PRINT Subcommand (HILOGLINEAR command)

PRINT controls the display produced for the subsequent designs.

  • If PRINT is omitted or included with no specifications, the default display is produced.
  • If any keywords are specified on PRINT, only output specifically requested is displayed.
  • HILOGLINEAR displays Pearson and likelihood-ratio chi-square goodness-of-fit tests for models. For saturated models, it also provides tests that the k-way effects and the k-way and higher-order effects are 0.
  • Both adjusted and unadjusted degrees of freedom are displayed for tables with sampling or structural zeros. K-way and higher-order tests use the unadjusted degrees of freedom.
  • The unadjusted degrees of freedom are not adjusted for zero cells, and they estimate the upper bound of the true degrees of freedom. These are the same degrees of freedom you would get if all cells were filled.
  • The adjusted degrees of freedom are calculated from the number of non-zero-fitted cells minus the number of parameters that would be estimated if all cells were filled (that is, unadjusted degrees of freedom minus the number of zero-fitted cells). This estimate of degrees of freedom may be too low if some parameters do not exist because of zeros.

DEFAULT. Default displays. This option includes FREQ and RESID output for nonsaturated models, and FREQ, RESID, ESTIM, and ASSOCIATION output for saturated models. For saturated models, the observed and expected frequencies are equal, and the residuals are zeros.

FREQ. Observed and expected cell frequencies.

RESID. Raw and standardized residuals.

ESTIM. Parameter estimates for a saturated model.

ASSOCIATION. Partial associations. You can request partial associations of effects only when you specify a saturated model. This option is computationally expensive for tables with many factors.

ALL. All available output.

NONE. Design information and goodness-of-fit statistics only. Use of this option overrides all other specifications on PRINT.