Measurement Scales in Accounting: A Case for Summated Scores

Abstract

Factor analysis is commonly used in accounting research to combine scale items. This practice is highly dubious and may be creating variable measures that do not accurately reflect the hypothesized constructs. This problem can explain why studies often get inconsistent results or lead hypothesized relationships to be non-significant. Our paper can act as an instructive guide for the reasons why factor analysis and the proceeding factor score is often inappropriate for combining scale items and makes a case for the use of summated scores. As evidence, this paper uses data from a paper using summated scores (Bailey, 2015) to explore how factor analysis and factor scores could have disrupted the integrity of the focal construct. This empirical testing exercise helps reveal the issues of an over dependence on empirical results and why summated scores can serve as an alternative to ensure confidence for researchers using well established multi-item measures.