William E Pelham1, Oscar Gonzalez2, Stephen A Metcalf3, Cady L Whicker3, Katie Witkiewitz4, Lisa A Marsch3, David P Mackinnon1. 1. Department of Psychology, Arizona State University, Tempe, AZ 85281. 2. Department of Psychology and Neuroscience, University of North Carolina at Chapel Hill, North Carolina, NC 27599. 3. Department of Psychiatry, Dartmouth College, Lebanon, NH 03766. 4. Center on Alcoholism, Substance Abuse, and Addictions, University of New Mexico, Albuquerque, NM 87131.
Abstract
OBJECTIVE: Nearly all studies treat the Five Facet Mindfulness Questionnaire as five independent scales (one measuring each of the five facets), yet almost no methodological work has examined the psychometric structure of the facets independently. We address this gap using factor analytic methods. METHODS: Exploratory and confirmatory factor models were fit to item response data from a sample of 522 adults recruited online. Findings were replicated in a sample of 454 adults receiving aftercare for substance use disorder. RESULTS: Parallel analysis suggested multiple factors for all five facets, in both samples. Exploratory factor models suggested the presence of method factors on the acting with awareness (items using the term "distraction") and describing facets (items that were reverse-scored). Confirmatory factor models fit poorly for all facets, in both samples. In follow-up analyses, model fit improved substantially on the acting with awareness and describing facets when method factors were included in a bifactor model. Model fit was also better for the facets of FFMQ short forms than for the full-length facets. The short-form facets and original facets correlated similarly with external criteria in both samples. CONCLUSIONS: None of the FFMQ facets fit a unidimensional factor model; yet, follow-up analyses suggested each can be considered substantively unidimensional. Initial tests suggest the facets' multidimensionality did not materially impact their relation to other psychological constructs, suggesting multidimensionality can be ignored for some purposes. The use of short-form facets or latent variable models (e.g., bifactor specifications) are both viable solutions for addressing multidimensionality when desired.
OBJECTIVE: Nearly all studies treat the Five Facet Mindfulness Questionnaire as five independent scales (one measuring each of the five facets), yet almost no methodological work has examined the psychometric structure of the facets independently. We address this gap using factor analytic methods. METHODS: Exploratory and confirmatory factor models were fit to item response data from a sample of 522 adults recruited online. Findings were replicated in a sample of 454 adults receiving aftercare for substance use disorder. RESULTS: Parallel analysis suggested multiple factors for all five facets, in both samples. Exploratory factor models suggested the presence of method factors on the acting with awareness (items using the term "distraction") and describing facets (items that were reverse-scored). Confirmatory factor models fit poorly for all facets, in both samples. In follow-up analyses, model fit improved substantially on the acting with awareness and describing facets when method factors were included in a bifactor model. Model fit was also better for the facets of FFMQ short forms than for the full-length facets. The short-form facets and original facets correlated similarly with external criteria in both samples. CONCLUSIONS: None of the FFMQ facets fit a unidimensional factor model; yet, follow-up analyses suggested each can be considered substantively unidimensional. Initial tests suggest the facets' multidimensionality did not materially impact their relation to other psychological constructs, suggesting multidimensionality can be ignored for some purposes. The use of short-form facets or latent variable models (e.g., bifactor specifications) are both viable solutions for addressing multidimensionality when desired.
Authors: Ian W Eisenberg; Patrick G Bissett; Jessica R Canning; Jesse Dallery; A Zeynep Enkavi; Susan Whitfield-Gabrieli; Oscar Gonzalez; Alan I Green; Mary Ann Greene; Michaela Kiernan; Sunny Jung Kim; Jamie Li; Michael R Lowe; Gina L Mazza; Stephen A Metcalf; Lisa Onken; Sadev S Parikh; Ellen Peters; Judith J Prochaska; Emily A Scherer; Luke E Stoeckel; Matthew J Valente; Jialing Wu; Haiyi Xie; David P MacKinnon; Lisa A Marsch; Russell A Poldrack Journal: Behav Res Ther Date: 2017-10-05