Time for a Measurement Check-Up: Testing the Couple's Satisfaction Index and the Global Measure of Sexual Satisfaction Using Structural Equation Modeling and Item Response Theory.

Abstract

Relationship and sexual satisfaction are two central outcomes in the study of relationships and are commonly used in both academia and applied practice. However, relationship and sexual satisfaction measures infrequently undergo specific psychometric investigation. Ensuring that measures display strong psychometric performance is an important but under-tested element of replication that has come under more scrutiny lately, and adequate measurement of constructs is an important auxiliary assumption underpinning theory-testing empirical work. A measurement check-up was conducted, including Confirmatory Factor Analysis (CFA) to test factorial validity, measurement invariance to test for group comparability, and Item Response Theory (IRT) to assess the relationship between latent traits and their items/indicators. This format was used to evaluate the psychometric properties of the Couple's Satisfaction Index (CSI) and the Global Measure of Sexual Satisfaction (GMSEX), two commonly used scales of relationship and sexual satisfaction with a sample of 640 midlife (40-59 years old) married Canadians who were recruited by Qualtrics Panels. Results of CFA suggested that both models were satisfactory. Invariance testing provided robust support for intercept invariance across all the groupings tested. IRT analysis supported the CSI and GMSEX, however, there was evidence that the GMSEX provided somewhat less information for those high on sexual satisfaction. This measurement check-up found that the CSI and GMSEX were reasonably healthy with some caveats. Implications are discussed in terms of replicability and meaning for scholars and practitioners.