Professor Effects on Student Trajectories : a Statistical Data Analysis in R
Abstract
Parameters quantifying the quality of instruction can be difficult to come across. One commonly used measure comes from course evaluations, which are largely subjective reviews that pertain to personality in conjunction with quality of teaching. Being optional surveys, course evaluations may not tell a full story. According to institutional research at Marquette University 83% of students offered in-class time to complete course evaluations did so, and only 59% not offered class time completed surveys. (1) Another common metric comes from online faculty review boards like https://www.ratemyprofessors.com/. According to a 2008 and 2009 study, these reviews are highly correlated with official institutional reviews. (2,3) As such, they may be unreliable for the same reasons. Student outcome data can serve as a good indicator of faculty effects. In this paper we construct two statistics from a large such dataset from the Math and Chemistry departments of Oakland Community College. We will observe the effects that faculty have on their students’ final cumulative GPAs, and analyze the distribution of withdraw rates between professors.