Enrollment Projection Modeling Methods

Estimates of program enrollment and degree production must necessarily be based on strong program input. Rather than require programs to generate these estimates from a blank sheet, initial estimates were provided to all programs using a consistent approach. In order to obtain the maximum value from program input, the estimates were provided at the academic plan (major) level since this is where real program knowledge resides.

Program Headcount Estimation Model

Using several years of Fall enrollment data, 10 year enrollment projections were developed using three separate models. A linear model (constant growth rate) and a logarithmic model (decreasing growth rate) were developed using a least squares fit to the data. The third model mirrored the overall university growth rate in the university level model. The average of these three projections was used as the initial enrollment estimates for each degree level (including certificates). Enrollment projections were also included for unclassified students in each academic plan (major) as well as undeclared students. These enrollment headcounts were used only as a factor to determine degree projections and are therefore not reported here.

The directions from the state indicated that counts should be first majors only. In practice, most students declare a second major or minor close to the time that they file an intent to graduate form. The number of declared double majors is small relative to the inherent enrollment projection error so no effort was made to separate those students. Additionally, there is no reasonable way to separate the associated FTE.

Program Degree Estimation Model

Initial estimates of degree production at the program were computed by applying the five year historical program degree production rate to the program enrollment projections. The degree production rate was computed by dividing the sum of the degrees awarded during the period by the sum of the fall enrollments during the period. This method provides the optimal estimate of the degree production rate, provided that the underlying rate is constant.

Model Implementation

The enrollment and degree estimates were developed and presented to the programs in an Excel format. Using a designed template, a VBA macro read five years of historical enrollment and degree data using a table lookup function, then computed the parameters for the prediction functions for all degree levels, created the associated graphs, and saved the worksheet under the academic plan name. A separate worksheet was created for each academic plan. The enrollment and degree projections for each degree level were read into a consolidated worksheet to provide a composite summary.

College Review and Modification of Program Estimates

The enrollment and degree estimates were delivered to the programs for review. Colleges were asked to review each program and either accept or modify the enrollment and degree estimates that had been computed. Some program representatives adopted the projections made with the model. In other cases, programs used their best judgment to make modifications that reflect local plans for the program. The program representatives delivered their changes either electronically by making changes in the Excel sheets, by making handwritten changes on paper printouts, or by communicating the changes in email text or verbal conversations with UAPS. The next two graphics describe the process flow.

Overall Process Flow for Program Headcount and Degree Projections

Example Degree Projection Model