The STEM Success Initiative was a grant supported program that entered its pilot year in 2014 at The College of Wooster. This program aims to particularly help students who are considered to be part of underrepresented (UR) groups in science, technology, engineering, and mathematics (STEM) courses, and keep them in STEM, as they have historically been at a higher risk to leave than their white peers. This paper aims to look specifically at students who are being retained and where they are persisting in introductory levels in three departments: chemistry, biology, and mathematics. We want to see if students are retained, in other words if they stay in STEM fields through graduation; this helps us see more long term which students are choosing to stay. Seeing where students persist: where they are continuing from one course to the next, gives us a much more in depth and specific look at our departments.

To analyze these measures, we used logistic regression to build models for binary outcomes (successes and failures/being retained or not, or persisting or not). Once we built many multiple logistic regression models, we used two evaluation processes for general linearized models: drop in deviance test and misclassification tables. In analyzing retention, we found that while ratios of male and female identifying students stays consistent from our initial pool of students to graduating majors, we did find a slight drop in students who are considered UR. We found a number of indicators as to if a student is retained or not, including if a student was a first generation college student, and how many STEM courses a student takes in their first year. We found a combination of grades, what term students took certain courses, gender, and UR status to all be indicators of persistence when analyzing Chemistry, Biology, and Mathematics Departments. Grades were the most consistent across the board, and gender came up very rarely, although twice in higher persistence levels in two different departments. Overall, analysis was conducted to provide more insight and understanding of the inner workings of STEM fields at The College of Wooster.


Frazier, Marian




Categorical Data Analysis | Multivariate Analysis | Statistical Models

Publication Date


Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis



© Copyright 2020 Harriet FC Sudduth