Divergent meta-analyses do not present uniform evidence that violent video game content increases aggressive behaviour.

Whether violent games increase aggression is a contentious issue. The relatively enduring disagreement in the literature about whether violent video games cause increased aggression is reflected in divergent meta-analyses. Though we applaud Mathur and VanderWeele (2019) for attempting to synthesise such divergent meta-analyses to determine an overarching view on the effects of violent media, we argue that their interpretation of the evidence is misguided. Underpinning the notion that the evidence, in general, favours a “violent game effect” lie two problematic assumptions: (a) that the analyses conducted within these meta-analyses are equally methodologically and statistically rigorous and therefore equally valid, and (b) that even tiny effects are veridical. Here, we show that the effects reported by Anderson et al. (2010) appear to overstate the evidence in favour of a relationship between violent game content and aggression, and that bias-corrected models produce only tiny effects (Hilgard et al., 2017). We then compare these smaller effects estimated by Hilgard et al. (2017) and Ferguson (2015) to show that they appear to be in close agreement. Finally, as a reminder that non-zero meta-analytic effect sizes do not guarantee that an effect is meaningful, we compare these effect sizes to the sizes of (significant) yet nonsense effects of Extra Sensory Perception to show that the effects of violent game content on aggression are so small that we should dismiss them as practically meaningless.