TRACK ADVISING

Every year, about 500 thousand students in Italy receive high-school track recommendations from their teachers. However, there is no clear evidence that these recommendations are effective or what exactly defines a “good” recommendation. Given the significant impact that these choices have on students’ educational trajectories and future opportunities, it is crucial to investigate how to improve the quality of track advising.

The project

This project aims to assess whether providing teachers with algorithmic information based on standardized test scores can help them make better, fairer track recommendations. We use a rigorous Randomized Controlled Trial (RCT) and advanced causal inference methods to address this question and ultimately enhance educational outcomes and equality. The provided information includes predicted educational outcomes constructed via machine learning algorithms applied to test scores from earlier school years. Teachers in the treatment group can use this information to assist their track recommendations for 8th-grade students choosing between vocational, technical, and academic high school tracks. We will measure the impact of this intervention on the quality and fairness of recommendations using principal stratification analysis, focusing on whether more students who can succeed in higher tracks receive appropriate advice and whether biases related to socioeconomic status, gender, or immigrant background are reduced. The study will track students’ subsequent performance and choices, providing robust evidence to guide educational policy on track advising. The project is funded through an Research Grant on Education of the Unicredit Foundation. The PI is prof. Andrea Ichino. Co-PIs are prof. Fabrizia Mealli and Davide Azzolini.

Results

The project is currently ongoing.