Increased Effect Sizes in a Mindfulness- and Yoga-Based Intervention After Adjusting for Response Shift with Then-Test

Abstract

Objectives: Response shift refers to variations in self-reported evaluations at different times from changes in one’s internal standards, values, and meanings. The current study explored the utility of the then-test to detect a potential mindfulness-based response shift occurrence during a mindfulness- and yoga-based intervention for student musicians, and to ascertain to what extent effect sizes could differ when adjusting for it. Method: Participants (n = 31) completed the Five Facet Mindfulness Questionnaire (FFMQ) halfway through the intervention (Time 1-FFMQ), post-intervention (Time 2-FFMQ), and immediately after Time 2-FFMQ with a then-test approach that asked participants to rate the FFMQ based on retrospective reflections on their mindfulness at Time 1 (then-test-FFMQ). Paired t-tests and Hedges’ g effect sizes were computed to estimate three potential effects: response shift (Time 1-FFMQ minus then-test-FFMQ), the conventional intervention effect (Time 2-FFMQ minus Time 1-FFMQ), and the effect after adjusting for response shift (i.e., actual intervention effect = Time 2-FFMQ minus then-test-FFMQ). Results: Response shift was significant for the FFMQ Observe subscale (g = 0.41) and total scale (g = 0.37). The adjusted scores in all subscales (Observe, g = 0.47; Describe, g = 0.25; Act Aware, g = 0.40; Non-judge, g = 0.28; Non-react, g = 0.57) and total scale (g = 0.60) achieved significance and yielded larger effect sizes than the conventional results, for which only Act Aware (g = 0.28), Non-react (g = 0.36), and total scale (g = 0.28) were significant. Conclusions: Notwithstanding some methodological limitations, this study lends support to the utility of the then-test to quantify response shift. When adjusting for it, effect sizes from a mindfulness- and yoga-based intervention were generally amplified. Preregistration: This study was not preregistered.