Response to Lunneborg: The Conditions for Interpretation of Regression Weights

Abstract

In a reply to McNeil (1990), Lunneborg (1991) indicates his strong desire to interpret regression weights. While taking such a stand, he hints at several conditions, but does not explore them deeply. Unfortunately, these conditions are seldom obtained in applications of the General Linear Model. Although these conditions can be obtained, most researchers do not obtain them and are often too impatient to restrain their interpretations. These conditions are R^2 close to 1.0 and predictor variables that have been manipulated.