Game Politics ([info]gamepolitics) wrote,
@ 2006-07-17 07:20:00
Previous Entry  Add to memories!  Tell a Friend  Next Entry
Entry tags:digg, doj, duke ferris, game revolution, legislation, research, violence

Youth Violence Waaaay Down - Games to Blame?

It was first published last summer, but a revealing piece by Game Revolution Editor Duke Ferris shows that youth violence is at historically low levels, even as politicians rush to enact laws targeting video game violence.

So why are we dredging this up now?

Duke's article, "The truth about violent youth and video games" turned up on Digg three days back, and has generated more than 1,500 votes, or "diggs," in that time. Numbers like those are a testament to how deeply this line of reasoning resonates with gamers.

Among Duke's points:

"There is no epidemic of youth violence in America.

The whole concept is a lie... Kids are not killing each other more frequently than they used to. In fact, it turns out the opposite is true.

Check out that ugly graph on the right... violent crime is at the lowest it has been in a good thirty years. For effect, I've also marked the release of the Playstation console, the first Grand Theft Auto game, the PS2 console, and the infamous GTA 3. Wow, look at those surges in violence!

Believe it or not, I got that graph... from the U.S. Department of Justice, Bureau of Statistics. All I added was the video game timeline..."

One cautionary note - a GP reader who is deeply into research and statistics offers this caveat to the Game Rev piece.

Want to talk about it? You can discuss this story via the "comments" feature (click below), or in the new GamePolitics Forums...


(Read 96 comments) - (Post a new comment)

An academic friend suggest caution...
[info]gamepolitics
2006-07-17 03:07 pm UTC (link)
A friend who is a stats whiz sent GP this:

"While I don't believe that playing violent video games causes violent crime (although some evidence suggests that exposure to media violence may be a risk factor for already at-risk children), I found Duke Ferris's graph and conclusions to be grossly misleading. To be honest, I thought it was satire when I read it last summer.

Mr. Ferris makes one of the biggest errors in statistics by not accounting for other factors that changed over the same period. In fact, I could plot the same graph showing a steady increase in youth incarceration rates beginning in the mid-1990s, but it would be equally flawed (although somehow I doubt that it would get as many diggs). The point is that an analysis of crime needs to use multivariate regression. Simply making a two-dimensional plot and attributing the subsequent drop in violent youth crime to playing video games, as some people have unfortunately done based on this graph, is simply wrong when more significant factors like economic conditions, youth incarceration, and passage of state laws that try children as adults dramatically increased over the same period. In fact, it's theoretically possible for exposure to media violence to cause a small increase in violent youth crime and yet to observe the same downward trend when these other factors have a larger and negative influence on violent crime rates. Just my two cents. "

GP's brainy bud adds: "...Please note that I am not motivated or persuaded by the unsupported claim that playing violent video games causes violent crime, but that inferring conclusions from this graph is the wrong way to examine the relationship."

(Reply to this) (Thread)
Re: An academic friend suggest caution...
[info]xlorep_darkhelm
2006-07-17 04:26 pm UTC (link)
I would consider that this graph, if anything, helps break the notion that there is a causal relationship between violent video games (or violent media in general) and violent actions. While it is, as your 'brainy bud' claims, an incomplete graph, it still makes a point -- that merely stating 'violent media causes violence', without any evidnce to support the claim, is rediculous.

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]cyn1c42
2006-07-17 06:04 pm UTC (link)
I know, however this shoots down the argument "violent games have caused an increase in juvenile violence

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]barfo
2006-07-17 06:35 pm UTC (link)
Your firend makes a very good point. I always felt the Ferris article was misleading for similar reasons, though not as well thought out. The one proper and valuable way to use this study is not to debunk the claim that video games cause violence (since neither side has any really scientifically valid evidence to support or deny the causal claim), but to refute the point that comes up in all the anti-game politicians statements that the issue is one that is very significant. Even if games were linked to causing violence, the fact that violence is so far down anyways implies that its not a big deal or a big contributing factor.

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]warriormouse
2006-07-17 07:20 pm UTC (link)
It's certainly true that the graph doesn't show a causal relationship. Graphs and statistics alone can never show causation; to show causation requires a scientific explanation of the actual method through which one factor influences the other factor.

For that matter, the graph doesn't even really show correlation, because it hardly has any data about games. And if we graphed the percentage of people under age 12 who played console or commercial PC games, it would be rising constantly (at a rate increasing over tiem) since the late 70's, whereas during that same time period the graph of violent crime fluctuates up and down.

And then, of course, there's the grand downfall of nearly all statistics, which is that it's basically impossible, outside of a laboratory setting, to control for all other potentially influential factors. (I'm a bit of a cynic when it comes to statistic in general, for precisely this reason.)

Now, the author of the article does use phrases like "the Playstation era has, in fact, produced the most non-violent kids ever," which narrowly avoids making the claim that the Playstation itself produced the most non-violent kids ever. I think the author is quite clear on the concept that his graph showed neither causation not correlation, but I also suspect he'd be entirely okay with people who read his article and DID start to wonder if maybe gaming decreases violence.

However! I think one of his major points, and the real thing the graph is intended to debunk, was that political arguments about videogames are typically predicated upon the claim that youth violence is on the rise, and that therefore video games should be investigated as a potential cause of this rise. The graph is meant to expose the falsehood of this supposed upsurge in violence, after which it becomes meaningless to wonder what caused a statistical change that, in reality, never took place.

(Reply to this) (Parent)(Thread)
Re: An academic friend suggest caution...
[info]denki_sasori
2006-07-20 03:33 am UTC (link)
Very correct, sir.

After reading the piece and seeing the graph, most people think that he infers videogames not only do not cause violence, they decrease violence, which is a complete misinterperetation. The purpose of the graph, I gather, is to show that releases of game systems and violent videogames do not cause surges in violence which the media,parent groups, and Jack Thompson say they do; if it were so, we'd see a rise after each system/game. But as we not only don't see a rise, we see a steady decline, the point here being that, if anything, videogames don't affect behaviour enough to either encourage or discourage violence, and are effectively a null factor.

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]the1jeffy
2006-07-17 07:27 pm UTC (link)
And your bud is totally correct. We shouldn't take this out of context either. But if video games were truly a major violent crime inducing factor, and are as important as the major anti-video game pundits would like us to believe, this graph wouldn't exist or owuld look differently.

You can say, with a fairly high degree of certainty, that if the pervasiveness of violent video games, entertainment, etc. was such a major factor in causing violent crime, it would outweigh the "lowering" effects of higher incarceration rates, trying as adults, etc.

My point is simple, violent media's effect on raising violent crime crime rates is statistically insignificant compared to the forces that exist to lower the rate of violent crime, whatever they may be. So, in essence, violent entertainment's effect on raising crime rates isn't as important as some outspoken crtics (Grossman, Jack Thompson, Kin Thompson, Hillary Clinton, etc.) would have us believe.

In fact, an interesting look at this data would be to theorize that violent video games can affect violent crime rates in either way. Perhaps, for some people, violent media does desensitize them, or train them, and this raises violent crime rates. But also, for some, violent media is an outlet for primal urges, causing violent crime rates to lower. So, whatever the particular M.O. might be, violent media can have either a positve or negative effect on crime rates. And at a glance, the record low crime rate would seem to say that whatever the cause, the forces that lower violent crime are more widespread in the general population.

In my mind, the particulars don't really matter - anyone trying to convince the populace that there is an epidemic of violence, whatever the scapegoat, is just plain wrong. You can't go as far as saying there is an epidemic of "peace" either. But the extreme anti-video game crowd doesn't have any factual leg to stand on.

(Reply to this) (Parent)(Thread)
Re: An academic friend suggest caution...
[info]kajex
2006-07-18 09:13 am UTC (link)
That's why they believe that convicts should have access to consoles, to be able to have an outlet for thier testosterone as an altenative to... er... prison love.

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]illspirit
2006-07-17 09:28 pm UTC (link)
Hmmm, I dunno how relevant the whole youth incarceration and (especially) try-them-as-adults thing really is. Even before the new wave of 'zero tolerance' and such, a 16 year old who killed someone would have still ended up in a juvenile correctional facility until they were 18. Trying them as adults and keeping them after they're 18 wouldn't have affected juvenile crime stats since they're adults then either way.

Granted, there's obviously a deterrent effect, but a large portion of youth crime are impulsive, heat of the moment, one-off affairs. Of which logical things like, say, thinking of the results, don't really factor into the decision.

And given that the game-grabbers main argument is that minors are too stupid to know right from wrong or to think of the results, and as such will-- inevitably and without fail --be forced to commit crimes by a game, the deterrent effect becomes largely irrelevant to the "debate."

Now, this isn't to say that games prevent crimes. Just that keeping a juvenile offender in jail after they turn 18 versus releasing them at 18 is a null factor in preventing them from committing more crimes before they turn 18. If games were an absolute cause of crime, the numbers should increase as everyone pretty much under the age of 18 would be in a detention center until they are 18 since almost all of them have played games.

Well, that's a bit of an exaggeration, as the numbers would taper off exponentially once 80% of all kids were in jail since they wouldn't have peers to attack anymore. :p

Either way, the graph does disprove the main lie of the game-grabbers. As a number of other people have pointed out, there is no growing epidemic of youth crime. Regardless of any other random correlations they can pull from their collective asses, the entire political house of cards is based on a lie. Take that lie away, and the rest of it crumbles.

This non-existent "epidemic" is the same crap a number of them tried (and keep trying) to spew as they attempt to undermine the Second Amendment as well. For the last decade or so, gun-control advocates (many of whom also want to control games) have been crying that the sky is falling, and promising that the 'streets will run red with blood' if we don't hurry up and ban all the guns. Yet the crime rate kept going down. I honestly think some of these politicians have stacks of pre-made think-of-the-children propaganda which they just fill in the blanks with whatever bogeyman is trendy...

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]noblebearaw
2006-07-18 09:37 am UTC (link)
I had'nt seen that particulsr graph before and thought it was interesting. As a gamer I'm inclined to agree with everyone else here that video games do not cause violence (or at least has never been shown to be so to a statisticlly signifigant extent); at least no more than , say, playing D&D leads to a cultist/satanic lifestlye (dead elwives, anyone?).


As for the drop in crime in the 90s, a book I read a couple of months back called Freakonomics seemed to have an interesting, albeit rather odd, explination. It seems the authors attribute the secondary cause to be the hiring of more police officers and the primary cause as being Roe v. Wade. According to the argument, right when the next generation of thugs were to hit thier prime and cause crime to surge again, it didnt, because they were never born.


i'll just be glad when politicans stop bothering the games industry and go look somwhere else for thier slam dunk.

(Reply to this) (Parent)(Thread)
Re: An academic friend suggest caution...
[info]dog_welder
2006-07-18 07:30 pm UTC (link)
Ann Coulter (you can love her or hate her, just hear this out) brought up very similar points in her latest book (though she's quicker to bash Roe v. Wade as she's very prominently anti-abortion).

Other contributing factors in the 90s:
1) More police were put on the streets (already mentioned).
2) We built more jails to house those who didn't deserve to remain a part of society.
3) Stricter mandatory sentencing guidelines were passed by a Republican-controlled Congress and signed into law by President Clinton (a "moderate" Democrat). This was actually a part of the actions taken under Newt Gingrich's "Contract With America" when the Republicans finally managed to gain control.

Part 1 allowed for us to better catch criminals. Parts 2 and 3 allowed us to prevent the frequent recidivist behavior that criminals often exhibit (i.e. you can't victimize innocent people if you're in jail already).

If you look at the graph, you start to see a dip in the 90s and the crime rate really takes a nosedive post 2000. Since any federal policies take time to show results, this is what you'd expect from this graph.

While the graph may not prove that "video games prevent violent crime," it certainly does show that there hasn't been a massive upswing in violence because people are "going GTA" on each other.

(As far as Coulter goes, her points that I've outlined above are pretty solid, you just have to read through her viciousness to get to them.)

(Reply to this) (Parent)
Re: An academic friend suggest caution...
[info]hemogoblin
2006-07-19 10:54 pm UTC (link)
As a student of econometrics, I know from experience the difficulty of trying to determine any sort of meaningful multivariable regressions. Econometrics is the art of choosing the correct variables (with the correct units) and the correct mathematical model to explain relations. Do you use gross, net, or rate? Do the variables have an linear, inverse, or logarithmic relationship? Are data points from past periods affecting future periods? Does a change in one of your dependent variables change another dependent variable, not simply the independent variable? Could your results be simple random error fluctuations? Finally, are you even sure that you collected your data correctly? I have an 800 page textbook upstairs with math so complicated that it would make your eyes bleed.

Unfortunately, what the author doesn't understand is that he is simply using the wrong model. He is comparing a country data rate to some binary variables. Hypothetically, if we looked at individual states, a very different conclusion could be reached. For example, perhaps most of the crime occurred in the poorer areas of less well-off states, where videogames are not prevalent and are not a factor. Perhaps more cops were hired in those areas and now crime is reduced there. At the same time, videogames have increased the crime rate in other richer states where they have better access to videogames. The violence may be increasing due to videogames, but the cops decreasing violence in the other state may be reducing the national level overall. This is a very hypothetical example, but you should see why choosing your variables matters. Clearly, the national crime rate would decrease, but videogames may be causing an increase.

Unfortunately, all you can conclude from his data is that the violent crime rate at the national level has decreased. No conclusions can be reached regarding the effect of videogames, because you didn't account for every other variable. All you know is that all the variables that effect crime have summed to the effect of reducing the national crime rate.

If I was personally going to try and determine the effect of videogames, you would need to do something like thus:
1) Find an isolated city where people didn't move around much, had a youth population of about 1000, and videogames were very uncommon.
2) Buy every family with children a videogame system and somehow force them to play it a reasonable amount (at least once a week)
3) Hope that no-one moves into or away from the city.
4) Hope that you chose a sample that represents all the residents of your country (not a chance, considering most people live in metropolis')
5) Hope that all other factors remain the same (no more cops are hired, no drug dealers move to the town on mass, people don't watch violent tv more, etc etc)
6) Record the crime rate every year for at least a few years.

See the problem why no-one knows the real effect of videogames on crime rate? You'd need some kind of divine intervention to keep all the variables constant and even then you wouldn't have a representative sample.

Disclaimer: I love videogames and I think its hogwash that they make people killers. I just can't prove they don't without resorting to case studies. ie: I play videogames and I've never killed anyone.

(Reply to this) (Parent)

(Read 96 comments) - (Post a new comment)

Create an Account
Forgot your login or password?
Login w/ OpenID
English • Español • Deutsch • Русский…