This paper estimates a model of the relationship between self-rejection and deviance, mediated by disposition to deviance, to examine the linear structural equations in data provided by a threewave panel study of junior high school students. The measurement models and latent structure more accurately reflect the theoretical basis of the postulated structure than do earlier multivariate analyses of panel data. The latent construct "self-rejection" (modeled in terms of global self-derogation, lack of desirable attributes, and experiences of rejection and failure in family and school) a Time 1 has the expected strong positive effect on the latent construct "disposition to deviance" (measured in terms of disaffection from, and readiness to adopt deviant alternatives to, conventional family, school, and community patterns) at Time 2 which in turn has the anticipated strong positive effect on the latent construct "deviance" (reflecting the commonality underlying the adoption of any of a range of deviant behaviors in three different groupings based on probability of occurrence). These findings differ from those of other multivariate analyses guided by the same theoretical framework that report negligible effects of self-rejection on deviance. The differences in findings are due to the theoretically indicated decisions in the present analysis to model (1) self-rejection in terms of the association between negative self-feelings and experiences of rejection in the conventional environment, (2) self-rejection as having major direct effects on acquired disposition to deviance, and (3) deviance as general behavioral outcome that might be expressed in terms of any of a number of behaviors that violate conventional norms. Further, the present analysis permits the relaxation of assumptions regarding infallible measurement.