The PSMATCH Procedure

PSMODEL Statement

PSMODEL treatvar <(trt-option)> = <effects> </ WEIGHT=weight> ;

The PSMODEL statement specifies the logistic regression model for computing propensity scores. If you omit the EWEIGHT statement, you must specify either the PSMODEL or PSDATA statement, but not both.

The treatment indicator variable treatvar must be a binary classification variable that is listed in the CLASS statement, and the effects are the explanatory effects, which can include variables, main effects, interactions, and nested effects for the logistic regression model.

You can use the following trt-option to specify the treated level for the binary treatment variable:

TREATED='level' | keyword

models the probability of the specified treated level. You can specify the value of the treated level in quotation marks, or you can specify one of the following keywords:

FIRST: designates the first ordered level as the treated group.
LAST: designates the last ordered level as the treated group.

By default, TREATED=FIRST.

You can specify the following option to fit a weighted logistic regression:

WEIGHT=weight: specifies a variable that contains the weight of each observation that is used in fitting the logistic regression model to derive the propensity scores. These weights should not be confused with weights that PROC PSMATCH derives from the propensity scores.

Last updated: December 09, 2022