| Accuracy of enrollment projections for school districts that are experiencing a rapid increase or decline in student population is difficult to attain. This study utilized five methods in the projection of enrollment for Georgia suburban school districts, as defined by their inclusion in the Standard Metropolitan Statistical Areas of the 1970 U.S. Census Report. These twelve districts ranged in size, in the 1983-84 school year, from 9442 to 68,984 students, with a total enrollment for all districts of 381,991.; The methods included the cohort survival method, linear regression, the standard compound interest formula, average numerical change, and the grade retention model. The latter two had been used by the Georgia Department of Education for five-year facilities planning and for annual State funded program allotments.; A post-projection approach was employed to project to known points in time for one, five, and ten-year projections. Five previous years of data were used to establish the enrollment trends. A mean percentage error for all of the districts in the sample was the statistical measure of accuracy for each method. The acceptable level for a good projection was 1% error per year in the projection.; Three one-year projections for each district yielded an overall mean percentage error as follows: standard compound interest, 1.95%; cohort survival, 2.23%; average numerical change, 2.26%; linear regression, 2.46%; grade retention method, 3.09%. Two five-year projections for each district yielded an overall mean percentage error of: cohort survival, 2.72%; average numerical change, 8.40%; standard compound interest, 8.43%; linear regression, 9.98%. One ten-year projection for each district produced a mean percentage error of: cohort survival, 3.25%; average numerical change, 19.35%; standard compound interest, 24.49%; linear regression, 26.78%.; The cohort survival method projected 22 of the 24 individual district five-year projections within 5% accuracy, and projected 11 of the 12 school districts' ten-year projections within 5% accuracy. This method was recommended for use by the Georgia Department of Education in developing five-year facilities plans for local school districts, and for making annual State funded program allocations based on projected enrollment. |
