Relación entre las calificaciones en la evaluación continua y en el examen final. Diferencias por género, nivel socio-económico y conocimientos iniciales (código 4137)

Abstract

We study the effect of a nudge in achievement in a midterm test on subsequent performance in Math by using administrative and survey data from students at the University of Alicante. We estimate the effect through a regression discontinuity design since only students obtaining a midterm score greater than or equal to 5 may interpret their score positively. The sign of the effect is unclear ex-ante as a nudge may induce students to believe that additional study effort is (un)important to ensure a pass in course. We find that the overall effect is small and not significant. We do not find differences by gender or socioeconomic status; some differences might exist in students who majored in Arts and Humanities, whose initial knowledge in Math is lower, but a rigorous conclusion cannot be drawn. However, our results point out that the nudge effect is heterogeneous by students’ confidence: a nudge for non-confident students leads to lower performance in subsequent assessment, while confident students, although reacting positively with respect to non-confident ones, do not change their performance. Our results shed light on the role of confidence in processing salient although noisy information on achievement and its consequences on subsequent performance.