Bloodlust spell from bloody video games (Barlett et al., 2008)

There are a variety of choices for my latest post and things I should do:

1- My comparison of the Byron Report with the NIMF’s and other reports, mainly on the presentation of the research on video games and their recommendations.
2- My post on the connection between violent video games and violent behaviours of people, i.e. violent crime and such. I thought of writing it for weeks, thinking of expanding on what is currently known, but never had the spare time.
3- An overhyped journal article about first person shooters and the third person effect.
4- My seminar paper for my class on Nick Yee’s Proteus effect.
5- Writing up my C.V. for a part-time job.
6- My reading of an issue in the American Behavioral Scientist.

However, compelling circumstances lead me to post about an interesting article from Christopher Barlett, a grad student at Iowa State University.


The current study utilized the General Aggression Model, with an emphasis on aggression-related priming, to explore the different effects on hostility, physiological arousal, and state aggression in those who played a violent video game (Mortal Kombat: Deadly Alliance) with differing levels of blood (maximum, medium, low, and off). Simple effects analyses showed that those in the maximum blood and medium blood conditions had a significant increase in hostility and physiological arousal, while those in the low blood and no blood conditions did not have such an increase in arousal and hostility. Further analyses indicated that those in the maximum and medium blood conditions used the character’s weapon significantly more often than those in the low and blood absent conditions. Implications and future research are discussed.

IMO, one of the blunders of research is the quantification of violence, Henry Jenkins had pointed a research study had found that all cartoons contained violence. One criticism of this study was its operational definition of violence was too broad. To this day, researchers had to clearly differentiate violent video games and non-violent video games. When I think about it, it seemed to be a persistent flaw across many studies. It’s like letting a participant run a puzzle task without carefully looking what strategies they adopt.

Christopher Barlett and colleagues had an ingenious idea in the quantification of violence, through the experimental manipulations on the amount of blood spilt in a violent behaviour in a video game. They had found that the bloodiest setting in a video game leads to an increase in state hostility, physiological arousal, state aggression, and aggressive priming.

They conducted two experiments to test the following hypotheses:

1- There will be significant differences in state hostility and physiological arousal between the various levels of spilt blood.
2- There will be significant differences in state hostility and physiological arousal between the no-blood condition and the blood condition.
3- There will be significant differences in aggressive thoughts between the no-blood condition and the blood condition.

Study 1


Participant: 74 undergrads (65 male), age (average = 19 years, SD = 1.14), play average per week = 16.51 (SD=20.77), mostly Caucasian (80%) and first-years (62%).

Video game: Mortal Kombat: Deadly Alliance. So this study could certainly be applicable to fighting games and maybe to first-person shooters at least… Actually, why not use FPS? I mean we could grab some modders to do some easy modding on manipulating the amount blood from Zero to Hollywood-type to Tasteless absurd. They chose this game because it has the option of varying the amount of blood. Another good thing is that they used one game to make comparisons in contrast to other studies that used multiple games to make such comparisons which may bring in confounding variables.

Other details include that a match last 60 seconds, there are four blood conditions (maximum, medium, low and none). The full set of characters to choose from (12 fighters) and three fighting styles (two martial arts and one weapon-based) from each character.

Trait aggression: the Buss and Perry Aggression questionnaire, 29 items to answer on a scale of 1 to 5 whether each aggression-related statement is descriptive of them.

State Hostility: They designated it as aggressive feelings/thoughts. I call it state hostility because that’s what the rest of article is calling it, unless I’m mistaken. Anyways, the State Hostility Scale asked participants 35-items on a scale of 1 to 5 on what they feel or think that is related to aggression.

State Aggression: They designated it as aggressive behaviors, my reason is the same as the previous. I find this variable weaker than the rest because it is measured through the total amount of time a player uses the weapon-based fighting style in proportion to the total amount of game play time. My problem is what if a player chooses to sadistically bitch slap the opponent to death, would that be more aggressive? Another confound is probably the player wants to compete as efficiently as possible given the contextual situation of the video game (i.e. time limit). Third, I’m not sure how they exactly measured it, is it the time when they actually use it or the time when they are in the weapon-based fighting style? Anyway, it is not translatable into the real world context of physical aggression.

Physiological arousal: to measure heart rate, participants are asked to press their index finger into a sensor. However it is limited in how they took heart rate measures, leading to inaccuracies.


Participants are given a pre-experiment questionnaire to measure their trait aggression, state aggression, state hostility and physiological arousal. That will be used as the control baseline. They are given a 3 minutes training session. Then they are randomly assigned to the four blood condition (maximum, medium, low and none) where they will be tested for 15 minutes. (What was the average play time for a fighting game in a natural setting?)


Trait aggression was not affected by the blood conditions. So that means that bloody video games don’t change our stable traits for 15 minutes at least.

There were significant effects in state hostility, starting at the medium setting, the bloodier the game the higher the state hostility score. The effect size are large and medium, respectively for the maximum (M = 76.21) and medium blood ( M = 73.88 ) condition. I am surprised that low or no blood condition did not differ significantly from baseline. Theoretically, it should have an effect since the violent act is there, perhaps not all seemingly violent act is perceived as violent. Another thing is while the effect size seemed large, it is not an extreme increase. It is subject to interpretation on what the scores of the state hostility scale means and whether a 10 point increase (in the maximum and medium blood condition) may actually mean something. Update (23/10/08): in retrospect, I can’t say that the no- or low- blood condition mean there’s no effect, since there’s no non-violent game condition to compare to.

There were significant effects on physiological arousal, but only those at the maximum blood with a medium effect size. Although, I wouldn’t take it with full confidence and won’t talk about it.

There were significant effects for state aggression, but the analyses were different from the previous ones. So those in the maximum/medium blood condition used weapon-based fighting style more often than the low/none blood condition. Effect size is small. I thought of an alternative explanation to this difference, spilt blood, as special effects, were rewarding visual stimuli and perhaps other effects might produce the same effect, say explosions. However, the authors argued the weapons effect that lead to an increase in state aggression. In combination with the violent content and the gore, it leads a higher increase. But they have not controlled for the weapons effect in the other factors, unless they did look at it and found that it has no effect. (or I’m not making any sense at all)

The results I mentioned were tested for the first hypothesis. The second hypothesis just combined all blood conditions in comparison to the no blood condition. I find it misleading because it simply state that the mere presence of blood is justifiable to say that there’s an effect, while further reading say that the low blood condition did not differ from the baseline measures. So why not a comparative analysis between high blood vs. low blood condition?

A limitation is the sample size in each condition is small and had to do some statistical corrections to get their results, why don’t they get more participants?

Study 2

Same as the first study, with exceptions: more female participants, non-gamers, two blood conditions (maximum vs. none), removed the state hostility factor (he should just remove the trait aggression factor), and added the cognitive priming testing as a means to test the third hypothesis.


Participants: 31 undergrads (16 male), age (average = 18.53, SD = 0.73), average play time per week = 5.18 (SD = 5.39), mostly Caucasian (84%) and first-years (87%).


Same as the first study, with the modification that there are only two blood condition and the addition of the Word Completion Task, so they fill in blanks in a word, and whether these words were related to aggression or not. Procedure is the same as the first study.

Trait aggression: main gender difference, male are more aggression than female. Nothing special.

Aggressive priming: Those in the maximum blood condition generated more aggressive thoughts than the no-blood condition. Effect size is medium. So, I guess that perhaps visual effects related to aggression (i.e. explosions and blood) might lead to greater priming of aggressive thoughts, but we need to conduct a study just to be sure.

This study provides an interesting perspective on how violence is quantified through the amount of blood spilled and how it show up in our psychological state. But, I wonder why they conducted two separate analyses to arrive at their conclusions, it bothers me a lot. There are some things to add in this experiment if someone were to conduct a replication study: varying the realism of blood, from puffs of red to anime-style blood; let the participants choose their own blood setting. This is an idea given by Christopher Ferguson and company. Finally, the context and realism of the situation, I recall that blood can be used as negative consequences that can lower aggression or that sanitizing the situation by removing the presence of blood can increase aggression (references needed). Anyways, the study is hardly generalizable to all bloody video games and other contextual factors need I mentioned before

Barlett, C. P., Harris, R. J., & Bruey, C. (2008). The effect of the amount of blood in a violent video game on aggression, hostility, and arousal. Journal of Experimental Social Psychology, 44(3), 539-546.


6 thoughts on “Bloodlust spell from bloody video games (Barlett et al., 2008)

  1. Pingback: The motivating role of violence in video games (Przybylski et al., 2009) « VG Researcher - Psychology

  2. Pingback: Technological advancement and violence exposure Level 2 (Barlett et al., 2008) « VG Researcher - Psychology

  3. Pingback: Knowledge about video games and its impact on priming aggressive cognition « VG Researcher – Psychology

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s