The keyword effect: A conceptual replication, effects on bias, and an optimization

Prompting learners to generate keywords after a delay is a promising means to enhance relative judgment accuracy in learning from texts. However, to date, conceptual replications of the keyword effect without the involvement of the researcher who originally proposed it are still scarce. Furthermore, it is unclear whether generating delayed keywords could reduce bias and whether the benefits of generating delayed keywords could be optimized by having learners compare their keywords with expert ones. Against this background, we conducted an experiment with N = 109 university students who read four expository texts and then were randomly assigned to one of three experimental conditions: (a) Generation of keywords after reading, (b) generation of keywords after reading and a comparison with external standards in the form of expert keywords, (c) no keyword generation (control condition). We found that generating delayed keywords significantly increased relative accuracy but did not reduce bias. Furthermore, we found that the comparison with expert keywords enhanced relative accuracy beyond the established keyword effect. However, we also found that the comparison with expert keywords increased bias (here: underconfidence). Overall, these findings suggest that generating and comparing keywords is an effective means to enhance relative accuracy.