Estimation of Method and Trait Variance in Ratings of Conduct Disorder

Abstract

A structural equation model designed to estimate the contributions of trait, method and error variance to maternal and teacher and child ratings of conduct disorder is presented. This model was fitted to data collected on a birth cohort of New Zealand children studied over a 3 yr period. On the basis of the fitted model it was estimated that between 28% and 40% of the variance in maternal and teacher ratings was ascribable to variations in the child''s generalized behavioral tendencies and the remaining variance to either method-specific factors or random errors of measurement. The implications of these results are discussed.