Video Games Causing Aggression in Children... Not quite
TL;DR: Academic studies need a to take a look at many variables in a Child's development to confirm any causes. An aggressive child may already take a liking to violent video games, and/or is shaped be their environment upbringing. I analyze 3 different studies.
Opinion: Media tends to gravitate to the "if, then" studies, as it is easy to understand. I do not mean to take away the meaning of these studies at all, as they have created progress in our understanding of this topic. Child development itself is too complicated, and these studies are necessary building blocks to the hard stuff. Note: I am not a Master or a PHD student. My focus is more into the psychology of human interaction with computers (not just personal computers). Although video games and that type of psychology are within my domain, child development is a new area I took interest in.
An Analysis and Criticism of Psychological Studies and Articles on Violent Video Games and its’ Aggressive effect on Children
Discussion of video game violence and its effect on children have been in dialogue for many years. This can be seen whether it is parents placing much blame of their children’s behavioural issues on violent games and media or national news coverage discussing violent crimes and aggressions are the cause of over exposure to violent video games and media. Prompted by the assumptions of the public, came many academic studies analyzing the potential harm violent video games can cause a child or teen throughout their maturity. As many studies and consensus of the psychological community, it is safe to say that a child’s environment is a strong factor for their development. Having a child being exposed to violence in video games is a part of that upbringing. The issue in regards to this specifically is that research articles on this topic are not asking the right questions, and rather comparing just the two factors: violent video games exposure and aggressive behaviours (Zhang and Zheng 2017). The impact of other factors will need to be included in studies to show the level of impact violent video game has to the child’s development. (Fikkers, Piotrowski, Valkenburg, 2017). Because having children be exposed to violent video games in controlled environments can be unethical, most studies are done through self-reporting surveys. The self-reporting itself is limited by the honesty and trust of each participant, so testing the reliability and validity can serve useful for all studies in regards to violent video game exposure. The linear findings of video game exposure and violent behaviour that many psychologists claim will need to be compared to other environmental factors and measures of these will need to be taken into consideration before any conclusion of results.
Any aggression that does result from playing violent video games can vary from gender and trait aggressiveness levels. Trait aggressiveness is a stable personality predisposition of an individual and acts out with aggression to certain situations. According to a study done by Zheng and Zhang for Social Behaviour and Personality 2016, Aggression can be enhanced through playing violent video games depending on an individual’s level of trait aggressiveness. They hypothesize first, that children who play violent video games will show stronger aggression than children who did not play violent video games.
Video game violence and gender are researched by many, in which that among adolescents who play violent video games, there are gender differences in relation to aggression. Males tended to be more aggressive and show higher levels of direct aggression than their female counterparts. The second hypothesis of this research article is: Among adolescents who play violent computer games, boys will show stronger aggression than girls.
Researchers use two popular games in their region, Virtual Cop2 and a computer version of Flight Landlord (also known as Dou Dizhu, a popular non-violent card game in China. Very similar to the Western equivalence to Crazy 8s and Uno). The first study (of two) first indicates to the researchers that their assumptions were correct for which game was violent and non-violent. The result proved correct: Virtual Cop2 being the violent video game variable and Fight Landlord the non-violent game.
The first study takes in 220 children age 9 to 12 from two elementary schools (50% females and 50% males), 90% Han Chinese ethnicity, and asks them play both games. They are then asked to rate them in terms of pleasantness, excitement, violent content, violent images, fear, interest, and reality; The Buss-Perry Aggression Questionnaire was the survey used. The stimulus materials are of course the two video games, Virtual Cop: a first-person shooter, and Fight Landlord: a virtual card game. Participants were then asked to play for 15 mins of each game and answer the questionnaire right after. The ratings used were a 5-point Likert scale ranging from 1 to 5 (1 the lowest and 5 the highest). Their second study takes in a separate group of 240 children (54% females and 46% males), ages 9 to 12, 95% Han Chinese ethnicity. This time the children were split into two groups: 120 children played Virtual Cop and 120 played Fight Landlord. Researchers selected 30 aggressive and 30 nonaggressive words at random from a review of a previous study (Anderson, Benjamin, and Bartholow, 1998). After the initial gameplay, the children were to complete a semantic classification task with those words. The children were to press 1 when they see an aggressive word or 2 for non-aggressive. Their reaction time is recorded here as well. Again, using the Buss-Perry Aggression Questionnaire, they measured the high, moderate, and low, levels of trait aggression using a Likert scale ranging from 1 to 5 (1 agree and 5 the strongly disagree).
The results from their first study evaluated the violence of the computer games. Using a one-way analysis of variance to compate participants’ ratings of the two games. Virtual Cop, the violent game, gained significant higher scores for violent content (F(1,75) = 21.61, p < .001) than Fight Landlord. Violent images for Virtual Cop scored F(1, 85) = 20.52 p < .001. The result for gender difference found that males found Virtual Cop, the violent game, to be more pleasant in terms of ratings than females. This does hint that males are more likely to act out on aggression when compared to females. Study 2 focussed on possible existing aggressive personality traits within individual children using semantic testing and the levels of their aggressive traits. The results from the semantic tests found that the reaction times for aggressive words were much shorter than were for nonaggressive words. According to the study, this means playing violent games primed aggression among the children. Researchers created a priming aggression score to show the different levels of aggression. These were calculated by subtracting the reaction time to aggressive words from reaction times of nonaggressive words, then comparing the scores between the two large groups of children. The average prime aggression score was -30 ms for the aggressively primed children that played Virtual Cop than the control group that played Virtual Cop. According to the data, playing violent video games is an indicator of higher levels of aggression. Gender type and gender interaction section was analyzed using a four-way ANOVA to see whether the priming semantic test was significant in relation to game type. There was significant increase among aggressively primed boys who played the violent game compared to the aggressively primed boys who played the non-violent game. There was no significance between the groups of females. Lastly is the testing of the different game type and the trait aggressiveness. Results found that there was a significant increase game type by trait aggressiveness F(1,47) = 4.18, p<.01. Expanding on this, it is found that the children who were primed with high aggressiveness traits and played the violent video game was significantly higher compared to those who played the nonviolent game.
Study 1’s findings determined which game was violent and which game was non-violent; Virtual cop is the violent game and Fight Landlord is the non-violent game. Study 2 found that playing aggressive games primed children and playing nonaggressive games had little to no effect, thus their first hypothesis was aligned with the results. Gender differences and the effects of violent video games were also consistent with their second hypothesis, where females had no significant increase in aggression after presented with the stimuli. Aggressive traits were used to determine children with higher aggression levels to be significantly increased, where children with middle and low had no significant increase.
Limitations that researchers mentioned were that they were using a cross-sectional design. This leaves unanswered questions in terms of actual aggression development. Using a longitudinal design would assist in being more accurate in regards to development. The types of aggressions were not mentioned at all within the study, so children may be experiencing different types of aggressions when testing them. The use of the Buss-Perry Aggression Questionnaire is very useful in the measure of aggression levels, but because this study was done in China, they must make a Chinese version of this study to avoid any cultural or translation issues.
The first article discussed in this paper may come from a line of articles that try to explain aggression and violent video games without taking on many reliable considerable factors. A child’s behaviour is strongly influenced by their upbringing and environment, so a child having behavioural and aggression issues caused by exposure to violent video games does seem very plausible. But the impact of video game violence compared to the impact of other factors of youth violence needs to be taken into consideration. Using video game violence as a measure to determine behavioural issues in young people is not strong enough evidence to say just that, so examining (in addition to violent video games) family, background and other factors on youth violence must be taken into consideration. Then the impact of the degree of exposure to violent video games can be measured (DeCamp, Ferguson, 2016). They hypothesize that video game violence is not a meaningful prediction of youth violence, and family and social variables are more influential factors.
The researchers are using the Delaware School survey (DSS), a survey of eighth and eleventh grade students within Delaware’s public and public-charter schools. 5133 eighth graders and 3886 eleventh graders are to be involved within this study. Violent game play exposure, violent game play propensity, and parental attachment are the independent variables. Violent game play exposure is the primary variable, measured by questions on the survey such as… “how often on average do you play violent video games that are rated M?” Answers to this question will be based on time per week played violent M rated video games. Violent game play propensity is a measure that most studies overlook and is needed to be used due to a self-selection bias where children choose or are allowed to play violent video games. A propensity score is used to predict the focus stimulus and is calculated for everyone using 171 other indicators (listed within the appendix of this study). These indicators can consist of questions in categories of age; gender; race; ethnicity; and specifics like “do you have parent’s in the military”, “which of the following people do you live with the most…” and the list goes on. Parental Attachment is also calculated and constructed using indicators like: “my parents/guardians are proud of me” or “my parent/guardian takes interest in my activities.” Additional points of interest for the survey are: whether parent or adult has hit child, youth disclosure and parental enforcement, and yelling and violence at home.
The results are split into their respective grade groups: grade 8 and grade 11. The bivariate cross-tabulation (bivariate: measuring only the two variables: violent video game play exposure and violent behaviour) results of the survey concluded that students who have played violent video games (male and female) have reported to once hitting someone. Generally, a linear increase in rates of violence according the time of violent video games played per week. This analysis of the information is the ordinal measure done by most studies about violent video games and violent behaviour. The result for the regression tests using the many indicator questions back track the results of the bivariate cross-tabulation. It is found that if a child has hit someone, they have more likelihood to have been exposed to violent video games. However, with the addition of propensity score of the individuals, it reduces the impact of the violent game exposure variable to behaviour. The violent game exposure variable has a 2.8% variance before the introduction of propensity score, which means the result of the effects drop further and generally results in no significance. The overall effect size across the samples in relation to violent video games and violent outcomes using the propensity score is non-significant (r=0.38, p= 0.279).
There should be a caution when claiming there is a relationship with violent video games and violent behaviours from the results of this study. Researchers mention that the increased time playing video games was also not a significant increase with the effect of violent behaviour while referencing the data. Social and family backgrounds have a larger role with an individual’s development in relations to violent behaviour. A youth who is suffering at home, for example: parent’s fighting verbally or even physical fights within their own home, are more likely to act out violently themselves. The data presented in this study gives an understanding that youth development must consider different developmental norms.
Some limitations the researchers encountered is the self-report data, so they are limited by honesty of the participants. There is no assumption that a teen would have any reason to lie on these surveys, but if one or more do, it can affect the overall result in a small but critical way. Propensity scores can be used for matching or as a control variable. The matching variable approach can assist in limiting spuriousness, but having propensity as a control variable makes it a continuous predictor and not just a simple dichotomy.
As stated before, self-report measure can be a limitation as it is limited by the honesty of the participants. In the next article, the study is considering just that: whether common self-report measure of television and video game violence exposure is reliable and holds up its validity (Fikkers, Piotrowski, Valkenburg, 2017). Three self-report measures: direct estimates, user-rated favourites, and agency-rated favourites are assessed in terms of test-retest reliability, criterion validity, and construct validity.
The way an individual interprets the effects of media violence (including video games), depends on two types of measurements: reliability and validity. Reliability: “the extent to which an experiment, test or any measure procedure yields the same results on repeated trials” (Carmines and Zeller, 1979). Validity: determines if the measure is an actual reflection of the concept it is intended to measure (Carmines and Zeller, 1979). This means that any measures of low reliability can result in error variance in any statistical models. There are not many studies that exist in which the focus is on the reliability and validity of exposure to violence in media exposures, which can be alarming when many studies are claiming a strong link of violence in video games and media to real violent behaviours.
Test-retest reliabilty is “the extent to which a measure provides similar results when tested again after a period of time” (Carmines and Zeller, 1979). It can be noted that having a participant be tested over and over may serve as unethical to participants if they are exposed to violent medias for long periods over time. Citerion validity is the comparison of the scores obtained with one measure to a validated measure. Lastly, Construct validity is the “extent to which a particular measure relates to other measure consistent with theoretically derived hypotheses.” (Carmines and Zeller, 1979). The researches have 4 hypothesis: "When measuring game violence exposure to direct estimates, user-rated favourites, and agency-rated favourites will be positively associated with aggressive behaviour; When measuring television game violence direct estimates, user-rated favourites, and agency -rated favourites will be positively associated with aggressive behaviour; When measuring game violence exposure to direct estimates, user-rated favourites, and user-rated, agency-rated favourites will show more exposure for boys than girls; When measuring television violence exposure to direct estimates, user-rated favourites, and user-rated, agency-rated favourites will show more exposure for boys than girls.” (Fickkers, Piotrowski, Valkenburg, 2017). Our focus will be on the hypothesis 1 and 3 in relation to video game violence.
Using 238 participants (53.8% boys and 47.5% girls, ages 10-14) with the average age of 11.9 years, standard deviation of 1.5. They used 2 waves of online surveys and “media diaries” to collect data. Participants are to answer direct estimates of exposure and favourite titles of television and games, and fill out online media diaries once at random a week. The answers for favourite titles are then coded for violent contents using official agency-rating systems: “Kijwijzer” for television and movie titles and “Pan European Game Information: PEGI” for video game titles. Media and game violence exposure was measured through the online diaries. Researchers adapted six items from the Direct and Indirect Aggression Scale (Bj.rkqvist, Lagerspetz, & Kaukiainen, 1992): call names, push in a rough way, kick or hit, threaten to beat up, fought with, and tripped on purpose.
Analysis of the game exposure results can conclude the following: Test-retest reliability (tau = .52 r -.73, p< 0.001); criterion validity, did not exceed the Researchers guidelines of .50 (tau .28, r = .43, p< 0.001). These two test can give insight into the first and third hypotheses (When measuring game violence exposure to direct estimates, user-rated favourites, and agency-rated favourites will be positively associated with aggressive behaviour; When measuring television game violence direct estimates, user-rated favourites, and agency-rated favourites will be positively associated with aggressive behaviour.), and were positively associated with aggressive behaviour. It is shown that boys are shown more exposure than girls. The validity of the construct validity for these two hypotheses can be confirmed from the user-rated favourites with aggression and gender (aggression: tau = .18, r = .73, p<0.001). The agency-rated game favourites met all the guidelines for reliability and validity and was achieved through the test-retest reliability and criterion validity, so construct validity for the two hypotheses. Researchers then had a “Post hoc assessment” of the media diary. They explain that there is no study that has used diaries to measure exposure to violence of media. For video game violence exposure, the media diaries and the correlations of the surveys were very low. Video game violence exposure when compared to the diaries resulted in the following: test-retest reliability (tau = 0.30, r= 0.45, p < 0.001) and construct validity (aggression: tau = 0.10, r=.15, p< 0.001; gender: tau = 0.23, r = 0.35, p< 0.001). The opposite is said for television violence exposure.
The results indicate that there is support for the reliability and validity of direct estimates for violent media exposures and that future studies can use these measures to conduct their experiments.
The researchers did not indicate any limitations within their study.
Exposure to violent video games does have impact when comparing only two variables: exposure and behaviour. When other factors are introduced it becomes less and less of a valid predictor of violent behaviour. This can also be interpreted in such that a child who is raised in an unsuitable hostile environment is more likely to play violent video games than compared to children who are not in such environments. The reason for this is unknown, but can be assumed that parents in these harsh situations are more likely to purchase violent games for children.
Video game violence is getting more realistic as time goes by, so it is important for psychologist to keep studying the impacts of such.
Included Article titles and Abstracts from the respected reasearch.
Article 1: “Priming Effect of Computer Game Violence on Children’s Aggression Levels”
Abstract 1: We investigated how aggression resulting from playing violent computer games varies by gender and trait aggressiveness level. In Study 1, 220 children rated 2 video games in terms of pleasantness, excitement, violent content, violent images, fear, interest, and reality. Results indicated that Virtual Cop2 and Fight Landlord games were perceived as violent and nonviolent, respectively. In Study 2, 240 different children responded to the Buss–Perry Aggression Questionnaire, played either Virtual Cop2 or Fight Landlord, and completed a semantic classification task involving rating whether 60 words were aggressive or nonaggressive. Results showed that boys, but not girls, displayed stronger aggression after playing Virtual Cop2, compared to Fight Landlord. Further, children with high trait aggressiveness exhibited strong aggression after playing Virtual Cop2, whereas those with moderate or low trait aggressiveness did not. Overall, our results indicate that gender and trait aggressiveness both affect aggression among children who play violent video games.
Article 2: “The Impact of Degree of Exposure to Violent Video Games, Family Background, and Other Factors on Youth Violence”
Abstract: Abstract Despite decades of study, no scholarly consensus has emerged regarding whether violent video games con- tribute to youth violence. Some skeptics contend that small correlations between violent game play and violence-related outcomes may be due to other factors, which include a wide range of possible effects from gender, mental health, and social influences. The current study examines this issue with a large and diverse (49 % white, 21 % black, 18 % Hispanic, and 12 % other or mixed race/ethnicity; 51 % female) sample of youth in eighth (n = 5133) and eleventh grade (n = 3886). Models examining video game play and violence- related outcomes without any controls tended to return small, but statistically significant relationships between violent games and violence-related outcomes. However, once other predictors were included in the models and once propensity scores were used to control for an underlying propensity for choosing or being allowed to play violent video games, these relationships vanished, became inverse, or were reduced to trivial effect sizes. These results offer further support to the conclusion that video game violence is not a meaningful predictor of youth violence and, instead, support the conclusion that family and social variables are more influential factors.
Article 3: Assessing the Reliability and Validity of Television and Game Violence Exposure Measures
Abstract This study evaluated whether common self-report measures of television and game violence exposure represent reliable and valid measurement tools. Three self-report measures—direct estimates, user-rated favorites, and agency-rated favorites—were assessed in terms of test-retest reliability, criterion validity (their relationship with coded media diaries), and construct validity (their relationship with aggression and gender). A total of 238 adolescents participated in a two-wave survey and completed two media diaries. For game violence, the three self-report measures were reliable and valid. For television violence, only direct estimates achieved test-retest reliability and construct validity. Criterion validity could not be established for the television violence measures because the media diary was not a valid criterion for television violence. Our findings indicate that both direct estimates and favorites are valid measures for game violence, whereas for television violence, only direct estimates are valid. We conclude with a discussion about ways to further improve upon and reconceptualize media violence exposure measurement.
Decamp, W., & Ferguson, C. J. (2016). The Impact of Degree of Exposure to Violent Video Games, Family Background, and Other Factors on Youth Violence. Journal of Youth and Adolescence,46(2), 388-400. doi:10.1007/s10964-016-0561-8
Zheng, J., & Zhang, Q. (2016). Priming effect of computer game violence on children's aggression levels. Social Behavior and Personality: an international journal,44(10), 1747-1759. doi:10.2224/sbp.2016.44.10.1747
Fikkers, K. M., Piotrowski, J. T., & Valkenburg, P. M. (2017). Assessing the Reliability and Validity of Television and Game Violence Exposure Measures. Communication Research,44(1), 117-143. doi:10.1177/0093650215573863