A CONDITION statement specifies the conditioning variables. The presence of a CONDITION statement requests that a conditional simulation be performed.

The lack of a CONDITIONAL statement simply means that an unconditional simulation for the VAR variables is to be performed.

If a CONDITION statement is given, the variables listed must be numeric variables in the DATA= data set. This requires a conditioning value for each of the CONDITION variables. This value is supplied by adding a _TYPE_=’COND’ observation for each CONDITION variable. Such observations are added to the DATA= data set by a DATA step.

Note that a data set created by the CORR procedure is automatically given the TYPE=COV, UCOV, CORR, or UCORR attribute, so you do not have to specify the TYPE= option in the DATA= option in the PROC SIMNORMAL statement. However, when adding the conditioning values by using a DATA step with a SET statement, you must use the TYPE=COV, UCOV, CORR, or UCORR attribute in the new data set. See the section Getting Started: SIMNORMAL Procedure for an example in which the TYPE is set.