Finding: This reanalysis demonstrates that apparent negative effects of positive emotions on performance in self-regulated learning tasks are likely methodological artifacts caused by unmodeled practice effects, not genuine psychological phenomena like "coasting."
Why it matters: This directly challenges fundamental assumptions about emotion-performance relationships in learning and highlights critical methodological issues for behavioral science research on motivation and goal-setting.
Method: The authors reanalyze existing data by adding log(trial) learning curves to computational models and implement leave-future-out cross-validation to replace WAIC model selection criteria.
Jargon: WAIC (Widely Applicable Information Criterion) - a model comparison metric; coasting hypothesis - the idea that positive emotions lead to reduced effort and performance.