Practical Significance and Statistical Models

Abstract

The results of an extensive analysis of the statistical methodology employed by researchers whose manuscripts were published in the first twelve volumes of the Educational Administration Quarterly are re ported. It is concluded that univariate analysis has been utilized pre dominately in the treatment of data ivhile tests for practical significance and significance of relationships have been largely ignored. A case is made for the application of the omega, squared statistical index to esti mate the proportion of the variance of the dependent variable accounted for by a given independent variable. James F. McNamara is Professor of Interdisciplinary Education and Educational Administration at Texas A &M University.