Not all confusion is productive: an investigation into confusion induction methods and their impact on learning

The current study was an attempt to discover the gold standard of inducing a state of confusion that is beneficial to the learning of complex science topics. Using a randomised controlled trial, participants received either one of three different types of confusion induction (deep-questions, intra-testing and breakdown scenarios) or a lecture-based information delivery control. Results revealed that breakdown scenarios were the most beneficial in terms of pretest to posttest learning gains. Additionally, significant interactions were discovered among learning, confusion induction methods, and measures of individual differences (i.e., goal orientation and attributional complexity). Interpretations and applications are discussed.