Title: Measurement invariance across educational systems in the First and Second International Science studies
Abstract: We investigate measurement invariance across educational systems in the First and Second International Science studies conducted by the International Association for Evaluation of Educational Achievement in 1973-74 and 1983-84, respectively. The study applied classical item analysis and multiple group item response theory modeling. Findings showed that some of the items in the assessments had negative point-biserial correlations in at least one educational system. A unidimensional 2PL model fitted well in all but three educational systems in FISS and all but one in SISS. However, we detected a lack of strong and partial invariance across educational systems in both assessments. We give an outline of a supplementary analysis which aims to examine the robustness of the findings with two alternative methods for evaluating measurement invariance.