HOW DOES PREDICTION ERROR IMPACT ON LEARNING? THE CASE OF COMPUTER GAMES

This is a study with children about the exploration of the motivational properties of reward uncertainty and its impact on learning in the environment of an educational computer game. The investigations focused on positive reward prediction error and its relationship to successful or unsuccessful learning in the specific computer game context. This study found that, on average, PE for successful learning (M = 35.70, SE = .88) was significantly higher than PE for unsuccessful learning (M = 33.65, SE = 1.00), with an effect of medium size [t(49) = 2.26, 1-tailed: p < .025, r = .31], suggesting that PE is linked to successful learning. This is an interesting finding due to its implications for education since it allows predictions as to when learning is most likely due to take place within such computer gaming environments.