That's a very deceptive study. The actual results were: non-players froze when faced with a gun while players reacted more actively (increased body sway):
From the discussion:
> Non-players exhibited an immobility-like reaction.
> Heavy players behaved as if “jumping the gun”, increasing mobilization, instead of immobilizing and waiting for the best chance to get rid from danger.
> Indeed, “... the simple sight of a weapon increases the likelihood of aggression if the person has mentally paired a weapon such as a gun with killing or hurting people...” [Anderson and Warburton (2012), p. 72].
> This is the case for heavy players of violent video games who are frequently exposed to weapon use for the purpose of killing and hurting others.
The authors are trying to say video games provoked more violent reactions. Well, to me it sounds like video games have educational value: players immediately recognized the danger represented by the weapon and reacted more aggressively. Did the non-players freeze because they did not know how to react to the danger?
Their arguments don't seem to be specific to video games. It seems to be about exposure to weapons in general. It's not clear what this result means. Games trained players to respond differently but is that good or bad? Does this mean players are less likely to survive violent encounters? Does this mean players will defend themselves more effectively? How do people who went through real firearms training respond in the same conditions? I don't know.
The PTSD stuff is also deceptive:
> Scores on the PCL-C hyperarousal cluster were also significantly higher for heavy players (Z = -2.08, p = 0.04).
That's the only significant difference. Emotional numbing, avoidance and recollections were not present. Hyperarousal includes symptoms like imsomnia, irritability, anxiety... The potential for confusion here is significant.
There's no way this particular study establishes any causal relationships.