Oh, you must mean this ( http://www.theesa.com/policy/scotus_chart.gif ) rise in violence.
Yes that's my main point: this plot does not prove the correlation.
If you look at a bigger picture e.g. (maybe here [1], altough it's not technically the same metric) there was a peak around 94. The peak and the relative low value before the peak cannot be explained by video games.
And that my second point: no one ever claimed that violence is only caused by video games. Clearly the blue curve is under the influence of (many) other factors, so it's very hard to determine whether there is a correlation without knowing these other factors.
But let me say it once again, it's a possibility that video games reduce violence. I am simply saying that if think this is the case because of this plot, you're probably biased.
[1] http://www.eji.org/files/juv%20murder%20rate%20drop_0.jpg
I don't get your line of reasoning. Let's go through all the scientific method steps:
The hypothesis is: Videogames cause violent behaviour in youngsters.
The test for this hypothesis is: If videogames cause violent behaviour in youngsters, then we should find a positive correlation between the rise in popularity of videogames and youngster violence.
The hypothesis failed the test. The correlation either is not there, or it is there but is negative. We can conclude, from the fact that we did not find a positive correlation, that the hypothesis is false.
We could if it was a lab experiment where the only factor that varies would be the popularity of violent video games. But it's not.
It's entirely plausible the the increase of home entertainment has indeed reduced the rate of juvenile crimes. (kids don't stay in the street any more). Yet it doesn't say much about how violent games induce violent behaviours.
Your explanation is obviously possible, but it does not preclude my conclusion. If your explanation for the signal correlation is true, then video games account for comparatively little youngster aggression.
I am terribly sorry for the ad-hominem, but I think, at this moment, you need to brush up on statistics 101. The arguments used are typical of pre-statistical studies, and HN is really not the place to be taught fundamentals.
That was a bit cheap but ok. Let me add more.
Here is the same data that I have reploted
This chart is less biased (y axis starts at 0) and the correlation is less obvious.
The correlation is actually -0.6557, which is significant but probably not as high as you would expect from the first chart. But ok the correlation between game software revenue and juvenile violence is probably here (we're talking about ~20 data points...). To go further and conclude anything about violent games, you would need at least the proportion of violent games among games, we don't have it here.
Here's the data that I've extracted from the pixel position. It's not 100% accurate, but it should be good enough.
1990 412.999995
1991 430.999995
1992 448.999995
1993 469.999995
1994 495.49999995
1995 487.99999995 3.2
1996 432.49999995 3.67407407407407
1997 384.49999995 3.74179894179894
1998 355.99999995 4.55449735449735
1999 322.99999995 5.43492063492064
2000 298.99999995 5.5026455026455
2001 291.49999995 5.90899470899471
2002 276.49999995 6.92486772486772
2003 285.49999995 7.06031746031746
2004 270.49999995 7.33121693121693
2005 279.49999995 6.92486772486772
2006 297.49999995 7.26349206349206
2007 286.99999995 9.43068783068783
2008 288.49999995 11.6656084656085
2009 261.49999995 16Your beef is with whoever told you that the world was becoming more dangerous, not those who post statistics that paint the opposite picture. Somebody lied to you. Find out why.
I haven't, I am not taf2. I was just saying that this inverse correlation that you see may be due to other factors or may not even be statistically significant. We're missing a lot of data here (e.g. how does the juvenile crimes rate fluctuates prior to 1990).
http://en.wikipedia.org/wiki/List_of_school_shootings_in_the...