[pygy_]

Yes, their evidence sugests a causal relation.

They define socialization/selection in the abstract:

    the socialization (violent video game play predicts aggression over time) 
    versus selection hypotheses (aggression predicts violent video game play over time)

[gojomo]

Unless they're missing some other factor that drives both aggression and the gameplay choices -- a factor other than the "previous levels of aggression" they mention as being a tested control.

Perhaps, some other childhood stressor or failure in coping mechanisms? And if this other factor tends to trigger first, escapism into violent gameplaying, but then later (if unaddressed), actual aggression, I think they could see the "steeper increases in adolescents' trajectory of aggressive behavior" that's reported, even though the gameplaying is just a waypoint or signal rather than cause.

(Under this hypothesis, it's possible the violent gameplaying aggravates the underlying issue, if it delays or prevents other remedies, but it's also possible the gameplaying serves to soften the aggression. You can't really tell from their sequenced relations analysis... you'd want some stronger random-like control on the amount of violent gameplaying.)

[jarofgreen]

> Unless they're missing some other factor

From Abstract: "and a comprehensive set of potential 3rd variables were included as covariates in each analysis".

Sounds like they certainly tried to control for that - you'd have to read the full study to see the details.

[gojomo]

Yes, too bad the full text is behind a paywall. So I haven't read the full study, but I was once upon a time a teenager.

I recall many aspects of my and my peer's psychological lives that would not be easily available to an outside researcher as part of their "comprehensive set of potential 3rd variables".