PRINT Subcommand (GLM: Univariate command)

PRINT controls the display of optional output.

  • Some PRINT output applies to the entire GLM procedure and is displayed only once.
  • Additional output can be obtained on the EMMEANS, PLOT, and SAVE subcommands.
  • Some optional output may greatly increase the processing time. Request only the output that you want to see.
  • If no PRINT command is specified, default output for a univariate analysis includes a factor information table and a Univariate Tests table (ANOVA) for all effects in the model.
  • If more than one PRINT subcommand is specified, only the last subcommand is in effect.

The following keywords are available for GLM univariate analyses. For information about PRINT specifications that are appropriate for other GLM models, see GLM: Multivariate and GLM: Repeated Measures.

DESCRIPTIVES. Basic information about each cell in the design. This process determines observed means, standard deviations, and counts for the dependent variable in all cells. The cells are constructed from the highest-order crossing of the between-subjects factors. For a multivariate model, statistics are given for each dependent variable. If the number of between-subjects factors plus the number of split variables exceeds 18, the Descriptive Statistics table is not printed.

HOMOGENEITY. Tests of homogeneity of variance. Levene tests for equality of variances for the dependent variable across all level combinations of the between-subjects factors. If there are no between-subjects factors, this keyword is not valid. For a multivariate model, tests are displayed for each dependent variable.

PARAMETER. Parameter estimates. Parameter estimates, standard errors, t tests, and confidence intervals.

ETASQ. Partial eta-squared (η2). This value is an overestimate of the actual effect size in an F test.

GEF. General estimable function table. This table shows the general form of the estimable functions.

LOF. Instruction to perform a lack-of-fit test (which requires at least one cell to have multiple observations). If the test is rejected, it implies that the current model cannot adequately account for the relationship between the response variable and the predictors. Either a variable is omitted or extra terms are needed in the model.

OPOWER. Observed power for each test. The observed power gives the probability that the F test would detect a population difference between groups that is equal to the difference that is implied by the sample difference.

TEST(LMATRIX). Set of contrast coefficients (L) matrices. The transpose of the L matrix (L') is displayed. This set always includes one matrix displaying the estimable function for each between-subjects effect that appears or is implied in the DESIGN subcommand. Also, any L matrices generated by the LMATRIX or CONTRAST subcommands are displayed. TEST(ESTIMABLE) can be used in place of TEST(LMATRIX).

MBP. Modified Breusch-Pagan test for heteroskedasticity. Tests the null hypothesis that the variance of the errors does not depend on the values of the independent variables. You can use the MBPDESIGN subcommand to specify the model on which the test is based. If the MBP keyword is specified without MBPDESIGN, then the model consists of a constant term, a term that is linear in the predicted values, a term that is quadratic in the predicted values, and an error term. For a multivariate model, tests are displayed for each dependent variable.

BP. Breusch-Pagan test for heteroskedasticity. Tests the null hypothesis that the variance of the errors does not depend on the values of the independent variables. You can use the BPDESIGN subcommand to specify the model on which the test is based. If the BP keyword is specified without BPDESIGN, then the model consists of a constant term, a term that is linear in the predicted values, a term that is quadratic in the predicted values, and an error term. For a multivariate model, tests are displayed for each dependent variable.

F. F test for heteroskedasticity. Tests the null hypothesis that the variance of the errors does not depend on the values of the independent variables. You can use the FDESIGN subcommand to specify the model on which the test is based. If the F keyword is specified without FDESIGN, then the model consists of a constant term, a term that is linear in the predicted values, a term that is quadratic in the predicted values, and an error term. For a multivariate model, tests are displayed for each dependent variable.

WHITE. White's test for heteroskedasticity. Tests the null hypothesis that the variance of the errors does not depend on the values of the independent variables. For a multivariate model, tests are displayed for each dependent variable.

Example

GLM DEP BY A B WITH COV
  /PRINT=DESCRIPTIVE, TEST(LMATRIX), PARAMETER
  /DESIGN.
  • Because the design in the DESIGN subcommand is not specified, the default design is used. In this case, the design includes the intercept term, the covariate COV, and the full factorial terms of A and B, which are A, B, and A*B.
  • For each combination of levels of A and B, the descriptive statistics of DEP are displayed.
  • The set of L matrices that generates the sums of squares for testing each effect in the design is displayed.
  • The parameter estimates, their standard errors, t tests, confidence intervals, and the observed power for each test are displayed.