Abstract
Watching a learning model acquire a skill can aid motor learning, but evidence for benefits compared to a correct model is mixed. One reason may be related to what an observer sees when watching a learner; including variation and/or errors. Moreover, seeing an error is thought to engage prediction processes based on the error between expected and actual outcomes. Therefore, if there is variability in the model’s performance; only when it is perceived as a ‘miss’ will this lead to ‘prediction error’ driven learning. We tested whether prediction errors drive learning, different to learning from watching variation. Three groups (n = 22/group, planned n = 34/group) practiced putting a golf ball to the centre of a 3 x 3 grid. Practice involved watching videos of a model putting, followed by physical putting blocks, without and with visual feedback. Self-predictions were captured in no-vision trials. Two groups watched the actor putt to different corners of the grid with instructions conveying ‘hits’ (variable) or ‘misses’ (errorful). The control group watched ‘correct’ centre putts. Retention and transfer putts to the corners were tested the next day. In practice, watching corner ‘hits’ and ‘misses’ increased putting variability compared to the control (p = .03). In transfer, the corner ‘hits’ group had greater consistency (p = .03), but not greater accuracy. Self-predictions only improved for the 'miss' group in retention and transfer (p = .059). While testing is ongoing, these data show weak evidence that prediction driven processes drive learning through watching.
