PDA

View Full Version : Violent video games desensitize people to real-life violence

arctangent
07-27-2006, 01:49 PM
Iowa State psychologists produce first study on violence desensitization from video games (http://www.eurekalert.org/pub_releases/2006-07/isu-isp072706.php)

Research led by a pair of Iowa State University psychologists has proven for the first time that exposure to violent video games can desensitize individuals to real-life violence.

"The results demonstrate that playing violent video games, even for just 20 minutes, can cause people to become less physiologically aroused by real violence," said Carnagey. "Participants randomly assigned to play a violent video game had relatively lower heart rates and galvanic skin responses while watching footage of people being beaten, stabbed and shot than did those randomly assigned to play nonviolent video games.

"It appears that individuals who play violent video games habituate or 'get used to' all the violence and eventually become physiologically numb to it."

Participants in the violent versus non-violent games conditions did not differ in heart rate or skin response at the beginning of the study, or immediately after playing their assigned game. However, their physiological reactions to the scenes of real violence did differ significantly, a result of having just played a violent or a non-violent game. The researchers also controlled for trait aggression and preference for violent video games.

They conclude that the existing video game rating system, the content of much entertainment media, and the marketing of those media combine to produce "a powerful desensitization intervention on a global level."

"It (marketing of video game media) initially is packaged in ways that are not too threatening, with cute cartoon-like characters, a total absence of blood and gore, and other features that make the overall experience a pleasant one," said Anderson. "That arouses positive emotional reactions that are incongruent with normal negative reactions to violence. Older children consume increasingly threatening and realistic violence, but the increases are gradual and always in a way that is fun.

"In short, the modern entertainment media landscape could accurately be described as an effective systematic violence desensitization tool," he said. "Whether modern societies want this to continue is largely a public policy question, not an exclusively scientific one."

LesJarvis
07-27-2006, 01:52 PM
Excellent, I can finally justify the shooting spree I've been planning.

Edit: while this is interesting and all, it doesn't really show that people who play violent videogames are more like to actually commit violent acts. Of course encouraging your study participants to go and beat the hell out of someone after playing violent games is probably unethical.

Gary Whitta
07-27-2006, 02:03 PM
This gets discussed in the new PC Gamer podcast that comes out later today, FYI. I believe the word I used to summarize this study's conclusions was "bollocks".

Ben Sones
07-27-2006, 02:06 PM
Galvanic skin responses are the only things keeping us all from beating each other to death.

SpoofyChop
07-27-2006, 02:08 PM
while this is interesting and all, it doesn't really show that people who play violent videogames are more like to actually commit violent acts.

Excellent point. Personally I believe that the entire global media system (news, entertainment, whatever) are all probably pretty effective at desensitizing people to any number of things...violence, suffering, poverty, political instability, etc.

What I mean by this is that if you never personally experience these things but you see them on TV all the time I think it's at least possible that you stop reacting to the images of these things as if they are real and you start reacting to them as if they are merely representations/stories. The only way to know what your real reaction to these situations would be however would be to actually be physically present during a violent crime, in a poverty stricken country, in a war torn region, or a political protest.

I think the chances are that the same tests conducted during these real life scenarios with the people physically present would not show the same "blase" reaction to these situations.

I'd also like to see a test in which somebody is exposed to violent images from news reports over and over and see if they have the same response the 10th, 20th or 100th time they see a newscast of violent images. My guess is that even just watching the newscast will desensitize you.

MatthewF
07-27-2006, 02:11 PM
Research led by a pair of Iowa State University psychologists has proven for the first time that exposure to violent video games can desensitize individuals to real-life violence.

I do not think that word means what they think it means.

LesJarvis
07-27-2006, 02:11 PM

ElGuapo
07-27-2006, 02:16 PM
Here's my study:

Control Group:
1. Plays chess with a researcher
2. Is taken to an inner city, given a baseball bat, and asked to kill a crack whore

Experiment group
1. Plays Grand Theft Auto with researcher and kills people on the screen
2. Is taken to an inner city, given a baseball bat, and asked to kill a crack whore

Or

Control group
1. Allow them to read some Emily Bronte in a comfortable chair
2. Slap them hard in the face out of the blue and see what they do

Experiment group
1. Plays Battlefield 2 for an hour
2. Slap them hard in the face out of the blue and see what they do

There you go, simple.

Zylon
07-27-2006, 02:17 PM
For their next trick, these researchers will prove that the second cup of coffee never tastes as good as the first.

MikeSofaer
07-27-2006, 02:26 PM
Doctors are as a whole highly phisically desentitized to other people's physical trauma, too.

metta
07-27-2006, 02:27 PM
I don't believe that playing violent games makes people violent, but I do think it gives them a context for their own violence.

To the people who claim playing violent games is completely neutral, I ask do you ever play or listen to mellow music to relax or get to sleep?

Does it work?

Jake Plane
07-27-2006, 02:28 PM
Serenity now, serenity now, serenity now...

stusser
07-27-2006, 02:29 PM
Read the article. The "real" violence they were exposed to was actually old episodes of Magnum P.I. I would argue that headshotting an alien in a videogame, being interactive, is less detached and more "real" than watching it happen on a television screen without my involvement or direction. Basically the study is bullshit.

Mehrunes
07-27-2006, 02:30 PM
Wait, I think something needs to be clarified...

After playing a video game, a second set of five-minute heart rate and skin response measurements were taken. Participants were then asked to watch a 10-minute videotape of actual violent episodes taken from TV programs and commercially-released films in the following four contexts: courtroom outbursts, police confrontations, shootings and prison fights. Heart rate and skin response were monitored throughout the viewing.

When viewing real violence, participants who had played a violent video game experienced skin response measurements significantly lower than those who had played a non-violent video game. The participants in the violent video game group also had lower heart rates while viewing the real-life violence compared to the nonviolent video game group.

At first I got the impression that the individuals were shown actual real violence. So what's the deal, are they being shown actual real violence, as implied by the second paragraph, or are they being shown fake TV "real violence" as implied in the first paragraph?

EDIT: ONE MINUTE LATE. Goddamn you, stusser!

Other Brendan
07-27-2006, 02:34 PM
I think one problem here is that the subjects were shown video of real-life violence. So really their evidence just shows that after 20 minutes of violent games, seeing something violent on the TV won't get as big a rise out of them physiologically as if they'd been playing non-violent games. There's a big difference between real-life violence and videotaped footage of it, just like there's a big difference between skydiving and videotaped footage of that. I think ElGuapo's experiments would be much more useful if we're studying connections between game violence and actual real-life violence.

EDIT: Also, the .pdf of the report I read said the real-life violence was footage of courtroom outbursts, prison stabbings, and other unpleasantness, not fake movie violence.

http://www.gamespot.com/news/6154696.html

EDIT 2: A quote from the researchers' study:
These were actual violent episodes (not Hollywood reproductions) selected from TV programs and commercially released films. In one scene, for example, two prisoners repeatedly stab another prisoner.

Morkilus
07-27-2006, 02:34 PM
According to the study, galvanic skin response actually decreases while watching "real violence" (read: videotaped) after playing violent games such as this:http://i28.photobucket.com/albums/c218/Morkilus/duke_nukem.gif

If that makes any sense to anybody, I'd like to hear it. The graphs in paper are pretty telling. Also, there was no control videos, with fluffy bunnies and pictures of Hillary Clinton.

skedastic
07-27-2006, 02:37 PM
The psych guys are barking up the wrong tree on this whole issue. Correlating attitudes towards violence with whether or not one has just played a violent video game tells us nothing about the underlying issue: whether violent video games cause violent behavior. Studies correlating video game playing habits with actual violence are somewhat better but still inadequate to speak to the underlying issue.

Other Brendan
07-27-2006, 02:57 PM
Correlating attitudes towards violence with whether or not one has just played a violent video game tells us nothing about the underlying issue: whether violent video games cause violent behavior.

I think they're just trying to take baby steps to get where they want to go. Like instead of proving that going to the beach causes cancer, first prove that going to the beach exposes you to more sun. Then prove that more sun = more UV rays. Then prove that too many UV rays can affect the way your body produces new cells. Then prove that an alteration to the way your body produces new cells can cause cancer. My lack of actual, reliable knowledge on how exposure to the sun gives you cancer aside, this is basically what the researchers in question are doing.

Tom McNamara
07-27-2006, 03:12 PM
The thing that gets me is that people can (and do) say the same thing about watching television. And from what it sounds like, the researchers don't examine how long this desensitization lasts. I think there's a psychological elasticity that's being overlooked, and I would also consider the influence of primal survival instinct; this "desensitization" might be nothing more than a subconscious trauma dampener that fades as the stimulus disappears.

Chris Woods
07-27-2006, 03:22 PM
This is a good study. Trying to attack it by saying "it doesn't prove violent video games leads to violence" is just a strawman argument. The study doesn't claim this at all. It claims to demonstrate that in this study people who played violent video games had less of a physical reaction to viewing violent scenes then those who didn't and that's all.

Research led by a pair of Iowa State University psychologists has proven for the first time that exposure to violent video games can desensitize individuals to real-life violence.

I do not think that word means what they think it means.

If by "They" you mean the website linked, sure. If by "They" you mean the researchers, no. The study doesn't claim to have proved anything. The strongest term they used is "demonstrates".

Don't be so gunshy. I don't think desensitization to violence is real news to people who play violent games.

Chris Woods

skedastic
07-27-2006, 03:23 PM
I think they're just trying to take baby steps to get where they want to go.

Which is, in general, a sound methodological approach. And it isn't being followed here: the whole mechanism is fabricated from precious little evidence, and no compelling evidence on causation. See, for example, this ridiculously biased meta-analysis (http://www.psychology.iastate.edu/faculty/caa/abstracts/2000-2004/01AB.pdf). There seems to be a small industry of folks churning out crappy papers doing a really bad job controlling for confounding influences and then leaping to the conclusion that correlation equals causation.

bigdruid
07-27-2006, 03:32 PM
Allow me to put my own spin on this:

"Researchers show that video game players are less aroused by real violence"

Being "desensitized" to violence doesn't mean "more likely to commit violence" - it means your basic, monkey reaction to violence (fight or flight) is diminished. I think that's a great outcome, and we as a society should require all citizens to play GTA for a few hours every day as a result.

Noise
07-27-2006, 03:38 PM
Allow me to put my own spin on this:

"Researchers show that video game players are less aroused by real violence"

Being "desensitized" to violence doesn't mean "more likely to commit violence" - it means your basic, monkey reaction to violence (fight or flight) is diminished. I think that's a great outcome, and we as a society should require all citizens to play GTA for a few hours every day as a result.

Or it means people are less likely to care about atrocities.

arctangent
07-27-2006, 03:41 PM
The real question is, who paid for the study?

Finally, although our main focus has been on
unintended desensitization and helping effects of violent
video games, a better understanding of video game effects
can be put to good use in other contexts. For example, if
violent video games designed primarily to entertain are
good at producing physiological desensitization, then it
should be feasible to design games to produce such
desensitization in desired populations and contexts. Can
we make better combat soldiers by desensitizing them to
some of the sights and sounds of combat? Can we help
medical students become comfortable with the types of
physical and emotional trauma they will experience in
emergency rooms? Can we use video games to systemati-
cally desensitize individuals who need to be desensitized
to specific stimuli that cause them problems (e.g., auto
accident victims afraid of riding or driving again)?

skedastic
07-27-2006, 03:42 PM
This is a good study. Trying to attack it by saying "it doesn't prove violent video games leads to violence" is just a strawman argument. The study doesn't claim this at all.

Sure it does. It also claims that violent video games cause less sympathy for victims of violence. The authors conclude:

In short, the modern entertainment media
landscape could accurately be described as an eVective systematic
violence desensitization tool. Whether modern societies
want this to continue is largely a public policy
question, not an exclusively scientic one (Anderson et al.,
2003; Gentile & Anderson, 2006).

which is a mealy-mouthed way of advocating more government restrictions on media content. The cites, incidentally, are to other papers by Craig A. Anderson. Craig A. Anderson just loves to cite Craig A. Anderson as evidence that what Craig A. Anderson is saying is valid.

stusser
07-27-2006, 04:17 PM
Again, the study actually tests whether violent videogames desensitize subjects from other forms of violent media.

Anyone who's ever gotten into a fight or even witnessed a real fight knows it's a very, very different experience than watching boxing on TV. That overpoweringly awesome feeling of dominating someone as you fuck them up bad or conversely the helpless feeling as you try to protect your face and groin from getting splattered, or when your face is pushed in the dirt, busting open your lip, tasting the grass and your own blood... you just don't get that from videogames, TV, movies. They are an entirely different experience. Not just varying degrees of the same thing, not just more visceral, different.

Tom McNamara
07-27-2006, 04:20 PM
Back in my day, we paid someone a nickel to get punched in the junk, and we liked it that way.

...Oh, and in other news, from Business Week: "Doctors are drawing on video-game technology to treat post-traumatic stress disorder among Iraq war veterans."

DustyTheHamster
07-27-2006, 04:57 PM
I did a study last night where I proved that by watching the news I desensitized my self to the stories they were reporting. So how did they find out that people got desensitized by playing violent video games?

did they have them play for a bit then have some random person get their arm lopped off by a piece of machinerary? or did they bring in a "pow" and have them executed in front of the panel? How was the virtual and nonrealistic violence portrayed in a video game any where close to the realism achived by movies or TV shows or even the news? Seen any footage from iraq recently?

using anecdotal evidence I prove that just by opening your eyes and/or listening you are desensitizing yourself.

Hans Lauring
07-27-2006, 05:14 PM
After game play, a second set of 5-min HR and GSR
measurements were taken. Next, participants watched a
10-min videotape of real violence in four contexts: courtroom
outbursts, police confrontations, shootings, and
prison fights. These were actual violent episodes (not Hollywood
reproductions) selected from TV programs and
commercially released films. In one scene, for example,
two prisoners repeatedly stab another prisoner.

Which tv and commercially released films contains real footage of prison stabbings?

Moggraider
07-27-2006, 06:33 PM
Come on guys, of course video games desensitize us to violence. Isn't that as clear as day? You don't need a study to tell you that. I'm only 20 and I've known I've been desensitized for about 10 years. If you can sit there and pound a guy into the ground in a videogame without thinking twice about it, you're desensitized. Period. Why do any of us have trouble admitting this? It's not like the study is saying we're more likely to commit violent acts. It's just saying we respond less to it, which is definitely true. TV, the Internet, and videogames all desensitize us constantly. Hell, I could eat lunch in front of my computer with Tubgirl in fullscreen on my monitor.

LesJarvis
07-27-2006, 06:35 PM
Hell, I could eat lunch in front of my computer with Tubgirl in fullscreen on my monitor.

That, my friend, is the difference between you and me.

PeterGinsberg
07-27-2006, 06:41 PM
Come on guys, of course video games desensitize us to violence. Isn't that as clear as day? You don't need a study to tell you that. I'm only 20 and I've known I've been desensitized for about 10 years. If you can sit there and pound a guy into the ground in a videogame without thinking twice about it, you're desensitized. Period. Why do any of us have trouble admitting this? It's not like the study is saying we're more likely to commit violent acts. It's just saying we respond less to it, which is definitely true. TV, the Internet, and videogames all desensitize us constantly. Hell, I could eat lunch in front of my computer with Tubgirl in fullscreen on my monitor.

The question, I think, is would you still be able to eat your lunch if you were in the bathroom with tubgirl?

What this study shows (regardless of the videos authenticity) is that violent video games desensitize viewers to violent videos. Would they still react differently from a control group if there was a real stabbing in front of them?

I've spent plenty of quality time with violent video games, and can pretty much ignore violence of any sort in in games, videos, etc. But what real life violence I've seen usually leaves me running around like a little girl. Yes yes, ancedote and all that, but this study doesn't seem to be answering any useful questions.

bago
07-27-2006, 06:53 PM
Cooking shows desensitize us to brutal acts of whipping, knife-work and butchery!

Justin Fletcher
07-27-2006, 08:34 PM
I'm no scientist, but I'd bet that any activity that gets the adrenaline pumping would "desensitize" a person in the same way. Time for more research, chaps.

Moggraider
07-27-2006, 08:53 PM
That doesn't make any sense, Fletcher. For example, looking at a girl while running makes you consider her more attractive than if you were standing still. This is similar to the suspension bridge effect - standing on a suspension bridge looking at a woman makes you give her higher attractiveness ratings (the suspension bridge is a dangerous situation) than if you were standing still in an office. That means that adrenaline pumping means increased arousal, not decreased. Desensitization is lack of increased arousal, just like porn makes you find your girlfriend less interesting. A slowed heartbeat means fascination or interest, and that's true from the time we are babies onward. Test it yourself - get on an elliptical exercise machine, and see how fast your heartrate is going. Now, continue on the elliptical, but start watching TV. Your heart rate goes down.

I'm not saying I'm anti-videogames, but as a psych major I do have to admit that videogames desensitize us to violence and make kids more likely to take violent courses of action, due to modeling.

Angrycoder
07-27-2006, 09:31 PM
I am going to kill that mother fucking reasearcher... once this round of counterstrike is over.

extarbags
07-27-2006, 09:40 PM
I'm not saying I'm anti-videogames, but as a psych major I do have to admit that videogames desensitize us to violence and make kids more likely to take violent courses of action, due to modeling.

Maybe consider an alternate career path, then. If you honestly consider yourself desensitized to violence and more likely to act violently because you feel no remorse at the thought of killing an unnamed pedestrian in GTA, the real psychological problem lies in your own inability to distinguish between pixels and people.

extarbags
07-27-2006, 09:42 PM
Desensitization is lack of increased arousal, just like porn makes you find your girlfriend less interesting.

And I just saw this part. Dude, what the fuck is wrong with you? You're betraying all kinds of psychological quirks of your own and projecting them onto everyone else. I'm half expecting your next post to be "Ok, you know how when you see your grandmother's socks in the dirty laundry, the natural response is to masturbate?"

skedastic
07-27-2006, 09:49 PM
I'm not saying I'm anti-videogames, but as a psych major I do have to admit that videogames desensitize us to violence and make kids more likely to take violent courses of action, due to modeling.

There is no compelling evidence that violent video games cause violent behavior.

Desensitization is lack of increased arousal, just like porn makes you find your girlfriend less interesting.

What the hell kind of porn are you watching?

extarbags
07-27-2006, 10:04 PM
In seriousness, here's the crux of this issue, which this and every study like it conveniently leave out: real violence and fake violence are fundamentally different. See, people who get up in arms about this want to make it about the violence itself, because that's all there is in a game or movie. But real-world violence is really about the consequences of that violence; consequences ranging from the immediate (the pain the subject of the violence feels) to the long-term (the effect on the victim's psyche and family, and on society as a whole).

In a game or movie, there are no real consequences. At best, the fictional consequences on the fictional victim/family/society might be examined, which happens more often in films but sometimes in games too. If a character in a game is well-developed and likeable, I'll often abstain from violence against them, if given the choice, and feel bad after committing violence against them if not. I know everyone isn't like that, but either way, anyone who can tell the difference between a real person and a video game character, and who isn't a psychopath, should be perfectly capable of distinguishing between acts of violence themselves, which are harmless in a vacuum such as a video game, and the consequences that give those acts meaning.

Or, to put it succinctly, violence is more than just blood and gore.

algahar
07-27-2006, 10:18 PM
i would love to.... second thought let's not, but what i am trying to say is, i think i will pass out even looking at a real bloody corpse. more than 10 years of Wolfenstein, DOOM, Half-life, CounterStrike etc could never have prepared me for the real deal. sorry.

not to mention shooting something alive, let alone shooting a person!

if they really wanna do this test, they should ship the lot of gamers to a friggin "Battle Royale" (http://www.amazon.com/gp/product/B000BBYMA2/sr=1-2/qid=1154063867/ref=pd_bbs_2/002-0319739-3253614?ie=UTF8&s=dvd)! and see how they react.

stusser
07-27-2006, 10:33 PM
It isn't about consequences, those come later. Not that it has any import to this discussion, but consequences stop psychos from going on killing sprees, not normal people. Normal people have empathy. I wouldn't kill anyone even if I could get away with it. Hell, I inadvertently shoved a chick in the subway this morning and felt bad about it.

Real world violence is about totally fucking some dude up or vice versa, which as I said earlier are inherently different experiences than watching a TV show.

Moggraider
07-27-2006, 11:54 PM
Maybe consider an alternate career path, then. If you honestly consider yourself desensitized to violence and more likely to act violently because you feel no remorse at the thought of killing an unnamed pedestrian in GTA, the real psychological problem lies in your own inability to distinguish between pixels and people.

Uh, thanks for ignoring my entire argument and what the psychological and anatomical literature says, then trying to redefine what it is I said. And I'm going to be a lawyer, not a psychologist (see you in court, pig lololol). Adults may or may not be more likely to act violently because of violent videogames (chances are they are more likely), but children are proven to be more likely to do so, which is what we should be worried about.

There is no compelling evidence that violent video games cause violent behavior.

Cause? No. Make more likely? Yes.

What the hell kind of porn are you watching?

Um, it's proven that men who watch a lot of porn rate their others as relatively less attractive on a scale of 1 to 10 than men who do not.

skedastic
07-28-2006, 12:14 AM
Cause? No. Make more likely? Yes.

"Make more likely" is what we mean, in this context, by "cause." And no, there is no compelling evidence that violent video games cause violent behavior.

Moggraider
07-28-2006, 12:16 AM
Yes there is.

Bill Dungsroman
07-28-2006, 12:16 AM
Um, it's proven that men who watch a lot of porn rate their others as less attractive on a scale of 1 to 10 than men who do not.
Um, don't use the word "proven" so freely.

skedastic
07-28-2006, 12:18 AM
Um, it's proven that men who watch a lot of porn rate their others as less attractive on a scale of 1 to 10 than men who do not.

Correlation, causation. I just know there's some relevant phrase I should bring up that uses both these words. (Not to mention that, "Desensitization is lack of increased arousal, just like porn makes you find your girlfriend less interesting" is a rather bizarre way of phrasing that finding.)

Moggraider
07-28-2006, 12:19 AM
What's the problem? The average is lower for one group than the other. That's proof that on average we'll see a relationship there. Social science is not freaking chemistry; you're never going to get a dead-on 1:1 relationship.

skedastic: I can see how you misinterpreted what I said. But there's no other viable explanation for the relationship.

Anyway, I can see you guys are upset by the original study. I know you all want to believe that it is wrong, because all of us are heavy videogame players, and some of us probably even have kids who we allow or plan to allow to play games. Still, it would be foolish and irresponsible to deceive ourselves and say that videogames don't affect the way people think. The key word here is simply responsibility.

Hans Lauring
07-28-2006, 01:20 AM
It seems more like people are upset, that you use the word proof as if you know what it means - maybe it's time for you to go back to your schoolbooks and post the studies they refer to, when they seem to make you believe it's so clearcut.

Just one silly case in point, the argument could easily be made that men with ugly girlfriends are more likely to watch porn. As Skedastic said: Correlation vs. causation.

Moggraider
07-28-2006, 01:34 AM
Attractiveness levels were controlled for in the study; I was talking about relative rates of devaluation. And even if I was wrong about the porn study, that wouldn't change anything about the topic itself. I just mentioned the study as an example of one way desensitization occurs. Aside from affecting perceptions of mate value, porn still has desensitizing effects in terms of the false expectations it encourages.

As for the use of the word "proof," well, things are being proven as far as psychological studies can possibly take them. Like I said, they can't be akin to laws of physics.

(An aside: in other studies on mate value, men with the higher numbers of sexual partners devalue their mates the most, so your alternative explanation for the porn study doesn't make any sense. See: Haselton and Buss, 2001.

Oh, and here's a summary of a study other than the one I mentioned that also proves your alternate explanation is wrong: http://failedtherorschachtest.homestead.com/files/Picture1.jpg).

Hans Lauring
07-28-2006, 01:44 AM
Attractiveness levels were controlled for in the study;

Cool, can I have that scientific standard for attractiveness that they used to do this? It would make so many discussions with my buddies go away.

I was talking about relative rates of devaluation. And even if I was wrong about the porn study, that wouldn't change anything about the topic itself. I just mentioned the study as an example of one way desensitization occurs. Aside from affecting perceptions of mate value, porn still has desensitizing effects in terms of the false expectations it encourages.

Again not proven. Recent studies in this country where parents and educators have been alarmed at the level of porn being seen by teens (as an example I think the rates were that something like 95% of all 15-year old had watched scenes of anal sex) showed that the teens were perfectly capable of distinguishing between on screen sex and sex with real partners. Allthough everyone agrees that porn was getting mkore hardcore and more teens were watching it, the surprise findings was, that the actual sex they engaged in, where pretty much the same fumbling stuff my generation got up to.

And your bad porn studies does invalidate the rest of your claims. At least your tendency to accept "proof" from your textbooks/teachers without questioning them or their agenda shows that we certainly shouldn't accept your "because I say so" argumentation without demanding that you actually source your claims.

Untill you do every case of "proof" and "proven" in your post should be replaced with "some experts and textbooks believes..."

Moggraider
07-28-2006, 01:50 AM
Attractiveness is typically standardized for in studies by aggregating raters' scores. After all, beauty is only what we say it is. Psych departments pay people to come in and rate pictures of people for studies all the time.

Can you cite the study you mention, please? That way I'll be able to take it to my profs with their ANTI-PORN AGENDAS and ask them to tell me what to think.

Hans Lauring
07-28-2006, 02:14 AM
Sorry, no. You're the one claiming to hold proof - I just referred to one study and even if I had the time to find it for you, I'd still have to translate it.
You go source your claims and we can talk about me doing the same.

Moggraider
07-28-2006, 07:19 AM
I already sourced two studies and provided results from them. Look up. I am shocked, shocked, that you can't hold yourself to the very standards that you try to impose on others! The study you're mentioning doesn't even sound like it's an experiment. It's an observational study/survey, so it can't disprove a thing.

RE: videogames, there is certainly no study that shows videogames don't increase likelihood of violence. All studies point to the same finding, in different degrees depending on how well they are designed. Everyone would be foolish to ignore all these findings and wildlly clamor for "proof." One study I've read about had college students play Tetris or Doom, then gave them the opportunity to administer a mild shock to a confederate. The Doom players gave longer shocks and resorted to them more quickly.

Look at these:

Anderson, C.A., & Dill, K.E. (2000). Video games and aggressive thoughts, feelings, and
behavior in the laboratory and in life. Journal of Personality and Social Psychology, 78, 772-790.
Bartholow, B.D., & Anderson, C.A. (2002). Effects of violent video games on aggressive
behavior: Potential sex differences. Journal of Experimental Social Psychology, 38, 283-290.

Now, if you can post that study you were talking about, that would be great, please, or you're full of crap.

extarbags
07-28-2006, 07:30 AM
Dude, can we pretty please edit that "Craziest poster on QT3 thread?" I think I want to change my vote.

Moggraider
07-28-2006, 07:53 AM
You should go nurse that groin now while we talk about things you can't seem to understand.

ElGuapo
07-28-2006, 08:02 AM
Don't take this as scientific fact, but just as interesting policy, which I thought it was. It certainly may be limited to an individual, or it may be a general institutional policy.

A friend of my brother just back a bit ago from Iraq. He wasn't in the rear with the gear either, he was a door smashing room clearing marine. Anyway, he has, supposedly some PTSS. The few times I've met him I haven't noticed it. He seems really nice and quiet, respectful, and doesn't drink much at all. Anyway, like about every male his age, he likes videogames. The doctors all know this about the returning marines.

His doctor told him not to play "realistic" shooter games, but fantasy stuff with non military themes is ok. So I guess something like Unreal Tournament would be ok, Battlefield 2 would not. The few times I played with him, we played non realistic games like Crimson Skies and sports games until he left. We did play Halo 2 a bit, but I was a bit wary with that one.

Interesting, no?

extarbags
07-28-2006, 08:06 AM
But that's a little different. I'd think that advice is meant to steer him away from games that might remind him of his time in Iraq. In other words, it's exactly the opposite; real-life violence may have oversensitized him to fake violence, which I have no trouble believing.

fuzzyslug
07-28-2006, 08:09 AM
I already sourced two studies and provided results from them. Look up. I am shocked, shocked, that you can't hold yourself to the very standards that you try to impose on others! The study you're mentioning doesn't even sound like it's an experiment. It's an observational study/survey, so it can't disprove a thing.

RE: videogames, there is certainly no study that shows videogames don't increase likelihood of violence. All studies point to the same finding, in different degrees depending on how well they are designed. Everyone would be foolish to ignore all these findings and wildlly clamor for "proof." One study I've read about had college students play Tetris or Doom, then gave them the opportunity to administer a mild shock to a confederate. The Doom players gave longer shocks and resorted to them more quickly.

Look at these:

Anderson, C.A., & Dill, K.E. (2000). Video games and aggressive thoughts, feelings, and
behavior in the laboratory and in life. Journal of Personality and Social Psychology, 78, 772-790.
Bartholow, B.D., & Anderson, C.A. (2002). Effects of violent video games on aggressive
behavior: Potential sex differences. Journal of Experimental Social Psychology, 38, 283-290.

Can someone slow him down by finding the crime statistics for the last 10 years? Video games are getting more popular; violent videogames are leading the way. Violent crime is down - way down. Explain.

Your concept of proof is certainly different from mine.

fuzzyslug
07-28-2006, 08:13 AM
In seriousness, here's the crux of this issue, which this and every study like it conveniently leave out: real violence and fake violence are fundamentally different. See, people who get up in arms about this want to make it about the violence itself, because that's all there is in a game or movie. But real-world violence is really about the consequences of that violence; consequences ranging from the immediate (the pain the subject of the violence feels) to the long-term (the effect on the victim's psyche and family, and on society as a whole).

You saved me a lot of typing. Thanks.

I can understand folks connecting the real world to games; that's often the point. But that doesn't mean it is. The actual and emotional separation between real life and what I see on my computer screen is what allows me to beat up that whore in GTA. If I thought she were real and the violence felt real, she's still have her money.

Moggraider
07-28-2006, 08:14 AM
Why do violent acts have to equate to crimes and arrests? Pushing, shoving, tantrums and other minor offenses are all violent acts not worthy of an arrest. They still qualify as violence. Your decision to redefine "violence" as "capital crimes" is unwarranted. And while overall crime has decreased, school crimes have increased.

Look, don't make me take this role of being the bad guy. I play videogames just like everyone else here; I'm just willing to admit that they affect the way we think.

extarbags
07-28-2006, 08:17 AM
I didn't. If you noticed, I specifically listed "pain" as one of the consequences of violence. Or if you want to go even smaller, the negative emotional reactions people have to being, say, shoved, even if it doesn't hurt or anything. The point is, real-life violence has consequences, game violence doesn't. That's why it's extremely different, and that's what every last one of these studies ignores.

fuzzyslug
07-28-2006, 08:18 AM
Why do violent acts have to equate to crimes and arrests? Pushing, shoving, tantrums and other minor offenses are all violent acts not worthy of an arrest. They still qualify as violence. Your decision to redefine "violence" as "capital crimes" is unwarranted. And while overall crime has decreased, school crimes have increased.

I'm not redefining anything. The statistics we have are limited but certainly appear to argue with the violent video games = violent acts correlation every politician wants to make. It's one thing to say someone has a lower heart rate while watching Magnum P.I. It's quite another to analyze wild swings in real world crime.

Do you have the school crime statistics? I'm honestly interested.

Moggraider
07-28-2006, 08:23 AM
Grandstanding politicians and real-world effects on the brain and behavior are two very different things. I am not a grandstanding politician.

extarbags
07-28-2006, 08:35 AM
And you also aren't demonstrating real world effects on the brain and behavior.

Moggraider
07-28-2006, 08:41 AM
Read the studies I provided the citations for on the second page.

ProStyle
07-28-2006, 08:44 AM
"In short, the modern entertainment media landscape could accurately be described as an effective systematic violence desensitization tool," he said. "Whether modern societies want this to continue is largely a public policy question, not an exclusively scientific one."

Sounds like someone has an axe to grind...
Or maybe heavily polarized language would actually get his study noticed by hacks in the political arena? It's certainly easier to attract their attention than say, I dunno, the respect of the scientific community - that might mean you'd have to produce solid evidence by making logical observations, and it's a lot harder to do that than it is to wire up some electrodes and record disparate data, correlating absolute bullshit to the results when you're finished. Fantastic work, Iowa State.

A public policy question? We pay for Americas Army, we pay for TV recruitment ads underscored by GodSmack, and we pay out the ass for illegitimate wars of aggression launched under false pretexts - but really, the downfall of "modern society" will definitely be due to free market video game publishers who continue to push the envelope of desensitization. I have no doubt in my mind, it's as clear as an azure sky of deepest summer.

extarbags
07-28-2006, 08:46 AM
You aren't paying attention. When you see the breach of netiquette that is about to be committed, remember that you drove me to this:

FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE

It proves nothing to study peoples reactions to violent TV shows after they play violent games. A person's reaction to fictionalized violence, in any form, is based on their level of sensitivity to the violent imagery depicted. A person's reaction to and attitude towards actual violence is based on their ability to identify with the consequences of that violence. They are completely different things. If you can't tell the difference, that's your issue, and it didn't come from playing video games.

metta
07-28-2006, 08:51 AM
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE
FICTIONAL VIOLENCE IS NOT THE SAME AS REAL VIOLENCE

Well argued.

skedastic
07-28-2006, 08:53 AM
Yes there is.

If you aware of any studies demonstrating a causal effect of violent video games on violent behavior, please cite them. Notice that data collected in a laboratory cannot in principle demonstrate such an effect. To the best of my knowledge, no one has used observational data to show such an effect either.

What's the problem? The average is lower for one group than the other. That's proof that on average we'll see a relationship there. Social science is not freaking chemistry; you're never going to get a dead-on 1:1 relationship.

Correlation does not imply causation. You may find this phrase useful in your studies.

Kitsune
07-28-2006, 09:03 AM
If games had a pervasive and relatively reliable effect in desensitizing people to violence, then don't you think there would have been some effect by now in other countries as well? I'm thinking, of course, particularly of mine, where despite games being played and consumed at levels that were at one time the top amount of saturation you could get in the whole world, nothing whatsoever has come of it. And we're talking shmups and fighters and beat-em ups and survival horror country here. There ought to have been some sort of reaction or effect by now, but there hasn't been.

Though they are obscure, Japan's also the only country in the world I know of that sells games where you can rape people. You'd think THAT would have had some effect on some of the rape cases we've seen here, but its never once been reported to be connected to a rape.

-Kitsune

Moggraider
07-28-2006, 09:05 AM
It proves nothing to study peoples reactions to violent TV shows after they play violent games. A person's reaction to fictionalized violence, in any form, is based on their level of sensitivity to the violent imagery depicted. A person's reaction to and attitude towards actual violence is based on their ability to identify with the consequences of that violence. They are completely different things. If you can't tell the difference, that's your issue, and it didn't come from playing video games.

Again, you're ignoring the other studies I posted, where subjects play either a violent or non-violent videogame, then are given the chance to administer shock. The players of the violent game administered shock for longer and resorted to it sooner. This was an experiment with college students, both male and female, so apparently effects last into adulthood.

If you aware of any studies demonstrating a causal effect of violent video games on violent behavior, please cite them.

Already did. See the second page.

Notice that data collected in a laboratory cannot in principle demonstrate such an effect. To the best of my knowledge, no one has used observational data to show such an effect either.

What are you asking for? Design for us the perfect experiment which would prove, once and for all, that games cause violent behavior. Do scientists have to provide real, loaded guns to subjects and then see if they do or do not shoot another person before you're satisfied? Let's be serious.

Though they are obscure, Japan's also the only country in the world I know of that sells games where you can rape people. You'd think THAT would have had some effect on some of the rape cases we've seen here, but its never once been reported to be connected to a rape.

Sex is a drive that is "filled" and "emptied;" violence is not. Japan has been very good about keeping its men happy with tons and tons of porn, everywhere. Then again, it's also the country with its people having by far the least amount of sex, which provides credence to the other studies I mentioned.

Chris Woods
07-28-2006, 09:14 AM

In all seriousness, you guys are going a little bonkers here. "Proof" in a psychological sense has never carried the same connotation it has in physical sciences, so please give the word some leeway. If you really don't like how Psychologists or Sociologists use the term proof you beef isn't with Moggraider.

Also, please stop trying to turn this into a "Video Games -> violence" thing. Even if Mogg tried to assert that the people doing these studies aren't. They're using proper terminology like "correlates"; "demonstrates"; "might" etc. since they know they are dealing with probabilities. The press always reports these things badly, just like any of the scientific stories that get dressed up by the media to sound more grand then they should. (Hell, I'm waxing political here, but the press is responsible for a good deal of our troubles.)

Skepticism is good, but this isn't skepticism happening here. You folks have an agenda. Historically, people with an agenda can't be trusted to give unbiased opinions. It looks really bad for gamers in general.

Chris Woods

extarbags
07-28-2006, 09:19 AM
Also, please stop trying to turn this into a "Video Games -> violence" thing. Even if Mogg tried to assert that the people doing these studies aren't.

Mogg is the person I'm arguing with. Why would I address somebody else's views?

skedastic
07-28-2006, 09:26 AM
Already did. See the second page.

No, you didn't. Yet again, there is no compelling evidence that violent video games cause violence. Laboratory experiments cannot provide compelling evidence on this question, and there are no observational studies using appropriate methods to recover estimates of causal effects.

What are you asking for? Design for us the perfect experiment which would prove, once and for all, that games cause violent behavior.

It is possible, given the appropriate data, to estimate the extent to which violent video games cause violent behavior. For example, in this paper (http://www.chass.utoronto.ca/cepa/stinebrickner_paper.pdf) the authors estimate the causal effect of video gaming on academic performance, exploiting a natural experiment to do so. Similar and other statistical methods could be applied to estimate the effect of video gaming on violence, but no one appears to have done so. This literature has not yet progressed beyond reporting partial correlations estimated from poor cross-sectional datasets.

Again, there is no compelling evidence that video games cause violent behavior.

Moggraider
07-28-2006, 09:35 AM
No, you didn't. Yet again, there is no compelling evidence that violent video games cause violence. Laboratory experiments cannot provide compelling evidence on this question, and there are no observational studies using appropriate methods to recover estimates of causal effects.

It would be nice if you took time to actually read the studies I posted instead of nonsensically insisting that lab studies are bullshit. If violent videogames caused people to shock others more, despite the laboratory setting and the presence of an experimenter, that sounds like enough of an effect to me. Lab settings typically cause subjects to act in more socially-acceptable ways, not less socially-acceptable ways.

It is possible, given the appropriate data, to estimate the extent to which violent video games cause violent behavior. For example, in this paper (http://www.chass.utoronto.ca/cepa/stinebrickner_paper.pdf) the authors estimate the causal effect of video gaming on academic performance, exploiting a natural experiment to do so.

That study shows that playing videogames lowers grades because students play games instead of studying. How is the method in this study applicable to proving that videogames cause violence? It sounds like what you're asking is impossible, because it's unethical to run a study asking people to cause real, physical harm to other people. So, I'm asking you again to put your thinking cap on and design an experiment that, if run and results came out, would shut you up.

skedastic
07-28-2006, 10:00 AM
It would be nice if you took time to actually read the studies I posted instead of nonsensically insisting that lab studies are bullshit.

I did not say that "lab studies are bullshit." I said that lab studies cannot provide compelling evidence on the effects of video games on violent behavior. I do not agree that what undergraduates do in a lab necessarily reflects behaviors outside of a laboratory, and that is a consensus view within the social sciences.

That study shows that playing videogames lowers grades because students play games instead of studying. How is the method in this study applicable to proving that videogames cause violence? It sounds like what you're asking is impossible, because it's unethical to run a study asking people to cause real, physical harm to other people.

Notice that the paper I cited did not do something that's "impossible, because it's unethical to run a study asking people to increase or decrease their time spent studying." Rather, what this study did was exploit random variation induced for a unrelated reason (random assignment of students to dorm rooms). If there were information in these data on violence, then they could also be used to study the question at hand: If video games cause violence, we would expect to find that students randomly assigned to a roommate who brought a video game console to school to engage in more violence than an otherwise comparable student. Notice this reasoning does not require the researcher to manipulate violence in any way, rather, the researcher can make causal inferences merely from observational data.

So, I'm asking you again to put your thinking cap on and design an experiment that, if run and results came out, would shut you up.

I trust you're satisfied.

Now, you might note that the psychologists you're citing don't actually agree with your opinions. If experimental results were all that's needed, why do they more often turn to observational data? And if the combination of lab results and some limited results from cross-sections were all that's needed, why does even Craig Anderson stress that longitudinal data is "badly needed"? (See page 359 of the paper I cited earlier in the thread.) Anderson repeatedly claims to have evidence of causation (he doesn't), but then implicitly acknowledges the case is weak by calling for data which are better able to shed light on the causal issues. Frankly, there is a lot of pretty awful statistical work being done in this area.

Moggraider
07-28-2006, 10:15 AM
I did not say that "lab studies are bullshit." I said that lab studies cannot provide compelling evidence on the effects of video games on violent behavior. I do not agree that what undergraduates do in a lab necessarily reflects behaviors outside of a laboratory, and that is a consensus view within the social sciences.

Of course a Hawthorne effect is in place, but it didn't affect behavior the way you're alleging it does, because the differences in behavior in the study I mentioned still popped up. A lack of a laboratory environment would presumably make those differences more exaggerated, not less.

Notice that the paper I cited did not do something that's "impossible, because it's unethical to run a study asking people to increase or decrease their time spent studying."

The two situations are not parallel at all, and you still haven't made it clear exactly what kind of experiment you're asking for.

Rather, what this study did was exploit random variation induced for a unrelated reason (random assignment of students to dorm rooms). If there were information in these data on violence, then they could also be used to study the question at hand: If video games cause violence, we would expect to find that students randomly assigned to a roommate who brought a video game console to school to engage in more violence than an otherwise comparable student. Notice this reasoning does not require the researcher to manipulate violence in any way, rather, the researcher can make causal inferences merely from observational data.

Your experiment-designing skills need work. You really expect videogames' effects to show up in people by making them beat up on their roommates? That's more simplistic thinking than even extarbags is guilty of.

Now, you might note that the psychologists you're citing don't actually agree with your opinions. If experimental results were all that's needed, why do they more often turn to observational data?

Observational data is universally acknowledged as being less compelling than experimental data. Scientists use it because they're settling for it, not because it's a straight alternative.

And if the combination of lab results and some limited results from cross-sections were all that's needed, why does even Craig Anderson stress that longitudinal data is "badly needed"? (See page 359 of the paper I cited earlier in the thread.) Anderson repeatedly claims to have evidence of causation (he doesn't), but then implicitly acknowledges the case is weak by calling for data which are better able to shed light on the causal issues.

Why is asking for longitudinal data a bad thing? Games have only gotten more and more realistic with time, and will continue to do so; it would be important to see if effects on behavior are changing or increasing. The field of study for this is still new, because videogames themselves aren't that old. Still, all the studies point in the same direction. You can't just ignore that.

extarbags
07-28-2006, 10:18 AM
Jesus Christ. I'm done. You're just too out there.

Moggraider
07-28-2006, 10:18 AM
Be seeing you.

skedastic
07-28-2006, 10:50 AM
Of course a Hawthorne effect is in place, but it didn't affect behavior the way you're alleging it does, because the differences in behavior in the study I mentioned still popped up. A lack of a laboratory environment would presumably make those differences more exaggerated, not less.

Yet again, changes in measures of aggressive behaviors which occur in a laboratory do not necessarily translate into changes in aggressive behavior outside of a laboratory. You are disagreeing not only with me, but with the psychologists you (and I) have cited and with basically every other social scientist.

The two situations are not parallel at all, and you still haven't made it clear exactly what kind of experiment you're asking for.

I don't see how the situations aren't directly analogous. The literature discussed in this thread is attempting to estimate the causal effect of gaming on violence. The paper I cited attempts to estimate the causal effect of gaming on academic performance.

Your experiment-designing skills need work. You really expect videogames' effects to show up in people by making them beat up on their roommates? That's more simplistic thinking than even extarbags is guilty of.

Who said anything about "beating up their roommates?"

Observational data is universally acknowledged as being less compelling than experimental data. Scientists use it because they're settling for it, not because it's a straight alternative.

Well, sort of: researchers use observational data when it is not feasible to perform controlled experiments. It is not feasible to perform controlled experiments to evaluate the effects of video games on violent behavior, except in a very limited sense. Because of that, everyone except Moggraider agrees that the story does not begin and end with lab experiments, and they also turn to observational data. But at the moment these studies using observational data are also very limited. Taken together, these observations tell us that there is no compelling evidence that violent video games cause violent behavior.

Why is asking for longitudinal data a bad thing? Games have only gotten more and more realistic with time, and will continue to do so; it would be important to see if effects on behavior are changing or increasing.

Of course, using longitudinal data is a good thing. One reason why there is no compelling evidence that violent video games cause violent behavior is that no one has used such data to study these issues. Noting that such data are "badly needed" is to implicitly acknowledge that existing results are not compelling.

You misunderstand why longitudinal data would be improvement over existing results. Repeated cross-sections would allow estimates of whether the correlations between gaming and violence are going up or down over time, which is of some interest but does not speak to the key question: do video games cause violence? Longitudinal data would allow us to see whether a given individual's violent behaviors rise or fall as his exposure to video games rise and fall. Evidence that changes in violence are correlated with changes in gaming is much more suggestive that the relationship is causal than evidence that gaming and violence are correlated across people. Better still would be quasi-experimental designs of the sort executed by the paper I cited on games and GPA. Since neither of these approaches has been attempted in the literature on games and violence, there is no compelling evidence that violent video games cause violent behavior.

The field of study for this is still new, because videogames themselves aren't that old. Still, all the studies point in the same direction. You can't just ignore that.

No, "all the studies" do not "point in the same direction." It is true, however, that there is a lot of evidence which is consistent with the idea that, for at least some non-trivial proportion of the population, video games cause violent behavior. But there are many relationships in the social sciences which appear to reflect plausible causal mechanisms which, on careful analysis, turn out not to exist. One reason to think that any causal effect of gaming on violence must be quite small is that gaming has exploded in the 20 years but violence has declined. It is possible that violence would have declined even more if gaming hadn't gone up, nonetheless, this piece of suggestive evidence should give us pause. The jury is still out on this question, pending the application of quantitative methods which can recover causation in a manner which is persuasive to researchers other than Craig Anderson and assorted undergraduates.

Moggraider
07-28-2006, 11:25 AM
Yet again, changes in measures of aggressive behaviors which occur in a laboratory do not necessarily translate into changes in aggressive behavior outside of a laboratory. You are disagreeing not only with me, but with the psychologists you (and I) have cited and with basically every other social scientist.

You're a real broken record. What is it you're saying the laboratory setting does to the behaviors being demonstrated? There is no reason to believe that the laboratory setting is putting a behavior in people's heads that wouldn't be there otherwise, especially considering the social nature of the experiment. Again, the fact that the difference in behavior still popped up despite the experiment being in a lab supports my claims, not yours.

I don't see how the situations aren't directly analogous. The literature discussed in this thread is attempting to estimate the causal effect of gaming on violence. The paper I cited attempts to estimate the causal effect of gaming on academic performance.

Parallels in your sentence structures don't equal parallels in the experiments.

Well, sort of: researchers use observational data when it is not feasible to perform controlled experiments. It is not feasible to perform controlled experiments to evaluate the effects of video games on violent behavior, except in a very limited sense. Because of that, everyone except Moggraider agrees that the story does not begin and end with lab experiments, and they also turn to observational data. But at the moment these studies using observational data are also very limited. Taken together, these observations tell us that there is no compelling evidence that violent video games cause violent behavior.

To directly, irrefutably prove that videogames cause violence, we'd have to take newborns and raise them from birth in two separate orphanage-simulating environments. One orphanage would have violent videogames, the other wouldn't, and we could measure how violent each group of kids is, thanks to all the psychologists pretending to be orphanage staff. But it obviously isn't possible to do that, so we go with what we can. What we can do has measured effects on behavior in numerous ways, and they all found at least a slight effect, including increased aptitude to cause harm to another and (in another study) increased likelihood of picking up a gun in an empty room the subject is presented with after playing a game, so that's what we draw conclusions from.

Of course, using longitudinal data is a good thing. One reason why there is no compelling evidence that violent video games cause violent behavior is that no one has used such data to study these issues. Noting that such data are "badly needed" is to implicitly acknowledge that existing results are not compelling.

You're starting to sound like the guys who try to claim there's no reason to believe in global warming.

You misunderstand why longitudinal data would be improvement over existing results. Repeated cross-sections would allow estimates of whether the correlations between gaming and violence are going up or down over time, which is of some interest but does not speak to the key question: do video games cause violence? Longitudinal data would allow us to see whether a given individual's violent behaviors rise or fall as his exposure to video games rise and fall. Evidence that changes in violence are correlated with changes in gaming is much more suggestive that the relationship is causal than evidence that gaming and violence are correlated across people. Better still would be quasi-experimental designs of the sort executed by the paper I cited on games and GPA. Since neither of these approaches has been attempted in the literature on games and violence, there is no compelling evidence that violent video games cause violent behavior.

I will take the consensus of America's psychologists much, much sooner than I would take your claim. Anyway, it sounds like what you're suggesting is a a month-by-month, or week-by-week, or day-by-day log of how much kids play videogames and how that affects the number of times they commit violent acts. In order to carry out such a study, a university professor would have to all at once get thousands of parents, school administrators and staff, security guards, and any walking passersby to report in every time a child commits a violent act. Either that, or it's back to my orphanage study. Good luck with that. It's back to the drawing board for you.

No, "all the studies" do not "point in the same direction."

Oh really? Find 1 (One!) study that suggests that violent videogames decrease, or do not affect, levels of violent behavior. I'll wait here.

The jury is still out on this question, pending the application of quantitative methods which can recover causation in a manner which is persuasive to researchers other than Craig Anderson and assorted undergraduates.

I don't think the question is as controversial as you portray it; you only get away with making such a statement because this is a message board made and posted in by hardcore gamers. And Anderson's studies have all been peer-reviewed and published in journals, and rather selective ones at that, which means that they have considerable, or majority, backing. Psychologists love tearing into colleagues's findings because such counter-findings are the easiest studies of all to create and get published, so if there was some kind of videogame-loving branch of the field at work, we'd've heard about it by now, and every gaming website would have trumpeted its findings across the internet.

Dave47
07-28-2006, 11:50 AM
I find it rather disturbing how quickly people are willing to throw out the social sciences in order to shield their hobby from criticism. I love bicycle racing, and I’m hoping that Floyd Landis turns out to be clean. But though I’m rooting for him, I’m nevertheless willing to admit that the smart money says he’s a doper. Demanding absolute proof as a criterion for forming judgments isn’t rational. We all like video games, but we should be willing to admit that they have some negative side effects.

Imagine a study in which groups of volunteers were made to drink diet sodas, and then exposed to a room full of snacks. If one test group ate significantly more than the other groups, we’d think that something in that soda acts as an appetite stimulant. And most of the time we would be right. Nothing really changes when “Coke and Pepsi” become “Tetris and Doom.”

bigdruid
07-28-2006, 11:55 AM
Allow me to rebut all your scientific studies with this unassailable fact:

I have been playing video games for close to 30 years now, which should have made me as crusty and desensitized as is humanly possible. And yet, I still cried like a baby when Boromir died.

Therefore, the studies are full of shit.

Thank you, feel free to cite my post in your future academic endeavors.

Lizard_King
07-28-2006, 12:03 PM
I find it rather disturbing how quickly people are willing to throw out the social sciences in order to shield their hobby from criticism. I love bicycle racing, and I’m hoping that Floyd Landis turns out to be clean. But though I’m rooting for him, I’m nevertheless willing to admit that the smart money says he’s a doper. Demanding absolute proof as a criterion for forming judgments isn’t rational. We all like video games, but we should be willing to admit that they have some negative side effects.

Side effects, sure. But I don't think desensitization to violence is a novelty, or anything to get excited about.

LesJarvis
07-28-2006, 12:05 PM
I find it rather disturbing how quickly people are willing to throw out the social sciences in order to shield their hobby from criticism. I love bicycle racing, and I’m hoping that Floyd Landis turns out to be clean. But though I’m rooting for him, I’m nevertheless willing to admit that the smart money says he’s a doper. Demanding absolute proof as a criterion for forming judgments isn’t rational. We all like video games, but we should be willing to admit that they have some negative side effects.

Imagine a study in which groups of volunteers were made to drink diet sodas, and then exposed to a room full of snacks. If one test group ate significantly more than the other groups, we’d think that something in that soda acts as an appetite stimulant. And most of the time we would be right. Nothing really changes when “Coke and Pepsi” become “Tetris and Doom.”

I'm skeptical of the social sciences aside from whatever specific criticisms of videogames they have to offer, but your point is a fair one. I think this is at least in part due to the fact that studies like these are often held up by politicians as proof that videogames are harmful to such a degree that we should restrict access to them. And while one irrational response doesn't justify another, it's certainly understandable.

My biggest problem with this study is that it doesn't really say anything that a reasonable person can't infer on their own. People become desensitized to certain subject matter after repeated exposure to it? Who'd have thunk it? I mean, I don't think anyone would really bother to argue with such a point, because it's obvious. Zylon's quip earlier in the thread brought to light this fact quite well:

For their next trick, these researchers will prove that the second cup of coffee never tastes as good as the first.

What this study doesn't tell us (which I pointed out in my initial response to the thread) is whether or not this desensitization can lead to a higher incidence of violent behavior, which is the question people are actually interested in.

skedastic
07-28-2006, 12:44 PM
You're a real broken record. What is it you're saying the laboratory setting does to the behaviors being demonstrated?

The behaviors that everyone except you are concerned with are violent actions which harm other people. The link that everyone except you acknowledges doesn't exist is between behaviors in a laboratory and violent behavior outside of the laboratory.

Parallels in your sentence structures don't equal parallels in the experiments.

Why don't you just help us and explain why think that measuring the effects of video gaming on GPA is fundamentally different the effects of video gaming on violent behaviors. Suppose that Stinebrickner had data in his survey on agressive behavior. If he ran exactly the same statistical models simply subsituting the measure of aggression for the measure of academic performance, it would seem that he would obtain an estimate of the effect of video gaming on aggression.

To directly, irrefutably prove that videogames cause violence, we'd have to take newborns...

"Directly, irrefutably prove" is a very high standard indeed. What I, along with basically everyone else who's even vaguely aquainted with this literature, am suggesting is that more evidence capable of recovering causation be amassed before we form conclusions. There are many methods available to social scientists to estimate causal relationships which stop far short of randomizing newborns. You are the only one who believes that the evidence on the table "proves" video games cause violent behavior.

You're starting to sound like the guys who try to claim there's no reason to believe in global warming.

I would hope that what I sound like is a professor with a lot of experience estimating causal relationships from large-scale observational microdata.

Anyway, it sounds like what you're suggesting is a a month-by-month, or week-by-week, or day-by-day log of how much kids play videogames and how that affects
the number of times they commit violent acts.

No, that is not what I'm suggesting, nor is it what Craig Anderson or anyone else is suggesting. Time log data such as this is expensive to collect and not required.

Oh really? Find 1 (One!) study that suggests that violent videogames decrease, or do not affect, levels of violent behavior. I'll wait here.

There are lots of studies which fail to find a statististically significant effect of gaming on various measures of violence. Look at Figure 1 in the Anderson and Bushnell meta-analysis. See how the lower bounds of all (but one) of the agressiveness measures contain zero? That means some fraction of the studies included in the analysis had point estimates less than zero, and many of the studies with point estimates above fail to obtain significance. Some less biased reviews of the literature come to far different conclusions than Anderson. For example,

Lillian Bensley & Juliet Van Eenwyk, "Video Games and Real-Life Aggression: Review of the Literature," J. Adolescent Health 29(4): 244-256.

reviews a number of studies which do and which do not find an effect of gaming on violence, concluding "Among young children (about aged 4–8 years), playing an aggressive video game caused increased aggression or aggressive play during free-play immediately after the video game in 3 of the 4 studies. For teenagers....it was not possible to determine whether video game violence affects aggressive behavior. Among college students, there is not consistent evidence that video game play affects aggression or hostility."

The results on violence and gaming are, in fact, so inconsistent that at least one meta-analysis concludes they are more consistent with a "catharsis effect" in which violent people exhaust their violent tendencies through gaming rather than actual violence,

Griffiths, M. (1999) "Violent Video Games and Aggression: A Review of the Literature," Aggression & Violent Behavior 203.

Boyle, Raymond and Hibberd, Matthew (2005) “Review of research on the impact of violent computer games on young people.'' manuscript, Stirling Media Research Institute,

which is linked to here (http://www.theesa.com/facts/third_party.php) and summarized:

- Key Finding: There are “many inconsistencies in the reported amount of research into media violence. Put simply there are a lot of myths, misinterpretations and mis-representations surrounding the quantity and quality of research on this issue.”

The authors reviewed academic research into violent video games “ensuring that all key studies examining the relationship between playing violent computer games and real-world violence in young people carried out between 1985-2004, are covered.” In their findings, Boyle and Hibberd write that “the research evidence of a direct link between video games and violent behaviour in society remains contradictory… there is a body of evidence that shows playing violent video games increases arousal and the possibility of aggression in some players. However, this evidence is often disputed and cannot be simply read as evidence that game playing translates into violent social behaviour.”

As you can see, there is no "consensus of psychologists," much less a consensus among social scientists more broadly, on the issue of video games and violence. As I just said, there is lots of suggestive evidence consistent with a causal effect, and some evidence inconsistent with a (non-trivial) causal effect, and a lack of evidence sufficient to support conclusions over causation. That opinion -- that the jury is still out -- is more or less consensus. The opinion that you're advancing, that the case is closed, is a radical position without support in the liertature.

I don't think the question is as controversial as you portray it; you only get away with making such a statement because this is a message board made and posted in by hardcore gamers.

The question on causation is every bit as controversial and unsettled as I portray it. Even Anderson, in unguarded moments, admits causation has not been established. As a review in the Loyola Law Journal put it,

“Numerous authorities, including the United States Surgeon General, the Federal Trade Commission, and the Federal Communications Commission, have noted the lack of consensus within the scientific community regarding a causal relationship between violent programming and violent behavior. Researchers, including those who frequently advocate in support of legislation restricting access to video games, such as Dr. Craig Anderson, likewise acknowledge the difficulty in establishing a causal link.”

skedastic
07-28-2006, 12:51 PM
Imagine a study in which groups of volunteers were made to drink diet sodas, and then exposed to a room full of snacks. If one test group ate significantly more than the other groups, we’d think that something in that soda acts as an appetite stimulant. And most of the time we would be right. Nothing really changes when “Coke and Pepsi” become “Tetris and Doom.”

You're overlooking the key point that the outcome of interest cannot be measured in a laboratory. Suppose that I go and collect data on people's BMI and whether they drink coke or pepsi. I find that, say, coke drinkers weight statistically significantly more than pepsi drinkers. Should we conclude that if we take a given individual and switch them from coke to pepsi, their BMI will fall?

Reldan
07-28-2006, 12:51 PM
Moggraider actually makes a more compelling point here. I think some of you guys are trying to strawman the hell out of him.

Everyone loves to play the "Correlation <> Causation" card, but that only strictly applies to the physical sciences, where causation can be directly proven. There's a lot that can be implied or indirectly "proven" with correlation, and with social science correlation is often all you really have to work with.

For those with semantics issues - get over it. Arguing over the meaning of the word "proven" doesn't change the data gathered by the study.

For the "No compelling evidence video games cause violent behavior" crowd, the study is showing a correlation between desensitization to violence and violent video games. You're making a pretty big leap when you assume that desensitization to violence leads to violent behavior. Your fallacy is that you're trying to change the argument from one you can't win to one you can. Me not going into shock at seeing a guy get hit by a car doesn't mean I'm going to start running people over. Desensitization to violence is not even necessarily a bad thing.

I like that Moggraider will refute or argue against anything you naysayers claim, but you pick and choose which of his questions to answer.

worm
07-28-2006, 12:59 PM
I've been totally desensitized to studies about media desensitizing people to violence.

skedastic
07-28-2006, 01:04 PM
Everyone loves to play the "Correlation <> Causation" card, but that only strictly applies to the physical sciences, where causation can be directly proven. There's a lot that can be implied or indirectly "proven" with correlation, and with social science correlation is often all you really have to work with.

Correlation does not imply causation in either the physical or social sciences. It is far more, not less, important to keep that in mind when considering social data because social data are typically subject to far more confounding than natural data. Of course, social scientists have developed a vast array of methods to make causal inferences from correlations, but they haven't been used in the literature we're discussing.

Chris Woods
07-28-2006, 01:26 PM
Mogg is the person I'm arguing with. Why would I address somebody else's views?

In seriousness, here's the crux of this issue, which this and every study like it conveniently leave out: real violence and fake violence are fundamentally different. See, people who get up in arms about this want to make it about the violence itself, because that's all there is in a game or movie. But real-world violence is really about the consequences of that violence; consequences ranging from the immediate (the pain the subject of the violence feels) to the long-term (the effect on the victim's psyche and family, and on society as a whole).

In a game or movie, there are no real consequences. At best, the fictional consequences on the fictional victim/family/society might be examined, which happens more often in films but sometimes in games too. If a character in a game is well-developed and likeable, I'll often abstain from violence against them, if given the choice, and feel bad after committing violence against them if not. I know everyone isn't like that, but either way, anyone who can tell the difference between a real person and a video game character, and who isn't a psychopath, should be perfectly capable of distinguishing between acts of violence themselves, which are harmless in a vacuum such as a video game, and the consequences that give those acts meaning.

Or, to put it succinctly, violence is more than just blood and gore.

Also, please stop trying to turn this into a "Video Games -> violence" thing. Even if Mogg tried to assert that the people doing these studies aren't.

Chris Woods

Flowers
07-28-2006, 02:05 PM
Why do violent acts have to equate to crimes and arrests? Pushing, shoving, tantrums and other minor offenses are all violent acts not worthy of an arrest. They still qualify as violence. Your decision to redefine "violence" as "capital crimes" is unwarranted. And while overall crime has decreased, school crimes have increased.
First, the nitpicking;
Pushing, shoving and tantrums do result in arrests.
Pushing and shoving are both battery and tantrums would fall under your state's disorderly conduct statute. Pushing and shoving are extremely mild forms of violence, a tantrum by itself is not violent at all. Striking during a tantrum would be violent because of the striking.

School crimes have probably increased because school reporting of crimes has increased. More schools have a police officer on duty or a police liason and regularly refer children to law enforcement officials much earlier in the disciplinary process. This would be a product of "zero-tolerance," and the war on drugs, not a product of increased drug use or violence. Parochial schools are some of the last places you can bust a lip and not have the cops called on you.

In public grade schools they are beginning to summon the police for fights. Do you think the first fight reported to police at my gradeschool was the first fight at my gradeschool?

Jason McCullough
07-28-2006, 02:20 PM
The "more likely to slap someone after playing Doom" study I hadn't heard about - that's the only result I've heard of that's made the direct jump from playing to more violence.

As to the thread, you guys are being ridiculous. It sounds about as good a result as you'll get in a social science.

skedastic
07-28-2006, 02:24 PM
You need to read more social science.

Matthew Gallant
07-28-2006, 02:31 PM
Maybe I'm desensitized, but you have to slap someone really hard for it to qualify as violence. Aggressive, sure. Violent, not really.

Flowers
07-28-2006, 02:31 PM
Sooo... can I start referring to myself as a scientist because of my degree in Political Science? Any guesses as to why I don't?

Morkilus
07-28-2006, 02:41 PM
Social science is to science as statistics is to math.

Is that close?

skedastic
07-28-2006, 03:09 PM
Obviously the important issue in this thread isn't whether social scientists are actually scientists. It's rather whether or not lawyers are actually mammals.

Jason McCullough
07-28-2006, 04:10 PM
Sooo... can I start referring to myself as a scientist because of my degree in Political Science? Any guesses as to why I don't?

Because you're a lawyer, not doing research in political science?

Moggraider
07-28-2006, 04:56 PM
The behaviors that everyone except you are concerned with are violent actions which harm other people. The link that everyone except you acknowledges doesn't exist is between behaviors in a laboratory and violent behavior outside of the laboratory.

If it was really me vs. the world, legislators wouldn't be so quick to action :).

Why don't you just help us and explain why think that measuring the effects of video gaming on GPA is fundamentally different the effects of video gaming on violent behaviors. Suppose that Stinebrickner had data in his survey on agressive behavior. If he ran exactly the same statistical models simply subsituting the measure of aggression for the measure of academic performance, it would seem that he would obtain an estimate of the effect of video gaming on aggression.

Would you listen to yourself for a second? It wouldn't be scientifically sound to conclude that all with Stinebrickner's setup. Games' effects are deeprooted and long-lasting, and taking or not taking a game console along to an experimental session wouldn't tell us who's really been affected by it.

"Directly, irrefutably prove" is a very high standard indeed. What I, along with basically everyone else who's even vaguely aquainted with this literature, am suggesting is that more evidence capable of recovering causation be amassed before we form conclusions. There are many methods available to social scientists to estimate causal relationships which stop far short of randomizing newborns. You are the only one who believes that the evidence on the table "proves" video games cause violent behavior.

I find it highly unlikely that I'm the only one; ask Leland Yee (a child psychologist) or countless other psychologists how they feel.

I would hope that what I sound like is a professor with a lot of experience estimating causal relationships from large-scale observational microdata.

You don't. You sound like a gamer with an agenda.

No, that is not what I'm suggesting, nor is it what Craig Anderson or anyone else is suggesting. Time log data such as this is expensive to collect and not required.

But you demanded longitudinal data. How else could we collect such a thing for accurate measurement?

There are lots of studies which fail to find a statististically significant effect of gaming on various measures of violence. Look at Figure 1 in the Anderson and Bushnell meta-analysis.

Which Anderson and Bushnell meta-analysis? This is the first time anyone here has mentioned it.

Lillian Bensley & Juliet Van Eenwyk, "Video Games and Real-Life Aggression: Review of the Literature," J. Adolescent Health 29(4): 244-256.

Your rather selective quote from this article is pretty unfair. Counting this meta-review, there are a total of four meta-reviews mentioned in the paper, all of which point to an increase in violent behavior in children, to some degree. The lack of a finding for teenagers is already explained in the paper as being caused by non-experimental designs. The conclusion of the paper agrees with me that the effects of video games on behavior will only increase as games get more realistic, which is why Anderson wanting longitudinal data is not a weakness.

Griffiths, M. (1999) "Violent Video Games and Aggression: A Review of the Literature," Aggression & Violent Behavior 203.

This meta-analysis also shows an increase in violent behavior in children, and it also states the opposite of what you claimed it to. From page 210: "Such evidence suggests that at a theoretical level, there is more empirical evidence supporting social learning theory than catharsis theory - particularly in younger children." Not to mention, of course, that catharsis theory is crap, and it hasn't had a day in the sun since Freud himself. Violence only breeds more violence, like studies on bullying have shown for decades.

Anyway, thanks for pointing me to five meta-analyses which all conclude that videogames cause violence in children. It looks like your little mantra will need some revisions, and this is really what I have been saying all along.

I'm not saying I'm anti-videogames, but as a psych major I do have to admit that videogames desensitize us to violence and make kids more likely to take violent courses of action, due to modeling.

Unicorn McGriddle
07-28-2006, 05:05 PM
The next time I do a post where I juxtapose quotes to create or emphasize meaning, and I have nothing else to say (except for the subject line, of course), but I still need to do the 5-character thing, I'm totally putting "Chris Woods."

Noise
07-28-2006, 05:52 PM
Social science may not be as "solid" as mathematics or physics, but it still beats assumptions and blind faith. It's not like researchers can simply ignore games and media on the hope they have no ill effect.

skedastic
07-28-2006, 05:57 PM
Would you listen to yourself for a second? It wouldn't be scientifically sound to conclude that all with Stinebrickner's setup. Games' effects are deeprooted and long-lasting, and taking or not taking a game console along to an experimental session wouldn't tell us who's really been affected by it.

I'm not sure why you think that "taking a game console to an experimental session" has something to do with Stinebricker's study. Nor am I sure why you wish to limit attention to "deep-rooted and long-lasting" effects -- since the entire psychometrics literature focus on short term or immediate effects, couldn't we follow their lead? Plus, "deep-rooted and long-lasting" effects are much more difficult to find statistically, which would explain why the psych guys don't try.

But you demanded longitudinal data. How else could we collect such a thing for accurate measurement?

Well, we could standard methods used to collect longitudinal data. For example, if I recall correctly there are repeated measurements on time spent playing video games in the National Longitudinal Survey of Children and Youth. You may not be aware that "longitudinal data" is not necessarily, or even commonly, time logs.

Which Anderson and Bushnell meta-analysis? This is the first time anyone here has mentioned it.

Anderson and Bushman, I meant.

The conclusion of the paper agrees with me that the effects of video games on behavior will only increase as games get more realistic, which is why Anderson wanting longitudinal data is not a weakness.

Again, longitudinal data is not a weakness, it's a strength. One of the reasons I am critical of the claim this body of research has established causation is that this strength has not yet been exploited. Anderson's comment that longitudinal data is "badly needed" is an acknowledgement that the sort of information they provide is badly needed to help resolve the causal issues. In the refereed literature, Anderson (almost always) chooses his language carefully to avoid directly claiming causality.

Anyway, thanks for pointing me to five meta-analyses which all conclude that videogames cause violence in children. It looks like your little mantra will need some revisions, and this is really what I have been saying all along.

Could you point me to exactly which sections of any of the papers I cited conclude that "videogames cause violence in children?" How about any other paper? When responding to this question previously you responded with papers showing experimental results, or observational studies which do not recover causal estimates. And, of course, anyone can read any of the papers I cited, or many other papers, to find that the only consensus in this literature is that more evidence is needed before causality can be be established.

It would not be very surprising if it turns out there is a causal effect of violent video games on violent behavior. We might expect that any such effect would be isolated to a fairly narrow subset of the population, as the literature suggests that such effects are similarly narrow in experimental settings. However, the literature is very clear, contrary to Moggraider's reading, that the results we have to date are not sufficient to make a compelling case for a causal effect of video games on social violence. That's the conclusion of the meta-analyses I cited, that's the conclusion independent bodies such as the FCC have made, and that's what we can infer by simply noting that there are no studies which so much as claim to have established this link.

In short, the idea that the average effect of video games on social violence is positive is plausible, and much of the evidence that can be amassed in a laboratory is consistent with that idea. But scholars have not as of yet amassed enough evidence to conclude that video games cause social violence. That's why, as I suggested, better data may resolve the issue.

That conclusion seems to be a common one in this literature, for good reason. I am accused of being a biased gamer. I don't think that's so; I am a rather jaded consumer and producer of quantitative studies very similar to many of those under discussion. I would be a bad lawyer, I mean, scientist, if I didn't reserve judgement on causality until we have evidence on causality.

Met_K
07-28-2006, 06:11 PM
Yeah. I totally broke my best friend's little brother's arm because of video games, not because we were rough-housing/fake-wrestling. I totally knocked my best friend's two front teeth out with a hockey stick because of video games. I totally got in fights all through high school because of video games. I totally did this and that because of video games.

Come on. Don't be a bunch of fucks. That's the thing I hate about you supposedly educated people. All theories and no bloody real life experience. You know what made me a violent person? The fact that I was a boy. That's what.

Jesus Christ.

Lizard_King
07-28-2006, 06:24 PM
If it was really me vs. the world, legislators wouldn't be so quick to action :).
That is the worst rationale given in a long time in this argument. You not only fail to support your argument, you undercut it by reinforcing that the argument behind it is fundamentally demagogic in nature. What Would Congress Do? I don't know, but probably something that people with brains should moderate or avoid altogether.

Moggraider
07-28-2006, 08:10 PM
I'm not sure why you think that "taking a game console to an experimental session" has something to do with Stinebricker's study.

That was the variable in the study, having a console during the study or not.

Well, we could standard methods used to collect longitudinal data. For example, if I recall correctly there are repeated measurements on time spent playing video games in the National Longitudinal Survey of Children and Youth. You may not be aware that "longitudinal data" is not necessarily, or even commonly, time logs.

The study you designed, though, still calls for time logs, just of an undefined interval. I asked you to design a specific experiment, and you predictably danced around and left your description open to interpretation. But anyway, you wanted to plot exact amount of time spent playing videogames vs. aggression, remember? That wouldn't be all that conclusive without time logs, or else posting the study here would just let posters pull the old "correlation =/= causation" card again. This is because presumably, reported violent incidents for most students will still be 0, so those students who did commit violence should have a more directly traceable cause.

Anderson and Bushman, I meant.

I still don't know what you're talking about; I ctrl+f'ed for Bushman and Anderson on this thread and nobody has mentioned such a study besides you. Please link to the study or provide a citation.

Again, longitudinal data is not a weakness, it's a strength.

Right, but you were saying Anderson's asking for it was a weakness, which it's not. It's just a logical next step in the research process. Or can you not remember what it is you say?

In the refereed literature, Anderson (almost always) chooses his language carefully to avoid directly claiming causality.

Right. We've established that it's the press that misrepresents studies like these all the time, not the study conductors themselves. Oh, and there's also people like you who straw-man constantly; that would be the second group of people guilty of misrepresenting the studies.

Could you point me to exactly which sections of any of the papers I cited conclude that "videogames cause violence in children?" How about any other paper?

Try the conclusions in the Griffiths paper and the section labeled (surprise!) "Other Meta-Analyses" in the Bensley paper.

It would not be very surprising if it turns out there is a causal effect of violent video games on violent behavior. We might expect that any such effect would be isolated to a fairly narrow subset of the population, as the literature suggests that such effects are similarly narrow in experimental settings. However, the literature is very clear, contrary to Moggraider's reading, that the results we have to date are not sufficient to make a compelling case for a causal effect of video games on social violence. That's the conclusion of the meta-analyses I cited, that's the conclusion independent bodies such as the FCC have made, and that's what we can infer by simply noting that there are no studies which so much as claim to have established this link.

That's the conclusion you're trying to advance, but it's not the conclusion the papers you cite reach, because they've decided that the studies point to increased violence in children. The Griffiths paper advances the idea that children learn the violence through social learning theory. That makes this the second time that you lie about what the papers you're citing say!

That is the worst rationale given in a long time in this argument.

Worse than the constant straw men and outright untruths that have been propagated in this thread? I don't think so. skedastic says I'm the only person with this view on videogames, but I proved him wrong by mentioning others. That's all there is to it.

skedastic
07-28-2006, 08:58 PM
That was the variable in the study, having a console during the study or not.

It wouldn't be scientifically sound to conclude that all with Stinebrickner's setup. Games' effects are deeprooted and long-lasting, and taking or not taking a game console along to an experimental session wouldn't tell us who's really been affected by it.

isn't even wrong, it's just a non sequitur. Notice that whatever you're trying to say here applies equally to GPA and violence. Are you simply saying you think it's impossible to measure any effects of video gaming using this sort of study?

But anyway, you wanted to plot exact amount of time spent playing videogames vs. aggression, remember?

No, I don't remember that at all, because that's not what I said. I was referring to regression methods based on fixed effects or difference-in-difference designs to remove time-invariant individual-specific heterogeneity. No, this does not require the sort of time logs you imagined, nor does it require the comprehensive third-party reporting of crime information. The sorts of methods and data I described are very commonly used by various social scientists to investigate these sorts of questions -- you do yourself a disservice by constantly reacting with horror and scorn to these ideas.

That wouldn't be all that conclusive without time logs, or else posting the study here would just let posters pull the old "correlation =/= causation" card again. This is because presumably, reported violent incidents for most students will still be 0, so those students who did commit violence should have a more directly traceable cause.

No, the sorts of studies I am talking about do not require time logs. No, that is not directly related to the "old correlation =/= causation'" card. No, it does not reflect well on what you've learned in school so far that you seem to think the difference between correlation and causation is some technical detail to be ignored. No, acts of violence do not have "directly traceable causes" in any relevant sense here. In summary: no.

I still don't know what you're talking about; I ctrl+f'ed for Bushman and Anderson on this thread and nobody has mentioned such a study besides you. Please link to the study or provide a citation.

Since the sequence of letters "bushman and anderson" was not mentioned early in the thread, it follows that no one linked to a paper with those authors? I'll betcha that isn't true.

Right, but you were saying Anderson's asking for it was a weakness, which it's not. It's just a logical next step in the research process. Or can you not remember what it is you say?

Why do you think Anderson thinks longitudinal data is "badly needed?" The lack of longitudinal data is, of course, a weakness. What questions can we address with longitudinal data that we can't with cross-sections? These must be important questions if badly need to be able to answer them.

Try the conclusions in the Griffiths paper and the section labeled (surprise!) "Other Meta-Analyses" in the Bensley paper.

I am beginning to think you simply don't know what people mean when they talk about "causation." Again, there are no studies demonstrating a causal effect of video gaming on violent behavior (as opposed to the sorts of aggressive behaviors which are ethical to elicit in controlled experiments).

That's the conclusion you're trying to advance, but it's not the conclusion the papers you cite reach, because they've decided that the studies point to increased violence in children. The Griffiths paper advances the idea that children learn the violence through social learning theory. That makes this the second time that you lie about what the papers you're citing say!

"Advances the idea that" is not equivalent to "offers evidence regarding that." What I am asking for is for a cite to a paper providing evidence that violent video games cause violent behavior outside a laboratory. Try again.

Worse than the constant straw men and outright untruths that have been propagated in this thread? I don't think so. skedastic says I'm the only person with this view on videogames, but I proved him wrong by mentioning others. That's all there is to it.

"Ask Leland Yee" is not exactly sufficient. Perhaps you can find someone who should know better who believes that the correlations and partial correlations presented so far in the literature "prove" causation. But they too would be wrong.

MikeSofaer
07-28-2006, 09:51 PM
That was the variable in the study, having a console during the study or not.Normally I would interpret that to mean "owning a console over the time period of the study" not "bringing a console to an experimental session" but I didn't read the thing, so who knows.

Justin Fletcher
07-28-2006, 10:08 PM
That doesn't make any sense, Fletcher...That means that adrenaline pumping means increased arousal, not decreased. Desensitization is lack of increased arousal...
The rest of the thread lost me in the middle of page 3, but I just wanted to pop in and say that I was referring to the time period after the adrenaline had stopped pumping, not during. My purely anecdotal, non-scientific experience has shown that my sensory arousal is generally muted when going through post-adrenaline "crash." I'd like to have this study repeated with the subjects playing football, visiting a haunted house, having sex, or engaging in other vigorous non-videogame activities to see if the results are similar.

Moggraider
07-28-2006, 10:17 PM
Notice that whatever you're trying to say here applies equally to GPA and violence.

That's incorrect. Studying and test scores are more directly related.

No, I don't remember that at all, because that's not what I said.

It's exactly what you said. Here it is again:

Longitudinal data would allow us to see whether a given individual's violent behaviors rise or fall as his exposure to video games rise and fall. Evidence that changes in violence are correlated with changes in gaming is much more suggestive that the relationship is causal than evidence that gaming and violence are correlated across people.

Continuing:

No, acts of violence do not have "directly traceable causes" in any relevant sense here.

That's the kind of study you're asking for when you bitch, over and over, for "proof" of "causation."

Since the sequence of letters "bushman and anderson" was not mentioned early in the thread, it follows that no one linked to a paper with those authors? I'll betcha that isn't true.

I ctrl+f'ed for that, then i ctrl+f'ed for "Bushman," and then I ctrl+f'ed for "Anderson." You're the first person to mention Bushman, and it was just in that last post! I can't find the study you're talking about, so I'm asking for you to tell me which it is, because it certainly isn't the OP study, which doesn't have data in the 0 range. Just because what you're saying makes sense in your own head doesn't mean it will still carry any cogence when you express your thoughts fumblingly on the internet.

I am beginning to think you simply don't know what people mean when they talk about "causation." Again, there are no studies demonstrating a causal effect of video gaming on violent behavior (as opposed to the sorts of aggressive behaviors which are ethical to elicit in controlled experiments).

What's your citation for that? "Skedastic et al, 2006?" The meta-analyses you posted admit that gaming causes violence in children. We're talking about social-science-caliber causation, of course, but still causation. There are also the studies I mentioned. That's three strikes, chief.

"Advances the idea that" is not equivalent to "offers evidence regarding that." What I am asking for is for a cite to a paper providing evidence that violent video games cause violent behavior outside a laboratory.

What I am asking for is for you to stop lying about what the papers you cite say. You said catharsis theory was supported in the Griffiths paper, when in fact he said the very opposite, that social learning theory is in place. That's what I said pages ago in this very thread! Then you said, twice, that there's no evidence that videogames cause violence, when in fact all five meta-analyses agree that they do so in children. Is this making any sense to you? You're being very intellectually dishonest for a professor.

"Ask Leland Yee" is not exactly sufficient.

All I had to do was show that I am not the only person with my view. I did that, so I've proven you wrong.

Normally I would interpret that to mean "owning a console over the time period of the study" not "bringing a console to an experimental session" but I didn't read the thing, so who knows.

Right, it means the latter.

The rest of the thread lost me in the middle of page 3, but I just wanted to pop in and say that I was referring to the time period after the adrenaline had stopped pumping, not during. My purely anecdotal, non-scientific experience has shown that my sensory arousal is generally muted when going through post-adrenaline "crash." I'd like to have this study repeated with the subjects playing football, visiting a haunted house, having sex, or engaging in other vigorous non-videogame activities to see if the results are similar.

Playing videogames is as adrenaline-pumping for you as having sex? Really? Gawrsh.

skedastic
07-29-2006, 08:00 AM
That's incorrect. Studying and test scores are more directly related.

Let me try again. Why do you imagine it's impossible to use methods such as in the Stinebrickner paper to study the effects of video gaming on violence?

It's exactly what you said. Here it is again:

You have a very odd definition of "exactly." I described well-known and often-used regression methods which do not involve "plotting" anything, nor the sort of data and attendent problems you imagined. The bottom line is that your indignation over the idea that these methods are feasible only underscores that you have lot to learn. These methods are commonly used by quantitative social scientists and there is no reason that they cannot be used to study violence and video gaming. Many dozens of large-scale longitudinal datasets of exactly the type you imagine are impossible to construct already exist.

No, acts of violence do not have "directly traceable causes" in any relevant sense here.

That's the kind of study you're asking for when you bitch, over and over, for "proof" of "causation."

No, it's what you imagine people are talking about when they refer to causation. When we are talking about something which increases or decreases the probability of some event we don't need, and in the social sciences can rarely attribute, "directly traceable causes." For example, we think that smoking causes lung cancer even though in any particular case we cannot "directly trace" any given individual's cancer back to its cause.

Another example, closer to the topic of the thread, is this paper on crime and police (http://ideas.repec.org/a/aea/aecrev/v87y1997i3p270-90.html). People may be familiar with this paper because Levitt talks about it Freakonomics. This is a good example both because it is a paper about uncovering the causal determinants of crime -- without "directly tracable causes" -- and because it illustrates that correlations are often misleading when we are interested in causation. Levitt shows that if we calculate the correlation, across time and space, between police forces and crime there appears to be little effect, or perhaps even a perverse effect in which more cops lead to more crime. But, as in the present case, that correlation reflects not only the causal effect of interest but also all sorts of other influences. He uses a method which is analytically identical to that in the Stinebrickner paper to estimate the causal effect, and finds that more cops leads to less crime.

Yet another example is the paper currently discussed in this thread. (http://www.quartertothree.com/game-talk/showthread.php?t=27979) The authors are interested in the causal effect of Fox News on voting patterns. Notice they do not simply calculate the correlation between Fox News and Republican vote share in any given year, which would be the analog of what psychometricians have thus far done with data on violence and video games. Rather, they exploited longitudinal data to make better estimates of the causal effect of interest. They note that one gets the wrong answer if one simply calculates the correlation in a given year. Notice too that they do not simply throw up their hands and announce the task is impossible because they cannot go back and "directly trace" the cause of each individual's vote.

I ctrl+f'ed for that, then i ctrl+f'ed for "Bushman," and then I ctrl+f'ed for "Anderson."

I'm sorry, but I guess you'll just have to break down and actually read the thread.

What I am asking for is for you to stop lying about what the papers you cite say.

I am hoping you may soon be able to go an entire paragraph without insulting me. The empirical evidence suggests I am overly optimistic. Back to the question you evaded again: please cite a paper which you think demonstrates a causal effect of video gaming on violence outside of the laboratory. Your previous attempts to answer this question have not turned up such a paper, although I suppose that your confusion over what researchers mean by "causation" makes this a particularly difficult task.

I think you will find, after you do some more reading on "causation" perhaps, that there are no such papers. We should be agnostic over the question of the causal effect of video games on violence because no one has presented evidence on the matter. That's why all those other researchers and observers have concluded that the jury is still out. It's also why even Anderson typically refrains from claiming causation in refereed papers -- he hasn't shown it, nor has anyone else.

All I had to do was show that I am not the only person with my view. I did that, so I've proven you wrong.

I do hope you don't write things like "Ask Leland Yee!" when your professors ask you to "cite." No, those three words do not "prove me wrong." It would be interesting if you could actually quote Leland Yee or anyone else explicitly claiming that correlation implies causation.

EDIT: Curiousity compelled me to google Leland Yee, who turns out to be a politician. I am willing to concede that politicians may make claims which are not supported by evidence.

Misguided
07-29-2006, 08:51 AM
I question their definition of "real-life violence". Their statement in the conclusion that those who play violent games eventually become psychologically numb is completely unsupported by the study.

I encourage people to examine the magnitude of the changes here.

Look at the data and ask questions. Note that the heart rate for BOTH groups went up after playing their games. Why does the heart rate of the group that played violent games DECREASE when watching the film? Why were statistical comparisons made to the baseline instead of post-game heart rates? Is the difference in a few beats per minute important?

But here's the important part:

I don't see any mention of correction for multiple comparisons, which potentially invalidates this entire study. They used .05 as their level of significance, which means that by random chance, they will get a positive result 1 out of every 20 times they conduct a statistical comparison. So, if they looked at 100 variables, 5 would give them statistically significant results BY CHANCE ALONE. You MUST correct for this number of comparisons (including all the ones they tested for and then "dropped because they weren't significant) and there seem to be a bunch in this study. There is no indication that they have done this.

Now I've only skimmed the study, so if I missed where this was discussed, someone please point it out. If they haven't applied this correction, every listed result may be bogus.

Moggraider
07-29-2006, 11:04 AM
Let me try again. Why do you imagine it's impossible to use methods such as in the Stinebrickner paper to study the effects of video gaming on violence?

I never said it was impossible. I attacked only your vague description of an ideal experiment.

You have a very odd definition of "exactly." I described well-known and often-used regression methods which do not involve "plotting" anything, nor the sort of data and attendent problems you imagined. The bottom line is that your indignation over the idea that these methods are feasible only underscores that you have lot to learn. These methods are commonly used by quantitative social scientists and there is no reason that they cannot be used to study violence and video gaming. Many dozens of large-scale longitudinal datasets of exactly the type you imagine are impossible to construct already exist.

You didn't say anything about regression at the time. Only when I took you to task about the design of the experiment did you start to be more specific.

I am hoping you may soon be able to go an entire paragraph without insulting me. The empirical evidence suggests I am overly optimistic. Back to the question you evaded again: please cite a paper which you think demonstrates a causal effect of video gaming on violence outside of the laboratory. Your previous attempts to answer this question have not turned up such a paper, although I suppose that your confusion over what researchers mean by "causation" makes this a particularly difficult task.

I didn't insult you; I said you were lying because you were. Multiple times. You stated that the papers said one thing, when in fact they said the exact opposite, because you didn't expect me to go and read them. I have already cited multiple papers in the past pages of this very thread which show causal effects; try going to read those. You don't consider it causation, but social science experiments do have limitations.

That's why all those other researchers and observers have concluded that the jury is still out.

Again, you're ignoring what the papers you presented in this thread claim. The meta-analyses all indicate that games do cause violence in children. The researchers are only unsure about other age groups. You repeatedly lie about what they state by ignoring the youngest age group that was under study. With the mainstream explosion of videogames in the mid-to-late 1990s, we should be able to see exactly how resilient the human brain is, but until then, the case looks pretty strong for what games do to children. Though realistically, we've known this for decades since Banduras's modeling experiments.

I do hope you don't write things like "Ask Leland Yee!" when your professors ask you to "cite." No, those three words do not "prove me wrong." It would be interesting if you could actually quote Leland Yee or anyone else explicitly claiming that correlation implies causation.

You're not my professor (are you even a professor at all? of what subject? at which university?) and my citations on papers are just fine, thank you very much. You claimed that I was the only person who held such a view on videogames. I pointed out that I was not, therefore you are wrong. Most psychologists nationwide probably hold similar views, or they would've knocked the claims down very quickly to snag easy publications. At the very least, the researchers mentioned here having been published in multiple journals indicates an acceptance of their methodologies and findings. As you say, Anderson wisely avoids using the language of causation, but that's what social scientists are trained to do. In my years of reading social science papers, I don't think I've ever seen a social scientist use "strong language" of the kind you're asking for, even in some of the most seminal papers of many fields.

Anyway, you've been saying all along that there is no compelling evidence that games cause violent behavior. What you keep ignoring is that the case looks pretty strong for children. Just amend your little mantra and we'll be done here :).

skedastic
07-29-2006, 12:49 PM
I never said it was impossible. I attacked only your vague description of an ideal experiment.

I guess someone else said, "It sounds like what you're asking is impossible." Followed by nonsensical criticism suggesting a complete failure to understand what Stinebrickner did. Since you've yet again evaded the question, let's try again: Why do think we can't use these methods to study the effect of video gaming on violence? You've claimed it's "impossible" and "scientifically unsound" and chided me to "listen to myself" for even suggesting such an obviously ridiculous study. I think maybe you now realize that these remarks were uninformed.

I have already cited multiple papers in the past pages of this very thread which show causal effects; try going to read those. You don't consider it causation, but social science experiments do have limitations.

If you continue in your studies, for example, if you take graduate courses in psychometrics, you will find that quantitative researchers in the social sciences categorically do not consider the sorts of studies that currently exist on video games and violence to provide compelling evidence of causation. So, yes, I "don't consider it causation," nor does the community of experts. In short, as I may have said before, there is no compelling evidence that video games cause violent behavior outside of a laboratory. Until better methods are applied to better data, the jury is still out on causation.

Again, you're ignoring what the papers you presented in this thread claim. The meta-analyses all indicate that games do cause violence in children. The researchers are only unsure about other age groups. You repeatedly lie about ...

No, you repeatedly misunderstand the literature, and I grow weary of your constant barrage of accusations and insults. The studies which make credible claims to causation occur in experimental settings in a laboratory. The studies which on real-life violent behavior do not show that there is a causal effect of video games on violence, in children or in adults.

You're not my professor (are you even a professor at all? of what subject? at which university?) and my citations on papers are just fine, thank you very much. You claimed that I was the only person who held such a view on videogames. I pointed out that I was not, therefore you are wrong.

Yes, I am a professor. No, I am not going to tell you who I am. My field of expertise is econometrics; I have a tenured position at a large, Ph.D.-granting institution, and I have a healthy list of refereed papers using or developing relevant statistical methods. I find the experience of being lectured by you on what quantitative social scientists do to be rather surreal.

Now, "Ask Leland Yee" is not a "just fine" citation. I didn't ask you to cite others who believe that video games cause violence, I asked you to cite others who think that the present correlational studies demonstrate causation, as you seem to fervently believe.

As you say, Anderson wisely avoids using the language of causation, but that's what social scientists are trained to do. In my years of reading social science papers, I don't think I've ever seen a social scientist use "strong language" of the kind you're asking for, even in some of the most seminal papers of many fields.

No, social scientists are not "trained to avoid using the language of causation." Social scientists frequently attempt to estimate causal relationships and describe the attempt as such. I have given you several examples in this very thread of papers estimating causal effects which explicitly make causal interpretations of their findings. Social scientists are, on the other hand, trained to be very careful about drawing causal inferences from correlations, particularly correlations estimated from observational data -- which is why your ongoing insistence that correlation implies causation is so distressing.

In psychometrics, methods to recover causal effects from observational data often fall under the umbrella of what they refer to as "structural equation modelling" or "path analysis." Why not be nutty and actually do some reading on these issues?

Justin Fletcher
07-29-2006, 07:24 PM
Playing videogames is as adrenaline-pumping for you as having sex? Really? Gawrsh.
Hey, I don't half-ass anything.

Moggraider
07-29-2006, 11:09 PM
skedastic, I think I've reached the end of meaningful exchanges with you. I criticized your original vague idea for a study until you added more specifics and I got what I asked for - a meaningful effort on your part in designing a realistic study. You can see that such a study would be expensive, and that's why it's not so easy for psychologists to carry it out at the behest of a message board member on the internet :). Hell, they don't even have to; politicians are already chomping at the bit with what they have. And no, I don't care who you are, though I figure you might be one of the CGM columnists or something. Am I supposed to be intimidated that you have a Ph.D? I'm not.

I've read the meta-analyses you provided, and the writers of all of them conclude that videogames encourage violence in children. That means what you've been saying all along in this thread is wrong, because you said it absolutely. You should learn to qualify your statements; maybe I should too. But you're still not excused for lying about catharsis theory vs. social learning theory, because that's pretty blatant, and dishonest, and I'm saddened that you neglect to address or apologize for your behavior. I hope your dishonesty doesn't get you in trouble during your academic career.

I will leave the rather pedestrian task of googling a politician's quote to you; I suggest you start with Yee and Clinton so you can see that I am, in fact, not "the only one" with such views. I don't agree with the politicans' actions, but it's clear that videogames do have effects on the brains of children. Don't try to make me seem like the overconcerned parent I most definitely am not.

Robert Sharp
07-30-2006, 09:05 AM
Mogg, you ASKED Skedastic if he was a professor, and now that he has answered your question you accuse him of beating you over the head with his PhD? That's bad debate form, my friend.

Apparently, all Skedastic is really saying is that these particular studies cannot be used to conclude causation in any reliable way. As far as I can tell, he's hoping for more information before reaching any solid conclusions here. You are the one making absolute statements by saying that video games definitely have effects on the brains of children. Of course, I didn't read every single word of this thread, so maybe skedastic slipped up somewhere and suggested, in absolute terms, that games DO NOT have such an effect, but I never saw that. I just saw some questions about these particular studies.

And yes, I'm sure there are others out there who share your views, even in the professional community. There are also, almost certainly, others who share skedastic's views on this.

I'm not trying to attack you Mogg (ummm..believe it or not!). I just think the two of you are arguing a little more vehemently than is healthy and from the tone of the posts, you are getting angry in a way skedastic is not. I think you have some good points here, btw. I just don't want to see them drowned out (or ended) on account of unfounded anger.

skedastic
07-30-2006, 09:18 AM
The major point I wanted to make is that there is no compelling evidence that video games cause violence outside of laboratory. I think, I hope, Moggraider now sees that that statement is correct, or at least he seem to have dropped the argument that correlation implies causation, and he cannot cite even one paper which provides something more compelling than correlations between gaming and real-world violence. He can't cite such a study since as far as I know there isn't one: we do not yet have compelling evidence regarding the causal effect of gaming on violence. That certainly doesn't mean that we should conclude there is no effect, but by the same token we cannot conclude that there is. Since that is the only substantive point I wanted to make, I am happy we now all in agreement.

The rest was all trying to clear up the debris of confusion Moggraider generates around tangential points. Anyone interested solely in the core point needn't bother with the rest of this.

Suppose it were impossible to get better evidence on causation than we have now. That would not mean that we should promote the existing weak evidence to "compelling" simply because it's all we can obtain. Luckily, there are many ways we can obtain better evidence on causation. They are not impossible, extremely expensive, or scientifically unsound. They do not require randomizing children in orphanages, polling security guards on weekly basis, "directly tracing" causes for each violent event, or any of the other silliness you dreamed up. These ridiculous criticisms are not products of my alleged vagueness -- what I said was more than sufficient for people who are familiar with these methods to see what I was suggesting. The papers I cited as examples should have been sufficient for others, like you, to at least see that such methods are feasible and to get some idea how they work. These studies will be done, not "at the request of some internet poster" but rather because everyone (well, almost everyone) knows that the existing evidence is not sufficient to get at the important issue.

No, I was not "lying" about what the meta-analyses say. If you read the thread again you'll find that there's a lot of wild-eyed screaming accusations from you which all turn out to be artifacts of your failure to understand the papers you're attempting to read. The fundamental problem would appear to be that you are very confused over what researchers mean when they talk about "causation." You idea that correlation is "social science causation" is very wrong.

I think I will decline to find out what politicians have said about the causal effect of violent video games on violence. It certainly wouldn't surprise if they've said silly things about what the scientific evidence says. I wonder if you realize how silly you in turn appear for continually citing politicians rather than scientists to bolster your argument.

skedastic
07-30-2006, 09:40 AM
Mogg, you ASKED Skedastic if he was a professor, and now that he has answered your question you accuse him of beating you over the head with his PhD? That's bad debate form, my friend.

To be fair, I mentioned that I am professor, in response to one of Moggraider's silly accusations. I was also hoping he might stop disdainfully lecturing me about what social scientists do and think. Alas, my hope was in vain.

Of course, I didn't read every single word of this thread, so maybe skedastic slipped up somewhere and suggested, in absolute terms, that games DO NOT have such an effect, but I never saw that. I just saw some questions about these particular studies.

No, of course I never said that. In fact, I said I wouldn't be surprised if it turned out that games do cause violent behavior. But we can't say one way or the other yet because the evidence doesn't yet exist. That's actually a consensus view in the professional community: even zealous advocates such as Anderson do not claim that we have compelling evidence that the observed correlations are causal. Only a certain poster in this thread (and some politicians. Oh, and a disturbing number of journalists.) thinks that correlation implies causation.

And yes, I'm sure there are others out there who share your views, even in the professional community.

Maybe, but I hope not. The particular claim that I was trying to get Moggraider to find support for in the literature is hs idea that the correlational studies produced so far "prove" causation. I hope that no one in the professional community shares that belief. On the other hand, obviously many people in the professional community do believe that video games cause violent behavior, but that isn't a belief formed from compelling evidence, it's a glorified guess. I wasn't asking Moggraider to find others who share that belief: I can name a number of them myself.

Robert Sharp
07-30-2006, 01:54 PM
Oh, I've never heard anyone say correlation is good enough for causation. I was referring to the video games affect the brain thesis. I don't read a lot of sociology anymore, but I can tell you that one of the things taught in logic is that correlation alone is never enough to prove causation. It's also taught when discussing Hume, who essentially used the point to dismiss the certainty of causation altogether.

As you have probably already noted, there are two central problems with assuming causation from correlation. First, there is no way to know which way the causality runs in many cases. Second, there could be external factors causing both (such as the fact that sales of ice cream correlate to murder rates in New York...which is true, btw).

Inactiviste
07-30-2006, 02:51 PM
In the fifties, didn't they tried to prove that rock'n'roll made you violent ?

The biggest problem with sociology is that it often appears like a mean to rationalize an opinion, with the use of carefuly selectionned data. The first question to answer is who is writting a study trying to prove that videogames can desensitize to violence (as the experiment is clearly designed as an argument to prove that), and why they want to question that. I guess it's from people who don't understand videogames, a new form of cultural expression, and therefore fear them, or feel estranged toward them. Where does the assumption that videogames as a form can influence your reaction to violence came from ? That should be at least interrogated before any experience.

And as noted above, the fact that a representation can desensitize you to another representation, which is the only thing the study proves, does not seem a big step ahead. Tangentially, it might just prove that videogames are just a good at mimesis than cinema.

Robert Sharp
07-30-2006, 03:25 PM
The music issue has been a question for years, and there are plenty of cases where a piece of music has directly incited violence. Of course, mob mentality may be partly to blame where concerts are concerned, but music can certainly work one up into a frenzy. It is meant to convey emotions.

With video games, everything is a bit more muddy.

Moggraider
07-30-2006, 05:55 PM
Rob, I know I asked skedastic if he was a professor, but only because he hinted about it first. His attitude makes it sound like you have to have a Ph. D to even have the right to disagree with him.

As you have probably already noted, there are two central problems with assuming causation from correlation. First, there is no way to know which way the causality runs in many cases. Second, there could be external factors causing both (such as the fact that sales of ice cream correlate to murder rates in New York...which is true, btw).

The better correlation to mention here is between warm weather and violence, though you're right.

The major point I wanted to make is that there is no compelling evidence that video games cause violence outside of laboratory. I think, I hope, Moggraider now sees that that statement is correct, or at least he seem to have dropped the argument that correlation implies causation, and he cannot cite even one paper which provides something more compelling than correlations between gaming and real-world violence. He can't cite such a study since as far as I know there isn't one: we do not yet have compelling evidence regarding the causal effect of gaming on violence. That certainly doesn't mean that we should conclude there is no effect, but by the same token we cannot conclude that there is. Since that is the only substantive point I wanted to make, I am happy we now all in agreement.

You've once again oversimplified what I've been saying all along to try to make yourself look better. There are laboratory experiments that show videogames lead to increased violence - these are, among others, the studies I talked about earlier in this thread, and the studies I cited on page 2 here in the thread. These laboratory experiments happened with college-age students, which shows that even with adults videogames have been shown to affect behavior, if nothing else at least in the context of the experiments.

The meta-analyses you cited show that the evidence is on the whole equivocal for teens and adults (something I didn't know), but on the whole, the experiments and observational studies on children are pretty clear - again, the meta-analyses show that the vast majority of children's studies listed in the analyses point in the same direction! From the beginning, I stated that I was unsure about non-child age groups, but that I was sure about children. The findings seem to agree with me on this.

You pooh-pooh the evidence we have because you don't think laboratory results are compelling, but nearly all psychology experiments are done in labs, so it's no surprise that the studies on videogames would continue using the field's traditional methods, especially in a relatively new area of study. You're not a psychologist, so you're demanding other methods be used. This isn't a bad idea at all, I just don't see it as too much of a weakness that it hasn't happened yet. I'm an undergrad at the fourth-best university in the country and I work for its psychology department, which has and has had probably more huge names in the field than any single other university; I don't see the school's faculty running real-world experiments, so I know not to expect such things. You stated on page 3 that "Laboratory experiments cannot provide compelling evidence on this question," and that's just ridiculous. If all psychologists thought like you, Banduras would never have been able to formulate social learning theory. This is just one example among countless potential ones. You also have little appreciation for what child psychologists do to make their experimental environments seem natural.

In summation, the papers and reviews that have been posted in this thread conclude that based on what we've found so far, the case for children is compelling, and that the case for older age groups is not compelling. You claimed the latter across the board, so you'd be wrong.

The rest was all trying to clear up the debris of confusion Moggraider generates around tangential points. Anyone interested solely in the core point needn't bother with the rest of this.

I don't consider your lying about the text of the meta-analyses you cited a "tangential point."

skedastic
07-30-2006, 08:21 PM
The question is whether video gaming causes violence outside of a laboratory. Oddly enough, the psychologists who work on this question use data that did not come from a laboratory. So, no it isn't some intarweb poster "demanding other methods be used," it's psychologists. I suspect that all of these researchers would be very surprised to learn that it's "just ridiculous" to think that laboratory experiments alone cannot provide compelling evidence on behavior outside of laboratories.

Some of the problem seems to be that Moggraider is apparently unaware that most quantitative social science is not experimental. I was never suggesting "real-world experiments" like he continues to imagine. I wasn't suggesting that we have a bunch of people bring game consoles to experimental sessions (to see if they beat up their roommates). More generally, I wasn't suggesting doing anything to experimentally manipulate the amount of video gaming people do in the real world. I am sure the faculty at his university don't commonly run such real-world experiments: they're very expensive. What social scientists, including the faculty in Moggraider's department, do in absence of enormous controlled experiments outside the laboratory is look to nature for experiments. For example, the psychologists using observational data to study violence and video games collect survey evidence which includes questions designed to measure both video game playing and violent outcomes. Some kid got a console for Christmas and plays way more games than some other kid who didn't? There's an experiment. The problem, of course, is that it isn't a controlled experiment, so we can't assume if we find that the console kid is more violent than the other kid that it's because he games more. That is, this sort of correlation does not imply causation.

So, no, there is no compelling evidence that video games cause violence outside of a laboratory. That's my interpretation of the literature, that's the interpretation of government agencies with research divisions like the FCC and FTC, and that's the interpretation of the researchers writing the studies. I think Moggraider doesn't really disagree, either. I think he thinks we're arguing about whether these outcomes are correlated. To set his mind at ease: I agree that there is compelling evidence that a randomly selected gamer is more likely to be violent than a non-gamer. But that isn't compelling evidence that gaming causes violence.

Of course, I'm not really writing any of this for Moggraider's benefit, but nonetheless his insults and attitude are grating. I remember when I too was an undergraduate and knew everything. (Now I know a little tiny bit more than jack shit.) I won't be replying again unless Moggraider actually cites a paper which claims to present evidence that gaming causes violence outside a lab. Which might exist, and I would be interested to see, but I haven't seen one yet, nor has Moggraider mentioned one.

And, in BREAKING NEWS: here (https://netfiles.uiuc.edu/dcwill/www/CMWilliamsSkoric.pdf) is a paper which uses longitudinal data to investigate the effects of video games on real-world violence. Exploiting longitudinal data collected from one of those expensive real-world controlled experiments, they find that the effect of gaming on violence is not statistically significant nor usually even of the expected sign (they expected to find a positive effect, they say, following Anderson's theories).

So I amend my previous comments. I am now aware of one study using data and methods that can provide reasonable evidence on the causal effect of video games on violence, and the evidence suggests that any such effect must be very small, if it exists. Of course, we should wait for more papers before we decide that the evidence is compelling, one way or the other.

EDIT: Damn do I ever talk too much.

Moggraider
07-30-2006, 10:53 PM
The question is whether video gaming causes violence outside of a laboratory. Oddly enough, the psychologists who work on this question use data that did not come from a laboratory.

What? The OP's study was in a lab, so were the ones I posted, and so were the ones in the reviews.

I suspect that all of these researchers would be very surprised to learn that it's "just ridiculous" to think that laboratory experiments alone cannot provide compelling evidence on behavior outside of laboratories.

The entire field of human psychology is built almost entirely on lab experiments, sometimes even when it shouldn't be, like with many (or more probably, most) studies in social psychology (What's the best place to study things like ethnic identity, collectivism, social facilitation, and competition? Psychologists think it's the lab!).

Some of the problem seems to be that Moggraider is apparently unaware that most quantitative social science is not experimental.

True, however the social science fields other than psychology have little or nothing to say on the subject, so psychologists study the issue, using their largely experimental ways. I'm sorry this isn't to your satisfaction.

What social scientists, including the faculty in Moggraider's department, do in absence of enormous controlled experiments outside the laboratory is look to nature for experiments.

The OP's study had subjects report their general media habits, if you'll remember, but you seemed too busy going on and on with apparently what is an intense dislike for Anderson, the researcher, to notice.

That's my interpretation of the literature, that's the interpretation of government agencies with research divisions like the FCC and FTC, and that's the interpretation of the researchers writing the studies.

That's the interpretation of researchers? Is it really? I'm surprised that you insist so doggedly on lying about what the papers you present to us say.

From the paper you cited in the post I am now quoting, on page 219 and onwards:

"Sherry's meta-analysis (2001) suggests that games do have some kind of aggression effect..."

"A second meta-analysis... reached the conclusion that exposure to violent video games is positively linked with agression..."

"Dill and Dill... suggested that the literature points to aggression findings..."

"... exposure to a violent video game can result in short-term increases in aggressive behavior..."

"In sum, researchers suspect a strong linkage between games and aggression..."

From the Griffiths paper you posted a citation for, on page 209:

"The majority of the studies on very young children... tended to show that children do become more aggressive after either playing or watching a violent video-game..."

Similar or stronger quotations can be found in any of the meta-reviews you posted, as I have already stated earlier in this thread. To use social science language, it appears that your interpretation and the researchers' interpretations are not the same thing. You can also add to these quotes decades of research on television viewing, which the latest paper you presented also speaks to: Longitudinal data on television-viewing does show long-term increased aggression, so there is little or no reason to believe that research on games will not (because I know you must be curious, read page 230).

In closing, your characterizations of researchers' findings were biased, minimizing, and wrong-headed. It appears that there is plenty of evidence from all those studies for some effect, even if you do not personally find the evidence "compelling." It sounds to me like these researchers do, and that's why the subject is continually under study. At the very least, there is more than sufficient grounds for editing your now oft-repeated claim that there is no compelling evidence that videogames cause violence. I think that the sum of these studies amounts to rather compelling evidence of at the very least some effect.

Of course, I'm not really writing any of this for Moggraider's benefit, but nonetheless his insults and attitude are grating. I remember when I too was an undergraduate and knew everything.

I never claimed to know everything, and I have admitted to gaps in my knowledge. I don't feel I spent much energy on insulting you, though you may feel insulted because I did not bow to your Ph.D like you wanted me to do. I did point out that you were lying in regards to the reviews you cited, because you were, and anyone else can read the papers like I did and discover this for themselves.

And, in BREAKING NEWS: here (https://netfiles.uiuc.edu/dcwill/www/CMWilliamsSkoric.pdf) is a paper which uses longitudinal data to investigate the effects of video games on real-world violence. Exploiting longitudinal data collected from one of those expensive real-world controlled experiments, they find that the effect of gaming on violence is not statistically significant nor usually even of the expected sign (they expected to find a positive effect, they say, following Anderson's theories).

I am willing to concede that this study's average subject, a 28-year-old white, educated, middle-class adult, will not be incited to aggressive behavior by playing an MMORPG as boring as Asheron's Call 2 for one month (my characterization of the subjects is not unfair - rather it is dead-on - read the study for yourselves if you feel like it). I would even be willing to concede that researchers will get the same finding for most any MMORPG. But that's not the question at hand. This was not an ideal study, or even a particularly useful study, for answering our questions, and you know that.

EDIT: Damn do I ever talk too much.

Agreed.

skedastic
07-31-2006, 09:11 AM
The question is whether video gaming causes violence outside of a laboratory. Oddly enough, the psychologists who work on this question use data that did not come from a laboratory.

What? The OP's study was in a lab, so were the ones I posted, and so were the ones in the reviews.

And again we are back to: Moggraider does not understand the professional literature he's attempting to read. Many of the studies we're discussing did not use data that came from a lab. Those are the studies the meta-analyses refer to as "correlational" as opposed to "experimental." Look, for example, at Table1 in Griffiths, 1999. See all those papers -- Lin and Lepper 1987, Rushbrook 1986, Kestenbaum and Weinstein 1985, Griffiths and Hunt 1993 and so on? All of those studies used survey methods of one sort or another to estimate correlations between video game play and measures of aggression. This is how Freedman, 2001 describes their methods:

Most of the non-experimental work consists of relatively small-scale surveys. People are asked about their exposure to video games, to violent video games, and to various other media. They are also asked about their aggressive behavior, or occasionally others provide information on the respondents’ aggressive behavior. Then the researchers conduct correlational analyses (or other similar analyses) to see if those who are exposed more to violent video games are more aggressive than those who are exposed less.

So, again, psychologists commonly use non-experimental data to try to shed light on behaviors that cannot be studies in a laboratory. It is dismaying that your lack of understanding of the papers you're trying to read is so profound that you aren't even vaguely aware of how the data was collected. In the case at hand, a substantial fraction of the studies on video games and violence use data that didn't come from lab experiments. The correlations estimated in these studies do not provide compelling evidence of causation, except in the mind of one earnest but very confused undergraduate:

there is no compelling evidence that video games cause violence outside of a laboratory.

That's the interpretation of researchers? Is it really? I'm surprised that you insist so doggedly on lying about what the papers you present to us say.

Yes, what I said is the interpretation of professionals in the field. You proceed to quote a bunch of papers talking about the effects observed IN LABORATORIES. Do try hard to understand that what we're talking about is behavior outside of laboratories. The reviews you accuse me of "lying" about are all quite clear that there is not yet compelling evidence on the effects of video games on such behavior:

there is a body of evidence that shows playing violent video games increases arousal and the possibility of aggression in some players. However, this evidence is often disputed and cannot be simply read as evidence that game playing translates into violent social behaviour

“Numerous authorities, including the United States Surgeon General, the Federal Trade Commission, and the Federal Communications Commission, have noted the lack of consensus within the scientific community regarding a causal relationship between violent programming and violent behavior. Researchers, including those who frequently advocate in support of legislation restricting access to video games, such as Dr. Craig Anderson, likewise acknowledge the difficulty in establishing a causal link.”

Correlational studies can tell us nothing about whether violent video games cause aggression. Even if we accept that there is a correlation between amount of time spent playing (violent) video games and aggressive behavior, there is no reason to think that games are the cause of aggression (Anderson & Dill, 2000; Colwell & Payne, 2000; Roe & Muijs, 1998). Furthermore, some correlational studies find no significant relationship with aggression (e.g., Sacher, 1993; van Schie & Wiegman, 1997).

In sum, researchers suspect a strong linkage between games and aggression but, with
the exception of relatively short-term effects on young adults and children possibly
caused by increases in arousal and/or priming [ie, lab results - Sked], they have yet to demonstrate this link.

The problem with all of this type of research is that correlational evidence is unconvincing not only because any observed positive correlations may be due to backward causation (aggressive individuals having a greater penchant for video games), but for the more plausible reason that the correlations may not be directly causal at all but may result from mediating factors (e.g., low educational attainment, low socioeconomic status, etc.) that may themselves be causally related both to video game playing and to aggressive behavior. This interpretation is well known in the literature on the effects of violent television viewing on aggressive behavior.

I cannot think of another important issue for which scientists have been willing to reach conclusions on such a small body of research. Even if the research had been designed and conducted perfectly, there is far too little evidence to reach any firm conclusions. And, as I shall discuss below, the research is far from perfect.

[...]

Thus, the existence of the correlation between playing violent games and aggressiveness does not prove that one causes the other. It provides no evidence for causality. While interesting, this research is not relevant to the central question whether violent video games cause aggression.

Notice Freedman is giving exactly the same argument our very own Moggraider has declared "just ridiculous." I could go on, as authors commonly point out in passing that correlations do not imply causation.

Those are the studies that Moggraider has been arguing "prove causation" outside a lab. As anyone can plainly see, professional quantitative social scientists absolutely do not share his opinion that the correlational studies thus far produced are compelling evidence of causation. I am afraid it is becoming painfully obvious that Moggraider is in way over his head.

I grow weary of the constant insults and lecturing. If you want to explain why you think that your unique interepretation of the extant evidence is better than the consensus of the professionals, please do so. I am somewhat entertained by your lengthy wild-eyed insistence that, damnit, correlation is too causation!

Incidentally, Moggraider, you might want to recall who brought up his credentials, such as they are, first, and then blathered on about them -- my excuse was that you demanded I elaborate on mine. I am afraid I am not intimidated, nor do I bow down to, the extensive expertise you've aquired as a psych major.

Moggraider
07-31-2006, 09:34 AM
Gee, I thought you said you weren't going to reply again.

Anyway, it's true that there were survey studies in the reviews, because it's the reviewers' jobs to include everything, but I find the lab studies a lot more salient, convincing, and relevant, as do most psychologists. This area of study has definitely shifted to experiments over survey work. The survey studies you mention are all 13 years old or more. It would be interesting to see what happens if the survey studies are left out of the reviews and only the experiments were considered.

Anyway, to review for everyone:

1. TV studies show the longitudinal effects skedastic was bitching about, but he's pretending that evidence isn't there.
2. Research on children is in considerable agreement. Only studies on older age groups are equivocal.

skedastic
07-31-2006, 09:57 AM
Gee, I thought you said you weren't going to reply again.

You claimed, again, to be citing papers demonstrating a causal effect of gaming on violence outside a lab. Lest anyone get the that that's actually what these papers say, I have to explain again that you've misread the literature.

Anyway, it's true that there were survey studies in the reviews, because it's the reviewers' jobs to include everything, but I find the lab studies a lot more salient, convincing, and relevant, as do most psychologists.

Surreal. You just got through indignantly insisting that psychologists have never used non-experimental data to study these issues, that all of the studies referred to in the meta-analyses use lab data. When you're confronted with the blatantly obvious fact that psychologists commonly use non-experimental data -- notably including many studies which you've just claimed actually used laboratory data -- this is all you can come up with? Why do you think so many psychologists turn to non-experimental data? Why do you think anyone should take your opinions seriously when you've so spectacularly demonstrated that you have absolutely no idea what the studies you're arguing about actually did, or why they did it?

This area of study has definitely shifted to experiments over survey work. The survey studies you mention are all 13 years old or more.

Which isn't particularly surprising, given that review is itself seven years old. Anderson has recently done some work with survey data, among many others. Would you like to explain why you think Anderson is using non-experimental data, when it's so blindingly obvious to you that no such data can be useful or relevant? Incidentally, now that you know that the correlational studies you thought were produced in a lab were actually produced using non-experimental data, are you sticking to your idea that they "prove causation," even if it's "social science causation," whatever that means?

It would be interesting to see what happens if the survey studies are left out of the reviews and only the experiments were considered.

Then the review would have nothing at all to say about behavior outside of the lab, would it?

We weren't talking about TV, we're talking about video games. And yes, as anyone can plainly see from the quotes in my post above, there is indeed considerable agreement in the literature that there is no compelling evidence, for any age group, that gaming causes violence outside of a laboratory.

Moggraider
07-31-2006, 12:03 PM
Surreal. You just got through indignantly insisting that psychologists have never used non-experimental data to study these issues...

Incorrect. I never said that. You are lying once more. Also, the distinction we had been drawing, as I saw it, was between the real-world studies and "natural experiments" you asked for, and the lab studies that did not meet this criteria, studies which may or may not be experiments. The phrase "lab study" does not mean "experiment" and the two terms should not be used interchangeably. Surveys are taken by people who go to a lab.

And yes, as anyone can plainly see from the quotes in my post above, there is indeed considerable agreement in the literature that there is no compelling evidence, for any age group, that gaming causes violence outside of a laboratory.

Also incorrect, for reasons already stated. It is entirely reasonable to expect that increased violence which has been proven by longitudinal data to occur due to television-watching will also apply to videogames. Also, while skedastic is quick to cast doubts on studies in a lab, everyone would be wise to take note how very much the field of psychology depends on work in labs. If we didn't trust lab studies, we'd have to throw out most of the work done in the entire field. I don't see how labs are less valid for testing videogame behavioral effects than they are for most other elements in psychology. skedastic is rather irresponsibly ignoring a quite strong body of work on children and is taking advantage of the relatively nascent state of this particular sub-field to cast doubt on claims with solid foundations for at least one age group under study.

skedastic
07-31-2006, 02:11 PM
Incorrect. I never said that. You are lying once more.

You're a real piece of work. Above:

What? The OP's study was in a lab, so were the ones I posted, and so were the ones in the reviews.

Heaping on more insults will not get you out of the deep hole you've dug yourself into. Psychologists frequently use non-experimental survey data to try to shed light on the real-world effects of gaming, amongst other questions, even if Moggraider doesn't approve. The correlations thus far produced from this approach are not evidence that gaming causes violence, even if Moggraider thinks otherwise.

Also, the distinction we had been drawing, as I saw it, was between the real-world studies and "natural experiments" you asked for, and the lab studies that did not meet this criteria, studies which may or may not be experiments. The phrase "lab study" does not mean "experiment" and the two terms should not be used interchangeably. Surveys are taken by people who go to a lab.

No, survey evidence used by social scientists is typically not collected in a laboratory. You have just got through making it crystal clear that have absolutely no idea how quantitative social scientists collect data. That does not prevent you, amazingly enough, from continuing to lecture me on what social scientists do.

By "lab studies" I mean experimental evidence generated in a laboratory. Even if surveys are physically completed in a laboratory, they ask questions about outcomes which occur in the wild, and thus they are not "lab studies" in any meaningful sense. I will try to remember to use clumsy phrasing such "controlled experiments in artificial settings" to avoid confusing you in the future.

Also incorrect, for reasons already stated. It is entirely reasonable to expect that increased violence which has been proven by longitudinal data to occur due to television-watching will also apply to videogames.

Amazing. Just amazing. Are you now just completely giving up the argument that there is compelling evidence that video games cause violence? No one except you thinks that if television causes violence it follows that video games cause violence. That's why all those researchers are doing studies on video games rather than just pointing to the somewhat related literature on television. I guess they're just lying irresponsible wrong-headed liars, too.

Also, while skedastic is quick to cast doubts on studies in a lab...

I have repeatedly noted that we cannot conclude that because aggressive behavior increases in a laboratory setting that real-world violence is causally affected by gaming. That's why, as I've noted, lots of psychologists turn to non-experimental data on real-life outcomes to try to shed light on this issue. Obviously (well, to almost everyone) I am not suggesting experiments done in laboratories are in general worthless. Nor have I denied that there is evidence from controlled laboratory experiments showing short-term increases in aggressiveness, and I have already said as much. I suppose this is where I'm supposed to accuse Moggraider of being a liar.

I don't see how labs are less valid for testing videogame behavioral effects than they are for most other elements in psychology. skedastic is rather irresponsibly ignoring a quite strong body of work on children and is taking advantage of the relatively nascent state of this particular sub-field to cast doubt on claims with solid foundations for at least one age group under study.

Look again at the post above with bolded quotes from a number of reviews of this literature. See how they're all saying pretty much exactly what I've been saying -- that there is no compelling evidence that video games cause violence outside of a laboratory? Such as the fact that "this research is not relevant to the central question whether violent video games cause aggression" so that "there is far too little evidence to reach any firm conclusions." Are all of these scholars also "lying" and "irresponsible" and "taking advantage of the relatively nascent state" of the literature? Possibly. Another explanation is that Moggraider is arguing about topics which he doesn't understand.

Believing that the evidence isn't in yet so we should not draw firm conclusions is more or less a consensus in this literature. Indeed, it would be not only simply wrong but also irresponsible to claim that there is compelling evidence.

Here, I'll play the trump card, again, and ask you to cite even one paper presenting evidence on the causal effects of video games on violence outside of a laboratory. Citing studies using experiments done in laboratories doesn't count. Nor does citing studies estimating correlations from cross-sectional surveys. Nor does citing papers on TV.

Moggraider
07-31-2006, 06:11 PM
Heaping on more insults will not get you out of the deep hole you've dug yourself into.

I did not insult you, though you may feel insulted.

Psychologists frequently use non-experimental survey data to try to shed light on the real-world effects of gaming, amongst other questions...

Of course they do. Everyone knows this.

The correlations thus far produced from this approach are not evidence that gaming causes violence...

The laboratory experiments are.

No, survey evidence used by social scientists is typically not collected in a laboratory.

Survey evidence can be collected in a college campus lab or departmental building, by mail, in a college class itself, or in a number of other locations and through methods both online and in person. My point was that survey data is not the "natural experiment" or "real-world data" that you keep harping about, so that there seemed to be in your mind two categories of data. (Though we do have real-world longitudinal data from TV studies that does not fall in your favor).

By "lab studies" I mean experimental evidence generated in a laboratory.

The phrase "lab study" does not mean "experimental evidence." You can quote me on that. Your rather idiosyncratic definition simply doesn't apply to the rest of the world. Observational studies can still be "lab studies," for example.

Amazing. Just amazing. Are you now just completely giving up the argument that there is compelling evidence that video games cause violence?

Again, you have misrepresented my beliefs. I stated from the beginning that it's clear that they cause violence only in children, as the vast majority of studies attest, and as the meta-reviews and analyses confirm. You keep pretending I believe they cause violence unilaterally, and that is simply dishonest. I only suspect this, as do researchers.

I have repeatedly noted that we cannot conclude that because aggressive behavior increases in a laboratory setting that real-world violence is causally affected by gaming. That's why, as I've noted, lots of psychologists turn to non-experimental data on real-life outcomes to try to shed light on this issue. Obviously (well, to almost everyone) I am not suggesting experiments done in laboratories are in general worthless. Nor have I denied that there is evidence from controlled laboratory experiments showing short-term increases in aggressiveness, and I have already said as much.

You have never explained why you feel that laboratory studies are less valid for studying videogames and violence than they are for studying any other subject in psychology. Is it because you are a gamer, and you simply want to fervently believe that they do not do so? That seems to be the case. You repeatedly cast doubt on the overwhelming evidence for videogames causing violence in children simply because the studies mostly take place in a laboratory. That is irresponsible. You ignore the well-known fact that the Hawthorne effect changes the behavior of subjects in more socially-acceptable ways, not less socially-acceptable ways.

As all the meta-analyses state, the sum total of the research on videogames and violence points strongly in the direction that games do cause violence in children. Again, the evidence is equivocal only for older age groups.

Look again at the post above with bolded quotes from a number of reviews of this literature. See how they're all saying pretty much exactly what I've been saying -- that there is no compelling evidence that video games cause violence outside of a laboratory?

The quotes you cite speak of the field as a whole, not solely in regards to children. This is because the evidence regarding children is quite strong. If you were to focus on the single age group, the sole group about whom I made a strong claim from the beginning, you would lose, and you know it. That is why you continue to straw man my position. It's quite pitiful. I have already provided a litany of quotes in regard to this group, and the papers in this thread are in agreement with me in regards to this group.

Here, I'll play the trump card, again, and ask you to cite even one paper presenting evidence on the causal effects of video games on violence outside of a laboratory. Citing studies using experiments done in laboratories doesn't count. Nor does citing studies estimating correlations from cross-sectional surveys. Nor does citing papers on TV.

Gee, do you have any more types of papers to categorically rule out? But, because you asked, Lynch (2001) - The effects of violent video game habits on adolescent aggressive attitudes and behaviors - does provide empirical support for Anderson's GAM. So does Bushman and Anderson's 2001 paper. It looks like you lose. Will you leave us alone now?

I am sorry that you personally feel that the longitudinal findings on TV and its proven effects on violence and aggression will not apply to videogames. I certainly can't think of any good reason to believe this for myself. For example, the non-interactive cutscene portions of videogames have everything in common with TV. They are TV. You should concede that at least these parts of games, by virtue of amounting to being TV content, have been longitudinally proven to cause an increase in violent behavior, because they are television (and in fact, are often much more violent than most TV programming). With such a concession, you lose. Again.

skedastic
07-31-2006, 06:59 PM
My point was that survey data is not the "natural experiment" or "real-world data" that you keep harping about, so that there seemed to be in your mind two categories of data.

You merely emphasize, as if more emphasis were needed, that you are not familiar with the methods scientists use to tackle these questions. Nor, apparently, did you read, or at least vaguely understand, any of the studies I cited as examples of ways to make causal inferences from non-experimental data. For example, the Stinebrickner paper uses survey data on real-world outcomes, exploiting a natural experiment to make causal inferences.

Again, you have misrepresented my beliefs. I stated from the beginning that it's clear that they cause violence only in children, as the vast majority of studies attest, and as the meta-reviews and analyses confirm.

Absolutely surreal. Look, yet again, at the quoted sections of the reviews above. How on earth do you manage to read these papers and somehow come to the conclusion that they're suggesting we know video games cause real-world violence in any age group? What part of the numerous variations of "causation has not yet been established" do you not understand?

You have never explained why you feel that laboratory studies are less valid for studying videogames and violence than they are for studying any other subject in psychology.

And you have never explained why all these psychologists keep using data that wasn't generated in a laboratory to study these questions. I suppose since you've just learned that not all of these studies came from laboratory experiments, you haven't had much time to think about it. Perhaps you could actually read some of these studies and find out why psychologists think that we need to use other data to figure out the real-world effects of gaming. Perhaps then you could mass email them with accusations they're pseudo-scientific wrong-headed lying liars (you wouldn't be insulting them, even if they feel insulted).

Is it because you are a gamer, and you simply want to fervently believe that they do not do so?

Oh, obviously it is. For example, when I said that I wouldn't be surprised if it turned out that there is actually a causal effect, but I agreed with the other scholars that we ought to wait for more evidence before we draw conclusions, what I meant was screw all this science and stuff, games R kewl, right?' It's likely that all the psychologists and other researchers who've noted that we don't have enough evidence yet are also horribly biased gamers, lying through their teeth to, uh, defend gaming, or something. You've discovered the conspiracy.

Gee, do you have any more types of papers to categorically rule out? But, because you asked, Lynch (2001) - The effects of violent video game habits on adolescent aggressive attitudes and behaviors - does provide empirical support for Anderson's GAM. So does Anderson and Bushman's 2001 paper. It looks like you lose.

It's like a traffic accident that just keeps going on and on and on. The love of all that's holy, get this through your head: CORRELATION DOES NOT IMPLY CAUSATION. I "categorically rule out" lab experiments and correlational studies, and Moggraider responds by citing one lab experiment and one correlational study! I think that all of this is just so far over his head he simply has no idea what any of these studies actually did or imply. The Lynch 2001 study estimates correlations between various measures of aggression and video game playing. Let's see what they say about what we can conclude about causality from those correlations:

Are
young adolescents more hostile and aggressive because they expose themselves to media violence, or
do previously hostile adolescents prefer violent media? Due to the correlational nature of this study,
we can not answer this question directly.[...] Although additional experimental and longitudinal
research is clearly needed, it is hoped that youth, parents, and educators can begin to use the results
of this research to modify video game habits.

Notice these irresponsible lying liars are noting that they have not provided compelling evidence of causation and, although they guess that much of the correlation is causal, they note that more research and better data are "clearly needed" before we draw conclusions over causality.

Anderson and Bushman 2001 describes lab experiments, not real-world outcomes. Everyone except our favorite earnest undergraduate here agrees that we cannot draw conclusions about how gaming affects outcomes in the real-world from these experiments. That's why Anderson and his coauthors have in other subsequent work turned to survey data on real-world outcomes.

I am sorry that you personally feel that the longitudinal findings on TV and its proven effects on violence and aggression will not apply to videogames. I certainly can't think of any good reason to believe this for myself.

Well, I suppose all those psychologists are just wasting their time collecting data on video gaming and various outcomes, doing analysis, writing studies, all that silliness. They should have just pointed to the somewhat related literature on television. Or maybe there are reasons, discussed in some of these papers, why the effect of video games may differ from that of television.

That said, I am curious which studies you think "prove" television causes violence. I am not saying there are none, but I am certainly not going to take your word for it, in part since you don't know what "cause" means.

Incidentally, the argument that gaming causes violence because of cut scenes: hilarious. I am glad to see that you have a sense of humour.

Moggraider
07-31-2006, 10:09 PM
You merely emphasize, as if more emphasis were needed, that you are not familiar with the methods scientists use to tackle these questions. Nor, apparently, did you read, or at least vaguely understand, any of the studies I cited as examples of ways to make causal inferences from non-experimental data. For example, the Stinebrickner paper uses survey data on real-world outcomes, exploiting a natural experiment to make causal inferences.

I misunderstand the methods? I'm not the one saying "lab study" means "experimental evidence."

I'm also not the one flat-out lying about what the papers I cite say.

The results on violence and gaming are, in fact, so inconsistent that at least one meta-analysis concludes they are more consistent with a "catharsis effect" in which violent people exhaust their violent tendencies through gaming rather than actual violence,

Griffiths, M. (1999) "Violent Video Games and Aggression: A Review of the Literature," Aggression & Violent Behavior 203.

Such evidence suggests that at a theoretical level, there is more empirical evidence supporting social learning theory than catharsis theory—particularly in younger children.

Absolutely surreal. Look, yet again, at the quoted sections of the reviews above. How on earth do you manage to read these papers and somehow come to the conclusion that they're suggesting we know video games cause real-world violence in any age group? What part of the numerous variations of "causation has not yet been established" do you not understand?

Look, yet again, at the quotations I have provided, and at the meta-reviews which you yourself posted, all of which indicate strong evidence for the finding that games cause violent behavior in children. If there was this much evidence on this or any other age group in most any other subject, something other than the gaming we all treasure, I bet you wouldn't be raising such a stink, and would rather accept what the studies are showing. I treasure gaming, but I don't get all cognitively dissonant when facts stare me in the face. Is it really so hard to believe that children's behavior is easily affected?

And you have never explained why all these psychologists keep using data that wasn't generated in a laboratory to study these questions.

You are merely answering a question with a question, because there is no valid reason to believe that lab studies are less valid for answering the question re: violence and videogames than they are for studying other elements in psychology.

It's like a traffic accident that just keeps going on and on and on. The love of all that's holy, get this through your head: CORRELATION DOES NOT IMPLY CAUSATION.

What? Correlation is entirely capable of implying causation. It just doesn't equal it.

I "categorically rule out" lab experiments and correlational studies, and Moggraider responds by citing one lab experiment and one correlational study!

Could this be because you have no right to dismiss centuries-old established methods of study with the wave of a hand?

Anderson and Bushman 2001 describes lab experiments, not real-world outcomes. Everyone except our favorite earnest undergraduate here agrees that we cannot draw conclusions about how gaming affects outcomes in the real-world from these experiments. That's why Anderson and his coauthors have in other subsequent work turned to survey data on real-world outcomes.

Let's say, for the sake of argument, that you're a completely biased summarizer, and see what another scientist in the field has to say about this.

"Anderson and Bushman's (2001) meta-analysis suggests that violent video games, in fact, do influence aggressive behavior, aggressive effects, aggressive cognition, and physiological arousal." - Kirsh 2001

"Violent video games negatively influence socio-emotional functioning during the adolescent period." - Kirsh 2001

"The increases in physiological arousal, aggressive cognitions, and hostile affects that follow violent video game play should interact with one another to negatively bias internal state variables." - Kirsh 2001

That said, I am curious which studies you think "prove" television causes violence. I am not saying there are none, but I am certainly not going to take your word for it, in part since you don't know what "cause" means.

I am supposed to do your research for you? Look at the papers you yourself cited. They contain textual and bibliographical references to these studies. I pointed them out in this very thread already, and I even did so immediately after you posted the relevant paper. You should read a thread carefully before you hastily jump to reply. I don't even have to prove anything; all I have to do is provide evidence, and then your mantra is wrong.

Incidentally, the argument that gaming causes violence because of cut scenes: hilarious. I am glad to see that you have a sense of humour.

I see that you have no viable answer to my contention. What is it about cutscenes that would make their impact different from that of TV? Can you provide any strong arguments or, more preferably, hard evidence regarding this? How do you think the heightened violence of game cutscenes affects violent behavior, relative to the often tamer, or usually tamer, violence seen on TV? Can you cite any papers that prove cutscenes are not TV? TV has proven longitudinal effects on behavior, so why not cutscenes?

worm
08-01-2006, 12:41 AM
This is a pretty clear case of two guys who are desensitized to one another's posts.

skedastic
08-01-2006, 01:11 AM
I'm also not the one flat-out lying about what the papers I cite say.

Well, I don't know that, Moggraider. You are flat-out wrong about what the papers say. I am willing to give you the benefit of the doubt and assume that you are confused instead of lying. Nonetheless, I find it increasingly difficult to explain how you manage to maintain a position which is not only wrong, but can be seen to be wrong by anyone simply by reading the quoted sections of the meta-analyses.

Look, yet again, at the quotations I have provided, and at the meta-reviews which you yourself posted, all of which indicate strong evidence for the finding that games cause violent behavior in children.

This is a good example. There is some evidence that video games cause increases in some measures of short-term aggression in controlled experiments. There is no compelling evidence that video games cause violence in the real world. I think I have been rather clear on this point. The papers are rather clear on this point. I no longer understand how you can continue to argue it.

Is it really so hard to believe that children's behavior is easily affected?

You continue to argue as if I am claiming that there is no effect of gaming on violence; apparently I "win" if that's the case. I am suggesting, rather, that we don't know yet one way or the other. One can believe that something is true whilst acknowledging that there is no compelling evidence that it's true. That is essentially what many of the papers argue: your difficulty seems to be in realizing that beliefs, even very plausible beliefs, are not in and of themselves compelling evidence.

You are merely answering a question with a question, because there is no valid reason to believe that lab studies are less valid for answering the question re: violence and videogames than they are for studying other elements in psychology.

I never suggested lab studies are not "valid," I noted that the domain of outcomes which lab studies can speak to is limited. As noted in many of the papers, lab studies in this context are limited to very short term effects of artificial gaming sessions on measures of aggression, whereas what we want to get at is the effect of real-world gaming on actual violence. For that reason, scholars, including psychologists, often turn to survey data. Both approaches are "valid." They do not provide evidence on the same questions.

What? Correlation is entirely capable of implying causation. It just doesn't equal it.

The meaning of "A implies B" is "if A is true, then B is true." If two events are correlated, it does not follow that one causes the other. So we say that "correlation does not imply causation." Only in special cases and after imposing particular models can we infer causation from correlation. In the case at hand, noting that people who tend to play games tend to be more violent does not imply that the first causes the second: even if there is no causal relationship between games and violence, we might plausibly find such correlation.

I again have to register dismay that I have to explain this to you. "What?" is not at all a good response to the remark that correlation does not imply causation. I hope you haven't taken a semester of mathematical statistics yet.

Let's say, for the sake of argument, that you're a completely biased summarizer, and see what another scientist in the field has to say about this.

"Anderson and Bushman's (2001) meta-analysis...

These quotes refer to lab outcomes. See above where all those nice psychologists quite clearly note that the causal effect of gaming on real-world violence has yet to be determined?

I am supposed to do your research for you?

If you want anyone to take you seriously, yes. If you make a claim about what the scientific evidence says, you better be able to cite it.

In my brief search, the couple of recent papers I skimmed, and the papers you cited, suggest that the evidence on television, like that on gaming, is only at best suggestive on the question of causation. Of course, I may be wrong, but a cite would be nice (I am certain you are aware that longitudinal data, when used correctly, mitigate but do not generally solve problems inferring causation).

It is true that the paper I read with some care on TV and violence was a much more sophisticated and well-executed piece of empirical work than any of the studies on video games and violence, which are pretty Mickey Mouse.

Moggraider
08-02-2006, 12:19 AM
Well, I don't know that, Moggraider. You are flat-out wrong about what the papers say. I am willing to give you the benefit of the doubt and assume that you are confused instead of lying. Nonetheless, I find it increasingly difficult to explain how you manage to maintain a position which is not only wrong, but can be seen to be wrong by anyone simply by reading the quoted sections of the meta-analyses.

The sections of the meta-analyses that I quoted all support what I say. The sections you have quoted do not state that there is no compelling evidence that videogames cause violence. That is your unfounded assertion. Rather, they state that their isn't enough evidence yet for them to be sure on the whole. Judging from the flurry of research on the subject, we must be able to infer that there is *some* compelling evidence on the whole.

This is a good example. There is some evidence that video games cause increases in some measures of short-term aggression in controlled experiments. There is no compelling evidence that video games cause violence in the real world. I think I have been rather clear on this point. The papers are rather clear on this point. I no longer understand how you can continue to argue it.

There is _lots_ of evidence that games cause short-term aggression increases in the laboratory, not just "some." The question again is, why are you so quick to say that these findings won't transfer to the real world at all? Do you feel that way about other subfields of psychology regarding their findings in laboratories? I don't think you do, and that would make your position inconsistent.

You continue to argue as if I am claiming that there is no effect of gaming on violence; apparently I "win" if that's the case. I am suggesting, rather, that we don't know yet one way or the other. One can believe that something is true whilst acknowledging that there is no compelling evidence that it's true. That is essentially what many of the papers argue: your difficulty seems to be in realizing that beliefs, even very plausible beliefs, are not in and of themselves compelling evidence.

Again, the meta-reviews state that the vast majority of studies on children point to increased aggression.

I never suggested lab studies are not "valid..."

You have stated that "Laboratory experiments cannot provide compelling evidence on this question." In other words, you were saying they're invalid.

I noted that the domain of outcomes which lab studies can speak to is limited. As noted in many of the papers, lab studies in this context are limited to very short term effects of artificial gaming sessions on measures of aggression, whereas what we want to get at is the effect of real-world gaming on actual violence. For that reason, scholars, including psychologists, often turn to survey data. Both approaches are "valid." They do not provide evidence on the same questions.

Since when is short-term violence not real-world violence, anyway? Are you saying findings like Lynch's of increased physiological arousal and norepinephrine levels due to videogames are artifacts of the laboratory environment? I would find this hard to believe. Anderson's college student studies use non-gamers just as well as gamers, and both show effects based solely on their experimental category.

The meaning of "A implies B" is "if A is true, then B is true."

imply

v 1: express or state indirectly [syn: connote]

Source: WordNet ® 2.0, © 2003 Princeton University

con·note Audio pronunciation of "connote" ( P ) Pronunciation Key (k-nt)
tr.v. con·not·ed, con·not·ing, con·notes

1. To suggest or imply in addition to literal meaning

Source: The American Heritage® Dictionary of the English Language, Fourth Edition

Oh, and since you asked, I'm actually all done with my major.

If you want anyone to take you seriously, yes. If you make a claim about what the scientific evidence says, you better be able to cite it.

Hold on there, hoss. Earlier in this very thread, you were referring to diagrams from papers nobody else had talked about, and when I asked you for the citation, you told me to go swim in a lake. Well, to that, I say go fish. Find the paper yourself.

skedastic
08-02-2006, 07:07 AM
The sections of the meta-analyses that I quoted all support what I say. The sections you have quoted do not state that there is no compelling evidence that videogames cause violence. That is your unfounded assertion. Rather, they state that their isn't enough evidence yet for them to be sure on the whole.

I see. So rather than "no compelling evidence" instead "there isn't enough evidence to be sure." You got me there, guy.

[ snip: funny attempt to find out what a technical word means by reading lay definition in a dictionary, other silliness. ]

Hold on there, hoss. Earlier in this very thread, you were referring to diagrams from papers nobody else had talked about, and when I asked you for the citation, you told me to go swim in a lake. Well, to that, I say go fish. Find the paper yourself.

So apparently there is no compelling evidence that TV causes violence either. (And: for the n-th time, I did cite that paper. You just couldn't find it by "hitting ctrl-F over and over." Those are some pretty keen research skills.)

Anyways, I find it very unlikely that there is anyone who even vaguely takes you seriously anymore, and I am not being paid to teach you this stuff. So I will now get back to designing this cool experiment I have in mind where I'm going to have undergraduates bring consoles to a lab, then see if I can get them to beat the living crap out their roommates. I will report back with my findings.

Flowers
08-02-2006, 08:32 AM
Because you're a lawyer, not doing research in political science?
Because it's an art.

GregB
08-02-2006, 08:33 AM
So I will now get back to designing this cool experiment I have in mind where I'm going to have undergraduates bring consoles to a lab, then see if I can get them to beat the living crap out their roommates. I will report back with my findings.

Maybe you can supply them with an inflatable Bobo the Clown to punch if they begin to feel aggressive.

Hans Lauring
08-02-2006, 10:14 AM
Anyways, I find it very unlikely that there is anyone who even vaguely takes you seriously anymore, and I am not being paid to teach you this stuff. So I will now get back to designing this cool experiment I have in mind where I'm going to have undergraduates bring consoles to a lab, then see if I can get them to beat the living crap out their roommates. I will report back with my findings.

Aww shucks, it's like a good show ending early. And I was really getting into it.

"What do you mean I'm insulting you? You lying son of a bitch!"

Destarius
08-02-2006, 11:02 AM
Can someone slow him down by finding the crime statistics for the last 10 years? Video games are getting more popular; violent videogames are leading the way. Violent crime is down - way down. Explain.

Your concept of proof is certainly different from mine.

Better enforcement.

MikeSofaer
08-02-2006, 11:32 AM
Judging from the flurry of research on the subject, we must be able to infer that there is *some* compelling evidence on the whole.

.....

imply

v 1: express or state indirectly [syn: connote]

Source: WordNet ® 2.0, © 2003 Princeton University

con·note Audio pronunciation of "connote" ( P ) Pronunciation Key (k-nt)
tr.v. con·not·ed, con·not·ing, con·notes

1. To suggest or imply in addition to literal meaning

Source: The American Heritage® Dictionary of the English Language, Fourth Edition

This thread is like watching someone try to beat his way through plexiglass with his head.

skedastic
08-02-2006, 11:57 AM
It's even funnier than it first appears when you ask Mr. Google about "heritage dictionary imply" and observe how Moggraider edited the dictionary definition to try to make it consistent with his now vastly entertaining insistence that correlation implies causation. Contrast the actual definition (http://www.yourdictionary.com/ahd/i/i0061700.html) with Mogg's quoting of it. Say, Mr. Undergraduate at the Fourth Best University in the Country, didn't anyone tell you that fraud is considered to be naughty?

imply

v 1: express or state indirectly [syn: connote]

im·ply Listen: [ m-pl ]
tr.v. im·plied, im·ply·ing, im·plies

1. To involve by logical necessity; entail: Life implies growth and death.

2. To express or indicate indirectly: His tone implied disapproval. See Synonyms at suggest. See Usage Note at infer.

3. Obsolete To entangle.

The American Heritage ® Dictionary of the English Language, Fourth Edition

MikeSofaer
08-02-2006, 12:11 PM
It's even funnier than it first appears ...Did you notice the way he uses the word "infer" in the formal logic sense earlier in the same post in which he professes ignororance of the formal logic usage of "imply"? Because that's my favorite bit.

The fact that he was arguing "if enough people are looking for it, it must be there" is only gravy.

Podunk
08-02-2006, 02:45 PM
Moggraider: F+

Met_K
08-02-2006, 02:48 PM
I hate to point this out but you guys are arguing with someone who a) hangs out with Jose Liz, b) probably has zero experience living in the real world, and c) is a graduate student (I believe he said so).

You expect logic from this person? Come on, folks. I expect he'll be spearheading a study in twenty years which says that the imagination is the leading cause of violence in 15-24 year old males. When was the last time a pasty white kid, who's probably never seen the ghetto and is now in an institution of higher learning trying to get a degree in an ultimately wonky field ("The science of thought!"), right?

Unicorn McGriddle
08-02-2006, 03:31 PM
Jose Liz knows a grad student? I always assumed that his only friend was some kind of sad little pet, like a turtle or a goldfish. Which he would constantly tell tales of his future success.

Moggraider
08-02-2006, 11:52 PM
I see. So rather than "no compelling evidence" instead "there isn't enough evidence to be sure." You got me there, guy.

The two phrases do not mean the same thing, and you're the only professor I've seen who phrases it as there being no evidence (if you are indeed a professor). Also, as I've stated over and over and have demonstrated here, the evidence for children is pretty much as strong as it can be at this point.

So apparently there is no compelling evidence that TV causes violence either.

As already mentioned, I pointed out the paper as soon as you posted it. You have now gone into the territory of lying about your own field. But because you are lazy, I will tell you that a reference is in the Van Eewyk paper.

"What do you mean I'm insulting you? You lying son of a bitch!"

I have already quoted directly from the paper skedastic was lying about. It was the part where he stated that catharsis theory was supported over social learning, when in fact the paper stated the exact opposite.

It's even funnier than it first appears when you ask Mr. Google about "heritage dictionary imply" and observe how Moggraider edited the dictionary definition to try to make it consistent with his now vastly entertaining insistence that correlation implies causation. Contrast the actual definition with Mogg's quoting of it. Say, Mr. Undergraduate at the Fourth Best University in the Country, didn't anyone tell you that fraud is considered to be naughty?

A word has multiple definitions, and there is not just one dictionary in the world. "Imply" does not necessarily mean a logical entailment, as the dictionary shows. It can also mean "to suggest." If you considered what I did a distortion, you'd be wrong. Even if it was a distortion, which it isn't, it would certainly be less of one than equating the phrase "lab studies" with "lab experiments," which leaves out observational studies and other methods. You'd be hard pressed to find a dictionary that gives you that one.

To Met_K: Not that it's any of your business, but I'm not a grad student, I'm not white or pasty, I'm not rich (my family is in fact quite poor, and I've actually been homeless in the past, if that's "real world" enough for you), and I don't hang out with Jose Liz. I'm not even the one who's working for a career in a "wonky field!" That's skedastic! If anyone else has any more erroneous characterizations, please keep them to yourselves.

I don't support the actions of politicians who grandstand about videogames and could even end up hurting the industry. Hell, I've hated them ever since I was a mature-game-buying little kid myself. But with the studies on children as unified as they are, it is irresponsible to state, as skedastic has done, that there is no evidence that games cause violence. The information he has trotted out for you does not speak to the level I have been talking about from the beginning: children only.

To close with quotes from the Van Eewyk paper:

"Dill and Dill concluded that "the preponderance of the evidence from the existing literature suggests that exposure to video game violence increases aggressive behavior..."

"Griffiths concluded that "the one consistent finding is that the majority of the studies on very young children—as opposed to those in their teens upwards—tend to show that children do become more aggressive after either playing or watching a violent video game."

"Sherry also found that more recent games, which contain human characters engaging in violence, registered greater effect sizes than games with more abstract violence."

extarbags
08-03-2006, 12:04 AM
As someone who took biology in ninth grade, I have to say that I think you have a neurological disorder.

Moggraider
08-03-2006, 12:07 AM
You'd be wrong.

extarbags
08-03-2006, 12:14 AM
Oh. Well damn, I guess that's what happens when people make wild assertions based on infinitesimal levels of relevant education.

Moggraider
08-03-2006, 12:16 AM
A B.A. is infinitesimal now? I know college graduation rates are increasing, but not that much... Wait, why am I even taking this kind of criticism from a police academy cadet?

extarbags
08-03-2006, 12:19 AM
Oh, I'm sorry, last week you were a psych major, but I have been out of town for a few days. Congratulations on getting your degree during that time.

Edit: also, "police academy cadet?" I don't get it... is that a crack about how you're going to make more money than me once you graduate?

Met_K
08-03-2006, 12:21 AM
Wait, why am I even taking this kind of criticism from a police academy cadet?

All you do is try to redirect people away from how wrong you are with stunts like this. You are exactly the type of person I had you pegged as. Get over yourself.

You lose at life. Have a good one.

Moggraider
08-03-2006, 12:22 AM
Thanks, Met. I think I will.

Oh, I'm sorry, last week you were a psych major, but I have been out of town for a few days. Congratulations on getting your degree during that time.

Edit: also, "police academy cadet?" I don't get it... is that a crack about how you're going to make more money than me once you graduate?

It's a "crack" about how you have no right to criticize others' levels of education. But I'm glad that the future disparity looms on your mind. And, because you asked: I have already completed my major.

extarbags
08-03-2006, 12:28 AM
It's a "crack" about how you have no right to criticize others' levels of education. And, because you asked: I have already completed my major.

Man, you really just do not get it. I'm not criticizing your level of education. I'm criticizing you. This is about you. I've only ever known one other poster here who routinely missed the point by that much; I don't think I need to tell you his name.

Further, the last I heard you were twenty years old, which is pretty young to be a college graduate, and you were going around referring to psychology as your "major," which is something people don't typically do after they get their degree. You also mentioned that you were going to be a lawyer, in which case, if you had finished your undergraduate program, one would think that you would identify "law" as your major. But again, all of this information is as of last week, so bring me up to speed if things have changed.

But beyond that, please explain why you called me a "police academy cadet." I really have no idea why you chose that phrase specifically, and I can't decifer what you're getting at here.

Moggraider
08-03-2006, 12:32 AM
All those questions, and I'm the dense one? What I'm saying is, a college degree's worth of classes amounts to something more than an infinitesimal level of education. I could take no relevant classes my last year, and still graduate. And that's what I'm going to do.

Yes, I'm younger than I should be. I skipped a year of school.

"Law" is not a major.

extarbags
08-03-2006, 12:35 AM
All those questions, and I'm the dense one?

Well, you didn't manage to answer them.

"Law" is not a major.

Really (http://www.worldwidelearn.com/online-education-guide/criminal-justice/law-major.htm).

Met_K
08-03-2006, 12:37 AM
Dude, I totally get it now. You're going to be one of those crazy defense lawyers on Law & Order that argue that free speech or video games or the internet caused their clients to rape, maim and kill. I get it. I totally get it.

That totally explains the psychology degree, BD Wong.

Wait, hold on, how the hell much does a BA even know about big boy psychology? Shouldn't leave the posturing to the grad students, the theses to the masters and the dissertations to the doctorates, Master P?

Moggraider
08-03-2006, 12:38 AM
Dude, I totally get it now. You're going to be one of those crazy defense lawyers on Law & Order that argue that free speech or video games or the internet caused their clients to rape, maim and kill. I get it. I totally get it.

Looks like you've got me pegged, Met_K!

Wait, hold on, how the hell much does a BA even know about big boy psychology?

Enough that we get a piece of paper that says we know something.

Well, you didn't manage to answer them.

Really (http://www.worldwidelearn.com/online-education-guide/criminal-justice/law-major.htm).

Yes. Really.

"Acceptance to a law school not only requires a bachelor's degree and a high score on an entrance exam..."

The bachelor's degree is what 99% of people are talking about when they say "major."

Met_K
08-03-2006, 12:40 AM
No, you are incredibly incorrect. No graduate goes around referring to their degree as a "major," they call it a degree. Thus, you would be a man with a bachelor's in psychology or who majored in psychology.

Past tense, my friend. You majored in psychology. You aren't a psych major - present tense. You were a psych major - now you're a bachelor's graduate. People who refer to being majors in something most often never graduated.

extarbags
08-03-2006, 12:41 AM
Okay, well, you're honestly the only person with a bachelor's degree I've ever heard refer to their degree that they already have as their "major." But then again, you're also the only person to accuse me of being a police academy cadet, which you seem unwilling to explain. Perhaps I'm too dense. Please psychoanalyze and defend me in a court of law when you're done making more money than me, a lowly police cadet (?).

Moggraider
08-03-2006, 12:44 AM
No, you are incredibly incorrect. No graduate goes around referring to their degree as a "major," they call it a degree.

When did I say I've graduated?

Met_K
08-03-2006, 12:45 AM
Screw you. I'm done with this game, you tart. You and Toutsuite can go off into from-the-behind troll heaven for all I care.

Sincerely,
Theodore Rex DX's Medulla Oblongata

Edit: Fixed so it's kid-friendly so I don't get banned. I'm trying to play nice but some of you people are crazy.

extarbags
08-03-2006, 12:46 AM
Oh, I see. So when I said you had an infinitesimal amount of education in the field, and you said

A B.A. is infinitesimal now?

you didn't mean that you actually, like, had a B.A. Is that it? So like, you just brought it up for the hell of it?

Moggraider
08-03-2006, 12:46 AM
Fuck you. I'm done with this game, you cunt.

How joyous it's been playing it with you.

Oh, I see. So when I said you had an infinitesimal amount of education in the field, and you said

you didn't mean that you actually, like, had a B.A. Is that it? So like, you just brought it up for the hell of it?

Why do I have to explain this again? A major is a set number of classes with certain restrictions on what you can take, all of the classes being in your chosen field. The rest of the classes you take in college are generally of your own choosing. I've completed my major. That means that for my senior year, I could take nothing but bullshit like ballroom dancing 101 and The History of Met_K's Mother's Sexual Conquests, and I would still get my degree. I have freedom now because I worked hard the other three years of my stay. Can we stop talking about me now?

Unicorn McGriddle
08-03-2006, 12:48 AM
As a community member in good standing in extarbags' field of study (ninth grade biology, where I got a gentleman's C), I must concur with his assessment in post #150.

extarbags
08-03-2006, 12:52 AM
Why do I have to explain this again? A major is a set number of classes with certain restrictions on what you can take, all of the classes being in your chosen field. The rest of the classes you take in college are generally of your own choosing. I've completed my major. That means that for my senior year, I could take nothing but bullshit like ballroom dancing 101 and The History of Met_K's Mother's Sexual Conquests, and I would still get my degree. I have freedom now because I worked hard the other three years of my stay. Can we stop talking about me now?

Ok, well congratulations on the degree you'll have in a year. In the meantime, I guess I'll get back to... um... police school.

Moggraider
08-03-2006, 12:53 AM
Wonderful.

Met_K
08-03-2006, 12:57 AM
jose liz Jose Liz Jose Liz! JOSE LIZ JOSE LIZ!

Podunk
08-03-2006, 07:45 AM
Man, this is the funniest shit I've seen in ages. Please, don't let it stop.

skedastic
08-03-2006, 10:43 AM
Can't resist. Keeps getting funnier. Plus, I wouldn't want to disappoint podunk. I note in passing that I'm quoting lots of psychologists because it's fun to contrast Mogg's claims regarding what they think with what they actually think. But even if most psychologists really did think that the extant evidence "proves" causation, they'd be wrong. In order to conclude that correlation between gaming and violence is evidence of causation, we'd have believe that there are no other variables affecting both gaming and violence and that violence cannot cause game playing, that is, "reverse" causation. There could be reverse causation, and it is highly plausible that there are third variables affecting both violence and gaming. Therefore, correlation between gaming and violence doesn't "prove" causation, nor does it imply causation, nor does it even "suggest" causation, and anyone claiming otherwise -- politician, psych major, others -- is simply wrong.

As already mentioned, I pointed out the paper as soon as you posted it. You have now gone into the territory of lying about your own field. But because you are lazy, I will tell you that a reference is in the Van Eewyk paper.

The paper you are presenting as "proving" that television causes violence is the Van Eewyk paper titled, "Video games and real-life aggression: review of the literature." Not quite following you there.

A word has multiple definitions, and there is not just one dictionary in the world.

Free tip to help you in your studies: don't write on a test that correlation implies causation. Your idiot professor, lacking your vast intellect, will probably take marks off.

To close with quotes from the Van Eewyk paper:

For those keeping score at home, this is a paper which concluded:

In conclusion, current research evidence is not
supportive of a major concern that violent video
games lead to real-life violence. However, well controlled
studies of adolescents are lacking. Also,
this conclusion might change as more research is
conducted on more recent and increasingly realistic
games.

That is, it says more or less exactly what I've been saying and flat-out contradicts Moggraider. So how does Mr. Moggraider manage to quote them claiming support for his idea that it's been "proved" that video games cause violence? Let's see!

"Dill and Dill concluded that "the preponderance of the evidence from the existing literature suggests that exposure to video game violence increases aggressive behavior..."

We could first note "aggressive behavior" isn't "real-world violence." But our spidey sense should go off at the ellipsis there. Let's see what Moggraider has cut out!

“the preponderance of the evidence
from the existing literature suggests that exposure to
video game violence increases aggressive behavior
. . . However, the paucity of empirical data, coupled
with a variety of methodological problems and inconsistencies
in these data, clearly demonstrate the

Oops. It's actually just another in the long list of comments that the existing evidence is weak, followed by a call for more research so we can actually get some good answers to these questions.

"Griffiths concluded that "the one consistent finding is that the majority of the studies on very young children—as opposed to those in their teens upwards—tend to show that children do become more aggressive after either playing or watching a violent video game."

Hmmm. Let's see what Griffiths goes on to say:

To briefly conclude, the
question of whether video games promote aggressiveness cannot be answered at present
because the available literature is relatively sparse and conflicting, and there are many
different types of video games which probably have different effects.

"Sherry also found that more recent games, which contain human characters engaging in violence, registered greater effect sizes than games with more abstract violence."

That's in reference to controlled experiments in a lab. The paper continues:

Including our review, each of the four reviews
identified major gaps in the existing research. These
gaps include a lack of randomized, well-controlled
research, particularly with adolescents; a lack of
research on possible long-term effects; and a lack of
research on subsets of individuals, possibly with
other risk factors, who may be more susceptible to
negative effects of game-playing. Three of the four
reviews (including our own) found that the current
evidence suggests a role that is either limited in size
or scope. Thus, at present, it may be concluded that
the research evidence is not supportive of a major
public health concern that violent video games lead
to real-life violence. However, this conclusion might
change as more research is conducted on more recent
and increasingly realistic games.

Surprise! It's simply another recitation of the consensus view that we don't yet have enough evidence to draw firm conclusions on causality. Everyone but Moggraider is a lying, intellectually dishonest, pseudo-scientific, irresponsible gamer suffering from cognitive dissonance ! It's a massive conspiracy to commit scientific fraud! Googling for the conspirator "van eenwyk" turns up more evidence of this heinous plot:

“Little kids do imitate video games, but
the question is whether it leads to real-world
violence,” said Juliet Van Eenwyk,
PhD, who studies the issue for the
Washington State Department of Public
Health, Seattle.

What!? Our very own Moggraider has already declared that we don't need to study real-world violence, we know everything we need to know from controlled experiments on aggression. It is "nonsense" to wonder if aggression observed in a lab translates into real-world violence, indeed, even asking the question means you think everything ever done in a lab is bullshit! What is this Van Eenwyk person thinking? And she goes on:

New, unpublished research from the
Washington State Healthy Youth Survey
supports the notion. The 10% of
6th, 8th, 10th, and 12th graders who
say they play video games for 4 or more
hours each day were more likely to be
aggressive than their less thumbstressed
classmates.
But when controlled for depression
and parental involvement, that association
disappeared, said Van Eenwyk.
“I think that playing games 4
hours a day is a symptom, and these
kids need help with social, emotional,
and family issues rather than regulation
for games.”

Translating: there is a correlation between playing video games and aggressiveness. But when depression and parental involvement are statistically held constant, that correlation disappears. That is evidence that kids who play tons of games have other problems and lack parental supervision, which leads both to gaming and aggression. The gaming is a symptom, not a cause, in Van Eenwyk's phrasing.

Recall that Van Eenwyk's paper is the one that Moggraider thinks "proves" gaming causes violence.

This article also makes reference to a landmark court case in which,

So far, the game developers are winning.
In June 2003, the US Eighth Circuit
Court of Appeals struck down the
St Louis law barring minors from purchasing
violent games. In the ruling, the
judges agreed with a brief filed by 33
media scholars and psychologists, writing
that the “conclusion that there is a
strong likelihood that minors who play
video games will suffer a deleterious effect
. . . is simply unsupported by the
record.”

These are very disturbing quotes indeed. Since Moggraider knows we've "proved" that video games cause violence, all these researchers claiming the evidence isn't in yet --- many noting their skepticism of the proposed causal mechanism while doing so --- are therefore part of the conspiracy. Or maybe they're all just ignorant, having failed to make sure that their opinions are consistent with those of a third-year pysch major. I mean, in only one year he'll "have a piece of paper which says he knows something." Let's hope "something" includes what we can or cannot conclude about causation from correlation. Mr. Moggraider may wish to find out what researchers mean by "causation" as a first step.

if you are indeed a professor

I am not going to similarly question Moggraider's claim that he's a 20 year old psychology undergraduate. I have no reason to suspect that he isn't really still learning the basics, with no real understanding of the professional literature. I do wonder a bit, though, about his claim that he's been reading social science papers for years and years. Perhaps he reads them like Kevin Kline's character in A Fish Called Wanda reads philosophy.

Arbit
08-03-2006, 11:20 AM
http://resources.wizards.com/Magic/Cards/TE/en-us/Card4833.jpg
It's true, the evisceration of one mogg[raider] always cheers up the rest.

Edit: as fitting as that flavor text is, I really wish he had chosen Mogg Fanatic instead of the raider. The name, picture, and mechanic (exploding and doing one point of damage) are much more fitting.

http://resources.wizards.com/Magic/Cards/TE/en-us/Card4832.jpg

McBain
08-03-2006, 11:49 AM
Of course violent media has a psychological effect on young boys.

Weren't ANY of you ever kids?

Shit.

I remember, one summer, taking the handle off of the big rectangular broom in the garage and beating the crap out of stuff with my god damn BO STAFF. Nothing to do with having just watched Teenage Mutant Ninja Turtles. Really.

You think I would have been doing that if I had just finished watching FRIED GREEN TOMATOES instead?

My question is, how's this a problem? Violence is part of the everyday world, and not always a bad one. What's going to happen to future generations of men if they grow up in some imagined utopian "sunshine and little green elves" daycare fantasyland? Where those who don't witness violence first-hand are never even exposed to it?

The fact that nobody even wants to raise these questions for debate anymore signals to me that the bus has already left -- and it's driving straight off a fucking cliff.

Bill Dungsroman
08-03-2006, 12:48 PM
OOoohhh, I wouldn't be too quick to judge someone else's educational level after earning a BA in Psychology, considering I got one in my spare time. Not exactly a PhD in Rocket Surgery, that.

Moggraider
08-03-2006, 01:15 PM
Bill, extarbags was the one who said something about it first.

Bill Dungsroman
08-03-2006, 01:27 PM
Bill, extarbags was the one who said something about it first.
Than be the bigger man, or something. Point is, it's pretty myopic to be getting into side arguments with people about their and your education level, while in the midst of banging your head against skedastic's respective educational level. Y'know?

extarbags
08-03-2006, 01:30 PM
Well, after sleeping on it, I have to admit that when he called me a "police academy cadet," that was a pretty good zinger.

Moggraider
08-03-2006, 01:31 PM
Than be the bigger man, or something. Point is, it's pretty myopic to be getting into side arguments with people about their and your education level, while in the midst of banging your head against skedastic's respective educational level. Y'know?

Well, I can't just stand mute, thanks to the Qt3 mob that loves to pop up. This is actually the last place I'd've expected class-based anger to emerge.

Jason McCullough
08-03-2006, 02:35 PM
Can we get a zombie thread declaration?

Arbit
08-03-2006, 02:54 PM
Well, I can't just stand mute, thanks to the Qt3 mob that loves to pop up.
This is exactly why the Qt3 mob pops up. Live and learn!

Moggraider
08-03-2006, 06:17 PM
The paper you are presenting as "proving" that television causes violence is the Van Eewyk paper titled, "Video games and real-life aggression: review of the literature." Not quite following you there.

Because you can't glean the thing I was pointing out from the title of a paper, you assume there's nothing there at all? I hope your research isn't usually that superficial. Look at the Huesmann papers I had already pointed out on Page 5.

Free tip to help you in your studies: don't write on a test that correlation implies causation. Your idiot professor, lacking your vast intellect, will probably take marks off.

I did not ask for tips from you, nor do I need them.

Translating: there is a correlation between playing video games and aggressiveness. But when depression and parental involvement are statistically held constant, that correlation disappears. That is evidence that kids who play tons of games have other problems and lack parental supervision, which leads both to gaming and aggression. The gaming is a symptom, not a cause, in Van Eenwyk's phrasing.

This is the first strong point you've made (well, not made, but rather "found") this entire thread. Bravo. This is new, meaningful information that actually says something and introduces an explanation for what's really happening. The rest of the thread is you over and over minimizing the positive results we do have. Get it through your head: I had expressed a strong opinion about only one age cohort, the one group we do have suggestive results for.

Recall that Van Eenwyk's paper is the one that Moggraider thinks "proves" gaming causes violence.

Incorrect, for multiple reasons: A) I was speaking in regards to only one group, 2) I was using it only to cite other study results, and 3) looking back through the thread, I realize I never said flat-out "prove." Rather, that has just been your straw-man of me the entire time.

Tom Chick
08-03-2006, 06:29 PM
Well, after sleeping on it, I have to admit that when he called me a "police academy cadet," that was a pretty good zinger.

I think what happened is that he confused you with Marcus nee ranvarian, who is a police academy cadet. And, frankly, I think being a police academy cadet is pretty cool. Not just because that's what Charlie's Angels were before they were Charlie's Angels.

-Tom

Moggraider
08-03-2006, 06:31 PM
Ooh, you're right. Sorry, extarbags. Please switch my insult to "collections officer."

extarbags
08-03-2006, 07:38 PM
I'm not that anymore either. The slander you're looking for is, I believe, "deadbeat."

worm
08-03-2006, 08:20 PM
At first I didn't care, I was desensitized.

MikeSofaer
08-03-2006, 08:24 PM
At first I didn't care, I was desensitized.I played the games where killing hookers was romanticized.

Unicorn McGriddle
08-03-2006, 08:32 PM
But then I spent so many days stocking my hideout with AKs
And I grew strong
And now a hooker's on my dong.

Bill Dungsroman
08-03-2006, 11:19 PM
Ooh, you're right. Sorry, extarbags. Please switch my insult to "collections officer."
So I guess among all that phat miggedity-mack education in psychology that evidently allows one to learn the intricate mental innerworkings of your fellow man, you failed to learn to fucking pay attention or even bother to check your notes (ie Qt3 Search Function/Archives).

Moggraider
08-03-2006, 11:21 PM
Yes, you're right, of course. Sorry.

Hans Lauring
08-04-2006, 12:59 AM
So I guess among all that phat miggedity-mack education in psychology that evidently allows one to learn the intricate mental innerworkings of your fellow man, you failed to learn to fucking pay attention or even bother to check your notes (ie Qt3 Search Function/Archives).

I think we established on page two of this thread that Moggraider wasn't paying particular attention to his chosen field of study - why on earth would he pay attention to what a bunch of nobodies on a board without (almost) M.A.'s did for a living?

The important point is that he's taking the time and arguing with people beneath him and his school has really cool podiums. Skedastic might be a professor (and a lying bastard), but before he posts a picture of him on a podium, his facts and knowledge is to be disregarded.

Moggraider
08-04-2006, 01:33 AM
Are you still mad because I like to bash the PSP? Sheesh. The PSP blows, get over it. I did. And I'm still waiting for you to cite that study, by the way, like I cited mine. I'd like to read it.

John Merva
08-04-2006, 01:39 AM
But then I spent so many days stocking my hideout with AKs
And I grew strong
And now a hooker's on my dong.

Hans Lauring
08-04-2006, 02:10 AM
Are you still mad because I like to bash the PSP? Sheesh. The PSP blows, get over it. I did. And I'm still waiting for you to cite that study, by the way, like I cited mine. I'd like to read it.

Since Prof. SKedastic has allready repeatedly showed your inability to read and understand the studies you cite yourself, I don't really see the point - but this being in Swedish (http://www.medieradet.se/upload/kall_pa_porr.pdf)probably won't make it any harder for you to misquote.

The Danish one I originally talked about apparently hasn't been published online yet (if ever). But since you claim causation has been proved and everybody else says no or the jury is still out - the burden of proof is still yours and not mine. I merely pointed to one dissenting opinion on something you claimed as fact.

And classy to bring past PSP discussions into this. It's always good to attack the poster when your arguments are weak ("you're a liar and probably not even a real professor", "you're just a police cadet", "you're mad because I don't like the PSP" - grow the fuck up)

Moggraider
08-04-2006, 07:25 AM
What, I didn't back up my porn study claim? No, I did. Look at post 50. And again, the evidence he was lying I quoted directly to show, and the mention of occupation was just a response in kind.

skedastic
08-04-2006, 09:15 AM
Look at the Huesmann papers I had already pointed out on Page 5.

You mean, the Huesmann papers using correlational methods unable, as they note themselves, to establish causation?

We don't have compelling evidence that television causes violence.

3) looking back through the thread, I realize I never said flat-out "prove." Rather, that has just been your straw-man of me the entire time.

Well, let's see. Moggraider claimed:

Adults may or may not be more likely to act violently because of violent videogames (chances are they are more likely), but children are proven to be more likely to do so, which is what we should be worried about.

After he was ridiculed for abusing the word "prove," he went on to assert that the standard for "proof" in psychology is really low,

As for the use of the word "proof," well, things are being proven as far as psychological studies can possibly take them. Like I said, they can't be akin to laws of physics.

Several pages later, he forgets about these remarks and exhibits his usual search skills trying to find out if he did or didn't claim that the effect was "proven." Also, the research has been demoted from "proving" games cause violence to only suggesting as much:

Get it through your head: I had expressed a strong opinion about only one age cohort, the one group we do have suggestive results for.

I guess I do have trouble wrapping my head around an incoherent and inconsistent argument. My bad.

This is the first strong point you've made (well, not made, but rather "found") this entire thread. Bravo.

It may very well be the first point made in this thread that you actually understood. Obviously I should not have used jargon (correlation, reverse causation, confounders, unobserved heterogeneity) to make my point and instead tripled the verbosity in order to dumb things down to your level. I wrongly assumed you would have been exposed to these ideas during your studies.

I have found this thread useful for two reasons. First, I am going to seriously consider doing some research on this topic, as the existing research is really badly done and the opportunity for cheap highly citeable papers abounds. Second, I have learned that it's possible to get a BA in psychology and have absolutely no idea whatsoever how to interpret statistical results. That is really quite disturbing.

Moggraider
08-04-2006, 09:31 AM
You mean, the Huesmann papers using correlational methods unable, as they note themselves, to establish causation?

We don't have compelling evidence that television causes violence.

What happened to your talk about wanting longitudinal data? The Huesmann papers control for things like SES and sex, and still find an effect. I'll stop there, though. I don't have to argue about this with you. Because the psych community is already citing the paper without reservation, as they did in the paper you posted, your protests mean very little.

After he was ridiculed for abusing the word "prove," he went on to assert that the standard for "proof" in psychology is really low,

You were straw-manning me for saying proof before I even said it, and I said it what, once? I do take it back, though, if that's what you want.

It may very well be the first point made in this thread that you actually understood. Obviously I should not have used jargon (correlation, reverse causation, confounders, unobserved heterogeneity) to make my point and instead tripled the verbosity in order to dumb things down to your level. I wrongly assumed you would have been exposed to these ideas during your studies.

Your points about better study methods being needed had merits, but the SES/bad-parenting finding is mountains more important, and would have been a better criticism to bring up in the first place, though it seems that a lot or all of the research in this thread, you did on the spot. Which is fine, of course, but you actually gave me most of the material I was quoting.

I have found this thread useful for two reasons. First, I am going to seriously consider doing some research on this topic, as the existing research is really badly done and the opportunity for cheap highly citeable papers abounds.

Awesome. I look forward to seeing what you can contribute, though I don't know how you'll fit this research into your usual job description for departmental approval.

Second, I have learned that it's possible to get a BA in psychology and have absolutely no idea whatsoever how to interpret statistical results.

That's quite the overstatement.

skedastic
08-04-2006, 11:11 AM
Because the psych community is already citing the paper without reservation, as they did in the paper you posted, your protests mean very little.

No, you just don't understand what the paper is saying nor how people are citing it. Why do you do this to yourself? Here you are, once again, leaping straight into an argument about how to interpret multivariate statistical evidence from survey data. Why does it never seem to occur to you to ASK what these results are and what they mean rather than simply assuming whatever guess you make off the top of your head must be correct?

The papers in question correlate youth TV viewing with adult aggressiveness. Finding a positive correlation, even after holding constant certain basic socioeconomic measures and even after differencing (ie, looking at how changes in aggressiveness are related to TV rather then levels of aggressiveness), is at best suggestive of a causal effect. Remember how holding depression and parental involvement constant made the video game and violence correlation go away? In the present case the estimates only reflect causal effects if this condition holds: There are no influences on changes in aggression which are correlated with childhood TV viewing, after holding basic socioeconomic controls and TV viewing constant. We have no reason to suppose that there are no such influences. Perhaps, for example, kids who will go on to develop psychological problems which increase violent tendencies later in life also tend to be kids who prefer to watch lots of television. Then we will find the sorts of correlations reported in this study even if there is no causal effect of childhood television viewing on violence. As the authors note,

In addition, the examination of the “third variable” hypotheses in this study, as always, is limited by the actual third variables included in the
study.

In their terminology, influences on changes in violence other than childhood TV viewing are "third variables," and this comment acknowledges that the analysis fails to reflect causality if there exist such variables which they cannot include in the statistical models. They go on to note this limitation again in the penultimate section.

This analysis is better than the cheap crappy little studies on games and violence because it uses better statistical models and exploits differencing to remove "third variables" which do not change over time, but nonetheless these methods do not establish causality, they can only suggest it, albeit more credibly than the analogous cross-sectional methods. The fundamental problem is that people who watch lots of TV when they're kids are not a random sample of kids: we can statistically reduce that problem, but other methods (and better data) are needed to actually overcome that problem.

I am continually amazed that you insist on arguing about these issues. Get ahold of a graduate level textbook on regression analysis. Find the sections on longitudinal data. You will discover that, surprisingly enough, what I'm saying can be formally proved mathematically without much difficulty, and is, for that reason, entirely uncontroversial. I do not expect that fact to prevent you from continuing to contest the matter.

You were straw-manning me for saying proof before I even said it.

Supposing that were true, it will still mean that your claim that you never said the effect was "proven" is wrong. But it isn't true. As anyone can verify for themselves, you claimed in post 43 that the effect of games on violence had been "proved." The first time I use a variant of the word "prove" is in post 108.

Your points about better study methods being needed had merits, but the SES/bad-parenting finding is mountains more important, and would have been a better criticism to bring up in the first place,

I DID immediately point out over and over that confounding influences imply that the correlation between games and violence does not establish causation. You just didn't understand what I meant. The depression and parental involvement result is merely one example of such confounders. No, it is not more important than better methods: even after controlling for basic sociodemographics we have every reason to believe that other confounders, and reverse causation, are present. Put another way: even after controlling for depression and parental involvement and other observed determinants of violence, the correlation between gaming and violence cannot be interpreted as reflecting causation or the lack thereof. Better data and better methods could do everything this study did and also provide compelling evidence on causation.

Awesome. I look forward to seeing what you can contribute, though I don't know how you'll fit this research into your usual job description for departmental approval.

Moggraider's willingness to lecture me what economists do is noted and filed for its future entertainment value. And, no, I do not need anyone's "approval" to undertake a research project.

Moggraider
08-04-2006, 12:56 PM
Blah blah blah.

The paper was cited as showing TV has longitudinally-validated effects. TV viewing is a strong predictor even when SES, sex, intellectual ability, and parenting are controlled for. This is the kind of thing Anderson and you were asking for regarding videogames. Let's not backpedal now that the evidence has been brought out for TV. The case just doesn't get much stronger than this.

From the last paragraph of the paper:

"Although some questions remain to be resolved about the exact extent of the effect of observed violence on aggressive and violent behavior and its importance relative to other causal factors, the current study provides compelling additional evidence that habitual exposure of children to violence in the media (or in the real world around them) does have lasting effects on their propensity to behave aggressively and violently."

So, it's compelling evidence. Your unresearched, uncorroborated claim that there is no compelling evidence that TV causes violence is wrong. We could dream up magical third variables all day, but I think that Huesmann covered the most obvious, powerful ones. This is something you essentially admit.

I DID immediately point out over and over that confounding influences imply that the correlation between games and violence does not establish causation.

You stated over and over that correlation does not mean causation, just like anyone can, but you did not provide evidence that there was actually a third variable underlying everything, which would have had weight. Hard evidence will always be a stronger argument (though I do wonder why that study you mentioned was never published).

Moggraider's willingness to lecture me what economists do is noted and filed for its future entertainment value. And, no, I do not need anyone's "approval" to undertake a research project.

You're able to take on research outside of your field and get grants and other funding for it without any kind of oversight from higher-ups at all? Wow. Or will this be some kind of pet project? Wouldn't you rather have institutional support? Isn't there some kind of review board at your university?

Squirrel Killer
08-04-2006, 01:09 PM
But then I spent so many days stocking my hideout with AKs
And I grew strong
And now a hooker's on my dong.
Seriously dude,
Unicorn sucks at haikus
Five seven five, bitch

Moggraider
08-04-2006, 01:11 PM
That's a continuation of the parody of the song "I Will Survive," not a haiku.

Squirrel Killer
08-04-2006, 01:20 PM
That's a continuation of the parody of the song "I Will Survive," not a haiku.
That's minus five points
Moggraider's fucked up again

skedastic
08-04-2006, 01:27 PM
This is the kind of thing Anderson and you were asking for regarding videogames. Let's not backpedal now that the evidence has been brought out for TV.

To repeat, I said longitudinal evidence would be better than what we have now regarding video games. I am not "back pedalling."

The case just doesn't get much stronger than this.

As far as you know. But then, you don't know, do you?

the current study provides compelling additional evidence that habitual exposure of children to violence in the media (or in the real world around them) does have lasting effects on their propensity to behave aggressively and violently."

This is misleading. A better editor would have forced the authors to write, "...does have long lasting associations with proposensity to behave...." Remember that whole "causation" and "correlation" thing? One way to badly phrase an observed correlation is to refer to as an "effect," thus making it sound causal, thus confusing undergraduates. But, hey: do continue to argue with math. I'm sure math is wrong and you're right.

You stated over and over that correlation does not mean causation, just like anyone can,

Funny, this remark came as shocking news to someone contributing to this thread.

but you did not provide evidence that there was actually a third variable underlying everything, which would have had weight.

I see. So the Moggraider Scientific Method holds that we assume that there are no confounders and thus that correlations are causal until such time as we prove otherwise. Everyone else seems to start by noting that because we can't rule out confounders, so the multiple explanations for the observed evidence preclude causal inference. I assume you simply don't understand that these "third variables" are anything that affects violence other than video games. Are you really going to try to argue that we know a priori that video games are the only factor affecting violence? Do you think a good scientist would simply assume that gaming is the only factor which affects violence until such time as he finds evidence otherwise?

Edit: The large multidisciplinary empirical literature on determinants of violence is evidence that these "third variables" exist. For example, see all these factors which predict violence (http://www.rand.org/news/Press/566-572.ellickson.pdf)? Those are all "third variables" which confound the relationship between gaming and violence. So, since you are impervious to logic, there's your evidence.

You're able to take on research outside of your field and get grants and other funding for it without any kind of oversight from higher-ups at all? Wow.

Look at the examples of research by economists in this thread. One is on the effects of studying on GPA. One is on the effect of changes in police force sizes on crime rates. And the last is on the effects of Fox News on voting patterns. Perhaps it surprises you to learn that these are topics being studied by economists. But perhaps you should have figured out that economics is a broader field than you previously imagined and not have been so surprised when an economist noted he might write something on gaming and violence.

That said, yes, I can choose to do research on whatever I please. Of course if I wanted to get funding for this particular project, like anyone else, I would have to apply for it. No, there is no "review board at my university" that I have to go to for permission to undertake a project.

Moggraider
08-04-2006, 01:43 PM
This is misleading. A better editor would have forced the authors to write...

Ho! You're qualified to edit these guys now? Slow down, champ.

I see. So the Moggraider Scientific Method... blah blah blah

What's the skedastic scientific method? Continuing to posit magical third variables when evidence contrary to your belief is staring you in the face? I think you should back down from this one. It isn't even your field.

Are you really going to try to argue that we know a priori that video games are the only factor affecting violence? Do you think a good scientist would simply assume that gaming is the only factor which affects violence until such time as he finds evidence otherwise?

I've already conceded videogames; you should concede TV. I guess we both learned a lesson in this thread.

No, there is no "review board at my university" that I have to go to for permission to undertake a project.

I see. There is at mine. Does that mean your way is better?

Met_K
08-04-2006, 02:30 PM
How's it going Jose Liz?

skedastic
08-04-2006, 02:34 PM
Ho! You're qualified to edit these guys now?

Yes, I am qualified to "edit these guys." In particular I am certainly qualified to point out bad phrasing of statistical results.

What's the skedastic scientific method? Continuing to posit magical third variables when evidence contrary to your belief is staring you in the face?

Now anything that affects violence other than video games is "magic," is it? You really can't think of anything that might affect violence other than video games? Other people can, and measure their effects statistically.

(Oops! See what I did there? I used "effects" to refer to correlations. See how easy it is to slip up and use language which someone who doesn't know better may intepret as invoking causation?)

You also need to be told, obviously, that it is in principle impossible using the methods in this paper to determine if such "third variables" exist. But do go on: I'm sure you're champing at the bit to quote some passage in the paper you don't understand, and we will again be back to Math vs. Moggraider. My money's on math.

Well, that's good, we're finally getting somewhere. But I am afraid I am not going to let you off the hook by agreeing with your equally uninformed assessment of the literature on television and violence. The question is: Can you learn from your mistakes?

I see. There is at mine. Does that mean your way is better?

Faculty at your university have to go to a review board to get papers they want to write "approved?" No, Mr. Moggraider, unless perhaps you go to some fringe religious "university" or something, they don't. At least, I have never heard of any university having such a board, and I cannot imagine faculty putting up with it. Perhaps you have misunderstood the role of ethics boards or something similar.

Squirrel Killer
08-04-2006, 02:40 PM
Ho! You're qualified to edit these guys now? Slow down, champ.
Thou hast much to learn
University profs edit
Mogg's mind is now blown.

Unicorn McGriddle
08-04-2006, 03:48 PM
Hey Squirrel Killer
Not everything's a haiku
Plus you are a dork.

Kitsune
08-04-2006, 04:02 PM
モッグレイダーは

ー狐

Unicorn McGriddle
08-04-2006, 04:20 PM
But how authentic!

Moggraider
08-04-2006, 05:28 PM
But I am afraid I am not going to let you off the hook by agreeing with your equally uninformed assessment of the literature on television and violence.

All hail skedastic, king of social scientists! Evidence is compelling only if he says so! Rah! Rah!

No, I don't think so. I've provided compelling evidence re: TV and violence, and your word alone to the contrary means very little.

Podunk
08-04-2006, 06:01 PM
There's this giant hole, and if I really squint, I can kind of make out Moggraider waaaaaay down at the very bottom. Man, is he ever going to town with that shovel!

Moggraider
08-04-2006, 06:29 PM
I don't follow you, Podunk. I conceded videogames once skedastic provided evidence of an underlying third variable that is the real cause of violence. Now though, he's railing against the strong, longitudinally-validated TV findings with nothing but his own words. I'm a reasonable man; if research shows something, I believe it.

extarbags
08-04-2006, 06:50 PM
I'm a reasonable man

Boy would I ever love to put this in my sig.

Met_K
08-04-2006, 07:19 PM
I thought this was like Candyman. How many times do I need to say your name before you come out of the mirrors and bore me to death? Jose Liz Jose Liz Jose Liz Jose Liz! JOSE LIZ!

skedastic
08-04-2006, 11:39 PM
No, I don't think so. I've provided compelling evidence re: TV and violence, and your word alone to the contrary means very little.

Math v. Moggraider: place your bets.

Moggraider
08-04-2006, 11:41 PM
I see you have no hard evidence to counter the paper that's been presented. I am amused by your insistence on your own infallibility. You will understand if I believe the published research before I believe you.

Athryn
08-04-2006, 11:49 PM
This thread has just come down to who gets the last word, imo.

skedastic
08-05-2006, 12:34 AM
I see you have no hard evidence to counter the paper that's been presented. I am amused by your insistence on your own infallibility. You will understand if I believe the published research before I believe you.

You mean, the published research which repeatedly acknowledges that it does not establish causation? Why do you persist in these arguments? Do you vaguely recall insisting for five pages that it's been proven that video games cause violence based on your (mis)understanding of the research? How'd that turn out? Don't you ever learn anything from your mistakes?

Consider the regression of (y_t - y_{t-t}) on X_t, where y_t is an outcome measured at time t (such as violence), X_t is a stochastic vector of covariates, and u_t is a disturbance term. Write the model,

y_t - y_{t-1} = X_t\beta + u_t

where \beta is a vector of parameters to be estimated. I claim:

1. The OLS estimate of \beta is consistent only if u_t is mean independent of X_t.

2. OLS estimates of the equation above cannot in principle provide evidence on whether u_t is mean independent of X_t.

If 1. and 2. are true, then the estimates in the paper Moggy thinks "proves" TV causes violence are only evidence of causation under the restrictive and probably false conditions I list in the post above, and it is in principle impossible to test those conditions from the estimated model. If so, then Moggraider is -- yet again -- completely and demonstrably wrong. Alternately, I am wrong -- foolishly thinking I am "infallible" for, you know, being able to trivially prove these results with "math," along with every advanced textbook on regression analysis ever written -- and Moggraider is correct. Perhaps we will never know who's right: could be everyone who knows what they're talking about, could be Moggraider. It's a mystery for the ages.

Moggraider
08-05-2006, 12:59 AM
You're backpedalling again, considering your earlier pining for longitudinal studies. Now, you have Huesmanns' on TV. Do you have any compelling evidence of relevant underlying third variables that Huesmann did not control for? Didn't he control for relevant factors? Do you know of any research or authorities which would agree with you that Huesmann's findings are weak? Is your next claim simply going to be that Huesmann is a hack like Anderson?

You claimed, without having done any research on your own and without even looking at the paper I told you about, that there was no compelling evidence on this issue. I proved you wrong. Does it really surprise you that you've made a mistake now, too? The underlying mathematical model teaches the concept in idealized textbook conditions and does not apply here. You know that. Can you even show us a social science study where the model applies flawlessly?

Unless you have any significant published findings to present, I urge you to let it go.

skedastic
08-05-2006, 08:28 AM
Do you know of any research or authorities which would agree with you that Huesmann's findings are weak? Is your next claim simply going to be that Huesmann is a hack like Anderson?

I didn't say that the findings in this paper are "weak." Recall what I said was that this paper is much better than the papers in the literature on games. I do agree with the authors of the paper that they face a serious limitation in that they unobserved heterogeneity is likely to be biasing their estimates, a problem which precludes them from being able to recover causal estimates. Any and every authority would agree with me on that, including the authors of the paper, because it's elementary. This isn't my opinion versus yours, Moggraider, it's you yet again railing against well-known, mathematically demonstrable results.

"He who refuses to do arithmetic is doomed to talk nonsense." -- John McCarthy

The underlying mathematical model teaches the concept in idealized textbook conditions and does not apply here. You know that. Can you even show us a social science study where the model applies flawlessly?

And here Moggraider is again lecturing me on how to interpret statistical evidence. The question isn't whether the "model applies flawlessly," it's whether the estimates can plausibly be interpreted as causal. If there are determinants of violence other than TV and the controls in the paper, then we have reason to believe that the correlation obtained is not a good estimate of the causal effect we are trying to find. And of course there are many such "third variables" -- that's, again, why they note "In addition, the examination of the “third variable” hypotheses in this study, as always, is limited by the actual third variables included in the study." They really need not have added this sentence as everyone who works with this sort of model is well aware of the problems unobserved heterogeneity pose. We do not start with the assumption that there are no variables affecting social outcomes other than those we are able to collect in our datasets. Instead, we acknowledge that that there of course many such variables and keep it mind when interpreting results. Longitudinal data can be used to mitigate but generally not solve these problems. Everyone who works with these methods is well aware of these facts.

Recall that anything that affects adult violence which isn't held constant in the analysis is a potential confounder, a variable which might lead to an observed correlation between childhood TV and adult violence even if there is no causal relationship between those two variables. Anything that happens between the time of childhood data were collected and the time the adult data were collected is such a confounder, in addition to many contemporaneous factors. Among the innumerable factors which are possibly confounders, I will simply note that the study could not control for maternal employment nor for adolescent mental health problems. Remember the two "third variables" the other study focused on which convinced you that maybe the correlation between games and violence may not be causal? Well, those two variables are left out of the current paper. What have we learned about the effect those variables may have?

Now, Moggy, what I've said is entirely uncontroversial. If you took even an undergraduate course in any flavour of statistics which focuses on regression analysis, you would learn as much. Screaming that I believe I'm "infallible" and the "king" and all this other silliness makes as much sense as screaming the same insults at your grade school teacher for her insistence that six times seven is 42 when you feel it really ought to be 65.7. These are just well-known facts, and your indignation at having them brought to your attention does not make them false.

But please, do go on. You seem to enjoy having both of your feet rammed into your mouth at all times, and I wouldn't want to spoil your (and everyone else's) fun.

Moggraider
08-05-2006, 09:16 AM
I provided compelling evidence on the issue of TV and violent behavior. The paper authors even say it's compelling, and the authors of the study you posted cite them without reservation, because these findings are much stronger than the videogame studies, something you admit. Your response was to copy a formula from your textbook and claim that proof is not established. When you define "proof" as something that can't be demonstrated, sure, nothing is proven. But you fail to provide research to the contrary, and I hope others can see how weak of an argument copying a formula from a textbook is.

Among the innumerable factors which are possibly confounders, I will simply note that the study could not control for maternal employment nor for adolescent mental health problems.

This is a weak criticism. There is no parallel to the study on games you mentioned earlier, because this paper controls for so much. The other paper made those two small factors seem so important because that's all they controlled for (at least it seems that way from the small excerpt from the paper you deigned to post). When you control for a large factor like SES, you're really getting at multiple effects. And anyway, the father's employment was controlled for, and an effect was still found. Father's employment is more variable than mother's employment: most everyone has a mother, but not necessarily a father. Also controlled for were overall SES and parental educational levels. So was IQ, which has strong effects on resilience re: mental health.

The parents' education and the child's intellectual ability as assessed during Waves 1 and 2 of the study were introduced into the longitudinal structural models as control variables with paths to every other variable. The resulting models for male and female participants both provide a good fit, as indicated by the nonsignificant chi-square statistics and the low RMSEs.6 Furthermore, for both male and female participants, the paths from childhood exposure to TV violence to adult aggression remain significant, whereas the paths from early aggression to adult TV-violence viewing are not significant. For males, the difference between the two paths is larger with the control variables introduced; for females, the difference is smaller. In both cases, the most plausible conclusion is that childhood exposure to TV violence is stimulating an increase in adult aggression regardless of the initial aggressiveness of the child, the intellectual ability of the child, or the educational background of the parents. For females, there may well also be a stimulating effect of childhood aggression on later TV-violence viewing, but it is a weaker effect. For males, there is little indication of such an effect. Finally, when father's occupation was again substituted for parents' education as a measure of SES, the key paths in the structural models did not change.

A longitudinal structural modeling analysis of the directionality of the effects suggested that it is more plausible that exposure to TV violence increases aggression than that aggression increases TV-violence viewing. These structural modeling analyses also demonstrated that the effects were not simply a consequence of lower SES children or less intellectually able children both watching more violence and being more at risk for aggressive and violent behavior. The structural models show that for both boys and girls, habitual early exposure to TV violence is predictive of more aggression by them later in life independent of their own initial childhood aggression, their own intellectual capabilities, their social status as measured by their parents' education or their fathers' occupations, their parents' aggressiveness, their parents' mobility orientation, their parents' TV viewing habits (including violence viewing), and their parents' rejection, nurturance, and punishment of them in childhood. Furthermore, the structural models suggest that being aggressive in early childhood has no effect on increasing males' exposure to media violence as adults and only a small effect for females.

Again, you said "there is no compelling evidence," and you're wrong. Whining about it won't get you anywhere. Unless you have some study or authority to corroborate your claims, just let it go. I hope this teaches you to stop making blanket statements, sight unseen of any literature on a subject. It appears that maybe even professors of econometrics don't know everything.

extarbags
08-05-2006, 09:27 AM
And anyway, the father's employment was controlled for, and an effect was still found. Father's employment is more variable than mother's employment: most everyone has a mother, but not necessarily a father.

Haha, love this logic. "We don't need to control for this, because we controlled for this other thing that's more variable!"

Moggraider
08-05-2006, 09:44 AM
They controlled for overall SES too, though.

skedastic
08-05-2006, 10:21 AM
Recall that the problem is that even after controlling for the stuff they are able to observe it's likely that there are other variables affecting both which kids watch TV and later violent outcomes. In the presence of such factors, the correlation between childhood and TV and violence is biased, that is, it does not reflect causation from TV to violence. We don't assume there are no such variables until someone finds one, we rather assume that there are and interpret results accordingly. That's the reason the authors should not claim this sort of correlation is compelling evidence of causality, they are wrong, and a better journal would not have let this claim stand.

Here is a very interesting and well done study on cognitive development and childhood TV (http://home.uchicago.edu/~jmshapir/tv012606.pdf). The topic isn't violence but a number of related social outcomes such as educational acheivement and labor market outcomes. Here is what to note: the authors do not just provide correlations between TV and their outcomes. In particular, in Section 3.2 ("Difficulties with Correlational Evidence") they note that even after controlling for parental education, income, and a variety of other factors, TV and cognitive development are negatively correlated, but they are skeptical that that correlation reflects causation:

Therefore, we would expect any unobserved variation in these characteristics to tend to bias an OLS regression of test scores on television viewing towards finding negative effects
of television.

Translating: On the basis of several regression models, they believe that correlations corrected for parental education and so on will tend to suggest negative effects of TV on outcomes such as educational acheivement even if TV does not cause such changes. The same characteristics which lead to kids watching lots of TV also lead to poor outcomes in terms of education, the labor market, and so on. Simply, the correlational evidence of this sort makes TV look like it's bad even if it isn't. This is exactly the sort of correlational evidence the Huessman paper presents.

The authors of this paper do not just present this correlational evidence, assume there are no "third variables," and insist that the correlation is "compelling evidence" that TV causes worse outcomes. They rather use a method which is analytically identical to the Stinebrickner paper mentioned earlier in the thread. In this case, they use variation in the timing of the introduction of television across regions in the U.S. circa 1950 to recover causal estimates. The basic idea is that kids who live in areas where TV was introduced first should have worse outcomes if TV actually causes worse outcomes.

The estimates of the models capable of recovering causal effects show that TV has basically no effect at all. The authors conclude that:

Our estimates
therefore cast significant doubt on the hypothesis that television was responsible for the post-
27 World War II declines in cognitive skills (Winn, 2002; Glenn, 1994) that Bishop (1989) links to
the productivity growth slowdown of the 1980s. Our findings also suggest that much of the recent
correlational evidence attributing negative developmental effects to childhood television viewing
may require reevaluation.

Simply, the correlational evidence of the sort in the Huessman paper presents is misleading if we want to find the causal effect of TV, and that's the question everyone is trying to answer. Data and statistical models capable of providing compelling evidence on causation (here, instrumental variables estimates exploiting variation induced by a natural experiment -- oh, and with over 300,000 observations rather than the 300 in Huessman paper) cast doubt on any such causal effect. It wouldn't be at all surprising in light of these findings that the correlational evidence on TV and violence is similarly misleading. Even if we had no such evidence that these correlations are misleading, we should still consider these problems and not claim that correlational evidence is "compelling." A better editor would never have let Huessman and colleagues make that claim, and it was frankly quite naive of them to so blatantly overstate their case.

Incidentally, notice this paper is about the effects of TV on social outcomes like cognitive development. And it was written by econometricians. I wonder if they got permission from the Study Approval Board at the University of Chicago before writing outside their field?

Moggraider
08-05-2006, 12:15 PM
Recall that the problem is that even after controlling for the stuff they are able to observe it's likely that there are other variables affecting both which kids watch TV and later violent outcomes. In the presence of such factors, the correlation between childhood and TV and violence is biased, that is, it does not reflect causation from TV to violence. We don't assume there are no such variables until someone finds one, we rather assume that there are and interpret results accordingly.

We know this, but I think that the researchers in question did their best to control for as much as they could. I mean, they even had several measures on parenting behaviors.

That's the reason the authors should not claim this sort of correlation is compelling evidence of causality, they are wrong, and a better journal would not have let this claim stand.

Are you saying that the Journal of Developmental Psychology is less academically rigorous than the journals from which you drew and draw your papers?

Translating: On the basis of several regression models, they believe that correlations corrected for parental education and so on will tend to suggest negative effects of TV on outcomes such as educational acheivement even if TV does not cause such changes. The same characteristics which lead to kids watching lots of TV also lead to poor outcomes in terms of education, the labor market, and so on. Simply, the correlational evidence of this sort makes TV look like it's bad even if it isn't. This is exactly the sort of correlational evidence the Huessman paper presents.

Granted, and of course watching a lot of TV also has that effect simply because you are watching TV instead of studying/working/etc.

The authors of this paper do not just present this correlational evidence, assume there are no "third variables," and insist that the correlation is "compelling evidence" that TV causes worse outcomes.

Neither Huesmann nor I assumed that there were no third variables. I happen to think Huesmann did a pretty good job of controlling for most everything, and so did the psychologist who cited him in the paper you posted.

They rather use a method which is analytically identical to the Stinebrickner paper mentioned earlier in the thread. In this case, they use variation in the timing of the introduction of television across regions in the U.S. circa 1950 to recover causal estimates.

You mean the whole paper wasn't just their copying a math formula out of their textbooks and then saying "QED?" Gee. I figured that if even you thought it was an adequate defense, it must hold up to very strong scientific rigor!

But thank you for doing some research on the subject, instead of expecting everyone on the board to just take you at your word. It looks like we're all demonstrating some personal growth here, if not due to our own willingness. I see your point regarding different methods of studying causation.

You will note, however, that the study you cite can't really speak specifically to the question of the induction of aggressive behaviors, considering the entirely different nature of TV programming in the 1950s. I also wonder if we can use the method of this study to examine TV's effects on violence at all, given the simultaneous and concurrent onset of national programming that has been popular for a number of decades now.

Some other relevant considerations involve the incredible homogeneity of the US population in 1950 relative to its present state, in many facets: ethnically, in terms of the divide between rich and poor, and in terms of rates of educational achievement. The study you present to us doesn't consider college enrollment at all, looking only at high school completion, which leads me to believe its measure of educational attainment is lacking. It is true that college enrollment was less common then than it is now, but I hope that you see the limitations of the authors' findings.

So, for the question of studying TV and its effects on violence in the modern day, maybe Huesmann's study is simply as good as we can get. If you look at his past work, you can see that he did do longitudinal studies on this question for a number of decades, if that means anything to you.

Incidentally, notice this paper is about the effects of TV on social outcomes like cognitive development. And it was written by econometricians. I wonder if they got permission from the Study Approval Board at the University of Chicago before writing outside their field?

Yes yes, very funny. I did mention a review board before because presumably you'd be using human participants in your future studies on videogames, if any (is this not the case?). You thought it would be funny to suggest I go to some kooky religious school. I don't, and in fact I don't think my university could be any more secular. I do also assume you'll seek funding for any studies you do undertake instead of paying out of your own pocket, an assumption which, if true, invalidates your attempt at sarcasm quite nicely.

skedastic
08-05-2006, 01:02 PM
Neither Huesmann nor I assumed that there were no third variables.

Why do you insist on doing this over and over and over -- making silly comments, then claiming you never made them when they're demonstrated to be silly? For example,

What's the skedastic scientific method? Continuing to posit magical third variables when evidence contrary to your belief is staring you in the face?

You're not merely assuming that there are no such third variables, you're mocking the very idea that there could be such variables and (erroneously) claiming that the Huessman paper somehow presents evidence that no such variables exist. You made repeated comments to the same effect.

You mean the whole paper wasn't just their copying a math formula out of their textbooks and then saying "QED?"

OK. Third time's the charm. That "formula" (which most people would refer to as an "equation") is very basic. It's the standard representation of a standard linear regression. Any semi-competent undergraduate who took a class on regression analysis and showed up to a lecture or two could write that equation down without "looking it up in a textbook." I could have written it down without copying it out of a textbook while you were still a toddler. You merely emphasize, again, how little you know about these topics when you make such remarks.

I note you make no attempt to address the point I was making: Gentzkow and Shapiro reproduce correlational results which are methodologically identical to those in Huessman's paper, note that these correlations may be misleading, and go on to show that better methods reveal that the causal effect is roughly zero despite the high negative correlations between TV and cognitive development. These results are a good example of correlations which are spurious -- they do not reflect causation -- in the context of media and social outcomes. We have no reason to believe that the analogous correlations in Huessman et al. are not also spurious. So we should not take the correlations they report as compelling evidence of causation. In light of the Gentzkow results, we should in fact be highly skeptical of any claim they reflect causation.

But thank you for doing some research on the subject, instead of expecting everyone on the board to just take you at your word.

Yeah, me and my crazy confidence in math. You also have my word that six times seven is 42. You can research that yourself if you like.

So, for the question of studying TV and its effects on violence in the modern day, maybe Huesmann's study is simply as good as we can get.

Maybe so, but, as I've noted before, simply because a given body of evidence is "as good as we can get" does not mean that it's automatically "compelling." It would mean rather that we have no good evidence on the question, and as good scientists we should remain agnostic until such time as good evidence is collected.

Anyways, I am glad we now, I hope, agree that the correlational evidence on TV and violence presented in Huessman and other similar papers does not "prove" that TV causes violence. The case that TV causes violence is only moderately stronger than the case that video games cause violence. That is: very weak indeed.

Yes yes, very funny. I did mention a review board before because presumably you'd be using human participants in your future studies on videogames, if any (is this not the case?).

No. If I write something on this topic I will do what Gentzkow and Shapiro did: I will use existing, large, nationally representative longitudinal surveys. I do not need ethics approval to use such data. I don't need anyone's approval at all.

Moggraider
08-05-2006, 01:37 PM
You're not merely assuming that there are no such third variables, you're mocking the very idea that there could be such variables and (erroneously) claiming that the Huessman paper somehow presents evidence that no such variables exist. You made repeated comments to the same effect.

Incorrect. You have shown a remarkable lack of reading comprehension in this thread. I know there are third variables, but as I have already stated, Huesmann covered the most salient ones, and the literature seems to agree. I asked you if you knew of literature that demonstrated otherwise, and you failed to present anything to that effect. Instead, you insist on positing some third variable that isn't covered that is serving as some kind of magical culprit and is the "true" cause of all this increased adult aggression. We can't disprove a negative. We can't prove that there aren't any other third variables, and as you have pointed out, Huesmann admits that there may be some, but we do our best to account for them.

You are asking us to disprove a negative, which is a basic fallacy of debate. Prove that there isn't an intangible, invisible pink unicorn in the sky whose chemical composition matches the rest of the sky! It can't be done.

OK. Third time's the charm. That "formula" (which most people would refer to as an "equation") is very basic. It's the standard representation of a standard linear regression. Any semi-competent undergraduate who took a class on regression analysis and showed up to a lecture or two could write that equation down without "looking it up in a textbook." I could have written it down without copying it out of a textbook while you were still a toddler. You merely emphasize, again, how little you know about these topics when you make such remarks.

How sad to see Qt3's most profound scientific mind degrade himself to nothing more than a condescending sophist asshole.

I note you make no attempt to address the point I was making: Gentzkow and Shapiro reproduce correlational results which are methodologically identical to those in Huessman's paper, note that these correlations may be misleading, and go on to show that better methods reveal that the causal effect is roughly zero despite the high negative correlations between TV and cognitive development. These results are a good example of correlations which are spurious -- they do not reflect causation -- in the context of media and social outcomes. We have no reason to believe that the analogous correlations in Huessman et al. are not also spurious. So we should not take the correlations they report as compelling evidence of causation. In light of the Gentzkow results, we should in fact be highly skeptical of any claim they reflect causation.

I told you I saw your point regarding inferring causation. However, as is blatantly obvious the paper can say very little or nothing as to the question of violence, considering the downright lack of violence on TV in the time period that is under study. Because the paper showed that 1950's TV made people only insignificantly less likely to graduate from high school, it is also showing that TV doesn't have any effects on violent behavior? That does not follow at all. The two questions are fundamentally different. But let's say we accept your rather questionable parallel between the two issues. The paper does "prove" (using the word in a sense that you seem to accept now that this study method has been used) that TV made 1950's students one-thousandth less likely to graduate from high school for every year of television exposure. If we accept the parallel you draw, we can conclude that violent TV would make people at least that much more violent, and your statement regarding the lack of compelling evidence is disproven.

Tell me, what would happen to the social sciences if correlational studies were not allowed at all?

Maybe so, but, as I've noted before, simply because a given body of evidence is "as good as we can get" does not mean that it's automatically "compelling." It would mean rather that we have no good evidence on the question, and as good scientists we should remain agnostic until such time as good evidence is collected.

The paper authors claim that the evidence is compelling, and the author of the paper you cited seemed to agree. I can't find any research that is pointing to some major inadequacies with Huesmann's paper; can you? Why do you always consider the papers you cite as evidence only of what you believe?

The case that TV causes violence is only moderately stronger than the case that video games cause violence. That is: very weak indeed.

Incorrect. You have failed to provide a single other researcher or authority's view on this question, and your assertion alone amounts to very little (much less, in fact, than even the laboratory experiments on violence and videogames)! The consensus so far seems to be that Huesmann's findings are strong. You have failed to show otherwise. And although one might not be able to accept Huesmann's findings as hard proof, one must say that his studies do at least suggest we should tone down the TV viewing, if nothing else. You can choose to be a stickler about the word "proof," but the literature does suggest that we act in ways that amount to proof being shown.

skedastic
08-05-2006, 02:52 PM
Incorrect. You have shown a remarkable lack of reading comprehension in this thread.

ROFL.

Instead, you insist on positing some third variable that isn't covered that is serving as some kind of magical culprit and is the "true" cause of all this increased adult aggression. We can't disprove a negative.

Yet again, your idea that we assume that there are not such "third variables" unless we have evidence that there are is just plain wrong. Attitudes, parenting, family structure, peer effects, neighbourhood effects, religious beliefs, wealth, mental health, physical health, body type, alcohol or drug abuse, marital status, past income and other labor market outcomes, and personality are things that immediately occur to me which may be "third variables" here. They are either uncontrolled altogether or partially proxied in the study. Any of these things or any of an innumerable number of other factors may affect both childhood TV viewing and adult violence. Social scientists do not just assume that there no such "magical" factors when doing these sorts of analyses. The potential that there even might be such factors is enough to preclude thinking that estimated correlations are causal. We do not assume that there are no such factors, claim causality has been "proven," and try to shout down any attempt to point that the correlations do no such thing.

The point has nothing to do with proving a negative. But please, do continue to bicker with me about regression analysis. Probably, your minutes of exposure to the subject have yielded many profound insights into why the conventional wisdom, and math, is wrong.

I told you I saw your point regarding inferring causation.

Good. The we agree, at long last, that the correlational studies on violence and on TV do not "prove," or even provide compelling evidence on, causation.

The paper does "prove" (using the word in a sense that you seem to accept now that this study method has been used) that TV made 1950's students one-thousandth less likely to graduate from high school.
If we accept the parallel you draw, we can conclude that violent TV would make people at least that much more violent, and your statement regarding the lack of compelling evidence is disproven.

You really know NOTHING at all about statistics, do you? Look at the size of the estimated effect in the Gentzkow paper. The point estimate -- our best guess -- of the effect of a year of TV viewing on the probability of completing high school is 0.002. That means that the probability of completing high school is estimated to fall by one-fifth of one percent for each year of TV viewing. Now look at the standard error for that estimate: 0.0135. That means that a 95% confidence interval for the estimated effect is (-0.0246 , 0.0285). I am sure Mr. Google will be able to define "confidence interval" for you. The estimates suggest that the causal effect is very small, zero plus or minus two percent, loosely speaking. And we fail to reject the hypothesis that the true effect is exactly zero. The estimates do not show, or "prove," that the true effect is exactly 0.002.

Tell me, what would happen to the social sciences if correlational studies were not allowed at all?

It isn't correlational studies aren't "allowed," I've written my share of them. It's rather how we should interpret such studies. In contexts such as the ones we're discussion we should not assume that we have estimated a causal relationship. You remember, "correlation does not imply causation" and all that?

The paper authors claim that the evidence is compelling, and the author of the paper you cited seemed to agree. I can't find any research that is pointing to some major inadequacies with Huesmann's paper; can you? Why do you always consider the papers you cite as evidence only of what you believe?

Uh oh. You're sticking to the "this is compelling evidence of causation" idea, huh? Look again at the Gentzkow paper: are the high negative correlations between TV and cognitive development compelling evidence that TV causes reductions in cognitive development? Following your reasoning, that is what we would have to conclude.

The "inadequacy" with Huessman's paper is that it only presents correlations and correlations do not imply causation in this context. You keep demanding I provide "authorities" to tell you stuff that you'd learn in any intermediate statistics course.

The consensus so far seems to be that Huesmann's findings are strong. You have failed to show otherwise.
And although one might not be able to accept Huesmann's findings as hard proof, one must say that his studies do at least suggest we should tone down the TV viewing, if nothing else.

You apparently "conceded" that the case video games cause violence is not compelling by casually changing "prove" to "suggest." Is this another such admission?

Moggraider
08-05-2006, 03:54 PM
No. I agree with Huesmann that he has provided compelling evidence of a causal relationship between TV and violence.

"Although some questions remain to be resolved about the exact extent of the effect of observed violence on aggressive and violent behavior and its importance relative to other causal factors, the current study provides compelling additional evidence that habitual exposure of children to violence in the media (or in the real world around them) does have lasting effects on their propensity to behave aggressively and violently."

The psychologist who cited him in the paper you posted agrees, and the consensus seems to be on his side.

I don't believe that because TV was not found to have a negative effect on half-a-century-old high school graduation rates, that that means TV does not affect aggressive behaviors. Neither do the authors of the study you cite. Unless you can provide some kind of hard evidence/research of some kind of third variable that is at the root of what Huesmann has discovered, the finding looks pretty solid. But you go and start looking into body type, marital status, and religious beliefs if you really think that will change the finding. Again, I'm not asking you to just be an asshole again; I already have extarbags and Met_K to do that.

skedastic
08-05-2006, 04:43 PM
I don't believe that because TV was not found to have a negative effect on half-a-century-old high school graduation rates, that that means TV does not affect aggressive behaviors.

Pay attention. The study showed that:

1. There is a large correlation between cognitive development and early TV viewing. That correlation persists even after holding parental education and a variety of other possible confounders constant.

2. The causal effect of early TV viewing on cognitive development is basically zero.

Obviously in this case we ought not conclude that result 1. is compelling evidence that early TV viewing causes diminished cognitive development. This is an example of a correlation which does not reflect causality.

Now substitute "violence" for "cognitive development. The Huessman study shows:

1A. There is a large correlation between violence and early TV viewing. That correlation persists even after holding parental education and a variety of other possible confounders constant.

We do not have an analog of point 2. The only evidence we have is this correlation. Should we conclude that 1A is compelling evidence that early TV viewing causes violence?

Notice that I am not arguing that because TV viewing seems to have had little or no effect on cognitive development in the immediate post-war period that there is no effect of TV viewing on modern kids. I do think, however, that the compelling demonstration that there were large spurious correlations between TV viewing and various social outcomes in that period should definitely increase our skepticism regarding claims that modern correlations reflect causality. Put another way, the Gentzkow paper shows that which kids watch lots of TV is not a random sample of kids, even conditional on a bunch of observable determinants such as parental education. Because television is not randomly assigned, we cannot conclude that correlations between television viewing and any social outcome reflect the causal effect of television.

But you go and start looking into body type, marital status, and religious beliefs if you really think that will change the finding.

Again, the burden of proof does not fall on others to show that there are "third variables," it falls on anyone claiming that some correlation implies causation to show that there aren't any such variables. If we cannot say whether there are any such variables, we should not claim that the correlation is compelling evidence of causation. All of the variables that Huessman et al. include as predictors of violence can together only explain about 10% of the variation in violence. 90% of the causes of violence are "third variables." We would have to somehow know that none of those variables are correlated with childhood TV viewing to conclude the correlation reflects causality. Since we do not know that, we cannot conclude that this correlation is compelling evidence of causality.

Suppose I calculate the correlation between the number of pairs of socks I own and the GDP of Mexico over the last twenty years. I expect that correlation will be large and positive and highly statistically significant. Do we conclude that this is compelling evidence that my sock-buying habits drive the Mexican economy? Would it be reasonable for me to insist that a skeptic provide evidence that some other "third variables" may be at work here, and if no such evidence is forthcoming, should I continue to insist that the Mexicans are at the mercy of my trips to the mall?

Suppose it were the concensus of psychologists that Huessman's correlation is compelling evidence of causation. That would simply mean that they're all wrong. Luckily, that isn't the consensus. What's been "longitudinally validated" is the correlation (sometimes misleadingly referred to as "effect") between TV and violence later in life. That doesn't mean TV causes violence later in life. The problem is, again, that Moggraider simply doesn't understand the professional literature he's trying to read.

Incidentally, Aristotle was not Belgian.

Moggraider
08-05-2006, 04:59 PM
I understand what the papers in this thread mean. I also understand that you are, again, harping on the definition of proof. However, you have, again, failed to provide evidence that Huesmann is wrong. It was done for Anderson and his videogame studies; why can't it be done for Huesmann? Could it be because his finding is in fact strong? You admitted before that longitudinal findings are convincing; you try to discredit this finding only because you made a misstep in your own words and were basically talking out of your ass without reading the literature on TV and violence. You also fail to cite any other researchers who would agree with you that the case for TV and violence is "very weak." Good luck finding that.

While Huesmann's study may not be "proof," it is "compelling evidence." You stated there was "no compelling evidence," so you are wrong. Huesmann's study is just about the best we can do, and controlling for blood type, hair color, or your sock collection would not change the finding. You may choose to continue to disagree, though it looks like the scientific community would be in disagreement with you. I know whose side I'm on. I hope you choose your future statements with more caution.

skedastic
08-05-2006, 09:22 PM
No content.

Oh well, fun's over. Move along folks, nothing to see here.

Moggraider
08-06-2006, 12:14 AM
Sunday at the Qt3-atorium!

skedastic versus the scientific community!

Who will win?

Sunday! Sunday! SUNDAY!

Glenn
08-06-2006, 12:36 AM
Moggraider is the scientific community? Can you get me grants?

Moggraider
08-06-2006, 12:37 AM
No. Just ask skedastic to study it for you; he can study anything he wants, anything he deigns to grace with his intellectual largesse.

MikeSofaer
08-06-2006, 12:39 AM
No. Just ask skedastic to study it for you; he can do anything he wants.I worry that in the wrong hands such power might be abused.

Bill Dungsroman
08-06-2006, 01:17 PM
Aw. Big nasty multi-page Qt3 blowouts aren't supposed to end this way. Matthew Gallant and I now skip through the daisied meadows, picking only the most perfect of flowers for our ladies fair, when once we'd each spat in our palms and sworn a blood oath. Alas.