Statscraft 0.7.0, release v0.7.0

ANOVA

Analysis of variance for factorial designs with covariates, the older form with MCA tables.

Syntax

ANOVA [VARIABLES=] dependents BY factors(min,max) [factors(min,max) ...] [WITH covariates] [/next list ...]
      [/COVARIATES=FIRST | WITH | AFTER] [/MAXORDERS=ALL | n | NONE]
      [/METHOD=UNIQUE | EXPERIMENTAL | HIERARCHICAL] [/STATISTICS=[MCA] [REG] [MEAN] [ALL] [NONE]]
      [/MISSING=EXCLUDE | INCLUDE].
A range follows the factors it applies to; factor values are truncated to whole numbers and cases
outside a range are left out, as is a case missing on any variable of its list. The model has the
covariates, the main effects and the interactions up to MAXORDERS (five by default); an order of
interactions with empty cells is dropped with a warning. UNIQUE (default) adjusts every effect for all
others (Type III); EXPERIMENTAL assesses covariates, main effects and each order of interactions in
turn, each adjusted for its own kind and the kinds before it; HIERARCHICAL adjusts each covariate and
factor only for those before it. COVARIATES sets where the covariates enter. Tables: ANOVA (with REG the
coefficients of the covariates where they enter), and, not with UNIQUE, Cell Means (MEAN), MCA,
Factor Summary and Model Goodness of Fit (MCA). STATISTICS without keywords is MEAN and REG. Weighted
when a weight variable is set.

Example

ANOVA VARIABLES=score BY region(1,4) sex(1,2) WITH age /METHOD=EXPERIMENTAL /STATISTICS=MEAN MCA.

statscraft help ANOVA prints the same text in the terminal. All commands