UNMEASURED variables;
The UNMEASURED statement specifies a list of variables that are treated as unmeasured. You might want to specify a variable as unmeasured in the following situations:
The variable corresponds to a latent construct that cannot be measured.
The variable was not measured or will not be measured in the data set for the causal effect estimation.
The data that are collected for the variable are not considered reliable enough to include in the causal effect estimation.
You can specify only one UNMEASURED statement.
If you specify multiple models by using multiple MODEL statements, each variable that you specify in the UNMEASURED statement is treated as unmeasured in every model. In addition, any variable that you specify in the IDENTIFY statement or in a TESTID statement cannot also be specified in the UNMEASURED statement. That is, the treatment and outcome variables, the adjustment variables, and the instrumental variables are all assumed to be measured or observed.