June 2018, VOLUME141 /ISSUE 6

Parental Desensitization to Gun Violence in PG-13 Movies

  1. Daniel Romer, PhD,
  2. Patrick E. Jamieson, PhD,
  3. Kathleen Hall Jamieson, PhD,
  4. Robert Lull, PhD, and
  5. Azeez Adebimpe, PhD
  1. Annenberg Public Policy Center, University of Pennsylvania, Philadelphia, Pennsylvania
  1. Dr Romer conceptualized and designed the study, conducted the statistical analyses, and drafted the initial manuscript; Drs P Jamieson and K Jamieson conceptualized and designed the study and reviewed and revised the manuscript; Dr Lull assisted in the design of the study and its analysis and reviewed and revised the manuscript; Dr Adebimpe assisted in the analysis of the study and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.


Video Abstract

OBJECTIVE: To assess effects of justified versus unjustified screen violence on parents’ willingness to allow children to view films that contain extensive gun violence.

METHODS: A national US sample of 610 parents with at least 1 child between ages 6 and 17 was randomly assigned to view a series of four 90-second video clips from popular films depicting violent gun use under either justified or unjustified conditions. Graphic consequences were removed to mimic the violence common in PG-13 movies. Parents reported their perceived justification for the violence, their emotional reaction while viewing it, the minimum age they would consider appropriate for viewing the film, and whether they would allow their own child to view it. Predictors included experimental condition, viewing order, child age, and recent moviegoing. Growth curve modeling determined the effects of the predictors on both intercepts and slopes of viewing order.

RESULTS: Parents were less emotionally upset and more accepting of child viewing for justified than unjustified gun violence, with perceptions of justification better predictors of parental restriction than emotion. Nevertheless, with the exception of parents with extensive moviegoing habits, parents viewed justified violence as appropriate for adolescents starting at age 15, older than the PG-13 rating suggests.

CONCLUSIONS: Parents are less restrictive of child viewing of gun violence in PG-13 movies when it features characters whose weapon use is seen as justified. The apparent acceptance of rising gun violence in PG-13 movies may be partly attributable to the perception that the violence in those films is justified.

  • Abbreviations:
    confidence interval
    Motion Picture Association of America
  • What’s Known on This Subject:

    Gun violence is increasingly popular in movies rated as acceptable for children over age 12, and previous research suggests that parents are becoming emotionally desensitized to such movies.

    What This Study Adds:

    Parents are not emotionally desensitized to such gun violence but are more willing to allow children to view it when it is seen as justified. Nevertheless, parents view such films as more appropriate for children older than age 13.

    The earliest studies of children’s responses to movies revealed that some movies can be disturbing to children.13 Because parents may want to shield their children from such content, the Motion Picture Association of America (MPAA) provides ratings that inform parents about potentially upsetting content.4 In particular, the PG-13 rating was introduced in 1984 in response to parents’ concerns that the more global parental guidance (PG) warning lacked the ability to warn parents about content that could potentially disturb younger children.2 Nevertheless, in recent years the PG-13 rating has been extended to films that were previously restricted to children older than 16 unless there was an adult in attendance5 and now includes films with levels of gun violence that exceed what is present in the more restrictive R-rating category.6,7

    Along with the increase in PG-13 violence, surveys also suggest that parents are becoming more accepting of violence in movies,8 reflecting a form of desensitization to movie violence.9 Given the long-standing evidence that heavy exposure to screen violence in children can lead to imitation,10,11 increased parental acceptance of movie violence is cause for concern. One explanation for this desensitization is that these films typically omit the consequences of violence that might be upsetting to children, such as graphic displays of blood and suffering12 that tend to receive R ratings.7 As parents view more of these movies, they may become emotionally desensitized to the violence9,13 and more accepting of its viewing by children.

    Another potential mitigating feature of gun violence in popular movies is the perceived justification for the violence. Violence conducted to defeat enemies or for purposes of self-defense may be seen as more justified and acceptable than violence that is random or without socially beneficial motives. This form of violence may be common in popular movies just as it has been on television14 and may also be less emotionally upsetting than unjustified violence.15 Thus, repeated exposure to justified gun violence in PG-13 films could lead to a form of normative desensitization in which the justified use of guns in movies leads to greater acceptance of its viewing by children.

    We sought to distinguish between these 2 forms of desensitization. If emotional desensitization were responsible for increased acceptability of gun violence in PG-13 movies, one would expect repeated viewing of bloodless gun violence to reduce the emotional response to such content and hence lessen its restriction of viewing by children whether it is seen as justified. However, if normative desensitization is important, then repeated viewing of justified gun violence should lead to greater perceptions that such violence is justified and correspondingly greater acceptance of its viewing by children. Furthermore, evidence of either form of desensitization should also be reflected in parents’ moviegoing practices. Parents who more frequently view recent movies should become desensitized to the violence in them.

    To test these alternative explanations, we examined both emotional responses and perceptions of justification for gun violence as parents of children ages 6 to 17 viewed a sequence of either justified or unjustified gun violence scenes from popular movies. To mimic the violence in PG-13 movies, we removed the graphic and potentially upsetting consequences of gun violence (blood and suffering) from clips of unjustified violence taken from R-rated movies. We asked parents whether they would allow 1 of their children to view the movie and the age at which the movie would be appropriate for viewing by children in general.



    As in a previous study,9 we recruited the sample of parents from Research Now, a firm that maintains a large representative online panel of US adults primarily for consumer research purposes. Panel members who were parents or guardians of children between the ages of 6 and 17 were invited to participate in a study to evaluate movies for their appropriateness of viewing by children. The study was approved by the Institutional Review Board of the University of Pennsylvania.


    Parents indicated how many children they had in each of 3 age categories: 6 to 9, 10 to 12, and 13 to 17. If they had a child in >1 category, they were randomly assigned to 1 category, and if they had >1 child in that category, they were randomly assigned to respond to the survey only in regard to that specific child. Recruitment was structured to overrepresent parents with children in the 13 to 17 age range (40%), with proportional representation of the 2 other age groups.

    Parents were assigned at random to view four 90-second movie clips that had been independently rated as containing either justified or relatively unjustified gun violence. Within each condition, participants were also randomly assigned to view 1 of 4 possible arrangements of the clips such that each appeared equally often in each viewing position. After viewing each clip and before moving on to the next 1, parents were asked the same series of questions regarding their emotion while viewing the clip, whether the major character shown in a still picture was justified in using violence, the appropriate age for children to view the movie, and whether parents would allow their own child to view it (see Table 1). To screen out respondents who did not attend to the videos or survey items, we asked a question after 2 of the clips that could best be answered if it was viewed and a third question that assessed following survey instructions. Respondents who could not answer all 3 questions correctly were removed from further analysis. Those who took longer than 45 minutes to complete the study were also removed.

    TABLE 1

    Parental Assessments


    The clips were taken from popular films rated by the MPAA as either PG-13 or R to children starting at the age of 17 unless accompanied by an adult (Table 2). Preliminary research with a range of violent clips from popular movies identified 4 that were rated as having a major character who was justified versus 4 that were rated as relatively less justified for engaging in gun use. The movies from which the clips were taken were popular at the box office, had a balance of victims of either sex, and could be edited if necessary to remove graphic consequences of violence. On average, both sets of films were previously seen by parents in the sample at an equivalent rate (34%).

    TABLE 2

    Characteristics of Movies From Which Clips Were Taken


    We tested the alternative explanations by using growth curve modeling with the program Mplus.16 Growth curve modeling is a multilevel analytic approach that permits the identification of effects on both the intercepts and slopes of individuals’ responses as a function of viewing order.17 We first tested the total effects of each predictor on the 4 dependent variables, each of which has a latent slope that could be related to its latent intercept. Often this relation is negative, reflecting either ceiling or floor effects. The method also permits tests of mediation. The causal model in Fig 1 served as the theoretical framework for the separate analysis of each child viewing outcome. The model predicted that experimental condition would lead to a cascade of effects beginning with judgments of justification that would further affect negative emotion and each of the measures of child viewing. A critical test in the model was whether the intercept of justification predicts the slope of allowing child viewing directly or as mediated by emotion. Child age and parent moviegoing were also potential predictors. Correlations between the slopes could also be examined to identify relations at the individual level. Tests of model fit and the plausibility of the solution assessed the adequacy of the model’s ability to explain the data. The direct and mediated pathways were tested by using bias-corrected bootstrapping with 95% confidence limits.18 We also conducted multilevel regression to identify parents’ judgments of appropriate ages for viewing of justified and unjustified violence.

    FIGURE 1

    Structural model for tests of mediation. Dashed path was added to improve model fit. Model parameters are standardized coefficients (P < .05) with different values for the allow and age models, respectively. Model fits for allow and age models, respectively: CFI = 0.98, 0.98; TLI: 0.98, 0.98; RMSEA = 0.052, 0.054. CFI, comparative fit index; RMSEA, root mean squared error of approximation; TLI, Tucker-Lewis index.


    Sample Characteristics

    Table 3 contains percentages of various characteristics of the parents in the study by condition. Random assignment was confirmed in that there were no significant differences by condition. Approximately 15% of those who started the survey were dropped from the analysis in both conditions either because of failure to qualify or to answer the quality control questions correctly. The sample was similar to a previous study of parents,9 with the exception that this sample had a higher proportion of fathers, and the age distribution of parents and/or guardians tended to be somewhat older than occurs in the US population.19

    TABLE 3

    Demographic Characteristics of Parents and/or Guardians by Condition

    Effects of Condition, Viewing Order, and Other Predictors

    As seen in Fig 2 and Table 4, parents who viewed justified clips reported higher levels of justification for gun violence (B = 2.27 [95% confidence interval (CI) = 2.05 to 2.44]) and a steeper slope as viewing continued (B = 0.32 [95% CI = 0.003 to 1.27]). There was no effect of child age on justification. However, the number of movies seen was positively associated with justification (B = 0.06 [95% CI = 0.004 to 0.120]).

    FIGURE 2

    Effects of experimental condition and age of child by order of presentation on (A) justification, (B) emotion, (C) allowance, and (D) age.

    TABLE 4

    Effects of Conditions and Other Predictors on Unstandardized Intercepts and Slopes of Dependent Variables

    Parents reported less negative emotion in the justified condition (B = −1.26 [95% CI = −1.49 to −1.03]). However, negative emotion increased with successive viewing in both conditions. There was no consistent effect of child age. The number of movies seen was positively related to emotion (B = 0.08 [95% CI = 0.002 to 0.152]), a difference that declined as viewing continued (B = −0.04 [95% CI = −0.070 to −0.023]).

    Parents were more willing to allow their own child to view justified violence (B = 0.44 [95% CI = 0.129 to 0.274]), and this effect increased over trials (B = 0.17 [95% CI = 0.010 to 0.476]). They were also more accepting the older their child was and the more movies seen.

    Regarding appropriate age for viewing, parents were again more restrictive for unjustified than justified violence (B = −0.85 [95% CI = −1.22 to −0.481]), and this effect increased with successive viewing (B = −0.08 [95% CI = −0.163 to −0.007]). Child age was positively related to this outcome and negatively related to movies seen.

    A mixed effects regression was used to estimate the age that parents thought appropriate for viewing each type of violence. We estimated this outcome for a family that was relatively unrestrictive (parent at age 40 who had previously seen the film and had an adolescent at age 14). Older parents who had not seen the film provided even more restrictive ages. The model shown in Table 5 indicated that such a parent would judge the justified violence as appropriate for a child at age 15 (±0.082). The corresponding age for unjustified violence was slightly younger than 16 (15.94 ± 0.082). For frequent moviegoing parents, the appropriate age for unjustified violence approached 13.

    TABLE 5

    Mixed-Effects Maximum Likelihood Regression of Appropriate Age of Viewing (N = 610)

    Mediation Analyses

    The structural model in Fig 1 provided excellent fits for both allowance and appropriate age. The standardized paths show that there was a strong effect of condition on the intercept of perceived justification, which in turn predicted the slopes of allowance and age, indicating less restriction as viewing continued. As seen in Table 6, justification mediated the relation between condition and slopes of both child viewing outcomes (0.130, −0.080). A test of the relation between emotion and the slopes of those outcomes was nonsignificant. However, the path from justification to emotion predicted less restriction on the intercepts of both outcomes (0.456, −0.752).

    TABLE 6

    Significant Mediated Paths Affecting the Intercepts and Slopes of Restricted Viewing for Each Analysis Using Bias-Corrected Bootstrapped SE With 95% CIs

    Although not shown in Fig 1, the correlation between the slopes of perceived justification and appropriate age was −0.287 (95% CI = −0.687 to −0.015). However, the correlation between the slope for justification and allowance was not significant (r = 0.171 [95% CI = −0.130 to 0.702]). None of the other slopes in the model were significantly correlated.

    Although child age predicted allowance and age intercepts directly, its indirect effects through emotion were not significant.

    Parents who watched more recent movies were less restrictive because of their tendency to view gun violence as more justified, which translated to less negative emotion (0.012, −0.020). However, the number of movies seen was also associated with greater negative emotion which transferred to more restriction (−0.044, 0.072). Effects on slopes were small as mediated by the inverse relation between intercepts and slopes. The unexpected relation between movies seen and the slope of emotion was unrelated to child viewing outcomes.


    The results go beyond previous research in revealing that parents are more sensitive to the justification for gun violence that is common in popular PG-13 movies than to their emotional reaction when deciding whether a film is appropriate for child viewing. Parents who viewed gun violence that they considered to be justified were more likely to allow their own child to view such content and were less restrictive regarding the age at which it would be acceptable for children. In addition, successive viewing of justified gun violence was accompanied by increasing perceptions of justification for the violence and less restriction of child viewing.

    Contrary to an emotional desensitization explanation, parents reported greater negative emotion with continued viewing of both justified and unjustified violence, and they were more accepting of justified violence with successive viewing of those movie segments, an effect that occurred apart from emotion. In total, the results suggest that parental desensitization to the gun violence in popular PG-13 movies is more the result of the normalization of this screen violence than of emotion reduction.

    The importance of justification for violence was also evident in the finding that parents who regularly viewed popular movies perceived greater justification for the violence in the segments they were shown, an effect that would be expected if exposure to such films increases normalization of screen violence. At the same time, they reported experiencing more negative emotion at the outset. Although this effect declined as viewing continued, this reduction was unrelated to judgments about child viewing and argues against the importance of emotion as the major source of parental desensitization to bloodless film violence.

    Although emotion was a direct predictor of average levels of judgments about child viewing, this effect was largely driven by differences in justification. This indicates that justification exerts some of its effects by influencing emotion. There were also effects of emotion that were independent of justification. In particular, some of the effects of movies seen were mediated exclusively by differences in emotion. Nevertheless, these effects were small.

    Aside from the important role of perceived justification for screen violence, the violence in PG-13 movies is mostly free of graphic consequences. Thus, the violence in these films may not be as upsetting to parents and their children as the more graphic violence in R-rated movies. However, this raises the concern that movie depictions of harmful behavior that omit the consequences, such as of smoking20 and alcohol21 use, can increase imitation in adolescents. A recent study revealed that even brief exposure to gun use in a film led children ages 8 to 12 to play with and pull the trigger of a real gun.22

    Our findings suggest that although the violence in PG-13 movies may be disturbing to parents, their acceptance of this violence may nevertheless account for the growing popularity of gun violence in these films. Even films with unjustified but bloodless gun violence were seen as appropriate for adolescents at age 16, which would place such films below the criterion for an R rating. Violence seen as justified was judged appropriate for those at age 15. Although these ages are consistent with the PG-13 rating, it may be appropriate for the MPAA to introduce a more nuanced rating system that gives parents more guidance than one that fails to warn them about content that may be appropriate for youth younger than 17 but not as young as 13.

    The increasing presence of comic book heroes in PG-13 films in recent years may also make their use of gun violence appear acceptable.7 Unfortunately, the approval of gun use in films with justified violence may teach the lesson that gun use is acceptable when responding to threats from others. Judging when violence is justified, however, is subject to personal interpretation. For example, sending the message that defending oneself is justification for use of force may increase its use. After Florida passed its “stand your ground law,” there was an increase in both justified and unjustified gun homicide in the state.23 Future research is needed to determine if repeated exposure to films with bloodless depictions of justified gun use can influence youth and their parents to view gun use as acceptable.

    It was somewhat surprising that we did not observe the strong increase in acceptance of child viewing with successive viewings that was found in an earlier study of parental reactions to screen violence.9 In the current study, there was only a reversed effect in the justified condition. It is noteworthy, however, that parents in this study were already reporting lower levels of restriction after the first video than in the earlier study. We also did not expose parents to as many videos as in the earlier study. Another study revealed that desensitization did not occur until after 4 viewings of violent content.13

    Despite our clear findings regarding parental reactions to bloodless screen violence, it is important to consider that we only showed participants a 90-second segment of each movie with a focus on a violent scene. This reaction might be different if they viewed each film in entirety. Nevertheless, parents who had previously seen the films only rated them as slightly more age appropriate (less than a quarter of a year), and pre-exposure was comparable between conditions. Our sample was somewhat older on average than parents in the population, but this also was controlled across conditions, and we took both previous viewing and age of the parent into account in estimating parents’ judgments of appropriate viewing age.


    Our findings suggest that the increasing parental acceptance of gun violence in PG-13 movies could be due to its relatively bloodless portrayal of violent consequences and its often-justified purpose. Giving such movies a PG-13 rating may be less restrictive than is warranted. Further research is called for to determine if viewing of violence in films and other violent media that appears to be justified affects children’s interpretations of such content and their attitudes toward the use of guns for self-defense.


    We thank Jeremy Quattlebaum for assistance in editing the movie clips and conducting the online survey and the team at Research Now for conducting the survey.


      • Accepted March 6, 2018.
    • Address correspondence to Daniel Romer, PhD, Annenberg Public Policy Center, University of Pennsylvania, 202 S 36th St, Philadelphia, PA 19104. E-mail: dan.romer{at}
    • FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

    • FUNDING: Supported by the Annenberg Public Policy Center of the University of Pennsylvania.

    • POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.