A multiregression model for predicting enrolments in the first seven grades of the queensland state school system

The application of mathematical model-building in educational research and planning is a very recent development. Planners have found that their ad hoc trend analysis procedures are not good enough to accommodate a very rapidly changing system. The multiregression model proposed in this paper has been derived from the class of straightforward least-squares applications of the general linear model, without any built-in statistical tests. The procedure used is a data search procedure in which the intent is to evaluate a number of variables substantively, rather than in terms of formal statistical procedures. The resulting multiregression model appears to be an excellent predictor of enrolment patterns in that it accounts for 99.49 per cent of the variance of the criterion variable. Its application in this example can be likened to a “moving least squares” model