RANDOM Subcommand (GLM: Univariate command)

RANDOM allows you to specify which effects in your design are random. When the RANDOM subcommand is used, a table of expected mean squares for all effects in the design is displayed, and an appropriate error term for testing each effect is calculated and used automatically.

• Random always implies a univariate mixed-model analysis.
• If you specify an effect on RANDOM, higher-order effects containing the specified effect (excluding any effects containing covariates) are automatically treated as random effects.
• The keyword INTERCEPT and effects containing covariates are not allowed on this subcommand.
• The RANDOM subcommand cannot be used if there is any within-subjects factor in the model (that is, RANDOM cannot be specified if WSFACTOR is specified).
• When the RANDOM subcommand is used, the appropriate error terms for the hypothesis testing of all effects in the model are automatically computed and used.
• More than one RANDOM subcommand is allowed. The specifications are accumulated.

Example

GLM DEP BY A B
  /RANDOM = B
  /DESIGN = A,B, A*B.

• In the example, effects B and A*B are considered as random effects. If only effect B is specified in the RANDOM subcommand, A*B is automatically considered as a random effect.
• The hypothesis testing for each effect in the design (A, B, and A*B) will be carried out by using the appropriate error term, which is calculated automatically.