What we think we learn from watching others: the moderating role of ability on perceptions of learning from observation.

Despite increased interest in the processes guiding action observation and observational learning, we know little about what people think they learn from watching, how well perceptions of learning marry with actual ability and how ability perceptions develop across multiple observation trials. Based on common coding ideas, we would think that ability and perceptions of ability from watching should be well matched. We conducted two studies to answer these questions that involved repeated observation of a 2-ball juggling task. After each video observation, observers judged if they could perform the skill and gave a confidence score (0-100%). In Experiment 1, an Observe-only group was compared to an Observe + Physical practice and No-practice group. Both observer groups showed a better physical approximation of the juggling action after practice and in retention and their confidence increased in a linear fashion. Confidence showed a small, yet significant relationship to actual success. In Experiment 2, we limited physical practice to 5 attempts (across 50 observation trials). In general, people who had high perceptions of ability following a demonstration were overconfident, whereas those with lower perceptions of ability were accurate in their assessments. Confidence generally increased across practice, particularly for trials following observation rather than physical practice. We conclude that while perceptions of ability and actual ability show congruence across trials and individuals, observational practice increases people's confidence in their ability to perform a skill, even despite physical experiences to the contrary.