This post is a part of series of study reviews written by professors whom I have an interest in joining as their grad student.
Kenneth Lachlan is a pupil of Ron Tamborini whom he’s also on the list. Given their research methodology concerning media research and more specifically video games, it seems that many of their works are on content analysis. It does seems interesting and insightful in careful analysis of what players do and see during game play, but data collection can be an arduous process. Moreover, many of their works mainly analyzed a medium of interest, say violent video games or violent television, without any input from the audience. So, most of their discussions revolved around the implications of said content, this left me a bit unsatisfied if I were to write this kind of study.
However, IMO content analysis can be classified as basic science, where it provides the foundations for applied science and therefore provide background information for future applied studies.
The current study is an indication of this opinion where content analysis and media effects research is combined to provide an insightful look into violent video game play and how players’ in-game behaviours varies according to personality and game content.
The methodological techniques used in past video game content analyses overlook player differences that might play an important role in generating violent content. The current study aimed to explore these differences in terms of personality type, trait hostility, perceived realism, and telepresence tendency. One hundred and sixty participants were assigned to play one of four video games after filling out a series of personality inventories. Content was then evaluated using coding techniques adapted from prior video game content analyses (Smith, Lachlan, & Tamborini, 2003). The findings indicated that game content is highly variable across player characteristics and telepresence tendencies.
It is said that the advisor(s) we choose partly determine our future; it’s like being in the character creation screen of many Western RPGs or choosing a locked research path in an RTS…
Besides content analysis, they’ve also examined telepresence: their argument is that while experiencing a heightened sense of telepresence would increase players’ identification with violent video game characters which leads to increased aggression and hostility. However, it is unknown whether this increased aggression and hostility would change game play which changes the content of the game as they argued. IMO, would it become a vicious cycle of increased aggression? Of course, holding that argument as true, this kind of cycle would be broken by progression in the game, i.e. cutscenes, loading screens, bathroom breaks, etc. Of course, these things won’t show up in this study.
Personality is another facet in this study where stable internal psychological variables play a role in the relationship between media and people’s psyche. Trait aggression is one of the notable personality trait in the relationship between media violence and aggression.
Combining all three facets, they argued that content analysis of interactive media should consider the interaction between player’s personality, game content and telepresence as more appropriate method (to put it bluntly), than just considering only the game content, due to the variability in the game content. So their overall question is what impact these three factors (separately or combined) will have on the frequency of in-game violence?
Participants: 160 undergrads. No demographic info given, except their justification in using undergrads, which I take this fact of research as granted (i.e. I almost don’t care) and they all come from the same intro to communications class. Although, most psych studies give out demographics like a convention, I find it unnerving that’s missing in this article.
I’m also pretty uptight in what versions of the questionnaire they used or any modifications they made.
Perceived realism: 5 items on a likert-scale. It’s a bad measure, in terms of reliability, since they had to drop two items from the analysis in order to have it reliable. A new perceived realism measure needs to be developed since the one they used is from the 1980’s.
Buss-Perry Aggression Questionnaire: measures trait aggression with physical aggression, verbal aggression, anger and hostility as subscale. A 29-items on a 7-point likert-scale. A pretty standard fare because I’m using it for my project…
Revised Eysenk Personality Questionnaire: a personality questionnaire measured for extroversion, neuroticism, psychoticism and social desirability.
Social Justice Questionnaire: measures for vigilantism, approval of punishment, and empathic concern for others in need.
Immersive Tendencies Questionnaire: measures for telepresence… how many items? How is it answered? I really want to know that.
Content analysis measure
To analyse video game content, they first start with an operational definition of violence as “any overt depiction of a credible threat of physical force or the actual use of such force intended to physically harm an animate being or group of being. Violence also includes certain depictions of physically harmful consequences against an animate being/s that results from unseen violent means.” Lifted off from the National Television Violence Study (a big study).
With that in hand, they defined a single violent interaction needing a unique perpetrator committing a harmful act against a unique victim. If there’s a change in say, perpetrator, act or victim then it becomes another violent interaction. So if I shoot a victim, and then fire with a different weapon at the same target, it counts as two violent interactions. If I shoot a bazooka at a horde of enemies, it counts as one. If I shoot at an enemy, then I get a return fire from the enemy, it counts as two.
Besides counting, they also looked at the means of violent interactions.
- Justified violence: protection of life, protection of property and retaliation.
- Unjustified violent: personal gain, mental instability, other.
- Natural means: kicking, punching
- Unconventional weapons: anvil, baseball bat
- Conventional melee weapons: knife, sword, mace
- Handheld firearms: rifle, pistol
- Heavy weapons: bomb, bazooka
Finally, consequences of violent interactions:
- No depicted harm,
- mild harm (indicators that harm will come, but has not followed)
- moderate harm (depiction of blood or visible injury)
- extreme harm (death, disfiguring injuries or massive amounts of blood spilled)
Video games used
Four video games are used: Rainbow Six, Grand Theft Auto 3 for the X-Box. SoCom: Navy Seals and Grand Theft Auto: Vice City for the Playstation 2. They chose these games because they were top selling games for their respective consoles, their similarities in game genre (first-person shooters and crime adventure). Okay, but what edition of Rainbow Six are they talking about?
Participants completed demographics and personality questionnaire in class, two weeks prior to the experiment. Participants are randomly assigned to play one of the four games. Play time is 20 minutes. Their game sessions are recorded for later analysis, using a VCR. I find it a big pain using tapes, takes a lot of space, can’t use your computer to view them at your own convenience or rewind the tapes to the beginning when you’re done. I wonder how much a DVD recorder costs?
I suggest that you’d get a copy of the article as the results section has lots of tables so you can follow the rest of this post. I’m not sure if the publisher would like it if I post the tables in this post. (Not that this blog is popular anyway).
Their initial analysis is looking at the averages and standard deviation of their content analysis (frequency, means and consequences) is that it looks there’s a great deal of variability between games. How significantly different from each other? They didn’t do any t-tests at this point or if there’s any need for one. Strictly looking at the number of violent interactions, the GTA games have lower averages than the other two games (Rainbow Six and SoCom).
A peculiar thing is that GTA: 3 has a very large standard deviation which the authors interpret this as content can be very variable across players in a, IMO, sandbox-type game. How it’s done makes a difference, GTA series has a large average and variability in the number of unjustified violence and the use of unconventional means. As for consequences, again GTA: 3 shows the same peculiarity when it comes to extreme harm, large standard deviation, although the average doesn’t look different from the rest. How do I interpret this? This shows that it’s a quite complex issue when it relates to video game controversy where a lot of fuss is made on the GTA series while other more violent games get an unremarked pass. A Slashdot post discusses this issue. However, there is still more to discuss, so keep reading.
Using regression analyses to examine the role of demographics, personality, perceived realism, telepresence, trait aggression, and social justice and its influence on game content. They’ve found three statistical significances:
- There’s a positive relationship between unjust violence with income and psychoticism, and negative relationship with hostility (from the Buss-Perry).
- There’s a positive relationship between gun use (to authors: did you mean handheld firearms?) with anger and empathy. A negative relationship with telepresence.
- There’s only one positive relationship between no harm depicted with psychoticism.
What to make of it? It means look at the individual games and don’t use the results to generalize to every video games. These results don’t make sense.
- Rainbow Six: A significant result was found for justified violence in that trait physical aggression is positively related while empathy is negatively related. Now it makes sense. No harm depiction is positively related to vigilantism and trait verbal aggression, while age is negatively related. No comment.
- SoCom: Significant result was found for total number of violent interactions: positive relationship with extroversion, social desirability and empathy. Negative relationship with perceived realism and psychoticism. Not quite the relationship one’s expected, what do personality characteristics that are primarily social have anything to do with in-game violent interactions? I would understand for perceived realism in that players would be more cautious.
- Grand Theft Auto 3: Two significances found: unjustified violence was positively related to hostility. Makes a bit sense. Natural means (punching and such) is positively related with perceived realism, verbal aggression, anger, and extroversion. It is negatively related with age, telepresence and vigilantism (nearly significant). Makes sense.
- Grant Theft Auto: Vice City: nothing, no significance found.
Earlier, the authors wanted to examine telepresence and its impact of game content. They haven’t found what they expected, but in the other direction where presence is negatively related to violent interactions. The authors remarked that, in my own words, that experienced players seemed to be methodological in their game sessions, like pros. Whereas less experienced players are more prone to shooting randomly, like amateurs. Such difference in gaming experience might explain some of the variances in violent interactions and the relationship with telepresence. Hmm… this is a weakness in content analysis while there are more violent interactions, does their emotional state affect their game performance or more importantly various outcome variables, such as aggression or frustration? Another question would be how do we increase telepresence without any other variables in play? Some of Blascovich’s work might provide clues.
The authors discussed personality and how it interacts with game content. In short, the premises of a game (genre, narrative, theme, realism, etc.) would form unique interactions with personality characteristics which leads to different kinds of game content and therefore indicative of players’ style. For example, the authors argued that SoCom emphasized stealth and that individuals with high psychoticism scores would find this kind appealing. If only they had a video game rating measure…
Okay, so this study does tell us that video games are very unique and the people playing are also very unique making game content very very variable. But a psychologist would ask is, would behavioural, cognitive and affective outcomes be different between these games when it comes to the number of in-game violent interactions? This study cannot answer that, but at least it’s the first to ask to this question and it does incite me in answering this question. Perhaps as my master’s thesis…
Lachlan, K. A., & Maloney, E. K. (2008) Game player characteristics and interactive content: exploring the role of personality and telepresence in video game violence. Communication Quarterly, 56 (3), 284-302.