An introduction to multi-battery factor analysis: Overcoming method artefacts.

Abstract

Examination of participants’ responses to factor or scale scores provides useful insights, but analysis of such scores from multiple measures or batteries is sometimes confounded by methodological artefacts. This paper provides a short primer into the use of multi-trait, multi-method (MTMM) correlational analysis and multi-battery factor analysis (MBFA). The principles of both procedures are outlined and a case study is provided from the author’s research into 233 teachers’ responses to 22 scale scores drawn from five batteries. The batteries were independently developed measures of teachers’ thinking about the nature and purpose of assessment, teaching, learning, curriculum, and teacher efficacy. Detailed procedures for using Cudeck’s (1982) MBFACT software are provided. Both MTMM and MBFA analyses identified an appropriate common trait across the five batteries, whereas joint factor analysis of the 22 scale scores confounded the common trait with a battery or method artefact. When researchers make use of multiple measures, they ought to take into account the impact of method artefacts when analyzing scale scores from multiple batteries. The multi-battery factor analysis procedure and MBFACT software provide a robust procedure for exploring how scales inter-relate.