Analysis of a cohort prediction model with applications to student enrollment forecasting
Raymond David Hager · Calhoun: The Naval Postgraduate School Institutional Archive (Naval Postgraduate School) · 1970
A model is presented for the prediction of future organization size based on the numbers of recruits entering the organization in the past. This model utilizes the correlation between populations of successive time periods in order to better estimate the future remaining personnel in the system. The number of personnel leaving from each recruit cohort is assumed to follow the same probability distribution, which is a function of the age of the cohort in the organization and the grade in which the cohort started. For large cohort sizes the total personnel in the system is approximately normally distributed. This result justifies the use of a best linear prediction method which takes into account past errors of estimating the continuing population from one period to the next. Sensitivity of predictions to errors in probability estimates is discussed. The model is applied to predicting university student enrollment. Comparison of predicted and actual student enrollment is included.