DOCUMENTATION
Sign In

Get access to My SAS, trials, communities and more.

Econometrics Procedures
 

DEEPCAUSAL Procedure

PSMODEL Statement

  • PSMODEL treatment-variable <=covariates> </options>;

The PSMODEL statement specifies options that are related to the propensity score model and the DNN to fit the propensity score model. The treatment-variable specifies the treatment variable of the propensity score model. It corresponds to one variable in the input data table; that variable must satisfy the following requirements:

  • The treatment variable in the input data table must be binary; that is, its value must be either 0 or 1.

  • The values of the treatment variable across all observations in the input data table must not be identical.

The covariates specify the covariates of the propensity score model. All covariate names are separated by spaces. If you omit the covariates, no DNN is trained for the propensity score model. For more information about the propensity score model, see the section Propensity Score Model.

You can specify the following option after a forward slash (/):

DNN=(dnn-options)

specifies the options of the DNN for the propensity score model. If you omit this option and specify the DNN statement, the DNN for the propensity score model has the options that you specify in the DNN statement. If you omit both this option and the DNN statement, the DNN for the propensity score model uses the default option values. If you specify both this option and the DNN statement, the DNN statement is ignored for the propensity score model. For the dnn-options, you can specify one or more options that are listed in the section DNN Statement. If no covariates are specified in the PSMODEL statement, the DNN= option is ignored.

Last updated: November 24, 2025