Parameters | Expression Parameters | Additional PROC MIXED Statements

Additional PROC MIXED Statements
Enter SAS PROC MIXED statements in this field to supplement those generated by preceding parameters.
You can specify any PROC MIXED statement except for the PROC, BY, or RUN statements. This specification provides you with maximum flexibility to fit various models .
Possible additional statements include WHERE and LSMEANS. A WHERE statement is useful for processing a subset of the data. Note that a Uniform(0,1) dummy response y variable is used, so statistical tests in the SAS Output are not meaningful. However, you can use a statement such as where response lt 0.5; to use a random subset of the input experimental design data set to reduce its sample size and hence power .
PROC MIXED statements are written in the following syntax:
Statement var 1 ... var n ;
where:
Statement is the PROC MIXED statement,
var 1 ...var n are the variables (or specific values) specified,
italicized items indicate a variable, or option to be specified,
the statement ends with a semicolon ( ; ), and
Note : Variables specified here can come from either the adsl.sas7bdat or ae.sas7bdat data sets.
Examples of commonly used PROC MIXED statements are listed in the following table.
CLASS var1...varn ;
names a single dependent variable and the fixed effects that determine the X matrix of the model
MODEL dependent = <fixed effects> ;
RANDOM random effects ;
LSMEANS fixed effects ;
specifies the R matrix in the model
REPEATED <repeated effect> ;
ESTIMATE ‘label’<fixed effect values> <|random effect values> ;
To Specify One or More Additional PROC MIXED Statements:
*
Type specific PROC MIXED statements in the Additional PROC MIXED Statements field.
For Additional Information
Refer to the SAS documentation for PROC MIXED for more information.