Validation and invariance of the Conceptions of Assessment-III Abridged (COA-IIIA) among pre-service and in-service teachers in the United States

As beliefs are well-known antecedents of teachers’ practices, including assessment practices, sound measurement of teacher beliefs is critical for scholarly research as well as practical purposes. The present study examined the validity of inferences derived from the Conceptions of Assessment III—Abridged (COA-IIIA) instrument with US PK-12 pre-service ( N = 554) and in-service ( N = 341) teachers. Prior research on the COA-IIIA in diverse contexts and with different teacher populations has yielded highly divergent findings concerning this instrument’s internal score structure. The study relies on confirmatory factor analyses to investigate the COA-IIIA’s score structure among US pre-service and in-service teachers and to test the measurement invariance of select COA-IIIA measurement models across these groups. Findings indicate poor fit of most prior COA-IIIA measurement models with US pre-service and in-service teachers, though we were able to specify several simplified models that were at least metrically invariant across these populations.