Title: Exploring measurement invariance with the alignment method in the First and Second International Science studies
Abstract: This study evaluates measurement invariance in the First and Second International Science studies (FISS and SISS) conducted by the International Association for the Evaluation of Educational Achievement in 1970-71 and 1983-84. We apply a traditional stepwise procedure and also utilize the alignment method, within the framework of item response theory. Findings show that configural invariance is established across most educational systems, while scalar or partial invariance (from a stepwise procedure) could not be established in either of the assessments. The alignment method identified a partial invariance model that fitted the data well but did not identify any items that had invariant parameters across all educational systems. Rankings of educational systems based on estimates from sum scores, a scalar model, and the alignment-derived partial invariance model showed some differences in both assessments. Finally, we outline future analyses to examine the feasibility of comparing science means over time in FISS and SISS.