The PLM Procedure

PROC PLM Contrasted with Other SAS Procedures

In contrast to other analytic procedures in SAS/STAT software, the PLM procedure does not use an input data set. Instead, it retrieves information from an item store.

Some of the statements in the PLM procedure are also available as postprocessing statements in other procedures. Table 69.1 lists SAS/STAT procedures that support the same postprocessing statements as PROC PLM does.

Table 69.1: SAS/STAT Procedures with Postprocessing Statements Similar to PROC PLM

	EFFECTPLOT	ESTIMATE	LSMEANS	LSMESTIMATE	SLICE	TEST
GENMOD	$\surd$	$\surd ^*$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$
GLIMMIX		$\surd ^*$	$\surd ^*$	$\surd ^*$	$\surd$
GLM		$\surd ^*$	$\surd ^*$			$\surd ^*$
LIFEREG	$\surd$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd \phantom{^*}$
LOGISTIC	$\surd$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd ^*$
MIXED		$\surd ^*$	$\surd ^*$	$\surd \phantom{^*}$	$\surd$
ORTHOREG	$\surd$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd \phantom{^*}$
PHREG		$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd ^*$
PROBIT	$\surd$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd \phantom{^*}$
SURVEYLOGISTIC		$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd ^*$
SURVEYPHREG		$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd \phantom{^*}$
SURVEYREG		$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd \phantom{^*}$	$\surd$	$\surd \phantom{^*}$

Table entries marked with $\surd$ indicate procedures that support statements with the same functionality as in PROC PLM. Those entries marked with $\surd ^*$ indicate procedures that support statements with same names but different syntaxes from PROC PLM. You can find the most comprehensive set of features for these statements in the PLM procedure. For example, the LSMEANS statement is available in all of the listed procedures. For example, the ESTIMATE statement available in the GENMOD, GLIMMIX, GLM and MIXED procedures does not support all options that PROC PLM supports, such as multiple rows and multiplicity adjustments.

The WHERE statement in other procedures enables you to conditionally select a subset of the observations from the input data set so that the procedure processes only the observations that meet the specified conditions. Since the PLM procedure does not use an input data set, the WHERE statement in the PLM procedure has different functionality. If the item store contains information about BY groups—that is, a BY statement was in effect when the item store was created—you can use the WHERE statement to select specific BY groups for the analysis. You can also use the FILTER statement in the PLM procedure to filter results from the ODS output and output data sets.