Joel Klein

Document Type

Research Study

Publication Date

Fall 2017


Research presented in articles such as “Who Gets to Graduate?” by Paul Tough in The New York Times Magazine, May 2014, shows that a student’s sense of belonging at a university affects their future performance and willingness to stay in school. Early studies conducted by Analytics and Institutional Research at St. Cloud State University (SCSU) confirm this finding. SCSU offers a survey to new students during their first term to measure their sense of belonging. The end goal is to leverage the survey information to predict likely future student behaviors within the university, and implement support strategies for at-risk students. In the fall semesters of 2013 through 2016, SCSU outsourced the survey to Skyfactor, a software consulting company, who delivered it to students through their Mapworks Survey Service. Mapworks is centered on improving student retention rates for universities. In fall 2017, SCSU discontinued using Mapworks and distributed its own belonging survey through Qualtrics, a survey delivery program from a different software consulting company.

A Belonging Research Group of eight students was formed in fall 2017 to analyze and identify factors in the Mapworks data that may predict academic performance, in order to prepare for analyzing the new Qualtrics data. This paper analyzes the differences between the two surveys’ response rates, assesses the relationship of belonging to poor academic performance, and attempts to propose a model for predicting which students in the 2017-2018 new incoming class are likely to perform poorly in or withdraw from classes.

There are several important insights resulting from this analysis.

  • The overall response rate to the Mapworks survey from the two cohorts starting 2014 – 2015 was 83 percent. Students who were female, students of color, not STEM majors, not Pell eligible and for whom the closest university to their hometown was not SCSU were the most likely to respond. The overall response rate to the new survey in 2017 was much lower at 46 percent, nearly half the Mapworks rate. Incentives offered for the Mapworks survey may be partially responsible for the difference.
  • Those with a lower belonging score are slightly more likely to receive a poor grade or withdraw from a class during the first term. As the level of belonging decreases from high to low, the probability a student will receive a D, F or W increases by roughly three percentage points.
  • Just fewer than seventy percent of low-belonging withdrawing students do so in the first 16 weeks of the semester, compared to roughly 65 percent of the students in the high- and medium-belonging groups.
  • Students who are male, students of color, STEM, Pell Eligible, or have a closer university to their hometown than SCSU have a higher likelihood of receiving a D, F, or W. Along with high-school GPA, the number of credits attempted, and belonging score, these demographics were combined to create a model that could be used to predict the probability of getting a D, F, or W for future cohorts.

These insights will help guide important decisions regarding the continual improvement of student belonging and academic performance at St. Cloud State University.


Edited by Shaya Kraut.