Modeling the Time-Varying Nature of Student Exceptionality Classification on Achievement Growth

Abstract

Our purpose was to examine different approaches to modeling the time-varying nature of exceptionality classification. Using longitudinal data from one state’s mathematics achievement test for 28,829 students in Grades 3 to 8, we describe the reclassification rate within special education and between general and special education, and compare four alternative growth models for students with and without disabilities with different specifications of disability classification as time-variant (TVC) or time-invariant (TIC) covariates. Although model fit statistics were inconsistent in endorsing a single model, we found that the TIC results were generally preferable to the TVC; however, the choice of model specification may rest on the purpose of the researcher and goals of representing the influence of covariates on growth.