Modifying Media Content for Preschool Children:A Randomized Controlled Trial

WHAT’S KNOWN ON THIS SUBJECT: Children have been shown toimitate behaviors they see on screen.

WHAT THIS STUDY ADDS: Modifying what children watch canimprove their observed behavior.

abstractBACKGROUND: Although previous studies have revealed that preschool-aged children imitate both aggression and prosocial behaviors onscreen, there have been few population-based studies designed toreduce aggression in preschool-aged children by modifying what theywatch.

METHODS: We devised a media diet intervention wherein parents wereassisted in substituting high quality prosocial and educational pro-gramming for aggression-laden programming without trying toreduce total screen time. We conducted a randomized controlledtrial of 565 parents of preschool-aged children ages 3 to 5 yearsrecruited from community pediatric practices. Outcomes werederived from the Social Competence and Behavior Evaluation at 6and 12 months.

RESULTS: At 6 months, the overall mean Social Competence and Be-havior Evaluation score was 2.11 points better (95% confidence interval[CI]: 0.78–3.44) in the intervention group as compared with the con-trols, and similar effects were observed for the externalizing subscale(0.68 [95% CI: 0.06–1.30]) and the social competence subscale (1.04[95% CI: 0.34–1.74]). The effect for the internalizing subscale was ina positive direction but was not statistically significant (0.42 [95% CI:20.14 to 0.99]). Although the effect sizes did not noticeably decay at12 months, the effect on the externalizing subscale was no longerstatistically significant (P = .05). In a stratified analysis of the effect onthe overall scores, low-income boys appeared to derive the greatestbenefit (6.48 [95% CI: 1.60–11.37]).

CONCLUSIONS: An intervention to reduce exposure to screen violenceand increase exposure to prosocial programming can positively impactchild behavior. Pediatrics 2013;131:431–438

AUTHORS: Dimitri A. Christakis, MD, MPH,a,b Michelle M.Garrison, PhD,a,c Todd Herrenkohl, PhD,d Kevin Haggerty,MSW,d Frederick P. Rivara, MD, MPH,a,b Chuan Zhou, PhD,a,b

and Kimberly Liekweg, BAa

aCenter for Child Health, Behavior, and Development, SeattleChildren’s Research Institute, Seattle, Washington; andDepartments of bPediatrics and cHealth Services, and dSchoolof Social Work, University of Washington, Seattle, Washington

KEY WORDSaggression, TV, preschool, prosocial, behavior

ABBREVIATIONSCI—confidence intervalSCBE—Social Competence and Behavior Evaluation

This trial has been registered at www.clinicaltrials.gov(identifier NCT01459835).

www.pediatrics.org/cgi/doi/10.1542/peds.2012-1493

doi:10.1542/peds.2012-1493

Accepted for publication Oct 26, 2012

Address correspondence to Dimitri A. Christakis, MD, MPH, 2001Eighth Ave Suite 400, Seattle WA 98121. E-mail: dimitri.christakis@seattlechildrens.org

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2013 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated they haveno financial relationships relevant to this article to disclose.

FUNDING: Funded by a grant from the National Institute for ChildHealth and Development (to Dr Christakis). Funded by theNational Institutes of Health (NIH).

COMPANION PAPERS: Companions to this article can be foundon pages 439 and 589, and online at www.pediatrics.org/cgi/doi/10.1542/peds.2012-1582 and www.pediatrics.org/cgi/doi/10.1542/peds.2012-3872.

PEDIATRICS Volume 131, Number 3, March 2013 431

ARTICLE

by guest on June 3, 2020www.aappublications.org/newsDownloaded from

Preschool-aged children in the UnitedStatesspendanestimated4.4hoursperday watching television at home and inday care settings.1 Although thatamount alone might give one pause,equally, and perhaps more concerning,has been the amount of aggressionthat they watch.2,3 Decades of researchrooted in observational theory haverevealed that children emulate behav-iors (good and bad) that they see onscreen.4–8 Considerable research hasestablished the adverse effects of violenttelevision programming on children’slevel of aggression.9–12 Cross-sectionaland quasi-experimental studies of tele-vision viewing among school-age chil-dren and adolescents have revealedtelevision viewing to be associated withaggression.13–15 Experimental designshave confirmed that reducing the amountof television children watch can re-duce aggression among 9-year-olds.12,16

Considerably less attention has beengiven to the effects of television onpreschool-aged children; however, lon-gitudinal studies of television viewingbefore age 5 have revealed it to bea potential risk factor for the sub-sequent development of bullying andaggression measured in early elemen-tary school.10,17–19 As aggressive behav-ior in the early childhood years has beenrepeatedly linked to violence in lateryouth and adolescence, interventionsthat might reduce early aggressivebehavior could have significant societalimplications.20–23

Research has also established thatcertain types of media programmingcan promote prosocial behavior.24–27

For example, high quality prosocialprograms can improve racial attitudes,their social interactions, and theirsharing propensities.28–31 This has ledmany researchers to emphasize thatfrom a public health standpoint, con-tent is as important as quantity in theongoing debate about screen time.Unfortunately, the current viewing

habits of most preschoolers, particu-larly those from disadvantaged fami-lies, lean heavily toward inappropriateprogramming (ie, noneducational orolder child/adult focused) at the ex-pense of higher quality shows.32–34 Todate, no randomized controlled trialconducted in naturalistic environ-ments with long-term follow-up hasbeen conducted in preschool-agedchildren. We developed and tested anapproach in which preschool-agedchildren’s viewing habits were alteredsuch that they substituted high qualityeducational programs for violence-laden ones.

METHODS

We conducted a randomized controlledtrial of a “media diet” intervention. Thetreatment group received the mediadiet intervention described later, andthe attention control group receiveda nutritional intervention designed topromote healthier eating habits. Thestudy protocol was approved by theSeattle Children’s Hospital InstitutionalReview Board.

Subjects

Letters describing the studies weresent to families with age-eligible chil-dren (3–5 years) enrolled in commu-nity pediatric practices without regardto whether the child had been seen inthe clinic recently. To be eligible, chil-dren needed to engage in some screentime each week and to have English-speaking parents. Families were giventhe opportunity to “opt out” of furtherrecruitment efforts and also had theoption to “opt in” by returning a postage-paid mailer. In a separate analysis, thedifferences between these 2 groupswere found to be minimal.35 Those whoneither opted out nor in were con-tacted by telephone and asked to par-ticipate. Attempts were made tooversample low-income families, asinitially identified by Medicaid status or

zip code of residence. At enrollment,parents were told only that “the studywas being done to better understandhow parents and children use televi-sion movies and computer games.”After enrollment, the survey and com-pleted media diary were collected bystudy staff during a home visit at thestart of the intervention. For all follow-up surveys, families had the option ofreturning materials by mail or sub-mitting them online.

Intervention

The intervention framework was basedon social cognitive theory36,37 andsought to increase parental outcomeexpectations and self-efficacy aroundmaking healthy media choices for theirchild, with a specific emphasis onreplacing violent or age-inappropriatecontent with age-appropriate educa-tional or prosocial content. The centralpremise informing much of the edu-cational approach, consistent with ob-servational theory, was that childrenimitate what they see on screen. Al-though the intervention addressed allscreen time (television, DVDs/videos,computers, video games, handhelddevices, etc), the primary focus was ontelevision and videos because thisaccounts for the vast majority ofscreen time in preschool-aged chil-dren. No attempt was made to reducetotal number of screen time hours;rather, the intervention focused oncontent and encouraging positive me-dia behaviors such as coviewing. In-tervention sessions began with theinitial home visit, in which the familiesassigned to 1 of 3 case managers whocollected assessment materials dis-cussed the child’s current media usewith the parent, shared interventionhandouts that were specific to thefamily’s needs, and engaged the parentin goal setting. The home visit was thenfollowed by mailings and follow-uptelephone calls with the case manager

432 CHRISTAKIS et al by guest on June 3, 2020www.aappublications.org/newsDownloaded from

for 12 months. The monthly mailingsincluded a program guide tailored tothe family’s available channels withrecommended educational and proso-cial television shows and schedules,and a newsletter with tips and re-inforcement of key messages (samplenewsletter available from correspond-ing author on request). The first 6mailings also included DVDs with 5- to10-minute clips of suggested educa-tional and prosocial shows to piquechildren’s (and parents’) interest.Shows were selected for recommen-dation via program guides and DVDsutilizing ratings publicly available fromCommon Sense Media, with attemptsmade to include shows featuring di-versity across gender, race, and eth-nicity.

Duringthe monthly telephone calls, thecasemanager reviewedprogressmadeon the parent’s goals since the lastencounter, coached the parent throughproblem-solving around barriers asneeded, and worked with the parent toset new goals as appropriate. Thecontrol group received a nutrition in-tervention, with analogous monthlynewsletters promoting healthy foodchoices and monthly check in calls.

In summary, the importance of reducingexposure to violent television andreplacing it as neededwith educational/prosocial programming in the in-tervention groupwas emphasized at theinitial visit, in the monthly newsletters,during the monthly telephone calls withthe research assistant, by monthlyprogram guides tailored to the partici-pating families’ television service, andby providing examples of the types ofprograms that we deemed age appro-priate and worthwhile. In addition, atthe initial visit, parentswere taught howto use the V chip on their television (ifthey wished) and how to set up kidzones on their DVR (where available).These practical strategies were alsoreinforced in newsletters.

Outcomes

Our primary outcomes were derivedfrom the Social Competence and Be-haviorEvaluation(SCBE), parentversion.The SCBE is a well-validated measurewithbothanoverall scoreandsubscalesfor internalizing (anxious, depressive,and withdrawn) and externalizing (an-gry, aggressive, oppositional) behaviors,as well as a subscale for social com-petence. Norms vary by age and gen-der.38 Higher scores indicate morepositive behavior, for both the overallscore and all subscales. We hypothe-sized that the intervention would in-crease the overall score and each of the3 subscale scores. The SCBE was col-lected at each time point (baseline and 6and 12 month follow-up). Based on ourexpected sample size, we estimated80% power to detect differences of 1/4of an SD in our primary outcomes.

Other Variables

Child and family demographic data werecollected via a parent survey at baseline.Giventhesubstantialnumberofmultiracialchildren, race and ethnicity were coded asnonmutually exclusive variables froma “check all that apply” question, so pro-portions do not add up to 100%. Familiesare considered to be low income if theirself-reported household income for thenumber of household members is below200% of 2009 Federal Poverty Guidelines.

Child media use and content wasassessed viamedia diaries at each timepoint. These diaries were modeled onones used in the Panel Study of IncomeDynamics Child Development Supple-ment.39 Parents were instructed toprospectively complete diaries by fill-ing in time of day, name of show, plat-form (eg, television, DVD, etc), and whowas watching with them (sample diaryavailable from authors on request).At baseline, parents completed themedia diaries prospectively for 1 week.For children who were in the care ofother adults during the day (child care,

relatives, etc), parents were asked tohave those adults help complete the me-dia diaries as well. Diaries captured time,content title, and co-use for television,video game, and computer use, and weresubsequently coded for ratings, content,and pacing. The protocol for coding vio-lence was based on that used for thePanel Study of Income Dynamics ChildDevelopment Supplement39 and catego-rized violence for each program by fre-quency (none, isolated, episodic, orcentral) and type (mild/slapstick, fantasyviolence, sports violence, realistic, orgratuitous). Prosocial programming wasdefined as that which role modeled non-violent conflict resolution, cooperativeproblem solving, empathy and recogni-tion of emotions, manners, and helpingothers. Coding for prosocial program-ming was further broken down into cat-egories of “primary” and “incidental,” justas was the case with educational pro-gramming. In the primary category, pro-social behaviorswere an explicit theme ofthe program and were consistently rolemodeled; examples include SesameStreet, Dora the Explorer, and SuperWhy. In the incidental category, prosocialbehaviors were role modeled, but in-consistently; examples include CuriousGeorge,MickeyMouse Clubhouse, and Sidthe Science Kid. For this article, we di-chotomized prosocial content as presentor not.

The first wave was coded by 2 re-searchers (MsLiekwegandDrGarrison)until 90% agreement was reached, withdisagreements resolved through con-sensus. Subsequent diaries were codedby 1 researcher (Ms Liekweg), with arandom 5% also coded by the secondresearcher toensure90%agreementwasmaintained.

At the study conclusion, parents in theintervention arm evaluated the pro-gram by responding to 2 questions: (1)would they recommend the program toother parents and (2) do they feel betterabout their child’s media use.

ARTICLE

PEDIATRICS Volume 131, Number 3, March 2013 433 by guest on June 3, 2020www.aappublications.org/newsDownloaded from

Analysis

Missing data for the questions contribut-ing to the SCBE scales were imputed byusing simple imputation with the Stata(Stata Corp, College Station, TX) imputecommand for subjects with no more than20%ofitemsmissinginthescale.Eachitemused in the analysis had ,2% of valuesmissing. Descriptive and bivariate statis-tics were calculated, with t tests used tocompare continuous variables betweenthe intervention and control groups, andx2 tests for categorical variables.

The main intervention effect was testedby using linear regression, with theSCBEoverall, externalizing, internalizing,and social competence scores as theoutcomes at both the 6- and 12-monthtime points. Each model controlled forthe child’s baseline score for the domainanalyzed and took into account casemanager as a random effect. Stan-dardized effect sizes were calculated bydividing the b coefficient for the in-tervention effect by the standardizeddeviation of the score (Cohen’s d). In

a secondary analysis, we tested for ef-fect modification by using the Wald testfor interaction terms by gender andlow-income status at the 6-monthfollow-up. All analyses were conductedby using Stata/SE, version 10.

RESULTS

Of the 3334 families contacted andassessed for eligibility (Fig 1), 1350(40%) did not meet inclusion criteria(742 did not watch at least 3 hours oftelevision per week; 31 did not meet the

FIGURE 1Consolidated Standards of Reporting Trials diagram.

434 CHRISTAKIS et al by guest on June 3, 2020www.aappublications.org/newsDownloaded from

age requirement; 453 did not meet thelanguage requirement; and 124 did notlive in Seattle), 35% declined to par-ticipate, and the remaining 25% (N =820) were randomly assigned; of these,

617 (75%) completed the baseline visit.A total 565 (92%) of those completingthe baseline survey completed at least1 follow-up survey and had sufficientdata to be analyzed, with 557 included

in the 6-month analysis and 539 in-cluded in the 12-month analysis. Thedemographic characteristics of the in-cluded children and their families wasrepresentative of the Seattle area, andthe only significant difference betweenstudy arms in baseline characteristicswas a somewhat higher proportion ofchildren in the control arm (43%) whohad an older sibling at home, as com-pared with the intervention arm (34%).

Total screen time did not vary betweengroups at baseline or follow-up (Tables1 and 2). Violence exposure did notdiffer between groups at baseline butwas significantly less in the interventioncomparedwith the control group both interms of minutes and as a proportion oftotal daily screen time (Table 2). Al-though exposure to prosocial contentwas somewhat higher in the in-tervention group than in the controlgroup at baseline (Table 2), posthoc re-gression analyses still revealed signifi-cantly increased prosocial exposure atthe 6-month follow-up both in terms ofminutes (P, .01) and as a proportion oftotal screen time (P = .02) after adjust-ing for baseline prosocial exposure.Between baseline and the 6-monthfollow-up, total daily screen time in-creased by 13.5 minutes in the controlgroup and 16.2 minutes in the in-tervention group; however, this increasewas accounted for by more violentminutes in the control group (increaseof 6.8 vs 0.3 in the intervention arm) andby more prosocial minutes in the in-tervention group (9.9 vs 1.9). No signifi-cant differences were observed in theproportion of screen time reportedas coviewing with an adult, whichremained stable at 68% to 70% acrossall arms and time points, and may re-flect social desirability bias in reporting.

In the primary outcome analyses, theintervention resulted in significantimprovements in the overall SCBE scoreat 6 months, as well as the externaliz-ing and social competence subscales

TABLE 1 Study Sample

Intervention, N = 276 Control, N = 289

Child demographicsGirl, % 45 46Age in mo, mean(SD) 50.9 (7.7) 51.6 (7.7)Race/ethnicity (not mutually exclusive), %White 82 81Black 8 8Hispanic 6 7Asian/Pacific Islander 14 18Native American 3 3

Family demographics, %Low-income 18 13One adult household 7 5Older sibling(s)a 34 43Respondent is mother 88 88Respondent education, %High school or less 19 18College degree 45 44Graduate or professional degree 36 38

Child baseline media useTelevision in bedroom, % 8 8Average daily total use, min, mean(SD) 73.9 (50.9) 70.4 (48.5)Average evening use, min, mean(SD) 13.7 (16.1) 13.7 (18.0)Average daily violent content, min, mean(SD) 22.1 (25.8) 22.9 (31.0)

a P value is ,0.05 for difference between groups.

TABLE 2 Screen Time and SCBE Scores by Study Group

Baseline T6 T12

Daily total screen time, minControl 70.4 83.9 81.3Intervention 73.9 90.1 78.2P .41 .25 .53

Daily screen time with central or episodic violenceControl 22.9 29.7 26.8Intervention 22.1 22.4 23.5P .76 .04* .30

Proportion of daily screen time with central or episodic violenceControl, % 29.2 30.0 30.1Intervention, % 28.1 22.9 25.2P .63 .01* .09

Daily screen time with prosocial contentControl 28.9 30.8 28.3Intervention 33.5 43.4 33.4P .06 ,.001* .12

Proportion of daily screen time with prosocial contentControl, % 42.4 40.0 37.7Intervention, % 47.9 49.3 44.8P .02* ,.01* .03*

SCBE overall scoreControl 106.03 106.38 107.93Intervention 105.96 108.36 109.57P .94 .047 .10

*P , .05.

ARTICLE

PEDIATRICS Volume 131, Number 3, March 2013 435 by guest on June 3, 2020www.aappublications.org/newsDownloaded from

(Table 3). At 6 months, the overall meanSCBE score was 2.11 points better (95%confidence interval [CI]: 0.78–3.44) inthe intervention group as comparedwith the controls, and similar effectswere observed for the externalizingsubscale (0.68 [95% CI: 0.06–1.30]) andthe social competence subscale (1.04[95% CI: 0.34–1.74]). Both subsca-les demonstrated a trend toward

improvement over time in each arm(Table 2), as might be expected withage during this developmental period,but with the regression results re-vealing greater improvement in the in-tervention arm (Table 3). Although theeffect for the internalizing subscale wasin a positive direction, it was not sta-tistically significant (0.42 [95% CI:20.14to 0.99]). Although the effect sizes did

not noticeably decay at 12 months, the

effect on the externalizing subscale was

no longer statistically significant (P= .05).

Although we detected no significant ef-fect modification by gender or low-income status, there was a trend to-ward an increased effect of the in-tervention in low-income boys for boththe overall SCBE score (P = .12) and theexternalizing subscale (P = .11). In Fig 2,we present regression results afterstratifying by gender and low-incomestatus, where we see statistically signif-icant effects for low-income boys on theoverall score (6.48 [95% CI: 1.60–11.37])and the externalizing subscale (3.95[95% CI: 1.53–6.37]). For the internalizingsubscale, we saw similar interventioneffects across gender, but the in-tervention effect was statistically signif-icant in low-income children (1.90 [95%CI: 0.23–3.56]) but not nonlow-incomechildren (0.18 [95% CI:20.41 to 0.77]).

TABLE 3 Regression Results From Primary Models

Outcome Effect Size b* 95% CI P

Overall SCBE score6 mo (N = 557) 0.19 2.11 0.78 to 3.44 ,.0112 mo (N = 539) 0.18 1.96 0.41 to 3.50 .01

Externalizing scale6 mo (N = 557) 0.14 0.68 0.06 to 1.30 .0312 mo (N = 539) 0.14 0.67 20.02 to 1.36 .05

Internalizing scale6 mo (N = 557) 0.09 0.42 20.14 to 0.99 .1412 mo (N = 539) 0.10 0.43 20.15 to 1.02 .15

Social competence scale6 mo (N = 557) 0.17 1.04 0.34 to 1.74 ,.0112 mo (N = 539) 0.14 0.82 0.02 to 1.62 .04

* b reflects the raw regression coefficient. Each model also controlled for the child’s baseline score.

FIGURE 2Regression results for SCBE scores at 6 months by gender and household income (b with 95% CI).

436 CHRISTAKIS et al by guest on June 3, 2020www.aappublications.org/newsDownloaded from

The intervention itself was well re-ceived. Overall, 77% of parents “would”recommend the program to otherfamilies, and 20% “might” recommendit. Thirty-four percent of parents feltmuch better about their child’s mediause than they did at the study’s start,and 35% felt “a little better.”

DISCUSSION

We demonstrated that an interventionto modify the viewing habits ofpreschool-aged children can signifi-cantly enhance their overall social andemotional competence and that low-income boys may derive the greatestbenefit. By focusing on content ratherthan quantity, this study is the first toour knowledge to employ a harm re-duction approach to mediating theuntoward effects of television viewingon child behavior. Importantly, we didnot see an increase in total viewing timein the intervention group comparedwith the control group. Both groupsincreased their viewing time, whichlikely reflects the fact that childrenwatch more television as they age.

Although they varied by group andoutcome, the overall effect sizes weachieved range from0.09 to 0.19,whichusing Cohen’s scale could be inter-preted as small. However, they areconsistent with what has been ach-ieved in the context of other inter-ventional trials designed to improvechildren’s behavior.40 Furthermore,the effects in the particularly high-risksubgroup of low income boys aresubstantial. Future studies may iden-tify and apply this approach toparticularly vulnerable populations.

Although we know that the roots ofaggression in later years begin inearly childhood, few studies to datehave focused on preschool aggressionprevention.41 Most prevention pro-grams begin at school entry40 andpreschool programs to date havelargely focused on secondary pre-vention and treatment.42

This study has a number of limitationsthat warrant mention. First, as with allbehavioral interventions, it was notpossible to blind the parents to studyarm. Althoughparentswere not told thepurpose of the study, they may havededuced it, and thismayhavebiased thereported results. However, the activeintervention period consisted of 6months, and our analysis included datafromasmuch as 1 year later. Further, ina previous article, we report that wealso found improvement in children’ssleep in the intervention arm (a findingthat is plausible based on previousstudies revealing that violent contentcan cause sleep problems in chil-dren).43 Given that sleep was not a tar-get of the intervention, these findingslend additional credibility to our be-havioral outcomes. Finally, we did finda difference between study arms inviolence exposure as measured by thecoded media diaries. Parents com-pleted these without knowledge ofwhich shows we categorized as violent.Given that there was no difference intotal screen time between groups,parents would have to have in-tentionally misrepresented the showstheir children watched (rather thanjust omitting violent ones), whichseems unlikely. Second, our samplemay not be representative of other

communities. However, our stratifiedanalyses revealed effects regardless ofincome. Third, we focused only on me-dia content in the home although weknow that ∼40% of preschool-agedchildren spend time in out-of-homecare arrangements where in manycases an additional 1 to 2 hours oftelevision is viewed.1 Interventionstargeting child care viewing practicesshould also be explored andmay in factenhance the effects achieved throughfocusing exclusively on home-basedviewing. Finally, given that we suc-ceeded in both increasing prosocial/educational content and reducing vio-lent content, we cannot ascertainwhether an increase in the former ora decrease in the latter was more im-portant.

In spite of these limitations, our studyhas some important implications.Population-based aggression pre-vention approaches for preschool-aged children are generally lackingin part because of the challengeposedby deploying a far-reaching andbroad-based approach outside of thestructured environments of schoolsor day cares. This approach we un-dertook by using a widely acceptedand highly used medium couldtherefore have broad public healthimpact, especially given the favorableratings it received from participants.Although television is frequently impli-cated as a cause of many problems inchildren, our research indicates that itmay also be part of the solution. Futureresearch to perhaps further enhancemedia choices particularly for olderchildren and potentially with an em-phasis on low income boys is needed.

REFERENCES

1. Tandon PS, Zhou C, Lozano P, Christakis DA.Preschoolers’ total daily screen time athome and by type of child care. J Pediatr.

2011;158(2):297–300

2. Christakis DA, Zimmerman FJ. The Elephantin the Living Room: Make TelevisionWork for Your Kids. Emmaus, PA: Rodale;

2006

3. Yokota F, Thompson KM. Violence in G-ratedanimated films. JAMA. 2000;283(20):2716–2720

4. Bandura A. Social Learning Theory. Engle-wood Cliffs, NJ: Prentice Hall; 1977

ARTICLE

PEDIATRICS Volume 131, Number 3, March 2013 437 by guest on June 3, 2020www.aappublications.org/newsDownloaded from

5. Bandura A. Social Foundations of Thoughtand Action: A Social Cognitive Theory.Englewood Cliffs, NJ: Prentice Hall; 1986

6. Bandura A, Ross D, Ross SA. Imitation offilm-mediated aggressive models.J Abnorm Soc Psychol. 1963;66(1):3–11

7. Bushman BJ, Huesmann LR. Short-termand long-term effects of violent media onaggression in children and adults. ArchPediatr Adolesc Med. 2006;160(4):348–352

8. Anderson CA, Sakamoto A, Gentile DA, et al.Longitudinal effects of violent video games onaggression in Japan and the United States.Pediatrics. 2008;122(5). Available at: www.pediatrics.org/cgi/content/full/122/5/e1067

9. Anderson DR, Huston AC, Schmitt KL,Linebarger DL, Wright JC. Early childhoodtelevision viewing and adolescent behavior:the recontact study. Monogr Soc Res ChildDev. 2001;66(1):I–VIII, 1–147

10. Christakis DA, Zimmerman FJ, DiGiuseppeDL, McCarty CA. Early television exposureand subsequent attentional problems inchildren. Pediatrics. 2004;113(4):708–713

11. MacBeth TM. Indirect effects of television:creativity, persistence, school achievement,and participation in other activities. In:MacBeth TM, ed. Tuning in to Young View-ers: Social Science Perspectives on Televi-sion. Thousand Oaks, CA: Sage Publications;1996:149–220

12. Robinson TN, Wilde ML, Navracruz LC,Haydel KF, Varady A. Effects of reducingchildren’s television and video game use onaggressive behavior: a randomized con-trolled trial. Arch Pediatr Adolesc Med.2001;155(1):17–23

13. Bushman BJ, Anderson CA. Media violenceand the American public: scientific factsversus media misinformation. Am Psychol.2001;56(6-7):477–489

14. Bushman BJ, Cantor J. Media ratings forviolence and sex: implications for policy-makers and parents. Am Psychol. 2003;58(2):130–141

15. Paik H, Comstock G. The effects of televisionviolence on antisocial behavior: a meta-analysis. Commun Res. 1994;21(4):516–546

16. Robinson TN. Reducing children’s televisionviewing to prevent obesity: a randomizedcontrolled trial. JAMA. 1999;282(16):1561–1567

17. Johnson JG, Cohen P, Smailes EM, Kasen S,Brook JS. Television viewing and aggressivebehavior during adolescence and adult-hood. Science. 2002;295(5564):2468–2471

18. Joensson A. TV–a threat or a complementto school? A longitudinal study of the re-lationship between TV and school. J EducTelevision. 1986;12(1):29–38

19. Zimmerman FJ, Glew G, Christakis DA,Katon W. Early cognitive stimulation, emo-tional support, and television watching aspredictors of subsequent bullying amonggrade-school children. Arch Pediatr Ado-lesc Med. 2005;159(4):384–388

20. Nagin DS, Tremblay RE. Parental and earlychildhood predictors of persistent physicalaggression in boys from kindergarten tohigh school. Arch Gen Psychiatry. 2001;58(4):389–394

21. Ladd GW, Burgess KB. Charting the re-lationship trajectories of aggressive, with-drawn, and aggressive/withdrawn childrenduring early grade school. Child Dev. 1999;70(4):910–929

22. Brame B, Nagin DS, Tremblay RE. De-velopmental trajectories of physical ag-gression from school entry to lateadolescence. J Child Psychol Psychiatry.2001;42(4):503–512

23. Kingston L, Prior M. The development ofpatterns of stable, transient, and school-age onset aggressive behavior in youngchildren. J Am Acad Child Adolesc Psychi-atry. 1995;34(3):348–358

24. Friedrich LK, Stein AH. Aggressive and pro-social television programs and the naturalbehavior of preschool children. Monogr SocRes Child Dev. 1973;38(4):1–64

25. Friedrich LK, Stein AH. Prosocial televisionand young children: the effects of verballabeling and role playing on learning andbehavior. Child Dev. 1975;46:27–38

26. Friedrich-Cofer LK, Huston-Stein A, KipnisDM, Susman EJ, Clewett AS. Environmentalenhancement of prosocial television con-tent: effects on interpersonal behavior,imaginative play, and self-regulation ina natural setting. Dev Psychol. 1979;15(6):637–646

27. Goldberg ME, Gorn GJ. Television’s impacton preferences for non-white playmates:Canadian Sesame Street inserts. J Broad-casting. 1979;23:27–32

28. Forge KLS, Phemister S. The effect of pro-social cartoons on preschool children.Child Study J. 1987;17(2):83–88

29. Mares M, Woodard EH. Prosocial effects onchildren’s social interactions. In: Singer DG,Singer JL, eds. Handbook of Children andthe Media. Thousand Oaks, CA: Sage; 2001:183–206

30. Silverman LT, Sprafkin JN. The effects ofsesame street’s prosocial spots on co-operative play between young children.J Broadcasting. 1980;24:135–147

31. Thakkar RR, Garrison MM, Christakis DA. Asystematic review for the effects of television

viewing by infants and preschoolers. Pediat-rics. 2006;118(5):2025–2031

32. Comstock G, Paik H. Television and theAmerican Child. San Diego, CA: AcademicPress; 1991

33. Huston AC, Wright JC, Marquis J, Green SB.How young children spend their time:television and other activities. Dev Psychol.1999;35(4):912–925

34. Mendelsohn AL, Berkule SB, Tomopoulos S,et al. Infant television and video exposureassociated with limited parent-child verbalinteractions in low socioeconomic statushouseholds. Arch Pediatr Adolesc Med.2008;162(5):411–417

35. Myaing MT, Garrison MM, Rivara FP, ChristakisDA. Differences between opt-in and activelyrecruited participants in a research study.J Clin Med Res. 2011;3(5):68–72

36. Bandura A. Social cognitive theory of masscommunication. In: Bryant J, Zillmann D,eds. Media Effects: Advances in Theory andResearch. Mahwah, NJ: Lawrence ErlbaumAssociates; 2002

37. Bandura A. Health promotion by socialcognitive means. Health Educ Behav. 2004;31(2):143–164

38. Kotler JC, McMahon RJ. Differentiatinganxious, aggressive, and socially compe-tent preschool children: validation of theSocial Competence and Behavior Evaluation-30 (parent version). Behav Res Ther. 2002;40(8):947–959

39. Vandewater EA, Bickham DS, Lee JH. Timewell spent? Relating television use tochildren’s free-time activities. Pediatrics.2006;117(2). Available at: www.pediatrics.org/cgi/content/full/117/2/e181

40. Hahn R, Fuqua-Whitley D, Wethington H,et al; Task Force on Community PreventiveServices. Effectiveness of universal school-based programs to prevent violent andaggressive behavior: a systematic review.Am J Prev Med. 2007;33(suppl 2):S114–S129

41. Webster-Stratton C, Herbert M. Strategiesfor helping parents of children with con-duct disorders. Prog Behav Modif. 1994;29:121–142

42. Wesbster-Stratton C. Aggression in youngchildren: services proven to be effecting inreducing aggression. In: Tremblay RE, BarrRG, Peters R, eds. Encyclopedia on EarlyChildhood Development. Montreal, Quebec:Centre of Excellence for Early ChildhoodDevelopment; 2005

43. Garrison MM, Christakis DA. Reducing ex-posure to violent TV improves children’ssleep. Pediatrics. 2012;130(3):492-499

438 CHRISTAKIS et al by guest on June 3, 2020www.aappublications.org/newsDownloaded from

originally published online February 18, 2013; Pediatrics Frederick P. Rivara, Chuan Zhou and Kimberly Liekweg

Dimitri A. Christakis, Michelle M. Garrison, Todd Herrenkohl, Kevin Haggerty,Trial

Modifying Media Content for Preschool Children: A Randomized Controlled

ServicesUpdated Information &

012-1493http://pediatrics.aappublications.org/content/early/2013/02/13/peds.2including high resolution figures, can be found at:

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on June 3, 2020www.aappublications.org/newsDownloaded from

originally published online February 18, 2013; Pediatrics Frederick P. Rivara, Chuan Zhou and Kimberly Liekweg

Dimitri A. Christakis, Michelle M. Garrison, Todd Herrenkohl, Kevin Haggerty,Trial

Modifying Media Content for Preschool Children: A Randomized Controlled

http://pediatrics.aappublications.org/content/early/2013/02/13/peds.2012-1493located on the World Wide Web at:

The online version of this article, along with updated information and services, is

ISSN: 1073-0397. 60007. Copyright © 2013 by the American Academy of Pediatrics. All rights reserved. Print the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on June 3, 2020www.aappublications.org/newsDownloaded from