The Modified Regression Technique (MRT)

What is MRT?

In 1996, the Modified Regression Technique (MRT) was developed by Dr. Richard S. Grip at Rutgers University (NJ). MRT was developed in response to complaints by superintendents and school business administrators of high growth school districts that the Cohort Survival Ratio was not providing accurate enrollment projections. After researching over 50 years of literature regarding school enrollment projections, it was determined that statistical regression provides more accurate projections that the Cohort Survival Ratio. To confirm his hypothesis that regression procedures would provide more accurate projections, Dr. Grip developed a multiple regression technique, MRT, which relied on three variables:

1. the number of current students
2. birth rates (only to predict kindergarten students)
3. migration rates.

Each variable was entered into a series of multiple regression equations to help project future enrollments.  MRT seeks to maximize the variance in a regression equation by using variables that help define the future number of students. The current number of students in each grade is the variable with the largest variance and therefore helps to project future enrollments better than the other variables.

Does the Modified Regression Technique outperform the Cohort Survival Ratio by providing more accurate enrollment projections?

In his research, Dr. Grip used data from the 1980’s from two high growth school districts in central New Jersey and projected enrollments seven years into the future using both CSR and MRT. Since the study was performed in 1998, enrollment for each district was known and therefore comparisons for accuracy could be made. The highlights of the results were: