Investigating the measurement of social and academic competencies for early late preadolescents and adolescents: A multitrait-multimethod analysis

The purposes of the study were twofold: (a) to test for evidence of construct validity related to social and academic (general, English, mathematics) competencies, as measured by self, teachers, parents, and peers, for children in Grades 3 (n = 175), 7 (n = 193), and 11 (n = 159), and (b) to examine the extent to which ratings by significant others were concordant with self-ratings for these four constructs. Based on a comparison of nested multitrait-multimethod covariance structure models, results demonstrated that despite evidence of global convergent and discriminant validities, method effects were instrumental in attenuating trait variances. In particular, self-ratings for early preadolescents and peer ratings for all ages were the most vulnerable to this anomaly. Overall, results suggest that researchers consider making greater use of parent ratings in the measurement of social and academic competencies for elementary children.