ANALYSIS Subcommand (MANOVA: Univariate command)

ANALYSIS allows you to work with a subset of the continuous variables (dependent variable and covariates) that you named on the MANOVA variable list. In univariate analysis of variance, you can use ANALYSIS to allow factor-by-covariate interaction terms in your model (see the DESIGN subcommand below). You can also use ANALYSIS to switch the roles of the dependent variable and a covariate.

  • In general, ANALYSIS gives you complete control over which continuous variables are to be dependent variables, which continuous variables are to be covariates, and which continuous variables are to be neither.
  • ANALYSIS specifications are like the MANOVA variables specification, except that factors are not named. Enter the dependent variable and, if there are covariates, the keyword WITH and the covariates.
  • Only variables that are listed as dependent variables or covariates on the MANOVA variable list can be entered on the ANALYSIS subcommand.
  • In a univariate analysis of variance, the most important use of ANALYSIS is to omit covariates from the analysis list, thereby making them available for inclusion on DESIGN (see the example below and the DESIGN subcommand examples).
  • For more information about ANALYSIS, refer to ANALYSIS Subcommand (MANOVA: Multivariate command).

Example

MANOVA DEP BY FACTOR(1,3) WITH COV
  /ANALYSIS DEP
  /DESIGN FACTOR, COV, FACTOR BY COV.
  • COV, a continuous variable, is included on the MANOVA variable list as a covariate.
  • COV is not mentioned on ANALYSIS, so it will not be included in the model as a dependent variable or covariate. COV can, therefore, be explicitly included on the DESIGN subcommand.
  • DESIGN includes the main effects of FACTOR and COV and the FACTOR by COV interaction.