Does Projecting School District Enrollments by Race Produce More Accurate Results?

Richard S Grip, Statistical Forecasting LLC

Since different races have unique fertility rates and migration patterns, performing school district enrollment projections by race and aggregating to a total may be more accurate than performing enrollment projections with all races combined. Twelve school districts in New Jersey of varying overall size and majority race percentages were used in this study. Using historical enrollment data for a five-year period, the Cohort-Survival Ratio method was employed to project enrollment for a four-year period, 2003-04 through 2006-07. Separate projections were completed both by race and with all races combined. Projected enrollments were compared to actual enrollments in each district for both methods used. Absolute Mean Error Rates (AMER) were calculated for each district for the prediction time period. The results showed that the projections with all races combined had lower AMERs for smaller districts than when projections were performed by race, but the results were mixed for larger districts.