Visualizing future enrollment and tuition revenue

December 2, 2019
  • Competencies
  • Interpretation and Application of Data
  • Meetings, Workshops, and Trainings
  • SEM Conference
session presenter talking into mic

by Betty J. Huff, Senior AACRAO Consultant

Auburn University wanted to analyze the probability of enrollment based upon a variety of data elements including high school demographics, test scores, financial need, and more.

Utilizing Qlik Sense as an analysis tool, the University was able to develop interactive visualization tables to inform users of the data and engage in predictive analytics.

A dataset was created for predictive models and included admitted students who applied over a period of two years. Because of the differences in tuition and the probabilities of enrollment of in-state and out-of-state students, separate models were estimated for these groups. Separate models were created for non-FAFSA filers, filers who are not eligible for Pell grants, and filers who are eligible for Pell grants. To account for differences in availability of options and scholarships for students depending on their home states, models for out-of-state students included the state of origin. This was done by using a multilevel approach with applicants (level 1) nested within states (level 2).   

 Overall, the models included: 

• random intercepts (i.e., the intercept is different for applicants from different high schools for in-state students and for applicants from different states for out-of-state students);

 • random effects of high school GPA (i.e., the effect of high school GPA varies depending on a high school for in-state students and depending on the state of origin for out-of-state students); and

• random effects of scholarship offers (i.e., the effect of scholarship offer varies depending on a high school for in-state students and depending on the state of origin for out-of-state students).

Using the model, the University is able to run simulations and analyze the probability of student enrollment by a variety of attributes and by moving the level of scholarships or financial assistance applied to the cost of tuition.

The predictive model assists enrollment management personnel in making decisions regarding the ability to affect student enrollment and understand the impact upon tuition revenue when scholarships are adjusted.