General Electric Company

unbalanced designs By means of three analyses, the operating characteristics of Levene s test ofhomoscedasticity in a 1-way layout are shown to be systematically affected by design imbalance when the error terms are normal. We consider a formal general linear hypothesis analysis, an empirical analysis, and an analysis of expected mean squares, employing the heuristic justification for an ordinary ANOVA. We propose a simple correction involving leverage and analyze its effect, as well as that of an ither adjustment, on power and level. Levene (1960) proposed a heuristic test for heteroscedasticity in the case of a balanced two-way layout, based upon an ANOVA of absolute residuals. Conover, Johnson, and Johnson (1981) verified empirically that Levene's test is one of the best tests in terms of power and robustness of level under normality violations. It is often presented as an option for testing heteroscedasticity in popular statistics packages, for example, BMDP Pro-grams 3D and 7D (Brown, Engelman, & Jennrich, 1990, pp. 147, 193) usin