Book Chapter

Selectivity and generalizability in longitudinal research: on the effects of continuers and dropouts

Little, T. D., Lindenberger, U., Maier, H.

In: Little, T. D., Schnabel, K. U., Baumert, J. (Eds.): Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples, 187-200 (2000)
Mahwah N.J.: Lawrence Erlbaum Associates.


Describes procedures for assessing the degree of selectivity and thereby representativeness and generalizability in longitudinal studies. Sample selectivity, or nonrandom participation, is a common but often neglected threat to generalizability in longitudinal research. In this context, selectivity refers to the potential for systematic differences between continuers and dropouts. This chapter deals with 2 selectivity questions referring to (1) the differences between participation subgroups on the means, variances, and covariances of observed variables and constructs that the subgroups share in common; and (2) the relationships between variables at previous time points and variables assessed at later time points, which are used to project, or adjust, the estimates of the multivariate relations in the completing subsample. Techniques for dealing with selectivity questions are illustrated using two subsamples from a group of 425 boys and girls (2nd-6th graders) randomly assigned to be dropouts or continuers and examined school performance, academic achievement, and personal agency constructs. (AUTHORS)