The Multidimensional Assessment of Interoceptive Awareness (MAIA)

Study at a glance

• Authors: Wolf E. Mehling, Cynthia Price, Jennifer J. Daubenmier, Mike Acree, Elizabeth Bartmess, Anita Stewart
• Year: 2012
• Journal: PLoS ONE, 7(11), e48230
• DOI: 10.1371/journal.pone.0048230
• N: 325 field-test respondents (309 in the CFA); 16 in cognitive pretest; instructor and patient focus groups
• Design: Mixed-methods scale development: literature review, focus groups with teachers and patients of body-awareness therapies, expert panel, cognitive interviews, then an online field test of 63 items in mind-body practitioners (meditation, yoga/tai chi, massage, somatic therapies). Item reduction by variable clustering (SAS PROC VARCLUS) and confirmatory factor analysis; convergent/discriminant validity against 21 subscales of 9 questionnaires; known-groups comparison; regression on trait anxiety.

Key findings

• Produced the 32-item, 8-scale MAIA: Noticing, Not-Distracting, Not-Worrying, Attention Regulation, Emotional Awareness, Self-Regulation, Body Listening, Trusting. Scales correlated .09 to .60 (median .35).
• An 8-factor model fit adequately by RMSEA (.060) and SRMR (.056), though CFI (.886) and TLI (.870) fell short of conventional cut-offs. A one-factor model fit poorly (CFI .615, RMSEA .108); a hierarchical model with one second-order factor fit almost as well as the 8-factor model (CFI .866, RMSEA .064).
• Internal consistency .66 to .87: three short scales sit below .70 (Noticing .69, Not-Distracting .66, Not-Worrying .67).
• 57 of 63 hypothesized correlation rank orders with other questionnaires were confirmed, all in the predicted direction. Not-Worrying carried the anxiety content (all pain-catastrophizing scales and trait anxiety about -.40 to -.46); in this sample Noticing correlated only -.33 with trait anxiety, which the authors read as Noticing not being an anxiety proxy.
• Experienced teachers (5+ years) scored higher than students on Noticing, Attention Regulation, Not-Worrying and Body Listening.
• Together the scales explained 41% of trait anxiety variance, with Noticing, Not-Distracting, Not-Worrying, Self-Regulation and Trusting each contributing; Emotional Awareness flipped to a positive (marginal) predictor once Self-Regulation was controlled.

Limitations

• Built and validated entirely in experienced mind-body practitioners (79% female, 85% white, mean age 48, most with a graduate degree, 62% with over 10 years of practice), with items chosen for face validity in that group. Scores were skewed high. The authors state that performance in body-awareness-naive samples, and sensitivity to change, were untested.
• Validity rests on a monomethod matrix: every validity measure is itself a questionnaire. The authors call self-report the MAIA’s ‘fundamental limitation’ and ask for validation against heartbeat and airway-resistance detection tasks, which this paper does not do.
• Several intended dimensions (Allowing, Judgmental/Narrative Awareness, Present-Moment Awareness) were lost in item reduction, so the instrument’s mindful-versus-ruminative distinction survives only indirectly, mainly through Attention Regulation.
• Cross-sectional, single sample; known-groups differences are confounded with age and sex (experienced group 11 years older, more men).

Connections

• Methods: Multidimensional Assessment of Interoceptive Awareness (MAIA)
• Concepts: Interoceptive taxonomy (Farb et al. Box 1), Interoceptive sensitivity, Anxiety sensitivity, Emotional granularity
• Debates: Is more interoceptive awareness better?, contemplative training increase objective interoceptive accuracy?
• Related studies: Farb et al. (2015), Ventura-Bort et al. (2021), Eusebio et al. (2022), Meessen et al. (2016)

The development paper for the Multidimensional Assessment of Interoceptive Awareness, the most-used interoceptive sensibility questionnaire in this wiki. Its aim was to separate adaptive body awareness (noticing, trusting, regulating) from the anxious body-focus older questionnaires measured, so that “more body awareness” could mean either.

How it was made, and why that matters

Items came from focus groups with teachers and patients of meditation, yoga, Feldenkrais, somatic therapies and similar practices, and were field-tested only in experienced practitioners. The authors did this on purpose: they wanted items that captured what practitioners learn, and chose to build the scales where people understood the questions best and test naive samples later.

That choice explains two findings the wiki has recorded as puzzles.

Noticing behaves differently in different populations. In these practitioners, Noticing went with slightly lower trait anxiety (r = -.33), and the authors argued it “cannot serve as a proxy measure of anxiety”, adding that this might not hold outside practitioners. In Ventura-Bort et al.’s (2021) German students, Noticing loaded on the attention-deployment factor that predicted worse emotional outcomes. Not a contradiction: the source predicted exactly this dependence on population. It does mean the adaptive reading of Noticing is established only in people trained to notice.

It was built to detect training, and mostly has not. The scales were selected to capture practice-related change, but the paper tests change only cross-sectionally (teachers vs students). The one randomized test in the wiki, Eusebio et al. (2022), found the MAIA rose equally in both yoga arms. See contemplative training increase objective interoceptive accuracy?.

What the source says about total scores

The Multidimensional Assessment of Interoceptive Awareness (MAIA) page concludes that total MAIA scores should not be used. The source is more mixed: a single-factor model fit poorly, but a model with one second-order factor over the eight scales fit almost as well as eight separate factors. That is some support for a shared higher-order dimension, so the case against totals rests on the later evidence that subscales relate to outcomes in opposite directions (Ventura-Bort) rather than on the original factor structure. The authors themselves present and interpret the eight scales separately.

What it never tested

Whether the MAIA tracks interoceptive accuracy. The whole validity matrix is questionnaires against questionnaires, which the authors flag and ask others to address with heartbeat and respiratory detection tasks. Later studies in the wiki that did pair it with a task, such as Meessen et al. (2016), found no relation, consistent with the Interoceptive taxonomy (Farb et al. Box 1) treating it as sensibility only.

Placement

The MAIA page now cites its source first-hand, with the original reliability figures and the second-order model. Authors include Cynthia J. Price, later the developer of Mindful Awareness in Body-oriented Therapy (MABT), and three of them co-wrote Farb et al. (2015). No contradiction with existing pages; one qualification to the total-score recommendation.