When using multiple sets of control samples, you must specify a subset of the columns entered in the
Variables Defining Analysis Groups field that define the analysis groups that should be used as control sets. The samples for which the control sets are used in reference to are decided by the distinct groups made by the
variables in the
Variables Defining Analysis Groups field that were not re-specified as
Variables Defining Control Sets. The values for these columns that indicate a sample as a control should be entered in the
Control Levels field.
For an example, with the
Drosophila Aging example, if you want to use males as the controls for females that are matched for
Line and
Age: specify
Line,
Sex, and
Age as the
Variables Defining Analysis Groups and specify
Sex in the
Variable Defining Control Sets field (with
"MAL" specified in the
Control Levels field below). This designates the values "ORE", "MAL", and "WK1" as controls for samples with Line="ORE" and Age="WK1", and designates the values "SAM", "MAL", and "WK1" as controls for samples with Line="SAM" and Age="WK1". Likewise, the controls for Line="ORE" and Age="WK6" would be samples with values "ORE", "MAL", and "WK6", and the controls for Line="SAM" and Age="WK6" would be samples with values "SAM", "MAL", and "WK6" for Line, Sex, and Age, respectively.
For more complex designs, such as
nested designs, or designs where the value of the control changes across analysis groups, it might be necessary to construct a new column to be included in the
Variables Defining Analysis Groups and the
Variables Defining Control Sets in the
Experimental Design SAS Data Set to indicate the analysis groups that should be used as controls.
The Available Variables field is populated with variables from the specified data set.