Behavioral Interventions and Students' Success at University - Evidence from Randomized Field Experiments

Abstract

Research over the past decades has documented the various individual and social returns to tertiary education. In many countries, however, the realization of those returns is jeopardized by high dropout rates and delayed graduation. Investigating the underlying reasons for the lack of academic success, the recent literature points to the critical role of behavioral biases and the resulting need for interventions that address them. Against this background, this thesis investigates the effectiveness of several low-cost and easy-to-implement measures that aim to increase students' success at university, thereby contributing to the young but growing body of research on behaviorally informed interventions in higher education.Using randomized field experiments for causal evaluation, the four main chapters of this thesis study i) whether social information about the past behavior of others can raise remedial math course participation and if increased course attendance translates into higher academic performance in the first year of studies (Chapter 2), ii) whether relative performance feedback on accumulated course credits can increase short-and long-term academic achievement (Chapters 3 and 4), and iii) whether a non-binding commitment to study according to the recommended study structure and reminder letters are able to help students succeed in university (Chapter 5).Taken together, this thesis provides a broad range of evidence for the versatility and potential of behaviorally informed measures. They can improve academic performance directly or indirectly, provide decision-relevant information that is frequently not available to students, or be used to address specific biases -such as procrastination and limited attention. Crucially, Chapters 4 and 5 provide some of the first evidence that behavioral interventions can have substantial positive effects on long-term academic success. Yet, the chapters of this thesis also paint a nuanced picture and show that effects are often heterogeneous. In many situations, a targeted implementation of the proposed measures may therefore be necessary to prevent unwanted effects. At the same time, examples of how econometric methods can be used to identify the beneficiaries of an intervention more precisely are provided.

Impact of the PATH Statement on Analysis and Reporting of Heterogeneity of Treatment Effect in Clinical Trials: A Scoping Review

“…The Study authors claimed findings of HTE in 15 risk modeling reports (Figure 2). For six, (49,60,64,71,75,77) heterogeneity was found on the absolute but not the relative scale (i.e., benefit magnification) and for nine (36,38,43,48,50,54,58,68,73), heterogeneity was also found on the relative scale (i.e., relative treatment effects also varied across levels of baseline risk). The four remaining reports (56,57,61,76) did not find clear overall treatment effects and none claimed HTE on either scale.…”

“…In the single report that was not found to be credible, multiple treatment-by-risk interactions were tested across a variety of outcomes and findings were inconsistent. (54) Effect Models. The 26 effect model analyses used diverse types of models and analytic strategies.…”