College internships and student academic outcomes: Immediately and dynamic effects

Abstract

This work examines the effect of internships on academic outcomes, academic performance and number of credits selected, using a panel dataset from a public university in Ecuador. My approach consists of a difference-in-differences design in which I use a new estimator robust to heterogeneous and dynamic effects. The treatment consists of whether a student takes an internship program during his or her semester of classes. The results indicate that, on average, internships have a positive effect on students’ academic performance of approximately 0.27 academic points (3.8% of academic performance mean). The evidence suggests that internships during class break are the best option for academic outcomes, since they have a greater average aggregate effect on academic performance and, although they decrease the number of credits selected, it is less than when internships are taken during classes. These results are robust across several specifications. In summary, as a policy alternative, the results suggest that it is preferable to take programs of this type in periods when students do not have to spend their time taking class hours.