In this study, children were shown one of three violent TV clips. Each clip showed an identical act of aggression, but the perpetrator's motivation and punishment for the violent act were manipulated. Children also filled out a questionnaire that asked about their family communication patterns (FCP). Overall, children who rated higher on the communication dimension were more likely to see motivated violence as more justified, whereas children who rated higher on the control dimension were likely to see punished violence as less justified. In addition, children who were more control-oriented and who had perceived the violent clip as justified chose aggressive story endings significantly more frequently than other children.
Three decades of research have lead to the conclusion that watching television violence is related to increases in aggression among viewers (Hamilton, in press). However, a growing body of research emphasizes that it is the type of violence and not simply the presence or absence of a violent portrayal that matters in terms of viewer tendencies to respond aggressively after watching media violence. For example, violence committed by attractive perpetrators is more likely to be imitated than violence committed by unattractive perpetrators (Bandura, 1994). Similarly, a recent meta-analysis of 217 studies showed that audiences are more likely to imitate programs that contain justified violence than those that contain unjustified violence, and they are more likely to think aggression is the norm (Paik & Comstock, 1994). It is clear that not all violent portrayals have the same effect on viewers. The type of violent portrayal influences an individual's likelihood to act aggressively after being exposed to the violence (Huesmann, 1986) and, as important, television violence influences an individual's likelihood to believe that aggression is the norm (Huesmann, Guerra, Miller, & Zelli, 1992, 1994).
But the arguments forwarded in this type of research tend to focus on program features. It is the program feature that is manipulated in the experiment; manipulation checks are performed to test if the manipulation "worked." When the manipulation does not work, it is often presumed that viewers were not attending or that they had some unusual or idiosyncratic interpretation. It is possible that interpretations of media violence are systematic and influenced by individual difference variables. While the effect of demographic variables such as gender and age on individual interpretations have been examined to some extent, more environmentally driven variables, such as family communication norms, have been examined to a lesser degree. It is the primary goal of this study to examine an individual difference factor--specifically, children's family communication patterns, on their interpretation of a violent act on television.
A second goal of this study is to determine whether children's interpretations of a violent depiction as justified or unjustified affect their likelihood to provide an aggressive solution to a hypothetical interpersonal dilemma. While evidence exists that exposure to justified violence increases the likelihood that viewers will act aggressively (Paik & Comstock, 1994), more research is needed that looks at the effects of `justified' violence on normative beliefs about aggression. In this study, choice of story ending was used to see if children would select an aggressive solution to a hypothetical interpersonal dilemma. In doing so, it was possible to test whether justified violence, and not just violence in general, can cause children to perceive violence as normative.
In sum, this research seeks to fill several gaps in the existing literature. First, because the justification for violence is seen as an important mitigating factor in viewers' tendency to act aggressively, it is important to see how perceptions of justification vary with individual differences in gender and family communication. Second, because television violence can cause viewers to perceive violence as normative, it is important to test if justified violence affects beliefs about aggression in the same way it affects tendency to behave aggressively.
Interpretations of Television Violence and the Role of Gender
There is little doubt that among children, boys are more likely to be affected by violent television than are girls: They imitate it more and may lower their inhibitions against aggression more as a result of watching violent television (Huesmann, et al., 1984). But besides the extensive work on the effects of media violence, there is some evidence that gender plays a central role in children's attention to and interpretation of violent television. For example, boys attend more to television than girls do, although girls attend more to low action than high action programs (Alvarez et al., 1988).
In addition to attentional differences, boys interpret television differently from girls. They are more likely than girls to have a difficult time distinguishing between television violence and real violence (van der Voort, Vooiijs, & Bekker, 1982), are likely to interpret television violence as more real and exciting (Eron, 1983), and are more likely to perceive television violence as being correct and normative (Thomas & Drabman, 1978). Therefore, boys attend to violent television and, once their attention is gained, they are more likely to interpret it as real, exciting, and normative. It is predicted that:
[H.sub.1]: There will be a main effect for gender across experimental conditions, such that boys will perceive the aggressive clip to be more justified than girls will.
Television Violence and the Age of Child
Age differences also play a role in children's responses to television violence. For example, children pay more visual attention to television, and to violent television as they age, with attention leveling off at approximately age 8 (Alvarez et al., 1988). In addition to greater attention to television, these older children are more likely to identify with the characters they see on television (Sheehan, 1983), perhaps because of their greater cognitive sophistication and increased ability to understand the characters and plots they see on television (Collins, 1983).
These increases in attention and comprehension are likely to be related to changes in response to television violence. A growing body of research suggests that age also plays a role in children's tendency to respond aggressively, in a variety of ways, to violent media images. First, there is an overall increase in the relationship between violence viewing and aggressive behavior as children age (Eron, 1983). Increases also occur in children's tendency to be desensitized by aggressive portrayals (Thomas & Drabman, 1975). In addition to these aggressive tendencies, there are also changes in children's attitudes towards aggression as they age, and these attitudes are also affected by exposure to violent television. Specifically, older children, as compared to their younger counterparts, are more likely to engage in aggressive-heroic daydreaming as a result of exposure to violent television (Valkenburg, Vooijs, van der Voort, & Wiegman, 1992; Valkenburg & van der Voort, 1995) and are more likely to believe that aggression is an adequate means of solving problems (Huesmann et al., 1994). So age plays a role, first, in children's ability to comprehend television and, subsequently, in their increased response to television resulting, in part, from their increased understanding.(1) Therefore, increases in children's comprehension of television and the subsequent belief that aggression is normative lead to the prediction that:
[H.sub.2]: There will be a main effect for age across experimental conditions, such that older children will perceive the aggressive clip to be more justified than younger children will.
Family Communication and Television Viewing
Although demographic variables such as gender and age clearly affect children's responses to violent television, family communication and interaction have also been found to mitigate the effects of television on children. This mitigation can occur both directly, when parents actively mediate television for their children (Messaris & Kerr, 1983) and, indirectly, through communication environments (Krcmar, 1996). The family environment, wherein children learn to interpret and perceive their surroundings, affects how they make sense of television.
Television mediation. Direct mediation, where parents comment on and explain television for their children, has been found to affect the knowledge and attitudes children develop as a result of a television viewing (Huston & Wright, 1996). In several studies, the negative effect of programs containing violent images and violent themes were minimized when adults offered commentary that derogated the violence (Corder-Boiz, 1980; Colder-Bolz & O'Bryant, 1978). Other studies have demonstrated that a coviewing adult who offers comments and interpretations of content can improve children's learning from educational programs such as Sesame Street and Mister Rogers' Neighborhood (Ball & Bogatz, 1970; Colder-Bolz, 1980; Collins, 1983; Friedrich & Stein, 1975; Salomon, 1977; Watkins, Calvert, Huston-Stein, & Wright, 1980), can enhance children's understanding of characters and events in action/adventure programs (Collins, Sobol, & Westly, 1981), and can make children feel more positive towards non-traditional sex roles (Colder-Bolz, 1980). It is evident, therefore, that parent and other adult interaction about television can impact the way children respond to it. However, parent mediation is only one way in which parents mitigate the effects of television. The broader communication environment in the home also seems to affect how children understand and are impacted by what they see on television.
Family communication patterns. The family communication environment teaches the child how to approach the world. For example, a family where open communication is stressed and ideas are readily shared may convey to the child that his/her ideas are valuable and should normally be shared with others, even with those outside of the family. In turn, that child may perceive others views as potentially valuable. A family where hierarchy is stressed and parents make most of the decisions may convey to the child that authority figures make rules which ultimately must be followed. The child may then value the interpretations and ideas of other authority figures. The two dimensions described reflect those measured by the Family Communication Patterns Inventory (McLeod & Chaffee, 1972).
Developed by Chaffee, McLeod, and Wackman (1966), the instrument examines the degree to which families encourage expression of autonomous opinions (concept orientation) and the degree to which families stress relational objectives (socio orientation). The construct of Family Communication Patterns has been used extensively over the past two decades by media effects researchers (McLeod & Chaffee, 1972; Meadowcroft, 1987) but has recently come under attack (Ritchie, 1991). In particular, the FCP has been criticized because the two-dimensional scales, socio and concept orientation, purport to measure family emphasis on relational objectives and objectives supporting the open sharing of ideas, respectively. Ritchie (1991) has demonstrated, however, that the traditional interpretation of the FCP may be inaccurate. Given his analysis of both the face validity of the items and his data, it appears that family members associate concept orientation with communication and supportiveness, and socio orientation with parental power and control. Therefore, the FCP has been reconceptualized to measure individuals' perceptions of communication and control orientation within the family. This reconceptualization will be utilized in this study. These dimensions and the family interaction styles they measure--have been found to be predictive of children's responses to the mass media. In short, the style of communication used within a family can influence the extent and type of effects the mass media have.
Communication orientation. In using family communication patterns as a predictor for media use, communication orientation, or openness, has been found to be related to adolescents' greater involvement in public affairs media (Weintraub-Austin, 1993). The same study found that high communication families were more likely to discuss the content of television programs among themselves (1993). Similarly, Chaffee and Tims (1976) and Meadowcroft (1986) found that greater communication orientation is related to children's greater interest in and knowledge about politics. In addition to affecting media use, communication norms also affect the kinds of statements parents and children make to one another regarding television.
Messaris and Kerr (1983) examined the effect of family communication patterns on parents' reported comments to their children during television viewing. Mothers who described the family as high in communication orientation reported providing more statements classified as non-directive statements. These included questioning the characters' motives and commenting on the characters' behavior. In short, these mothers shared their opinions about the program with their child without telling the child what to think and how to react themselves in response to a particular depiction. Similarly, Krcmar (1997) found that parents who scored higher on communication orientation were more likely than their low communication counterparts to ask questions of their child. Interestingly, children from high communication families also asked more questions, apparently perceiving parent-child discussion as a true process of give and take where their opinions would be taken into consideration.
Therefore, communication orientation impacts the kinds of media used by children, the types of statements made by parents and children, and to some extent, communication orientation acts as a mediating variable between television stimuli and media effects. For example, Austin, Roberts, and Nass (1990) found that high communication families were more likely to mediate television for their children, which in turn resulted in children's social reality perceptions being affected by parent perceptions. In this way, high communication families discuss television and help children interpret it.
Because of this greater verbal openness among high communication families, it is possible that children from these families see all situations as being open to interaction. That is, they perceive an egalitarian ethic in which each member of any dyad has the right to take an active part in an interaction, to engage in the interaction, and to react to it. It seems possible, too, that this ethic can be applied to other situations. For example, when a high communication child engages in conflict, s/he may perceive that the most important element of the conflict is the content of the altercation. Therefore, regardless of the status of the two individuals, each has an equal say. Ultimately, the most important factor is whether or not the conflict is fair and, similarly, whether or not the impetus for the conflict was fair. It is arguable, therefore, that high communication children may place greater emphasis on motive in a conflict situation than children who rate lower on this dimension. Based on the argument earlier, that age and gender affect perceptions of media violence, and based on the argument forwarded above, it is predicted that:
[H.sub.3]: Controlling for age and sex of the child, children who score higher on the communication dimension will perceive motivated aggression as being more justified.
Control orientation. Control orientation has also been found to be related to a variety of outcome measures in terms of children's viewing patterns and in terms of parents' communication styles. Additionally, control orientation has been found to act as a mitigating factor between child exposure to media violence and the outcome of watching that violence. For example, children in high control families tend to mirror their parents' behavior in terms of TV viewing. Chaffee and Tims (1976) found that students from families high in control orientation were more likely to view programs with their parents (28% of their viewing) than children from families low in control orientation, who watched only 15% of their programs with their parents. Abel (1976) also attempted to index the extent to which family communication patterns predicted children's viewing behavior. The results indicated that in families ranking high on the control dimension, children's viewing was related to the children's perceptions of their parents' viewing preferences. Although it is possible that similar viewing patterns between parents and children are related to factors such as availability of only one television or lack of other leisure-time activities, the fact that these parallel viewing patterns occurred only with high control families suggests that children imitate parents' viewing because parents have higher status in these families.
Finally, Chaffee, McLeod, and Atkin (1971) examined the effect of family type on children's imitation of parental viewing behavior. Families who scored high on the control dimension were most likely to share similar television viewing patterns in terms of both content and amount of time spent viewing. So, high control orientation relates to greater imitation of parental behavior, thus stressing the ideology in those families for children to imitate the high status members of the family, specifically, the parents.
Apparently, status and authority are also important in control-oriented families during parent/child discussion. Both Messaris and Kerr (1983) and Krcmar (1997) found that control orientation was related to greater instances of parental commands. Messaris and Kerr (1983) found that families with higher scores on control orientation were more likely to have mothers who reported providing directive comments or advice about television. Krcmar (1997) found that greater control orientation led to more instances of parental commands during interaction sessions.
The results above suggest that control-oriented families stress a less egalitarian ethic. In fact, rules and authority govern these families, and authority is seen as an important factor in both behavioral outcomes (such as TV viewing) and in verbal interactions. It is possible, therefore, that for children from high control families, authority and rules govern decision-making. This is different from high communication children who may examine the variables involved in a decision when they make choices. Therefore, it is predicted that:
[H.sub.4]: Controlling for age and sex of the child, children who score higher on the control dimension will perceive unpunished aggression as being more justified.
Despite the fact that children from high control families use authority and rules to guide their perceptions, it is somewhat paradoxical that when making their own decisions, these High control children are also more likely to engage in noncompliance. In fact, Krcmar (1997) found that children from high control families were most likely to disobey their parents when they had an opportunity to make a decision when their parents were not present. Similarly, it is the parents who are most authoritarian (Baumrind, 1971) who seem to have the least compliant children, and it is the mothers who use the greatest incidence of commands whose children disobey them more frequently (Kuczynski & Kochanska, 1990). An explanation for this seeming paradox may come from reactance theory (Brehm, 1981), which suggests that the very attempt to limit or control behavior is likely to result in that person's counter attempt to assert his/her freedom. This attempt usually comes in the form of performing the restricted behavior. In accord with reactance theory, children are less likely to comply with their parents in families who exhibit higher levels of control. Therefore, it is likely that high control orientation will result in greater likelihood to select the aggressive story ending.
But another factor--exposure to the intervening experimental stimuli--is also likely to affect children's choice of story ending. Previously, it has been demonstrated that justified violence increases the likelihood to act aggressively (Paik & Comstock, 1994). Similarly, it is expected that violence perceived as justified will result in the perception that violence is normative. In sum, children who come from control-oriented families, especially those who perceive the violent clip as justified (regardless of experimental condition) will be more likely to choose the aggressive story ending. It is predicted that:
[H.sub.5]: Controlling for age and sex of the child, children who score higher on the control dimension will be more likely to choose the aggressive story ending, especially after viewing aggression they perceived as justified.
Children in grades K-6 were recruited for participation in a study of television violence. The parents of children in three schools were contacted by letter and asked if their child might participate in the study. In one of the three schools, parents were informed that children would participate after school; in the other two schools, parents were told children would participate during normal school hours. The response rate from each of the three schools was similar, with an average of 23% of the children returning consent forms and agreeing to participate.(2)
A total of 191 children participated in the study. An equal number of boys and girls took part in the study. Overall, 67% of the children self-identified as white, 28% as African American, 1% as Asian, 2% as Hispanic, and 2% as "other." The mean age was 8.2.
Children scheduled to participate in the experiment were brought from their classroom to the cafeteria or an empty classroom. There, they were randomly paired with an undergraduate experimenter who had previously been randomly assigned to one of 3 experimental conditions.(3) Undergraduate experimenters lead the children to the appropriate VCR, which was showing either clip 1, 2, or 3. VCRs were set up facing a corner of the room. Children sat on the floor in front of the television. In this way, children could see only the screen of the television in their own experimental group. The undergraduate experimenters did not sit with the child while the clip was being shown; however, children were reminded by the person operating the VCR that they were not to talk to each other during the clip. In the rare instances where this occurred, the child was quickly reminded to remain quiet.
After watching the clip, children were reunited with their undergraduate partner. The experimenter read the child some questions from the brief questionnaire and filled out the child's responses. Next, the child was read a brief story book with pictures. The child was asked to choose from one of four possible story endings. This was done to obtain a proxy measure of children's willingness to provide an aggressive solution for a potential interpersonal problem. Next, children were debriefed about the harmful effects of aggression and the positive consequences of "talking it out." Lastly, children were given a colorful sticker and thanked for their participation.
Video clip. The 5-minute clip was taken from Walker Texas Ranger and was edited to show a brief story about Walker and his partner, who are searching for a young man thought to be in trouble. They interview several people who ultimately lead them to the young man. When the man is found in the desert, he is involved in a fight with another man. In the first condition, the clip shows one man hitting the other. The second man then hits back in retaliation. Finally, the second man is taken away by the police for fighting. Therefore, both motive and consequence are shown in condition one. In the second condition, the portion in which the perpetrator is then taken away by the police was edited from the clip. Therefore, a motive but no consequence is shown in version 2. In the third clip, the first punch is edited from the clip, and the police are shown taking away the perpetrator. Therefore, there is consequence, but no motive in the third clip.
Questionnaire. The questionnaire contained several parts. First, children were asked simple demographic information such as their age and whether they were a boy or a girl. Next, they were asked about the clip they had just seen. They were first reminded of the two men fighting. Each man was described to them in a neutral way (the man in the pink shirt with the beard and the man with the very long hair). For each man they were asked first if he was "right" or "wrong" for hitting the other man. Although "right" and "wrong" are obviously not synonymous with justified and unjustified, an earlier pilot study revealed that this was the most complex language understood by 5-year-olds and, for them, was the best synonym.
Once they responded that the man was either "right" or "wrong," they were asked "how right?" Children were asked to indicate if the man was just a "tiny bit right/wrong," a "medium amount right/wrong" or "really, really right/wrong." Pretesting had indicated that breaking the scale down this way allowed even 5-year-olds to use a 6-point scale. The scale was organized so that "really, really right" was coded as "6" and "really, really wrong" was coded as "1." Next, they were asked to explain why the man was right/wrong for what he did. Children's responses were recorded verbatim by the experimenter. These questions were repeated for each of the two men involved in the altercation in the clip unless, as in the final condition, both men were not shown hitting. In that case, only the relevant question was asked. In this way, all subsequent statistical analyses could be computed on children's responses regarding both the perpetrator and the retaliator.
Next, the questionnaire contained the 9 items from the Family Communication Patterns inventory (McLeod & Chaffee, 1972). These questions are designed to measure two dimensions of family communication: openness/warmth (communication dimension) and parental control (control dimension). A factor analysis revealed that, in fact, a 2 factor solution best fit the data. Factors were retained if the factor loading was above .60 for the retained dimension and below .30 for the other dimension. All but 1 of the items loaded on one of the 2 dimensions. One item, however, seemed to pose problems, particularly to the younger children. "You should always look at both sides of an issue," was part of the original McLeod and Chaffee (1972) scale and normally taps into the communication dimension. However, this item has been used primarily with older children, making its utility with younger children questionable. Lastly, the questionnaire contained questions about the child's own TV viewing and about his/her parents' involvement in and control over TV viewing.
Story book. A small story booklet (including captions and pictures) was created for the purpose of the experiment. The story completion task was designed to test if children would be more likely to select an aggressive or non-aggressive solution to an interpersonal problem as the result of their perceptions of the stimulus clips. Each child was asked to choose an ending for the story from four possible endings. The plot of the story contains a character who has his lunch stolen by the class bully. Faced with the dilemma about what to do about the stolen lunch, the main character may proceed one of 4 ways: "tell an adult" (coded as "0," nonaggressive choice); "confront the bully and nicely ask him to return the lunch" (coded as "0"); "chase the bully, push him down, and retrieve the lunch" (coded as "1," aggressive choice); or "tell all your friends so they can gang up on the bully, beat him up, and get the lunch back" (coded as "1"). The four endings were rotated to control for order effects.
[H.sub.1] predicted that there would be a main effect for gender such that boys would see the violent clip as more justified than girls would across all three conditions. An ANOVA was conducted with sex as the independent measure and perceptions of justification as the dependent measure. Across all 3 conditions, there was no main effect for sex (F(1, 190) =.25, p[is greater than].10). [H.sub.2] predicted a main effect for age, with older children perceiving the violence as more justified. Similar analyses revealed that older children did perceive the violence as more justified across the 3 conditions (M = 2.30, SD = .91) than did younger children (M = 1.97, SD = .79), although this result only approached significance (F(1, 190) = 2.77, p [is less than] .10).(4)
Hypothesis 3 predicted that, controlling for the age and the sex of the child, children who rated higher on the communication dimension would be more likely than low communication children to interpret motivated violence as justified. To test this, only those participants who saw the `motivated' clip were used in the analysis. Because the FCP produces continuous variables, a regression was conducted. Sex and age were entered on the first step. As a set, these were not significant predictors of perceptions of justification ([R.sup.2] = .07, F (57, 2) = .16, p [is greater than] .10). Communication orientation was entered on the next step. The change in [R.sup.2] was significant ([R.sup.2] change = .04, F (56, 3) = 3.45, p [is less than] .05), and the beta was in the predicted direction (B = .20, p [is less than] .05). Children whose families rated higher on the communication dimension perceived motivated violence as being more justified (M = 2.79, SD = .81) than those who rated low on that dimension (M = 2.03, SD = .92). Therefore, Hypothesis 3 was supported.
Hypothesis 4 predicted that controlling for the age and the sex of the child, children who rated higher on the control dimension would be more likely than those low on the control dimension to interpret punished violence as less justified. Only those participants who saw the 'punished' clip were used in the analysis. Next, a regression was conducted entering sex and age on the first step. These were not Significant predictors of perceptions of justification ([R.sup.2] = .12, F (63, 2) = .46, p [is greater than] .10). Control orientation was entered on the next step. The change in [R.sup.2] approached significance [R.sup.2] change = .03, F(62, 3) = 2.95, p [is less than] .10), and the beta was in the predicted direction (B = -.19, p [is less than] .10). Therefore, Hypothesis 4 was not supported, although results approached significance.
Hypothesis 5 predicted that children who rated higher on the control dimension and who interpreted the violent clip as justified would be more likely than other children to choose an aggressive story ending. An ANOVA was conducted to test this hypothesis. The dependent measure, the story endings, was coded as either aggressive ("1") or nonaggressive ("0"). The independent measures were control orientation (high/low via a median split) and perceptions of the violent clip as justified (yes/no). Overall, there were no main effects for either control orientation (F (189, 2) = .61, p [is greater than] .10) or perceptions of justification (F (189, 2) = 10, p [is greater than] .10). However, the interaction between control orientation and perceptions of justification on choice of story ending was significant (F (188, 3) = 3.2, p [is less than] .05). A post hoc Scheffe revealed that children who were more control-oriented and who had perceived the violent clip as justified chose the aggressive story endings significantly more frequently (M = .18, SD = .36) than the other participants. Children who were more control-oriented and who had perceived the clip as unjustified (M = .09, SD = .24), children who were less control-oriented and perceived the clip as unjustified (M = .08, SD = .42), and those who were less control oriented and perceived the clip as justified (M .08, SD = .27) all chose the less aggressive story endings more often.
This study demonstrated that there are individual differences in children's interpretations of television violence as justified or unjustified. First, although earlier research has found consistent age and sex differences, the present study found that boys and older children do not perceive all acts of aggression as justified, although the result for age approached significance and was in the predicted direction. One possible explanation for these findings, which ran counter to earlier results in this area, is that the test may have been under powered. A larger sample size may have resulted in support for [H.sub.2] This possibility is further supported by the fact that the results were in the predicted direction.
In addition to these demographic variables, family communication patterns were also investigated. It was found that the FCP does in fact account for some of these interpretive differences. For example, children who rated higher on the communication dimension were more likely to see motivated violence as justified, whereas children who rated higher on the control dimension were likely to see punished violence as less justified, although this only approached significance. In addition, children who were more control-oriented and who had perceived the violent clip as justified chose aggressive story endings significantly more frequently than their counterparts who were either less control-oriented and who had perceived the clip as justified, children who were less control-oriented and perceived the clip as unjustified, and those who were less control-oriented and perceived the clip as justified.
It is clear from the present study that family interaction factors mitigate interpretations of TV images and that those interpretations affect child beliefs about the aggression as justified. The present study has two main theoretical implications.
First, family communication mitigates the relationship between TV viewing and interpretations of TV images. Past research has shown that the context in which TV violence occurs affects how those images are understood by audiences. For example, Potter and Barry (1997) found that violence that appears motivated or is performed by an attractive perpetrator is perceived as justified by viewers, whereas violence that is not motivated or is performed by an unattractive perpetrator is seen as unjustified. These findings, although important to media effects scholars, leave out the issue of individual differences in viewers' interpretations, and especially, the role of family environment on children's interpretations of television images. Therefore, future theory building should include attention to family factors when drawing conclusions about how TV is interpreted, especially by children.
The second main implication of this research is that interpretations of images, and not simply the images themselves, impact the ways in which television affects us. Recently, researchers have begun to consider the link between interpretations of images and the effects of those images. For example, in another report based on the present data, it was found that age affected children's interpretations of violent images, which in turn impacted the effect that the image had on them. Interestingly, this study suggested that it was the interpretation of the image, in addition to the content of the image, that affected children's beliefs about aggression. Simply put, an identical image, seen by two different children, was interpreted in two different ways and had different effects on them. Therefore, interpretations of a violent image make a difference in the effect of that image, but more importantly, interpretations differ from individual to individual. The TV image alone does not determine its effect.
What can be inferred then, is that the increase in aggression, the increased desensitization to aggression, or the increased belief that violence is the norm should be seen as a multi-stage process. First, viewers watch the violent images on television: those images depicting violence in a specific context may then begin to affect interpretations. But the process is not complete and not fully determined by the image itself. Viewers then interpret those images as justified or unjustified, based on what they have just seen and based on their store of information created through family factors, as well as individual difference factors. Lastly, based on those interpretations, viewers react by either increasing their aggression, showing no change, or in some cases, decreasing their aggression. This study, then, suggests that although program images are important in determining differences in interpretation, other factors must be considered as well. In fact, current theories used to explain increases in aggression resulting from watching TV violence tend to neglect these family factors.
For example, Social Learning Theory has been used to explain many media effects, from increases in aggression (Potter & Barry, 1997) to eating disorders (Harrison & Cantor, 1997), but Social Learning Theory places greater emphasis on the context of the act depicted than on the variations in individual interpretations of those acts, suggesting that tendency to imitate is based on the image, and not the interpretation of that image. Therefore, models of media effects must be expanded to include factors that mitigate interpretations of images, and not just the tendency to imitate those images.
Limitations and Future Research
This study found that family factors mitigate the relationship between TV viewing and interpretations of television, which in turn affects whether or not violent TV images cause increases in aggression. However, several limitations exist.
First, and perhaps most important, information on family communication
patterns was collected only from the child. Previous research has found
that the correlation between parent and child scores on the FCP are quite
low (Ritchie, 1991) and that these low correlations seem to be due to parents'
and children's tendency to focus on different instances of communication
when drawing their conclusions concerning the family environment (Krcmar,
1997). Therefore, this study may actually measure the relationship between
children's interpretations of their family and their interpretations of
media violence. While this may be interesting in and of itself, that set
of results would be quite different from those suggested here: that family
environment affects children's interpretations of TV. For example, it is
plausible that a child who perceives his/her family as more controlling
and rigid may do so because s/he is particularly focused on rules and authority.
Similarly, s/he might also perceive violence that is punished as less justified
because of this attention to rules and authority. This explanation would
be consistent with the findings of this study but would suggest that parallel
interpretations on the part of the child, and not family influences, were
explanations for children's interpretations of the TV violence. Therefore,
future research should measure both parent and child scores on the
FCP to rule out this alternative explanation.
A second limitation of this study was the operationalization of increased aggression. In order to test if children who perceived the violence as justified acted in a more aggressive manner, children were asked to choose an ending to a story book that was read to them. A more valid test of aggression, such as an increase in actual aggressive behavior, might be a better operationalization of the outcome variable. In fact, although the method used may have tested beliefs about aggression as normative, additional methods must also be used to test this variable. Future research might attempt to link interpretations of media violence with increased aggression and with normative beliefs through more valid tests.
Overall, this study has attempted to demonstrate that the process of effects resulting from TV viewing is actually a multi-step one in which viewers encode, interpret, and compute what they see. Only then do they show attitudinal and behavioral effects. Many factors mitigate the interpretation of the image. For example, family communication factors affect how children interpret what they see on television. Therefore, future models of media effects must take into consideration the factors that mitigate the relationship between TV viewing and interpretations of TV images. Furthermore, future models must be built with emphasis on the fact that the link between TV viewing and imitation occurs through a multi-stage process.
(1) For a complete discussion of the relationship between development and age differences on children's response to justified and unjustified television violence, see Krcmar (1997). In another paper based on this data set, a detailed argument is forwarded concerning children's moral development and their responses to the experimental stimuli.
(2) The response rate in this study was relatively low, calling into question the representativeness of the sample. However, because consent forms had to be returned by children, it is likely that some did not reach us because they were lost. In fact, of the consent forms that were returned, approximately 97% gave affirmative consent. Therefore, it is assumed that the sample, although small, was not different in any important way from those who did not participate.
(3) It was expected that random assignment would result in equal numbers of boys and girls per experimental cell. In condition 1 there were 28 girls and 40 boys. Condition 2 had 25 girls and 35 boys. Condition 3 had 30 girls and 29 boys.
(4) Lack of significance may have been due to lack of power for the test. Power for the test of H2 was .37.
Abel, J. D. (1976). The family and child television viewing. Journal of Marriage and the Family, 38, 331-335.
Alvarez, M. M., Huston, A. C., Wright, J. C., & Kerkman, D. D. (1988). Gender differences in visual attention to television form and content. Journal of Applied Developmental Psychology, 9(4), 459-475.
Austin, E. W., Roberts, D. F., & Nass, C. I. (1990). Influence of family communication on children's television interpretation processes. Communication Research, 17, 545-564.
Bandura, A. (1994). Social cognitive theory of mass communication. In J. Bryant & D. Zillmann (Eds.), Media effects (pp. 61-90). Hillsdale, NJ: Earlbaum.
Ball, S., & Bogatz, G. A. (1970). The first year of Sesame Street: An evaluation. Princeton, NJ: Educational Testing Service.
Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology Monographs, 4, 61-83.
Brehm, S. S. (1981). Oppositional behavior in children: A reactance theory approach. In S. S. Brehm, S. M. Kassin, & F. K. Gibbons (Eds.), Developmental Social Psychology: Theory and Research (pp. 117-124). New York: Oxford Press.
Chaffee, S. H., & Tims, A. R. (1976). Interpersonal factors in adolescent television use. Journal of Social Issues, 32(4), 98-115.
Chaffee, S. H., McLeod, J. M., & Atkin, C. K. (1971). Parental influences on adolescent media use. American Behavioral Scientist, 14, 323-340.
Chaffee, S. H., McLeod, J. M., & Wackman, D. B. (1966). Family communication and political socialization. Paper presented at the annual meeting of the Association for Education in Journalism, Iowa City, IA.
Collins, W. A. (1983). Interpretation and inference in children's television viewing. In J. Bryant & D. R. Anderson (Eds.), Children's understanding of television (pp.125-150). New York, NY: Academic Press.
Collins, W. A., Sobol, B. L., & Westby, S. (1981). Effects of adult commentary on children's comprehension and inferences about televised aggressive portrayal. Child Development, 52, 158-163.
Corder-Bolz, C. R. (1980). Critical television viewing skills for elementary schools. Television and Children, 3(2), 34-39.
Corder-Bolz, C. R., & O'Bryant, S. (1978). Can people affect television: Teacher vs. program. Journal of Communication, 28(1), 97-103.
Eron, L. D., (1983). Age trends in the development of aggression, sex typing, and related television habits. Developmental Psychology, 19(1), 71-77.
Freidrich, L., & Stein, A. H. (1972). Aggressive and prosocial television programs and the natural behavior of preschool children. Monographs of the Society for Research in Child Development, 38(4, Serial No. 151).
Hamilton, J. T. (Ed.). (in press). Television violence and public policy. Detroit: Michigan Press.
Harrison, K., & Cantor, J. (1997). The relationship between media consumption and eating disorders. Journal of Communication, 47, 40-67.
Huesmann, L. R. (1986). Psychological processes promoting the relation between exposure to media violence and aggressive behavior by the viewer. Journal of Social Issues, 42(3), 125-140.
Huesmann, L. R., Guerra, N. G., Miller, L., & Zelli, A. (11994). The normative beliefs about aggression scale (NOBAGS). Chicago: University of Illinois.
Huesmann, L. R., Guerra, N. G., Miller, L., & Zelli, A. (1992). The role of social norms in the development of aggression. In. H. Zumkley & A. Fraszek (Eds.), Socialization and aggression (pp. 139-152). New York: Springer.
Huesmann, L. R., Lagerspetz, K., & Eron, L. D. (1984). Intervening variables in the TV violence-aggression relation: Evidence from two countries. Developmental Psychology, 20(5), 746-775.
Krcmar, M. (1997). Justified and unjustified television violence and the development of moral reasoning in children. Paper presented at the National Communication Association conference, Chicago, IL.
Krcmar, M. (1996). Family communication patterns, discourse behavior, and child television viewing. Human Communication Research, 23, 251-277.
Kuczynski, L., & Kochanska, G. (1990). Development of children's noncompliance strategies from toddlerhood to age 5. Developmental Psychology, 26, 398-408.
McLeod, J. M., & Chaffee, S. H. (1972). The construction of social reality. In J. Tedeschi (Ed.), The Social Influence Process (pp. 50-59). Chicago: Aldine-Etherton.
Meadowcroft, J. (1986). Family communication patterns and political development: The child's role. Communication Research, 13, 603-624.
Meadowcroft, J., & Fitzpatrick, M.A. (1988). Theories of family communication: Toward a merger of intersubjectivity and mutual influence processes. In R. P. Hawkins, J. W. Wiemann, & S. Pingree (Eds.), Advancing communication science: Merging mass and interpersonal processes (pp. 253-275). Beverly Hills, CA: Sage.
Messaris, P., & Kerr, D. (1983). Mothers' comments about TV: Relation to family communication patterns. Communication Research, 10, 175-194.
Paik, H., & Comstock, G. (1994). The effects of television violence on anti-social behavior: A meta-analysis. Communication Research, 21, 516-546.
Potter, W. J., & Barry, M. (1997). Component factors of justification of television violence. Unpublished Manuscript.
Potter, W. J., & Ware, W. (1987). An analysis of the contexts of anti-social acts on primetime television. Communication Research, 14, 664-686.
Ritchie, L. D. (1991). Family communication patterns: An epistemic analysis and conceptual reinterpretation. Communication Research, 18, 548-565.
Rothschild, N., & Morgan, M. (1987). Communication and control: Adolescents' relationships with parents as mediators of television. Journal of Early Adolescence, 7 (3), 299-314.
Sheehan, P. W. (1983). Age trends and the correlates of children's television viewing. Australian Journal of Psychology, 35(3), 417-431.
Surbeck, E., & Endsley, R. C. (1979). Children's emotional reactions to TV violence: Effects of film character, reassurance, age, and sex. Journal of Social Psychology, 109(2), 269-281.
Thomas, M. H., & Drabman, R. S. (1978). Effects of television violence on expectations of other's aggression. Personality and Social Psychology Bulletin, 4(1), 73-76.
Thomas, M. H., & Drabman, R. S. (1975). Toleration of real life aggression as a function of exposure to televised violence and age of subject. Merrill Palmer Quarterly, 21(3), 227-232.
Valkenburg, P. M., Krcmar, M., & de Roos, S. (in press). The impact of adult mediation and a cultural children's program on children's knowledge of an attitudes toward opera. Journal of Broadcasting and Electronic Media.
Valkenburg, P. M., & van der Voort, T. H. (1995). The influence of television on children's day dreaming styles: A 1-year panel study. Communication Research, 22(3), 267-287.
Valkenburg, P. M., Vooijs, M. W., van der Voort, T. H., & Wiegman, O. (1992-1993). The influence of television on children's fantasy styles: A secondary analysis. Imagination, Cognition and Personality 12(1), 55-67.
van der Voort, T. H., Vooijs, M. W., & Bekker, P. A. (1982). Children and violence on television. Psycholoog, 17(6), 370.
Watkins, B. A., Calvert, S. L., Huston-Stein, A., & Wright, J. C. (1980). Children's recall of television material: Effects of presentation mode and adult labeling. Developmental Psychology, 16, 672-674.
Weintraub-Austin, E. (1993). Exploring the effects of active parental mediation of television content. Journal of Broadcasting and Electronic Media, 37, 147-158.
Marina Krcmar is an Assistant Professor in the Department of Communication at the University of Connecticut. Her research interests include television and children's moral development and the relationship between television and family communication. This study was made possible by a summer research grant to the author by East Carolina University while the author was an Assistant Professor there.
Please direct inquiries about this page to Dale Vidmar.