Displayed Output
The displayed exact conditional analysis output of the LOGISTIC
procedure includes the following:
- the "Conditional Exact Tests" table provides two tests
for the null hypothesis that the parameters for the specified
effects are zero: the exact probability test and the exact
conditional scores test. For each test, the test statistic, an
exact p-value (the probability of obtaining a more extreme
statistic than the observed, assuming the null hypothesis), and a
mid p-value (adjusts for the discreteness of the distribution)
are displayed.
Individual hypothesis tests for the parameter of each continuous
effect, and joint tests for the parameters of classification
variables, are generated by default. A joint test for all effects
may be requested with the JOINT or JOINTONLY options.
- if you specify the ESTIMATE, ESTIMATE=PARM, or ESTIMATE=BOTH
options, the "Exact Parameter Estimates" table displays
individual parameter estimates for each parameter conditional on
the values of all the other parameters in the model. These are
either the exact conditional maximum likelihood estimates (MLE)
or, in cases where the conditional MLE does not exist, the median unbiased
estimates.
Also displayed are one-sided or two-sided confidence limits for
the estimate, and a one-sided or two-sided p-value for testing
that the parameter estimate is zero. The one-sided p-value is
the smaller of the left and right tail probabilities for the
observed sufficient statistic for the parameter under the null
hypothesis that the parameter is zero, while the two-sided
p-value is twice the one-sided p-value.
- if you specify the ESTIMATE=ODDS or ESTIMATE=BOTH options, the
"Exact Odds Ratios" table displays odds ratios for
individual parameters, confidence limits, and a p-value for
testing that the odds ratio is 1.
- if you request an OUTDIST= data set, the "Sufficient
Statistics" table is displayed before printing any of the exact
analysis results. The table lists the parameters and their
observed sufficient statistics.
Copyright © 2000 by SAS Institute Inc., Cary, NC, USA. All rights reserved.
