Dataset Information

Testing Measurement Invariance Across Unobserved Groups: The Role of Covariates in Factor Mixture Modeling.

ABSTRACT: Factor mixture modeling (FMM) has been increasingly used to investigate unobserved population heterogeneity. This study examined the issue of covariate effects with FMM in the context of measurement invariance testing. Specifically, the impact of excluding and misspecifying covariate effects on measurement invariance testing and class enumeration was investigated via Monte Carlo simulations. Data were generated based on FMM models with (1) a zero covariate effect, (2) a covariate effect on the latent class variable, and (3) covariate effects on both the latent class variable and the factor. For each population model, different analysis models that excluded or misspecified covariate effects were fitted. Results highlighted the importance of including proper covariates in measurement invariance testing and evidenced the utility of a model comparison approach in searching for the correct specification of covariate effects and the level of measurement invariance. This approach was demonstrated using an empirical data set. Implications for methodological and applied research are discussed.

SUBMITTER: Wang Y

PROVIDER: S-EPMC7797957 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:AimTo investigate whether items of the SF-12, widely used to assess health outcome in clinical practice and public health research, provide unbiased measurements of underlying constructs in different demographic groups regarding gender, age, educational level and ethnicity.MethodsWe included 23,146 men and women aged 18-70 of Dutch, South-Asian Surinamese, African Surinamese, Ghanaian, Turkish, or Moroccan origin from the HELIUS study. Both multiple group confirmatory factor analyses (MGCFA), with increasingly stringent model constraints (i.e. assessing Configural, Metric, Strong and Strict measurement invariance (MI)), and regression analysis were conducted to establish comparability of SF-12 items across demographic groups.ResultsMI regarding gender, age and education was tested in the ethnic Dutch group (N = 4,615). In each subsequent step of testing MI, change in goodness-of-fit measures did not exceed 0.010 (RMSEA) or 0.004 (CFI). Moreover, goodness-of-fit indices showed good fit for strict invariance models: RMSEA<0.055; CFI>0.97. Regarding ethnicity, RMSEA values of metric and subsequent models fell above 0.055, indicating violation of measurement invariance in factor loadings, thresholds and residual variances. Regression analysis revealed possible age-, education- and ethnicity-related DIF. Adjustment for this DIF had little impact on the magnitude of age and educational differences in physical and mental health, but ethnic inequalities in physical health-and to a lesser extent mental health-were reduced after DIF adjustment.ConclusionsWe found no evidence of violation of measurement invariance of the SF-12 regarding gender, age and educational level. If minor DIF would remain undetected in our MGCFA analyses, we showed that this would have negligible effect on the magnitude of demographic health inequalities. Regarding ethnicity, the SF-12 was not measurement invariant. After accounting for DIF, we observed a reduction of ethnic inequalities in health, in particular in physical health. Caution is warranted when comparing SF-12 scores across population groups with various ethnic backgrounds.

Dataset Information

Testing Measurement Invariance Across Unobserved Groups: The Role of Covariates in Factor Mixture Modeling.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets