The study titled “The Multidimensional Assessment of Interoceptive Awareness, Version 2 (MAIA-2)”, examined whether a popular questionnaire (MAIA) for measuring body awareness could be improved.
The original MAIA, published in 2012, is widely used. It had 32-question across 8 body-awareness skills.
- Noticing – Awareness of physical sensations.
- Not-Distracting – Tendency not to ignore or suppress discomfort.
- Not-Worrying – Ability to sense discomfort without spiraling(over-worrying).
- Attention Regulation – Staying focused on bodily sensations.
- Emotional Awareness – Noticing how emotions show up physically.
- Self-Regulation – Using body awareness to calm distress.
- Body Listening – Seeking information from the body.
- Trusting – Feeling safe in one’s body.
But two of the scales—Not-Distracting and Not-Worrying— had lower reliability. This paper reports how researchers refined the tool to make it more accurate and useful.
Researchers collected data from 1,090 adult visitors (ages 18–69) to the Science Museum in London. Most were healthy members of the general public, and about 60% were native English speakers.
Participants completed:
- The original MAIA (32 items, 8 scales)
- 6 new items designed to improve the two weaker scales
The new questions were developed by the same team who built the original MAIA, based partly on earlier focus group input.
The researchers split the data into two groups:
- Exploratory analysis (like looking for patterns) in one half
- Confirmatory analysis (testing if the new structure holds) in the other half
They kept the 8-scale structure but dropped one of the new items that didn’t fit well.
What They Found
Adding five of the new items improved the reliability of both target scales:
- Not-Distracting: Reliability improved from 0.66 to 0.74
- Not-Worrying: Reliability improved from 0.58–0.65
The updated tool, MAIA-2, now has 37 items and all eight scales retained their original meaning.
Why This Matters
Having a more reliable measure of body awareness means:
- Better evaluation of interventions like mindfulness, yoga, or somatic therapy (therapy that focuses on the mind–body connection)
- More precise tracking of progress over time
- Stronger research data to support your programs
MAIA-2 offers steadier readouts on two practical pain-and-stress domains: “distraction from discomfort” and “worrying about sensations.”
That means clearer pre/post changes when you teach or learn breath work, body scans, or pain-coping skills.
Why Did This Happen
Improving reliability often comes down to:
- Adding more items that capture different facets of the same construct
- Reducing ambiguity in wording
- Ensuring items align in scoring direction (so “high” always means more of the trait)
In MAIA-2, the extra items gave a fuller picture of how people manage discomfort and worry about bodily sensations.
Implications / Benefits
- More dependable results for research and clinical trials
- Better suited for general population samples, not just mind-body practitioners
- Free, public-domain access makes it easy to adopt
How You Can Try This
- Download MAIA-2 or use our online test — it’s completely private and runs locally on your device.
- Use the test before and after mindfulness or yoga training. Allow at least a couple of weeks between tests to detect measurable change.
- Track which domains shift (e.g., Attention Regulation, Self-Regulation) and connect them to specific practices (breathwork, body scan).
- Don’t rely on a single “total” score — use the subscale results in line with your goals calming, listening, trust.
Final Thoughts
The MAIA-2 shows that even well-established tools can be improved. In this case, a few thoughtful tweaks made body awareness measurement sharper and more reliable—good news for anyone studying or cultivating mindful connection to the body.
Technical Summary
TL;DR
- Study: Psychometric upgrade of the MAIA questionnaire using 1,090 adult museum visitors; split-sample exploratory clustering and confirmatory factor analysis.
- Finding: Five new items were added; MAIA-2 (37 items) kept the 8-factor structure and improved reliability for Not-Distracting and Not-Worrying. Fit: RMSEA = .055 (95% CI .052–.058), SRMR = .064; χ²(601)=1597.7, p < .0001.
- Why does it matter: If you track body-awareness training (e.g., mindfulness, yoga, pain programs), MAIA-2 gives more dependable subscale scores for “distraction” and “worry” about sensations, with the same familiar 8 domains.
Feeling your body accurately can shape stress, focus, and self-regulation. MAIA is one of the most used self-report tools to measure this across eight domains.
Researchers split a large adult sample (N=1,090) into two halves: one for exploratory clustering (PROC VARCLUS) and one for confirmatory factor analysis (CFA). Five of six candidate items were retained, producing a 37-item MAIA-2. The eight-factor model showed acceptable fit (RMSEA = .055, SRMR = .064), though CFI/TLI were below usual cutoffs (.860/.845).
Reliability (Cronbach’s alpha) improved for Not-Distracting (.74) and modestly for Not-Worrying (.67). The tool remains public domain and available for research and clinical evaluation.
Methods Overview
Visitors (18–69 years) at the Science Museum London completed the original 32-item MAIA plus six new items. The sample was split (odd/even IDs) for exploratory clustering vs. CFA; missing data were minimal. Analyses used SAS PROC VARCLUS for clustering and Mplus for CFA, with predefined fit criteria and Cronbach’s alpha; Feldt tests compared alphas with prior samples.
Study Design Snapshot
| Item | Details |
|---|---|
| Design | Psychometric validation with split-sample exploratory clustering and CFA |
| Participants | N = 1,090 adults; convenience sample of museum visitors; 47% female; mean age 30.6 (SD 11.3); ~60% native English speakers. |
| Intervention / Exposure | Completion of MAIA (32 items) + 6 candidate items; five were retained, yielding 37 items. |
| Comparator | Internal comparison across versions/scales; Feldt test vs. prior samples (original, primary care, German validation). |
| Outcome Measures | Factor structure (RMSEA, SRMR, CFI, TLI, χ²), standardized loadings, Cronbach’s alpha, scale-scale correlations. |
| Follow-up Length | Not reported |
| Statistical Analysis | SAS PROC VARCLUS (quartimax) for clustering; Mplus ML CFA with factor covariances; a-priori fit thresholds (e.g., CFI >.90, RMSEA < .06, SRMR < .08); Cronbach’s alpha; Feldt tests for alpha comparisons. |
Key Findings
- The eight-factor structure was retained after adding five items (total = 37).
- Reliability improved notably for Not-Distracting and modestly for Not-Worrying.
- Model fit was acceptable by RMSEA/SRMR; CFI/TLI were below conventional targets.
- Scale-scale correlations were moderate; strongest r = .52 (Self-Regulation with Body-Listening; Emotional Awareness with Body-Listening).
Numbers (selected):
- CFA fit: RMSEA = .055 (95% CI .052–.058); SRMR = .064; CFI = .860; TLI = .845; χ²(601)=1597.7, p < .0001.
- Cronbach’s alpha (MAIA-2): Not-Distracting =.74; Not-Worrying =.67; Noticing =.64; Attention Regulation =.83; Emotional Awareness =.79; Self-Regulation =.79; Body Listening =.80; Trust =.83.
- Alpha improvement (Feldt test): Not-Distracting improved vs. original and two comparison samples (p ≤ .001); Not-Worrying improved vs. primary care (p = .001), not vs. original/German.
- Lowest item loading: original item 5 in Not-Distracting (λ = .30); removing it raises alpha only to .76; all six items reverse-scored.
Limitations Stated by Authors
- Self-report only; no objective criterion to validate scales.
- Convenience sample of museum visitors; limited characterization.
- Fit indices like CFI/TLI debated in the field.
- High proportion of non-native English speakers; sensitivity analyses suggested minimal impact.
Critical Notes
- CFI (.860) and TLI (.845) fell short of stated target thresholds (CFI > .90; TLI > .95), despite RMSEA/SRMR meeting criteria—so model fit is “acceptable,” not excellent.
- Retained a weakly loading reverse-scored item to preserve content coverage; reverse scoring itself can reduce reliability.
- No test–retest stability or responsiveness reported here (important for program monitoring). Conclusive takeaway: MAIA-2 improves internal consistency for Not-Distracting and keeps the 8-factor structure in a large adult sample, but some fit indices remain below ideal.
Glossary
- Interoception: Sensing and interpreting signals from inside the body.
- Interoceptive awareness (sensibility): Conscious, self-reported aspects of interoception (what you notice and believe about your body).
- Cronbach’s alpha: A number (0–1) showing how consistently items on a scale measure the same idea.
- CFA (Confirmatory Factor Analysis): A test of whether data fit a hypothesized factor structure; common fit indices include RMSEA, SRMR, CFI, TLI.
- Not-Distracting / Not-Worrying: Subscales reflecting less tendency to distract from or worry about discomfort.
Paper Quality (Moderate-High)
Solid upgrade of a questionnaire in a large sample, but convenience sampling and some model indices below ideal reduce certainty.
| Criterion | Rating | Notes |
|---|---|---|
| Randomisation & blinding | 🔴 Low | Observational psychometric study; convenience sample at a museum (no randomization/blinding). |
| Sample size & power | 🟡 Moderate | N=1,090; target ≥500 stated for factor analyses; no formal power calc reported. |
| Attrition / adherence | 🟢 High | Minimal missing data (12/1,090). |
| Conflicts of interest | 🟢 High | Authors declared no competing interests. |
Suggested Reads
- Mehling W.E. et al. (2012). The Multidimensional Assessment of Interoceptive Awareness (MAIA). PLOS ONE.
- Bornemann B., Herbert B.M., Mehling W.E., Singer T. (2014). Differential changes in self-reported interoceptive awareness through 3 months of contemplative training. Frontiers in Psychology.
- Mehling W. (2016). Differentiating attention styles and regulatory aspects of self-reported interoceptive sensibility. Philosophical Transactions of the Royal Society B.