The GLMPOWER Procedure

Syntax: GLMPOWER Procedure

The following statements are available in the GLMPOWER procedure:

  • PROC GLMPOWER <options>;

  • BY variables;

  • CLASS variables;

  • CONTRAST ’label’ effect values <…effect values> </ options>;

  • MANOVA ’label’ <test-options> </ detail-options>;

  • MODEL dependent-variables = independent-effects;

  • PLOT <plot-options> </ graph-options>;

  • POWER <options>;

  • REPEATED factor-specification;

  • WEIGHT variable;

The PROC GLMPOWER statement, the MODEL statement, and the POWER statement are required. If your model contains classification effects, the classification variables must be listed in a CLASS statement, and the CLASS statement must appear before the MODEL statement. In addition, CONTRAST and POWER statements must appear after the MODEL statement. PLOT statements must appear after the POWER statement that defines the analysis for the plot.

If you specify one or more MANOVA or REPEATED statements, then the model is assumed to be multivariate. Otherwise, a univariate model is assumed, in which case multiple dependent variables represent cell means scenarios for a single response.

You can use multiple CONTRAST , MANOVA , REPEATED , POWER , and PLOT statements. Each CONTRAST statement defines a separate between-subject contrast. Each MANOVA or REPEATED statement defines a separate within-subject contrast for a multivariate model. Each POWER statement produces a separate analysis and uses the information that is contained in the CLASS , MODEL , WEIGHT , CONTRAST , MANOVA , and REPEATED statements. Each PLOT statement refers to the previous POWER statement and generates a separate graph (or set of graphs).

Table 47.3 summarizes the basic functions of each statement in PROC GLMPOWER. The syntax of each statement in Table 47.3 is described in the following pages.

Table 47.3: Statements in the GLMPOWER Procedure

Statement

Description

PROC GLMPOWER

Invokes procedure and specifies exemplary data set

BY

Specifies variables to define subgroups for the analysis

CLASS

Declares classification variables

CONTRAST

Defines between-subject linear tests of model parameters

MANOVA

Defines within-subject linear tests of model parameters for multivariate models, in terms of contrast matrix coefficients

MODEL

Defines model and specifies dependent variables; for univariate models, multiple dependent variables represent cell means scenarios for a single response

PLOT

Displays graphs for preceding POWER statement

POWER

Identifies parameter to solve for and provides one or more scenarios for values of other analysis parameters

REPEATED

Defines within-subject linear tests of model parameters for multivariate models, in terms of common repeated measures transformations of the dependent variables

WEIGHT

Specifies variable for allocating sample sizes to different subject profiles