Avatar Race Affects White Video Game Players

[Bara_no_Hime]

The gist of the results of the study is that white players become more aggressive and have stronger negative biases against black people AFTER they play a violent game with a black avatar.

There was already a thread about this.

As I said there, I really don't think that Hot Sauce is a good gauge of anything other than what people like spicy food.

 

[Raikas]

What I mean is that the test doesn't really tell us about how that aggression plays out in the long term.

Does it matter in this context? This study doesn't appear to purport to talk about long-term changes to behaviour, but rather the immediate reaction to the colour of the avatar, so why would they want to expand the study along those lines?

Mainly because as it stands, what use do the results have?
I mean, if the effect of the aggression increase only lasts for, say, 15 minutes, unless you're sitting in the same general area as the player - and if you are that usually suggests a prior connection to the person in question that would most likely effect the chances of aggression towards you - It's not really going to have a chance to affect you. By the time Influenced McAgressive comes into contact with you, their 15 minute spurt of aggression is probably either up or almost up.

I took the point of the study to be about the subconscious racial stereotyping rather than an affect on aggression afterwards. I think we all know that people have subconscious stereotypes, but the study seemed to be primarily giving us some actual stats on it. The length of time that it affects a person seems like it's probably secondary (or at least the stuff of a differently-focused study) to the fact that it happens at all and at statistically significant rates.

 

[persephone]

I have a bachelor's in psychology, and after reading this article, I see some red flags right away.

One: Two studies do not news make. This experiment (and variations thereupon) would need to be repeated a solid fifty-plus times by different researchers, using different subjects, in different places, and the collected meta-data would need to be used to draw any kind of conclusion.

Two: Even if a conclusion is drawn, the article should be reporting that the data lends itself to the hypothesis, not that the results are in some way fact. Fortunately, while it seems the researchers are being good and restraining themselves to statements like "this suggests" rather than "this means," whoever wrote the article title and the big bold quote on the right didn't stick to that all-important attitude. Correlation does not causation make, and ruling out confounds in psychological experiments (i.e., getting rid of variables and factors you can't control) is notoriously difficult.

For example, it's well-established that:
(a) People react differently when they're being observed.
(b) People will often try to be "good participants" by doing what they think the researchers want rather than what they would naturally do.
As you can imagine, both of these confounds are pretty difficult to get around, and they can seriously skew your data.

Three: Where's the baseline here? Maybe the actual study goes into this (I didn't read that, only the article), but how is it meaningful to measure all of this without knowing how your participants felt about black people before your study? Did they believe in these stereotypes, consciously or unconsciously? What if for at least some of your participants, their attitude wouldn't have changed at all, or it only changed a little, and you don't know it because you have no baseline?

Four: They only tested white college students, presumably from a single university. Is this university known for being racist? Were these students all first year psychology students? A common problem in psychology is that too many experiments rely on making first year students participate in order to get their participants, thereby skewing the data -- not only are they not drawing from a varied pool, but they're not getting true volunteers. Where and how you obtain your participants can vastly affect your data, and I have to wonder if, for example, we're basically seeing the results of testing the Entitled Snob Man-Children sorority and fraternity.

Five: Why didn't they also test black students??? That seems like a HUGE oversight to me. What about Asian or other ethnicities? Why didn't they also use some baselines such as additional avatar ethnicities or even allowing participants to create their own avatars? A black avatar might've seemed a lot less (or more) stereotypical if they'd created him/her themselves. Did they line up genders between participants and avatars? Did they even use female avatars? How were the avatars dressed? Named? Those things can influence stereotypical perception.

TLR: Don't worry about this experiment, guys; it needs to be repeated a lot, varied up quite a bit, and done under many different circumstances before it could actually mean anything.

 

[Muspelheim]

Yes, well... No. I doubt my individual levels of racist tendencies is affected enough to mean anything at all.

But perhaps it's entirely true, and I'm a few percents more racist than I was before when I've had another go at GTA: San Andreas. But what does it mean, in the long run? Furthermore, is there enough games with black protagonists to create a problem? Except characters I've created in roleplaying games, I can only think of CJ from GTA. Admittadly, I was not kind to that character, but that was because he was such a terrible git.

And in all honestly, I doubt the effect, such as it is, is applicable to only video games. I imagine I might be slightly more racists towards Japanese after seeing a documentary of the warcrimes commited in Nanjing.
Or if we spin it, it'd not be unreasonable to assume that someone black might feel slightly more racists against whites after visiting a museum about the Atlantic slave trade.

It's temporary, if anything, and I doubt the slight boost effect to already present racist tendencies are that much of a problem, if they immediately fade.

What happens when I play a violent lizardman?

I do wonder the same thing...

http://i.imgur.com/AJHjjLX.jpg

Damn cowardly sugar-addicted caravan-dwelling furfags stealin' our jerbs... And everything else...

(I love that picture, he's got such an innocent, lovable "Have you heard the good news?" sort-of face...)

 

[Angelous Wang]

Could be true, could be bullshit. It' certainly bullshit with me.

My Trevor play time in GTA V is way more aggressive than my Franklin play time. With Trevor I'm blowing shit up or beating strangers to death every 20 seconds, my Franklin just drives around recklessly most of the time.

I'd say it's more of an individual thing really.

 

[Darkmantle]

I have a bachelor's in psychology, and after reading this article, I see some red flags right away.

One: Two studies do not news make. This experiment (and variations thereupon) would need to be repeated a solid fifty-plus times by different researchers, using different subjects, in different places, and the collected meta-data would need to be used to draw any kind of conclusion.

Two: Even if a conclusion is drawn, the article should be reporting that the data lends itself to the hypothesis, not that the results are in some way fact. Fortunately, while it seems the researchers are being good and restraining themselves to statements like "this suggests" rather than "this means," whoever wrote the article title and the big bold quote on the right didn't stick to that all-important attitude. Correlation does not causation make, and ruling out confounds in psychological experiments (i.e., getting rid of variables and factors you can't control) is notoriously difficult.

For example, it's well-established that:
(a) People react differently when they're being observed.
(b) People will often try to be "good participants" by doing what they think the researchers want rather than what they would naturally do.
As you can imagine, both of these confounds are pretty difficult to get around, and they can seriously skew your data.

Three: Where's the baseline here? Maybe the actual study goes into this (I didn't read that, only the article), but how is it meaningful to measure all of this without knowing how your participants felt about black people before your study? Did they believe in these stereotypes, consciously or unconsciously? What if for at least some of your participants, their attitude wouldn't have changed at all, or it only changed a little, and you don't know it because you have no baseline?

Four: They only tested white college students, presumably from a single university. Is this university known for being racist? Were these students all first year psychology students? A common problem in psychology is that too many experiments rely on making first year students participate in order to get their participants, thereby skewing the data -- not only are they not drawing from a varied pool, but they're not getting true volunteers. Where and how you obtain your participants can vastly affect your data, and I have to wonder if, for example, we're basically seeing the results of testing the Entitled Snob Man-Children sorority and fraternity.

Five: Why didn't they also test black students??? That seems like a HUGE oversight to me. What about Asian or other ethnicities? Why didn't they also use some baselines such as additional avatar ethnicities or even allowing participants to create their own avatars? A black avatar might've seemed a lot less (or more) stereotypical if they'd created him/her themselves. Did they line up genders between participants and avatars? Did they even use female avatars? How were the avatars dressed? Named? Those things can influence stereotypical perception.

TLR: Don't worry about this experiment, guys; it needs to be repeated a lot, varied up quite a bit, and done under many different circumstances before it could actually mean anything.

Well here I was scrolling down to say this and ya took the words right out of my mouth. I imagine most of those oversights are a result of working with such a small sample size. Perhaps they couldn't get enough minority students for it to be meaningful, larger versions of this study with more variety need to be done.

On a personal level, it bothers me when a study focuses exclusively on one race, then tries to make a statement about race. How can you say it's a white people thing, without first checking if it's a human thing.

 

[ThatDarnCoyote]

Huh.

I shall now dismiss this study as biased quackery because I do not like the results, despite knowing nothing more than the average layman about psychology or the reliability of the methods used in this study.

Well, there's always the standard disclaimer that applies to most studies like this one:

"Results valid for 18-year-old college freshmen in the United States. Other results may vary."

 

[SFMB]

Well, only time my avatar "race" affected me in violent way was when I played GTA: San Andreas. I rage-quited after the bicycle mission after being fed up with the gangland slang all of the characters spewed all the time. In other words, I could not relate to the characters at all. Actually, when I think of this now, it was not a race-issue, it was a cultural thing.
If we go to the "black people make you violent"-way, I must admit that when playing my Drow warlock in Neverwinter Nights 2, I went all evil.
Does that count?

 

[WouldYouKindly]

Quick question of methodology and incomplete data that the article gave us.

First, we need to rule out insane coincidence for these things. They should have been given the IAT before they played the game and then again after to detect any changes in the individual. The questions they asked shouldn't be asked twice because it's not likely to change just because of a game. Second, was there also a mirrored increase for the group that played a white character? It could be an increased association just due to the fact that it's fresh in their minds. That doesn't make a good headline though. Third, they should follow up to see if there's a lasting impression. I consider this unlikely. It might bring some subconscious racism to the surface for a while, but I don't think it'll stick around. Finally, the whole hot sauce thing. It might be that a difference in race brings more distance between the character and player and allows them to do meaner things without feeling guilty.

 

[Grouchy Imp_v1legacy]

The first study used 126 students, the second used 141. So between two studies the total number of test subjects did not exceed 270. This isn't enough of a subject base on which to build a solid theory - and while we're on the subject there was no mention of control groups that began the test with the camera pointed away from the players so that they couldn't tell which colour their avatars were.

Besides, all of this is academic. We all know that it's the videogames themselves that cause the violence, not the avatar in the videogame. C'mon people, we're alone here. We can admit it amongst ourselves. Whisper with me: It's the games.

Please note that the sample sizes are perfectly acceptable for the scope of this study. One does not build a theory based on a single or even a pair of experiments anyway, so I don't think there's a danger of that. This serves a simply a part of a whole, and people far too often dismiss single studies as meaningless because they don't serve as conclusive proof of anything. We work off of bodies of evidence over lots of experiments, but I see this over and over on The Escapist where someone jumps to the conclusion that the study is flawed because it a) does not have sufficient sample size, or b) the study is proposing a theory that it isn't.

Whilst I accept your point that one does not build a theory from only two studies, I still maintain that <150 people is a ridiculously small target group. I have spent the last fifteen years of my life as a statistical analyst and I guarantee you that this study would not hold up in a court of law due to it's limited scale. I know what you mean in that recent threads on this very site along the lines of 'X% of Americans believe this' or 'X% of Brits believe that' are always greeted with the argument that the sample size wasn't large enough, but this case taken on it's own merits has pretty much zero validity. Show me multiple studies taken in multiple countries encompassing thousands of subjects - then we have a subject for debate. Until then these findings are meaningless.

 

[frizzlebyte]

Oh no, people assign negative stereotypes to things different from themselves. If only we didn't know this by looking at all of human history. I don't know why people try to argue this point. To make themselves seem 'progressive' by claiming to have no bias against anyone at any time, I guess?

I think that's close. The thing that a lot of people can't stand is the idea that they might actually have racist/sexist/irrationally biased tendencies, so they try to say it isn't happening.

From where I sit, the first step to conquering these biases is to admit that I have them, and then reflect on why that is. By understanding my biases, I can say that I've grown past a ton of them. It's just a shame that so many of the PC brigade don't do the same. It's a great growth experience.

 

[Gorrath]

The first study used 126 students, the second used 141. So between two studies the total number of test subjects did not exceed 270. This isn't enough of a subject base on which to build a solid theory - and while we're on the subject there was no mention of control groups that began the test with the camera pointed away from the players so that they couldn't tell which colour their avatars were.

Besides, all of this is academic. We all know that it's the videogames themselves that cause the violence, not the avatar in the videogame. C'mon people, we're alone here. We can admit it amongst ourselves. Whisper with me: It's the games.

Please note that the sample sizes are perfectly acceptable for the scope of this study. One does not build a theory based on a single or even a pair of experiments anyway, so I don't think there's a danger of that. This serves a simply a part of a whole, and people far too often dismiss single studies as meaningless because they don't serve as conclusive proof of anything. We work off of bodies of evidence over lots of experiments, but I see this over and over on The Escapist where someone jumps to the conclusion that the study is flawed because it a) does not have sufficient sample size, or b) the study is proposing a theory that it isn't.

Whilst I accept your point that one does not build a theory from only two studies, I still maintain that <150 people is a ridiculously small target group. I have spent the last fifteen years of my life as a statistical analyst and I guarantee you that this study would not hold up in a court of law due to it's limited scale. I know what you mean in that recent threads on this very site along the lines of 'X% of Americans believe this' or 'X% of Brits believe that' are always greeted with the argument that the sample size wasn't large enough, but this case taken on it's own merits has pretty much zero validity. Show me multiple studies taken in multiple countries encompassing thousands of subjects - then we have a subject for debate. Until then these findings are meaningless.

Indeed, they wouldn't hold up in court, but they aren't meant to either. The sample sizes are acceptable for the scope of this study, not for determining a point of law or generating a theory. But that doesn't make the findings meaningless, it just makes the work incomplete. Multiple studies conducted by different scientists in different places is what needs to happen, but until it does we neither treat this data as the end-all-be-all, nor do we discount it as meaningless. We simply look at the study and it's findings, make sure the study was properly conducted (and again, for this study the sample sizes are fine, Universities give guidance on these sorts of things and the general rule seems to hover around 80-120), and wait for more studies to be conducted before drawing a conclusion/forming a theory.

 

[Saetha]

What happens when I play a violent lizardman?

Damn cold-blooded bipeds. Coming over here, stealing our Skoomba.

Umm, pardon me, sir, but I don't think you realize how problematic statements like these are against Argonian kind. Need I remind you that skooma addiction is a HUGE problem among Argonians, that many Argonians get hooked on Hist sap AT BIRTH, that racism against Argonian immigrants is still alive and well in places like Skyrim and Cyrodiil? It's comments like these that send us back to 3rd Era Morrowind.

Yeah, check your privilege, human scum.

As for the OP... didn't we already have a thread for this? I could've sworn we did. Anyway, yeah, I wouldn't exactly trust any of this until follow-up research is done. I mean, the whole study just kinda seems pointless to me, without any data to compare how minorities react or anything.

 

[Grouchy Imp_v1legacy]

The first study used 126 students, the second used 141. So between two studies the total number of test subjects did not exceed 270. This isn't enough of a subject base on which to build a solid theory - and while we're on the subject there was no mention of control groups that began the test with the camera pointed away from the players so that they couldn't tell which colour their avatars were.

Besides, all of this is academic. We all know that it's the videogames themselves that cause the violence, not the avatar in the videogame. C'mon people, we're alone here. We can admit it amongst ourselves. Whisper with me: It's the games.

Please note that the sample sizes are perfectly acceptable for the scope of this study. One does not build a theory based on a single or even a pair of experiments anyway, so I don't think there's a danger of that. This serves a simply a part of a whole, and people far too often dismiss single studies as meaningless because they don't serve as conclusive proof of anything. We work off of bodies of evidence over lots of experiments, but I see this over and over on The Escapist where someone jumps to the conclusion that the study is flawed because it a) does not have sufficient sample size, or b) the study is proposing a theory that it isn't.

Whilst I accept your point that one does not build a theory from only two studies, I still maintain that <150 people is a ridiculously small target group. I have spent the last fifteen years of my life as a statistical analyst and I guarantee you that this study would not hold up in a court of law due to it's limited scale. I know what you mean in that recent threads on this very site along the lines of 'X% of Americans believe this' or 'X% of Brits believe that' are always greeted with the argument that the sample size wasn't large enough, but this case taken on it's own merits has pretty much zero validity. Show me multiple studies taken in multiple countries encompassing thousands of subjects - then we have a subject for debate. Until then these findings are meaningless.

Indeed, they wouldn't hold up in court, but they aren't meant to either. The sample sizes are acceptable for the scope of this study, not for determining a point of law or generating a theory. But that doesn't make the findings meaningless, it just makes the work incomplete. Multiple studies conducted by different scientists in different places is what needs to happen, but until it does we neither treat this data as the end-all-be-all, nor do we discount it as meaningless. We simply look at the study and it's findings, make sure the study was properly conducted (and again, for this study the sample sizes are fine, Universities give guidance on these sorts of things and the general rule seems to hover around 80-120), and wait for more studies to be conducted before drawing a conclusion/forming a theory.

Ok, perhaps "meaningless" was too strong. But "next to meaningless" would still apply - for two reasons:

1) The lack of a control group. The absence of a control experiment invalidates all studies based on empirical evidence. Without a control group any analysis is flawed.

2) I'd dispute the claim that universities give guidance that 120 us an ideal test size. Given the current population of the US, a sample of such a size represents only 0.00003% of the nation as a whole.

Please note that I'm not trying to invalidate this study, merely raising the issue that it is massively incomplete. Like I said, show me comparisons with a control group, show me a larger test size (in the region of 10,000 people - still only 0.003% of the US) and show me a larger age range - then we will has a viable study. Until then, no dice.

 

[Azure23]

Could be true, could be bullshit. It' certainly bullshit with me.

My Trevor play time in GTA V is way more aggressive than my Franklin play time. With Trevor I'm blowing shit up or beating strangers to death every 20 seconds, my Franklin just drives around recklessly most of the time.

I'd say it's more of an individual thing really.

Same here, all my best rampages (including a twenty-two minute running firefight with three swat teams which sadly will never be immortalized) are with Trevor, just makes more sense. Franklin is the level headed, voice of reason guy. What game were they testing this on anyway? Postal?

EDIT: oh, Saints Row 2, nice.