How predictive are self-reported strategies for their actual use?

Following some basic thoughts concerning the quality of retrospective self-reports the necessity of analyzing the predictive validity of learning-strategy-questionnaires is pointed out. The study presented is aimed at such comparison between students' retrospective self-reports about strategic learning and their actual use of these strategies. For this purpose we analyzed the use of learning strategies of 270 4(th), 6(th), and 8(th) graders while studying a text in individual sessions and compared it to students' self-reported use of strategies in a questionnaire referring to the same domain. Neither for single strategies nor for strategy-scales substantial relations between retrospective self-reports and the actual use of strategies could be found. Taken together the participating students - relatively to their actual use - tended to overestimate their strategic competence and performance. The presented results indicate that the predictive validity of questionnaire-based strategy data of students of this age can be seriously questioned.