by observing aggression only if they displayed positivelachl affects while viewing.

Evidence supporting the assertion that televised violence can reduceaggression is scant and is directly at odds with correlational data basedon widely varied samples as tapped by many different researchers.

Almost all experimental studies have some methodological flaws. TheLeifer and Roberts (1971) study injected an intervening test betweenexposure to violence and their measures of aggression; the stimulusmat eriak employed by Liebert and Baron (1971) preserved a story linein the. aggressive program but not in the noncg;;ressix e one; Stein andrriedrich's control group observed films while the other groups watchedactu:11 television Fograins. An ex tended list would encompass almost allthe ;Wm exNuimenial studies cited in this paper. It is important, how-evei . that each of these Ilaws tends to be iiniqu e rather than shared by allof 1;ie invo.tiv;ttors..The studie.: are quite consistent in the overall direc-tion of their findings and, in the aggregate, may be defended soundlyagainst charges of confounded elf ects.

:78

e 6. a 00 1.0 .

29

While it is possible to thoughtfully analyze any given investigationregarding the question of adequaey of design, there is no easy answer tothe question of their external appiicability. 1:1 fact, sophisticated re-searchers have long no forsaken the concept of the "critical experi-ment" regarding any sort of final knowledge about general processes inthe social sciences. Instead, it is widely agreed that the best solution liesin considering the accumulated weight of evidence on a particuka issue,coming variously front surveys, correlational studies, and experimentsconducted both in the laboratory and in the field. If the balanee is suffi-ciently tipped v. hen a variety of methods, approaches. and laboratories

are considered, then the researcher can draw conclusions of social appl-icability with some confidence.

The present writer believes that the findings discussed here. in con-junction with the considerable body of earlier research, warrant formal-ly advancing some tentative conclusions into the arena of public debate.Specifically, the fgllowing summary is suggested by the data in aggre-

gate:I. It has been shown convincingly that children are exposed to a sub-

stantizd amount of violent content on television, and that they can re-member and 'earn from such ex posu re.

2. Correhiiiw,al studies have disclosed a regular association betweenaggressive television viewing and a variety of measures of aggression,employing ktressively broad samples in terms of range of economicbac kgrou od and geographic mid fam ily characteristics.

3. .Experimetital studies preponderantly support the hypothesis thatthere is a directional, causal link between exposure to television viol-

ence and an observer's subsequent aggressive behavior.It has been repeatedly shown in experimental studies of observing

anression in film and television formats that, under some circum-stances. disinhibitory effects can mew in samples covering the agegamut from preschool children to mature adults. At almost every agerange, such findins have been found by at least two or more independ-

ent research wams. These results ceneraily mesh with the now numer-ous cot relational studies which are able to approach more closely the

situations in which viewing, and aggression, occur naturalistic:dly. Stud-ies failing to produce statistically reliable results are a distinct minority,and those sugl;esting that seeihg aggression !educes 4giessioa are rareenough to be called anomalous.

If a probabilistic view of the accumulated evidence is taken, as it typi-

cally is in fte health sciences, the v. ei;:ht of the evidence to date would

seem to represent real progres in determinin the effects of violence on

television upan youngiters. Specifically, there k more than a ttivial lan-

sis for a "best guess" conclusion which is central to the major question:At least under some eireumstmwes, exposure 10 telerim.d aggession

Cali lead children to accept ;that they have seen Ps a partial guiLle for

;7.19

;

30. . . . .

TELEVISION AND SOC:IM. LEARNINO

their own actions. Asa result. thc present entertainment offerings of the

televisim medium may he contributing, in some measure, to the aggres-

sive behavior of many normtd children. Such an effect has now bccn

shown in a wide variety ofsituations.; do. .

FOOTNOTES

1. Sincere thanks are due to Emily Davidson, Diane Lichen, John

Neak, Jacqueline Portnoy, Rita Poulos. and Michael Sobol for their

assistance in the preparation of thk paper. The opinions expressed

herein remain the responsibility of the author, not of the staff of the

Television and Social Behavior program nor necessarily of the inves-

tigators whose work is described.2. According to the usage of Bandora and Walters. inhibitory and dkin-

hibitory effects may iNclude responses which are dissimilar to those

displayed by the model, but the emphasis is placed upn the ft:et that

a class of behaviors can be affected. Thus. for these theorists, dis-

similarity is not a definingproperty of the processes.

3. It may not be immediately obvious that this last question is a correla-

tional one. It is. however, merely an alternate form of the question:

Do the elects of tele'vision go !wilier with the sex of t.he observer?

4. It should be emphasized th:It this expkmation is intended only to illus-

trate the reasoning involved in this technique, and does not fully re-

flect its complexities. A complete discussion may be found in Neale

and Lieber t (in press).5. An analot;ons experiment conducted by Larder (1962) showed a simi-

lar increase in preference for an aggessive toy by children .Aho had

merely heard an aggressive story, as compared with.children who

had been exposed to a nonaggressive story. .

6. While many interesting correlations reported by Stein and Friedrich

will not be reviewed in this paper. it is of some interest that overall

measures of interpersonal aggression and interpersonal prosocial

behavior were positively related: i.e., they tended to go together.

Thus, children who were more aggressive were also more likely to

engage in prosocitd types of activities. It is extremely important to

note that this result does not mean that the two categories will be

similarly susceptible to such additional influences as the effects of

television. For example, in atinus, height and weight (ns Illeiisur:s of

body size) are also positively !elated. Nenetheless, if we consider he

eff ects of a chr!nge of diet (in this case nutritional diet rather thzan tel-

evision diet). it k clear that three milkshakes daily would readily in-

cre::sc the weight of most people but wovld not influence the height

ei any of them. Thus. potie. relationships between two measures

do not mean that both of them will be equally influenced by the pres-

ence or absence of some sort of treatment.

0

.

31

REFERENCES

Ball. S., and Bogatz, E. A. The first year of Sesame Street. An evalua-tion. Princeton: Educational Testing Service. 1970.

Bandura, A. Influence of models'. reinforcement contingencies on theacquisition of imitative responses. Journal of Personality and SocialPsychology.I965, 1, 589-95.

Bandura, A. Prineipks of behavior modilkation.New V ork: bolt,Rine-hart &Winston, 1969.

I3andura, A., and Menlove, F. L. Factors determining vicarious extinc-tion of avoi.lance behavior through symbolic modding. Journal ofPersonality and Sovial Psycholo;y.1967 , 5, 16-22.

Bandura, A.. and Mkehe!, W. Nlodificntion of self-imposed delay ofreward throuth exposure to live and s} mbolie models. Journal ofPersonality and Social PAyehology.1965. 2(5). 69S-705.

Bandura. A., Ross, D., and Ross, S. Imitation of lilm-mediated aggres-sive models. Journal of Abnormal and Soda! Psychology, 1963,66(1),3-11.

Bandura, A., Ross, D., and Russ, S. Vicarious teinforcement and imita-tive leamilly,. Journal of Abnormal and Social Psychology, 1963,67(7), 601-07.

Bandura. A.. and Walters. R. Social karning and personality dory!.opment. New Y oik:Holt.Rinehart &Winston. 1963.

BerKowitz. L. Experimenhl investigations of hostility catharsis. Journalof Consuhing and Clinical Psye1,ology.1970 35(I), 1-7.

Bryan, J. and Schwartz, 'I'. Effects of film material upon children's be-havior. Psychological liirlkqin, 1971. 75(1), 50-59.

Bryan, J.. and Walbek, N. Preaching and practicing generosity.Children's actions and reactions. Child Developmen1,1970, 41, 329-53.

Coates, 11:, and Hartup. W. Age and verbalization in observationallearning. Developmental Psychology. 1969, 1, 556-62.

Dominick, J. R., and Greenberg. B.. S. Attitudes toward violence: theinteraction of televbion exposure...family attitudes, and social class. .

Television and Social Iteharior: a report to the Surgeon General'sScientific Advisory Committee, Vol. 3. Washington,.D. C.: U. S.Government Pi intik: Office, 1971.

Ekman, P., Lichen, R. M., Friesen, W. V., Harrison, R., Zlatchin, C.,Malmstroi t. E. J.. and 11aron. IL A. Fzial Expressions of emotionwhile watching televked violence as predictors of sul.equent aggresiion. Tekvision a;kt Ih.havior, Vol 5. Washington, D. C.:U. S. Government Printing Oilicc. 1971.

Eron, L. Relationship of P.' viewing habits and aggressive behavior inchiklren. Jourwil of .41 normal and Psychohvy, 1963, 67(2),

. 193-94. 1963.

. ,

32. I TIMPVISION AND SOCIAL LEARNING

Feshbach, S. Effects of reality versus fantasy in filmed violence.

Television and Social lkhavior, Vol. 2 (this volume).Feshbach, S., and Singer, R. Tolevifion and Aggresskm. S;in Francisco,

California: Jossey-Bass, Inc., 1971.Grainer, G. Violence in television drama: a study of trends and symbol-

ic functions. Television and Social lkhavior, Vol. 1. Washington,

D. C.: U. S. Government Printing Office, 1971.

Goranson, R. E. A review of recent literature on psychological effects

of media portrayals of violence. Report to the National Commission

oa the Causes and Prevention of Violence, 1%9.

Greenberg, 11. S., and Gordon, T. F. Children's perceptions of tekvi-

sion violence: a replication. Television and Social Behavior, Vol. 5.

Washington, D. C.: U. S. Government Printing Office., 1971.

Hanratly, M. A. !mit:Ilion of film- meeiated agiression nainst live and

inanimate victims. Unpublished master's thesis, Vanderbilt Univer-

sity, 1969.Hanratty. M. A.. Liebert, R. M., Morris, L. W., and Fernandez, L. E.

_Imitation of film-inedia!ed aggression against live and inanimate vic-

tims. Proceedings of tke 77th Annual Convention of the American

Psychological Association, 1%9. pp. 457-58.

Haaratty. M. A., O'Neal. E., and Sulzer. J. 1.. Thc effect of frustra-

tion upon imitation of aggression. Journal of Personality and Social

Psychok;;:y, in press.Hartman, D. P. Influence of symbolic:My moth led instrumental aggres-

sion and pain cues un aggrctisiVe behavior. Journal of Personality

and Social Psychology. 1969, 11(3). 10-S8, 1969.

Hicks, D. J. hnitation and retention or film-mediated aggressive peer

and adult models. Journal of Personality and Sociul Psychology,

1965, 2(1). 97-100.Hicks, D. J. Effects of co-observer's sanctions and adult presence on

imitative aggression. Chik1.1%.vehTmcnt, 1968, 38(1). 303-09.

Hill, J. L. Lieber!, R. M., and Mott, D. E. Vicarious exthiction of

avoidanCe behavior through films: an initial test. Psychological

Reports, 1968, 2. 192.Larder, D. L. Effect of aggressive story content on nonverbal play be-

havior. P.sychohgieal Reporls, 1962, 11. 14.

Lefkowitz, M. M., Eron. L. D., Walder. L..0., and Iluesmann, L. R.

Television violence and child aggression: a followup study..

Television and Social Behavior, Vol. 3, Washington, I). C.: U. S.

Government Pi inting Nice, 1971.Lcsscr, G. S. Dcsirnin a pi egiam for hroadcast television. In Korten,

F F., Cook, S. W., and Lucy, 0. L. (Eds.) psychohTy ,id the

probkos of society. Washington: AmericimPsycholovical Associa-

tion. 1970. 20%-14.

Lk-bert, R. M.. and I.:l('fl 1. A..Short-terin ejects of tele\ ked anre-

Sion on ehildNn's ar.essive behavior. 1elevi4on wid Swial lk-

havior, Vol. 2 (this

1 33

Hebert, R. M., and reinandez. L. R. Imitation as a function of vreari-ous and direct reward. Developmental Psychology, 1970. 2(2), 23°-32.

Licbert, R. M., Odom. R. D.. Hill, J. 11.1 and Hull, R. 1.. liffects of ageand rule familiarity on the production or modeled language con-structions. Developmental Psychology. 1969. 1(2), 108-17..

Lichen, R. M.. and SpiegIcr, M. D. Personality: an introduction to theo-ry and research. liornewood, TheDorsey Press, 1970.

Leifer, A., and Roberts, D. Children's responses to television violence.Television and Soeial Behavior, Vol. 2 (ibis volume).

1.8vans, 0. Ivan. Effect of exposure to symbolic aggression on aggres-sive behavior. Child Development, I 961. 32, 37-44.

Lyle. J., and it off mnn, 11. R. Children's use or television and other me-dia. Television and Swial Behallor.Vol. 4.

McIntyre, J. J., and Teevan, J. J. Television and deviant behavior.rckuision and Soda Behavior, Vol. 3.

Mcf.eod, J. M., Atkin, C. K and Chaffee, S. H. Adolescents, parents,an4 wkvision use. Tek vision and Social Ikhavior. Vol. 3.

Mallick. S. K.. ald MeCatidless. B. R. A study of catharsis of aggies-Sion. Journal of Personallty and Sock3 cholo1y,1966, 4. 591-56.

Martin, M. F., Gelft.nd, D. M.. and Hartmann. D. P. Effects of adultand peer observers Oa children's responses to an aggressive model.MN Dew/iv:nen/. 1971, in press.

Mussen, P., and Rutherford. E. Fleets of aggresNive cartoons on chi!-

dren's a:!gressive piny. Journal of Abnormal and ,SL.cial Psychology,1114111/MNI01.010."°. 1IP 111.11

Neale. J. M., and 1.ichert, R. M. Science ma social behavior; An intro-duction to methoefs of rescitin.h, Englewood Cliffs, New Jersey:

Prentiee-llall, in press.Osbot n. D. K., pnd Endsky. R. C. Emotional reactions of young chil-

dren to TV violence. Child Development, 1971, 42, 321-31.

Reid, J. B. Reliability assesc.inent of observation data: a possible metho-dological problem. Child Dc velopnwnt, 1970, 11.13-50.

Rosenhan. D.. and White. G. M. Observation and rehearsal as determi-nants of pro-social behavior. Journal of person.gity and Social Psy-choloi!y. 191.7, 5, 424-31.

Rosenthal, R. Experimenter effects in behavioral research. New York:Appktoa, 1%6.

Rosenth:d. P.., and Gaito. The inic.rprct.!; ion of levels of sienificance bypsycholojeal researchers. Journal of Psychology,11:0, 55. 33-38.

Savitsky- !togas. R. W., 1%mA. C. F., and 1jeber1. The roleof frnstrrition vud a:!.,:er in the imitation of filmed rlession agninsta hinnan vhnim. P.:yd;o!of,:kd Rcpeos. 1.471. n,

Schramm. Lyle. J.. and Parker, I. Television in the lives of ourchildion. Stanford: Stanford Univerity Press, 1961.

'13

am. yr

,

34TELEVISION AND SOCIM. LEARNING

Stein, A. 11., and Friedrich. I. K. Television content and young chil-

dren's behavior. Television and Social Behavior, Vol. 2 (this vol-

ume).Stevenson, 11. W. Television and the behavior of preschool children.

Televhion and Social list.havior, Vol. 2 (this volume).

Walder, L. O. Abelson, R., Eron, LI'D., Banta, T. J., and Laulicht, 3.

H. Development of a peer-rating measure of aggression.

Psychological Reports. 1961, 9, 497-556.

Walters, R. 11., and Llewellyn-Thomas, E. Enhancement of punitive-

ness by visual and audiovisual displays. Canadian Journal of Psy-

chology, 1963, 17, 244-55.

35

Appendix A:TV and Social Learning:

A Summary of the Experimental Effectsof Observing Filmed Aggression

Gloria D. Strauss and Rita W. Poulos

This appendix presents the major results of studies appearing in Tele-vision and &vial Behavior: an annotated bibliograph:' of research fo-cusing on television's impact on children (AtLin, Murray, , and Nayman,1971), rrepered under the anspices of the Natiomd Institute of MentalHealth. All research reports included here are experimental in natureand were &doled to explore the effects of filmed aggression on the atti-tudes and/or behavior of the audience. Additionally, each report met thefolloying requirements:

1. One or more groups of subjects were exposed to films or televisionprograms displayhig violence or aggression.

2. At kast one group of subjects was not exposed to film or televisionaggressive content, thereby constituting a control condition. .

3. At least one dependent variable was examined which might tipsubjects' aggression, whether measured behaviorally or through attitudequestionnaires.

4. A statistically significant result was obtained.The material is pre::ented in two lists. The first includes those studies

which lend support to the hypethesis that observing filmed viofence caninstigate aggression by the viewer. The second group of studies includesthose which support the hypothesis that observing filmed violence can..educcaggressive behavior by the viewer.

It should be noted that this is not an independent literature search,since even studies conducted more recently for the Television and So-cial Behavior program or many citeil in Lieberes review are not includ-ed. Rather, it is an attempt to summarize the findings of experiments in-cluded in the Athin. Murray, and Nayman biblio;:traphy according totheir relevance to a pioticular pair of oppoNing hypotheses.

35

cs's

The

follo

win

g st

udie

s pr

ovid

e su

ppor

t for

the

hypo

thes

isth

at v

iew

ins

aggr

essi

on c

an in

stie

ate.

avre

ssio

n:

Rel

eren

ceS

ubje

cts

Inde

pend

ent v

aria

bles

vr.r

inu!

s;R

esul

tS

I A. :

a nd

ura

,an

d71

c..3

2A. S

andu

ra.

Rot

s an

dR

oss

3A.

Ber

kow

itz,

Con

vin,

end

Pei

roni

rnus

(19:

33)

4A, f

lerk

ewitz

and

G-c

ri(I

MC

)

Nur

sery

scho

olch

ildre

n

Nur

sery

.sch

ool

child

ren

Col

lege

mal

es

Col

lege

mat

es

...:V

.agt

4V$4

,

Fou

r co

nditi

ons:

Eeh

r.vi

ora

l1

) hu

man

.1.1

gres

5ive

ratin

2 in

mod

eiliv

ep:

ay s

ituat

ion

2) h

...m

an e

nres

siue

3) c

arto

on a

gcre

ssei

re

4) n

o tr

eetr

rnnt

Fou

r co

nditi

ons:

I) a

gere

s!.:v

e m

odel

ree.

arc7

ed21

agn

rcs:

ive

mod

elpu

nish

ed3)

act

:.q.:

bu: n

on-

ae.)

:!:S

ilve

mod

al4)

no

expo

sure

tom

odel

Thr

ee fa

dtor

s:1)

anp

red

vs. n

eutr

ally

trea

ted

at21

ase

reqs

ive

vs.

neut

ral f

ilm3)

just

i1ip

d vs

.un

j:izt

ified

film

edas

ress

;on

Thr

ee fa

ctor

s:1)

ang

red

vs. n

eutr

ally

trea

ted

fis

Beh

avio

ral

ratin

g in

play

si =

Van

Atti

tude

mea

sure

s

Ele

ctric

sho

ckad

min

iaer

edto

con

fede

rate

Ss

who

&ne

rved

age

ress

ive

mod

els

exhi

bite

d m

ore

acer

essi

on th

an d

1/41

thos

e w

ho d

idno

:. F

ilmed

inst

ance

s of

ags

ress

ion

tr..e

re e

s ef

fect

ive

aSse

quen

cas.

Ss

who

obs

erve

d th

e as

eres

sive

mod

elre

war

ded

disp

laye

d m

ore

eggr

eseo

n th

an th

ildre

n in

the

othe

r th

ree

grou

ps. S

s w

ho o

bser

ved

the

amre

nive

mod

el p

unis

hed

ded

not d

iffer

from

thos

e w

hovi

ewed

the

nona

egre

ssiv

e m

odel

or w

ho w

ere

not

expo

sed

to a

mod

il.

Ss

who

wer

e an

gere

d an

d vi

ewed

a fil

m in

volv

ing

just

ified

aeg

reas

ion

indi

cate

d gr

eate

r ho

stili

tyto

war

d th

eir

torm

ente

r th

an E

s w

ho v

iew

ed th

e.n

eutr

al fi

lm. S

s w

ho v

iew

ed :h

e ne

utra

l film

dis-

play

ed m

ore

"indi

rect

agg

ress

ion"

than

cgd

thos

ew

ho v

iew

ed a

film

incl

udin

g I s

eque

nce

of r

elat

hrel

yun

just

ified

agg

ress

ion.

.

Ang

ered

.1F

s w

ho v

iew

ed th

e re

gres

sive

film

and

who

inte

ract

ed w

ith th

e cu

e-va

lenc

ed c

onfe

dera

ted

wer

em

ore

aggr

essi

ve th

anas

in a

ny o

ther

gro

up.

ON

O

Cel

.5;

tn - .0

The

follo

win

g st

udie

s pr

ovid

e su

ppor

t for

the

hypo

thes

isth

e: v

iew

ing

agcr

essi

on c

on in

stig

ate

aggr

essi

on: ;

Con

tinue

d)

Ref

eren

ceS

ubje

cts

inde

pend

ent v

aria

bles

Mai

n de

pend

ent

varia

ble

Res

ults

5A. B

erko

witz

and

Cee

n05

167)

GA

. Gee

n an

dB

erko

witz

1111

0

7A. 1

4:1n

-etty

.

Mor

ris, a

ndF

erna

ndez

(150

0)

SA

. Har

tman

n(1

0ZS

)

CO

MV

.)m

ales

Col

lege

stud

ents

4 en

d 5

yeer

-old

mal

es

Ado

lesc

ent

ma:

es(d

elin

quen

t)

2) e

7.7.

.esc

ive

vs.

neut

ra! f

i:m3)

con

fede

rate

did

or

did

not h

ave

cue

valu

e

Thr

ee fa

ctor

s:1)

rm

-er.

sive

vs.

neu

tral

Elm

2) *

*jus

tifie

d vs

. unj

ustif

ied

I il;c

J ag

3rer

sion

3) c

onfe

dere

te d

id o

r di

dno

t hav

e cu

e ve

lenc

e

Thr

oe fa

ctor

s:1)

frtis

trai

ed v

s. in

sulte

dvs

. neu

trai

ly tr

eate

d S

s2)

aee

ress

ive

vs. n

eutr

alfil

m3)

con

fede

rate

did

or

did

rot h

ave

cue

%/e

lm

One

thct

or:

1) f:

lm in

whi

ch a

n ad

ult

aore

ssed

aga

inst

acl

own

vs. n

o fil

m

Tw

o fa

ctor

i:1)

eng

ered

vs.

neu

tran

ytr

eate

d S

s.

Ele

ctric

sho

ckad

rnin

iste

res.

! to

conf

t.der

Lte

Ele

ctric

sho

ckad

min

iste

red

toco

nfed

erat

e

Ss

who

saw

the

just

ified

agg

rers

ive

film

and

inte

r-ac

ted

with

a c

ue-v

eien

ced

conf

eder

ate

wor

e n-

42re

angr

essi

ve th

en th

ose

iNho

saw

the

neut

re: r

iim a

ndin

tera

cted

with

the

cue-

Val

ance

d co

nfed

erat

e.

Es

who

wer

e in

sulte

d pr

ior

to o

bser

vMg

an c

ores

-si

ve fi

lm b

elie

ved

mor

e ac

gres

sive

ly th

an d

id S

sw

ho w

ere

insu

lted

prio

r to

obs

ervi

ng th

e ne

utra

lfil

m.

Beh

dvio

ral r

atin

gC

hild

ren

who

vie

wed

the

film

wer

e m

ore

aggr

el-

in a

p:r

y si

tuat

ion

sive

tita

n th

ose

who

did

not

.

Adm

inis

trat

ion

As

who

witn

esse

d th

e ag

gess

ive

film

adm

inis

tere

dof

ele

ctric

sho

ckm

ore

shoc

k th

an th

ose

who

saw

the

neut

ral f

ilm. .4

0

The

fol

low

ing

=tid

ies

prov

ide

supp

ort f

or th

e hy

poth

esis

that

view

ing

azar

essi

on c

an in

stig

ate

aggr

essi

on: (

Con

tinue

d)

Ref

eren

ceS

ubje

cts

Inde

pend

ent v

arie

b:es

Mai

n de

pend

ent

varia

ble

Res

ults

9A. H

z;nr

icn

(193

1)

10A

. Lov

nes

11A

. Mus

sen

Rut

herf

ord

(126

1)

12A

. Vy'

Llte

rsan

dP

arke

(1:3

4)

Chq

r Ir

en,

ard

12 to

16 Nur

Fer

y

chiid

ren

S7x

- an

dse

ven-

ye:3

*-ol

dch

ildre

n .

Pre

t;cho

olm

ales

2) fi

lm in

voM

ng a

ncu

trai

gr:.m

e vs

. film

of a

bor

sac

cres

sive

beh

avio

r vs

.fil

m o

f the

pai

n re

actio

nsof

a v

:ctim

of c

:ort

ssio

n

One

fr...

:tor:

Atti

tude

1) a

7.r.

!ssi

on-a

rciu

sing

vs.

mea

sure

ser

p::5

sing

vs.

am

lival

ent

film

s

One

fact

or:

1) a

m(!

ssiv

evs

. non

-ac

c,es

sive

car

toon

Tw

o fo

ctor

s:1)

frus

trat

ed v

s. n

eutr

ally

trea

ted

Ss

2) a

ggre

us,v

eca

rtoo

n vs

.no

nagm

.siv

e ca

rtoo

n vs

.no

cart

oon

Fou

r co

nd:ti

ons:

1) d

isob

edie

ntm

odel

rew

arde

d2)

dim

:et:c

lient

mod

elpu

n i

ld3)

Pth

e:cl

ient

mod

elno

een

scou

ence

s4)

no

film

Cho

ice

of e

sgre

s-si

ve v

s. n

eutr

al to

y

Oes

:re

to d

este

oyon

inan

imat

eob

ject

Rer

ista

nce

tote

mpt

atio

n to

diso

bey

,41 A

g:re

ssiv

e at

titud

es w

ere

stim

ubte

d by

som

eag

:res

sion

-arc

usin

gfi

lrre

edu

ced

byon

e o?

the

appe

asin

g fi

ims,

and

not

infl

uenc

ed b

y th

eam

-b

iva

len

t

Ss

who

hod

soo

nth

e ag

wes

sive

film

pre

ferr

ed th

e0.

1111

v:si

ve to

y, w

hile

.as.

who

had

view

ed th

e no

n-an

rass

;ve

film

pref

erre

d th

e no

nags

ress

ive

toy.

Sw

ho s

aw th

e ag

rjres

sive

film

exp

ress

ed m

ore

avzs

ress

ive

impu

lses

than

thes

e in

otn

er g

roup

s.

as w

ho s

awth

e m

odel

pun

ishe

d en

d th

ose

who

mw

no

film

exh

ibite

d lit

tle d

isob

edie

nce.

Chi

ldre

nw

hoS

awdi

sobe

denc

e re

war

ded

or r

ecei

ve n

oco

nseq

uenc

es d

isob

eyed

mor

e.i.

Ira 00

The

fono

win

g st

udie

s pr

ovid

e su

ppor

t for

the hypothesis

thet viewine agsre:n can instirlteascression : (Continued)

Reference

Subjects

Independent variables

Moi

dep

ende

ntResults

13A.

7:3l

ttr5

Thorn:1s

(1963)

14A. 'In: tcrs

and S.7:31

15A. Waitrs.

Leal. and

kro.: Zei

(1:"CP:

1GA.

r.'alters,

Lirnllyn-

Thumas, and

Ac'rxr

Hieh

-chool

male: end

adults

7-11 year-old

males

(eruot

dA1'cd z:nd

nor maf I

Kinder-

garten

males

Adult

males

One factor:

ecAresrive film vs.

neutral film

Three conditions:

1) tvoress.ive model.

frn

2) rtcnvmre7sive model-

3) neutrai film

Three conditions:

1) disoltt.dient model

rewarded

2) d:se:ze.tlient model

puni

shdA

3) no film

Two conditions:

1) einressive film

2) neutral film

Adm:ni:tration

of electric shock

Behavioral

rating in a

play situatien

Resistance to

temntation to

disobey

Shock to

confederate

Ss

who

had

see

n th

e ag

gres

sive

film

adm

inis

tere

dmore shocks than they had donepreviously.

a; r

hoobserved the aggressive model dsplayed

more

apre

ssio

n th

an d

id S

s w

ho w

ere

expo

sed

toth e. nona:gressive model or neutral fiim.

Ss

who

sew

the

mod

el r

ewar

ded

wer

e m

ore

aggr

es-

sive

than

thos

e w

ho s

aw th

e m

odel

pun

ishe

d: Is

who

f:aw

the

mod

el p

unis

hed

aggr

esse

d 'e

ssth

anth

ose

who

did

not

see

a fi

lm.

Is w

ho v

iew

ed th

e ag

rese

ive

film

wer

e m

ore

aggr

essi

ve.

de.

^

The

foilo

win

g st

ud:e

s pr

ovid

e su

ppor

t for

the

hypo

thes

isth

at v

ier:

inn

crcs

son

Cal

red

uce

acT

essi

on

Ref

eren

ceS

ubje

cts

Inde

oend

ent v

aria

blus

Mai

n de

pend

ent

varia

ble

Res

ults

1 B

.F

eshb

ach

Col

lege

Tw

o fa

ctor

s:(1

061)

mal

es1)

inw

Ited

vs. n

eutr

ally

tre.

.:mi

2) a

gr;fe

f.siv

t: vs

. neu

trel

film

Atti

tude

mea

ures

Ss

who

had

bee

n in

sulte

d pe

or to

vie

win

g th

ean

gres

sive

film

sho

wed

less

cgg

ress

ion

then

sim

ilar

Ss

who

vie

wed

the

neut

ral f

ilm.

23.

Fes

hbac

hP

read

oles

One

fact

o':

Dai

ly b

ehav

ior

Exp

osur

e to

agg

ress

ive

tele

visi

on c

onte

nt r

educ

ed .

(196

9)ce

nt a

ndst

eady

die

t of a

ggre

ssiv

era

ting

by in

stitu

-or

con

trol

led

expr

essi

on o

f agg

ress

ion.

adol

esce

ntT

V v

s. s

tead

y di

et o

f'S

one;

per

sonn

el:

me:

esne

utra

l TV

sho

ws

pers

encl

ity te

sts;

ettit

ude

:met

es.

ea'

3

Gem r12 : yam C

6

-41

REFERENCES (APPENDIX A)

Atkin, C. K., Murray. J. P., and Nayman, 0. B. Television and sorb/behavior An annotated bibliography of researa foenung on televi-sion's impact on chihhvn. National Institute of Nlental Ilealth, Pub-lic !lean Service Publication, No; 2099, 1971.

Bandura, Albert, Ross, Dorothea, and Ross. Sheila A. Imitation of film-mediated aggressive models. Journal of Abnormal and Social Psy-chology. l963, 66(1), 3-11.

13andura, Albert, Ross, Dorothea. and Ross, Sheila A. Vicarious rein-forcement and Hitative learning. Journal of Abnormal and SocialPsychology,1963.67((i), 601-07.

Berkowitz, Leonard, Corwin, Ronald, and lieironimus, Mark. Filmviolence and su)'sequent aggressive tendencies. Public OpinionQuarterly, 1963, 27, 217-29.

Berkowitz, Leonard, and Geen, Rusgell G. Film violence and the cuepropeaies of avaibble targets. Journal of Person:My and Social

, Psycho loey, 1966, 3(5), 525-30.13erkowitz, 1.conard. and Geen. Russell G. Stimulus qualities of the tar-

get of anression: a further study. Journal of Pers Ottality and SocialPsychology, 1967, :M. 364-68.

Feshbach, Seymour. Stimulating versns cathartic divcts of a vicariousanressiveactivity. Journal of Abnormal and Soda! Psychology.1.961, 63(2), 3S1-85.

Feshbach. Seymour. The cvtharsk effect: research and another view. InD. L. Lange. R. K. Baker. and J. S. Ball (Eds.). Violence and themedi.v. Wash h. :on. D. C. National Commission un the Causes andPrevention of Violence, 1969, pp.:Al-472.(See also Feshbach, S.. and Singer, R. Tekvision and aizgression.San 1:1:tncisco: Jossey Bass, 1971.)

Gun, Russell G., and Berkowitz, Leonard. Some conditions faci1itatio2.the oeeturence of aggression after the obser% Atm of violence.Journal Of Personality,1967, 35, 666-76.

Ilanratty. M. A., Liehert. R. M., Morrk, L. W., and Fernandez: I.. E.Imitation uf film-mediated ag:.,.ression against live :did inanim::te vic-tims. Proceeding% of the 77th Annual Convention of the AmericanPsychological Association, 1969, pp. 457-58.

Hartmann, Donald P. 4 olluence of symbolically modeled instrumentalaggrescion and pvin clies on agrresive behavior. Journal of Person-ality and Social Psychology, 1969, 11(3), 280-88.

Heinrich. karl. Filmerieben, wirkung, Fihnerziehung: Der EinliaNsdes Villas au! die A:,:!re.%sivitac! furendlichcn: ExperinientelleUntet-m:cinini!en und ihre Icrn-pychologiN:..htn Ko1;seeven/01.(Fil.n experience. OP2 effects, liba education: the 1!):!::em:e of filmson og:.rcs.citelle.s of youth: e Ypc.-iments ;tad coaccqueuces fcr thepsyclio:ogy of loarnk2;.)Berlin: 11. Schroedcl, 061.

ts.

42 TELEVISION AND SOCIAL LEARNING

LAMM, 0. lvar. Effect of exposure to symbolic aggression on aggres-

, sive behavior. Child Development, 1961, 32, 37-44.

Mussen, Paul. ant' 'tutherford. Eldred. Effects of aggressive cartoons

on Children's alTressive p hi y . Journal of Abnormal and Social lisy-

chokgy. 1961. 6(2), 441411.Waiters, Richard II.. and Parke, Ross D. Influence of response conse-

quences to a social mad on resistance to deviation. Journal ofExperimental Child Psychology, 1964: 1, 269-80.

Walters, Richard 11.. and Llewellyn-Thomas, Edward. Enhancement of

punitiveness by visual and audiovisual displays. Canadian Journal

of Psychohwy. 1963. 17, 244-55,Walters. Richard H.. and Willows, Donna C. Imitative behaVitir of dis-

turbed and nondisturbed children folkwing exposure to aggressive

and nonaggressive !yodels. Child Development. 1901, 39. 79-89.

Walters, Richard II., Le.it. Marion, and Mezi, Louis. Inhibition and dis-

inhibition of responf.es through empathetic learninB. Cmadian.Jour-

na! of Psychology. lt,63. 17, 235-43.

Walters, Richard II., Llewellyn-Thomas, Edward, and Acker, C. Wil-

liam. Enhancement of punitive behavior by audiovisual displays.

Science, 1C;62, 136(3519), 872-73.