Despite the vast amount of admissions data we collect, there is still limited understanding of what factors are most predictive of admitted students’ decision to attend Duke. Traditional yield models have focused on a small set of variables that have limited power in predicting students’ decisions, and there may be student attributes that these models have failed to consider. This project would involve various approaches to analyzing up to 25 years of admissions data to develop a more reliable model for first-year undergraduate yield. Possible outputs could include a predictive model leveraged by the Admissions Office to optimize their admissions decisions, a dashboard to visualize “what-if” scenarios, or any number of other possibilities.
Project Lead: Heather Mechler, Associate Vice Provost and Director of Institutional Research
Project Manager: Diego Rodriguez
