PRINT Subcommand (LOGISTIC REGRESSION command)

PRINT controls the display of optional output. If PRINT is omitted, DEFAULT output (defined below) is displayed.

  • The minimum specification is PRINT followed by a single keyword.
  • If PRINT is used, only the requested output is displayed.

DEFAULT. Goodness-of-fit tests for the model, classification tables, and statistics for the variables in and not in the equation at each step. Tables and statistics are displayed for each split file and METHOD subcommand.

SUMMARY. Summary information. This output is the same output as DEFAULT, except that the output for each step is not displayed.

CORR. Correlation matrix of parameter estimates for the variables in the model.

ITER(value). Iterations at which parameter estimates are to be displayed. The value in parentheses controls the spacing of iteration reports. If the value is n, the parameter estimates are displayed for every nth iteration, starting at 0. If a value is not supplied, intermediate estimates are displayed at each iteration.

GOODFIT. Hosmer-Lemeshow goodness-of-fit statistic 1.

CI(level). Confidence interval for exp(B). The value in parentheses must be an integer between 1 and 99.

ALL. All available output.

Example

LOGISTIC REGRESSION VARIABLES=PASS WITH GPA,GRE,MAT
 /METHOD FSTEP
 /PRINT CORR SUMMARY ITER(2).
  • A forward stepwise logistic regression analysis of PASS on GPA, GRE, and MAT is specified.
  • The PRINT subcommand requests the display of the correlation matrix of parameter estimates for the variables in the model (CORR), classification tables and statistics for the variables in and not in the equation for the final model (SUMMARY), and parameter estimates at every second iteration (ITER(2)).
1 Hosmer, D. W., and S. Lemeshow. 2000. Applied Logistic Regression, 2nd ed. New York: John Wiley and Sons.