Since working as a research assistant with NCID two years ago, Fernando Sols is pursuing a Masters in Economics at University College of London, and currently works with Banco de España as a visiting student. He returned to NCID to present his research on the reliability of student evaluations in determining instructor performance in the classroom.  

Across many geographical and institutional contexts, administrators and agencies struggle to fairly and effectively rate and incentivize good instructors. One of the most common tools in this area, end-of-term student evaluations, have conditional validity. Many biases impact these results, ranging from the time of day that class is held, to the age of the professor. In reviewing the validity of this exercise, Sols examined data first-year cohorts in four fields at Carlos III University in Madrid, in a paper titled Do Students’ Evaluation reflect how effective their Teachers have been? Evidence from a Natural Experiment.

Taking advantage of a quasi-random allocation of students within required courses, his research estimates value added of professors and then compares this student evaluations and student performance. Controlling for many factors, such as teacher and class characteristics, his conclusions ultimately challenge the validity of end-of-term evaluations as a measure of instructor performance. “Student evaluations do not necessarily measure teaching effectiveness. Mostly what they report is perceived utility.”

Applying these findings, “If you are a principal or the dean of a university and you are concerned about the performance of teaching, you should be looking more at the output of the students and at their scores.”