Abstract
A large and rapidly growing literature suggests that playing action video games may enhance a range of cognitive capacities (Powers et al, 2013). This literature, however, has recently been criticized for methodological shortcomings, including inadequate blinding (e.g. Boot et al, 2012). Two of the most highly cited studies reported successful training of mental rotation (MRT; Feng et al, 2007) and multiple object tracking (MOT; Green & Bavelier, 2006). The only published attempt to replicate these findings found no training effect (Boot et al, 2012), yet this study did not document sufficient reliability in its MRT and MOT tests to detect a training effect. We attempt here to replicate the original training effects with highly reliable MRT (alpha=0.82) and MOT (alpha=0.82) tests. We used the same action video game as the prior studies (Medal of Honor, MOH, a first-person shooter), plus matched action (Wii Sports) and non-action (World of Goo, strategy game) control training games and a no-training control (NTC). We carefully blinded participants to our hypotheses, and an extensive post-training questionnaire confirmed that this blinding succeeded. Conditions did not differ significantly in their pre/post difference for either MOT (F(3,76)=0.93, p=0.43) or MRT (F(3,77)=1.96, p=.13), and no active training condition improved significantly more than the NTC on either measure. Surprisingly, MOH improved numerically (though not significantly) less than all other conditions for both measures. In sum, our carefully controlled, blinded study failed to replicate two key benefits reported for action video game training, despite highly reliable measures. It is possible that the original reports of positive training effects for MRT and MOT were due to insufficient blinding or random chance. Our results suggest that action video game playing has less influence on MRT and MOT than those prior results implied.
Meeting abstract presented at VSS 2014