Being Mindful About Overuse of Total Scores: a Comparison of Total Scores and Moderated Nonlinear Factor Analysis Scores in Assessing Mindfulness Across Race/Ethnicity, Age, and PTSD Diagnosis

TL;DR

This study compares total scores and a more advanced method for measuring mindfulness, finding that the advanced method reveals important differences across groups like race, age, and PTSD diagnosis.

Contribution

The study introduces moderated nonlinear factor analysis to detect measurement bias in mindfulness scores across demographic and clinical groups.

Findings

Measurement noninvariance was found across race/ethnicity, age, and PTSD diagnosis.

Hispanic participants scored higher on acting with awareness but lower on describing and observing.

Total scores overestimated differences for Hispanic participants and those with PTSD.

Abstract

Although mindfulness-based interventions show initial positive results on a range of substance use behaviors, evaluations of mindfulness-based interventions would benefit from state-of-the-art alternative approaches to the ubiquitous use of sum or total scores. Sum scores do not reflect “true” underlying mindfulness as they do not consider differences in the relative weight of each item and/or the possibility that measurement may differ across groups. The purpose of this study was to identify measurement noninvariance/differential item functioning (MNI/DIF) across racial and ethnic groups, age groups, and those with PTSD diagnoses and differences in inferences on the factors of the Five Facet Mindfulness Questionnaire between scale scores estimated using moderated nonlinear factor analysis (MNLFA) and a total score analog model (TSA). Age, PTSD diagnosis, non-Hispanic Black race/ethnicity, Hispanic race/ethnicity, and other race/ethnicity showed statistically significant MNI/DIF. In the MNLFA model, PTSD diagnosis and Hispanic race/ethnicity contributed to significant MNI/DIF on the “true” acting with awareness, describing, and observing latent factors such that Hispanic participants were higher on average on acting with awareness scores and lower on average on describing and observing scores. The TSA model failed to estimate significant differences on acting with awareness score for participants with PTSD diagnosis. Additionally, in the TSA model, there was an increase in the effect size of Hispanic participants’ baseline describing and observing estimates, thus overestimating differences in respective scores for Hispanic participants. Failing to correct for MNI/DIF in mindfulness scale scores can impact inferences and effect sizes for group differences in mindfulness thus creating bias in characterizing mindfulness, particularly for Hispanic individuals and those with PTSD diagnoses.