Does learning happen and remain stable over time? A longitudinal assessment of entrepreneurship education using situational judgment tests

This study aimed to longitudinally assess the learning outcomes of an entrepreneurship course. Research was performed in two stages. The first stage, qualitative, consisted of elaborating and obtaining evidence of validity of the situational judgment tests. Situational tests are behavior-based and ask respondents to inform how they would act in a given situation. Also, a self-report competence instrument was presented, as a way to obtain evidence of convergent validity between the two instruments. The second stage was a quasi-experiment. Students completed the situational judgment tests and the competence self-report questionnaire in three different moments (pre-test, post-test, and follow-up). Hypotheses were tested through t-tests, analyses of variance, chi-square tests, and logistic regressions. Results point that learning happened and remained stable over time. Among items, those reporting the best improvement rates were those that refer to educational objectives of higher complexity levels, as the competence to develop business models. Positive correlations between the situational tests results and those of competence tests were identified. The results of logistic regressions suggest that with situational tests, knowing to which extent respondents have improved or worsened their results, there is more chance of knowing which group the respondent belongs to (training or control) than when using the self-report instrument. •Students learned, and learning occurred as a result of the entrepreneurship course.•Learning remained stable over time, considering three months measure.•Items reporting the best improvement rates were those that refer to educational objectives of higher complexity levels.•Learning improvement was followed by improvement in self-perception of entrepreneurial competences.•Logistic regressions showed that SJTs were better measures to differentiate respondents than the self-report competence test.