One-Way MANOVA (multivariate k groups)
VerifiedMultivariate
Run this test straight away on a free built-in teaching dataset — no data of your own needed — or bring your own. Either opens the guided workspace: variable setup, assumption diagnostics, results with effect sizes and confidence intervals, figures, and APA-ready reporting.
Loading teaching datasets…
Or use your own dataset
Loading your datasets…
Multivariate omnibus test for k ≥ 3 independent groups on p ≥ 2 continuous DVs.
Returns FOUR multivariate statistics (Wilks' Λ, Pillai's trace, Hotelling-Lawley, Roy's largest root) that triangulate the evidence — Pillai's is most robust when assumptions are uncertain. Per-DV univariate follow-ups with Bonferroni correction localise which DVs drive the multivariate effect. Pre-checks multivariate normality (Mardia), homogeneity of covariance (Box's M), and outliers in p-dimensional space.
Worked example
Do three diets differ across several health outcomes at once?
Weight, blood pressure and cholesterol (correlated outcomes) were compared across three diets with MANOVA, followed by univariate tests.
The diets differed on the combined outcomes, Pillai's trace = .27, F = 4.5, p < .001.
A MANOVA showed a significant multivariate effect of diet on the health outcomes, Pillai's trace = .27, F = 4.5, p < .001.
Try it yourself: Load this ready-made sample and follow the run above.
When to use it
- k groups × p continuous correlated DVsManufacturing study compares 4 production shifts (n = 25 each) on 3 quality metrics (defect rate, cycle time, yield).
- Psychometric battery comparison across k groupsBig Five subscale comparison across 4 occupational groups (n = 50 each) on 5 trait scores.
When NOT to — use instead
- Two groups onlyUse Hotelling's T² — MANOVA reduces to it for k = 2. → Hotelling's T\u00b2 (multivariate two-group)
- Single DVUse one-way ANOVA — multivariate machinery isn't needed. → One-Way ANOVA (3+ groups)
- Within-subjects / repeated measuresUse multivariate RM-ANOVA or LMM with multivariate response. → One-Way RM-ANOVA (3+ measurements)
- p > n_j (singular within-cells covariance)Standard MANOVA fails. → One-Way MANOVA (multivariate k groups)
Hypotheses
Parameter tested: set of group mean vectors {μ_j}
Assumptions (and what to do if they fail)
Check: Per-DV Shapiro-Wilk per group.
If violated: MANOVA is fairly robust to mild MVN violations with balanced n; severe non-normality inflates Type I error.
Check: Report Box's M χ² and p.
If violated: Heterogeneous covariance matrices inflate Type I error under unbalanced n.
Ready to run a One-Way MANOVA (multivariate k groups) on your own data?
Guided setup, automatic assumption checks, effect sizes, figures and an APA write-up.
Run this test →