Determinants of Higher Education Enrollment in the United States: 1946-1980.

Abstract

The future of higher education enrollment in the Unied States is clouded. Until recently college enrollments were ever increasing, and the only pertinent question was "how fast?" During the last decade, however, two factors have combined to cast doubt on the prospects of further enrollment increases: the decline, from 1961 to the present, of total fertility, and , particularly among males, diminished enrollment rates. Further research into the future of higher education enrollment is dictated not just by the great divergence among existing projections and forecasts, but also by the limitations of these works. Most assessments of the prospects for higher education enrollment are based on projections--linear extrapolations of existing or past enrollment rates applied to the future population composition. They reify rates, capitalize on their chance fluctuations, and generally ignore changing conditions that might alter enrollment patterns of the present and past. Some explicit forecasting models do exist--models in which rates of enrollment are endogenously determined--but they suffer from a different limitation: Many of the indicator variables have future values that are as much in doubt as the enrollment rates that they are predicting. This dissertation will involve the construction of a forecasting model for higher education enrollment that stresses prospective indicators--variables whose future values are already reasonably well known. These include variables such as sibship size, where smaller families make for more financial resources per child, and "sibling squeeze," where tight birth-spacing creates a strain on the within-family resources necessary for paying for college. Data analysis is primarily at the macro level. It involves linking the annual series on enrollment provided by the Census Bureau from 1946 through 1980 to macro indicators representing the constructs described briefly above. Time-series equations are estimated by sex and age, with consideration given to substantive linkages within and between these equations.