Skip to content

ANOVA / ANCOVA / MANOVA

🎯 Purpose

This checklist provides targeted visual EDA steps for preparing to run ANOVA-family models. It ensures that key assumptions like normality, homogeneity of variances, and linearity are visually inspected before formal testing.

🟦 ANOVA (1 DV, categorical group)

  • [ ] Outcome distribution checked by group
  • [ ] Boxplot or violin plot created
  • [ ] Group sizes reviewed for balance
  • [ ] Levene’s or Bartlett’s test for equal variances
  • [ ] Normality Check: QQ plot or histogram of residuals (after a preliminary model fit) reviewed.

🟨 ANCOVA (1 DV, 1+ covariate + group)

  • [ ] Scatterplot of DV vs covariate, colored by group
  • [ ] Group × covariate interaction checked
  • [ ] Covariate distributions compared across groups
  • [ ] Linearity: Ensure the relationship between the covariate and DV is linear across all groups.

🟩 MANOVA (2+ DVs, 1 categorical IV)

  • [ ] Pairplot or correlation matrix of DVs
  • [ ] DVs standardized if scales differ
  • [ ] Visual group separation assessed (PCA optional)
  • [ ] Multivariate Normality: Check histograms for each DV by group.

🟥 MANCOVA (2+ DVs, 1+ covariates, 1 group)

  • [ ] Scatterplot matrix with grouping/covariates
  • [ ] Partial plots or adjusted means reviewed
  • [ ] Covariate balance across groups verified
  • [ ] Homogeneity of Slopes: Visually check that the relationship between covariates and DVs is consistent across groups.

🧠 Final Tip

"Visual EDA for these models is about confirming that groups are comparable before you test if they are different."