Establishing measurement invariance across time within an accelerated longitudinal design

Abstract

The chapter provides a review of the potential value of accelerated longitudinal designs (ALD) for research, through (i) a systematic literature review, and (ii) an original investigation showing in particular the importance of evaluation of measurement invariance. It empirically draws on a large-scale UK study with a substantial aim to map teaching practices with students’ learning outcomes (including dispositions) in mathematics during secondary school. The focus here is on the methodological implications of combining the separate cohorts of students’ outcomes into a longitudinal model of progression for the whole secondary school trajectory and in particular the measurement methods involved to make this possible. Other challenges are also explored in the context of a comprehensive procedure that covers the validation of the construct of mathematics disposition and subsequent multilevel modelling to account for the complex longitudinal design.