Multiple Classification Analysis (ANOVA command)

STATISTICS=MCA displays the MCA, the Factor Summary, and the Model Goodness of Fit tables.

  • The MCA table presents counts, predicted means, and deviations of predicted means from the grand mean for each level of each factor. The predicted and deviation means each appear in up to three forms: unadjusted, adjusted for other factors, and adjusted for other factors and covariates.
  • The Factor Summary table displays the correlation ratio (eta) with the unadjusted deviations (the square of eta indicates the proportion of variance explained by all categories of the factor), a partial beta equivalent to the standardized partial regression coefficient that would be obtained by assigning the unadjusted deviations to each factor category and regressing the dependent variable on the resulting variables, and the parallel partial betas from a regression that includes covariates in addition to the factors.
  • The Model Goodness of Fit table shows R and R 2 for each model.
  • The tables cannot be produced if METHOD is omitted or if METHOD=UNIQUE. When produced, the MCA table does not display the values adjusted for factors if COVARIATES is omitted, if COVARIATES=FIRST, or if COVARIATES=WITH and METHOD=EXPERIMENTAL. A full MCA table is produced only if METHOD=HIERARCHICAL or if METHOD=EXPERIMENTAL and COVARIATES=AFTER.