PROC VARCOMP <options> ;

This statement invokes the VARCOMP procedure. You can specify the following options in the PROC VARCOMP statement.


specifies the input SAS data set to use. If this option is omitted, the most recently created SAS data set is used.


specifies the convergence value of the objective function for METHOD=ML or


specifies the maximum number of iterations for METHOD=ML or METHOD=REML. By default, MAXITER=50.

METHOD=TYPE1 | MIVQUE0 | ML | REML | GRR <(options)>

specifies which of the five methods (TYPE1, MIVQUE0, ML, REML, or GRR) you want to use. By default, METHOD=MIVQUE0. METHOD=GRR provides a specialized analysis only for certain designs, whereas the other four methods apply to any random-effects model and design. You can specify the following options in parentheses after METHOD=GRR.


specifies the specification limits for the first random factor, which is regarded as the product being tested in the gauge R&R study. The lower limit (LSL) must be smaller than the upper limit (USL). The value k is optional. The default value is 6, which corresponds to the number of standard deviations between the "natural" tolerance limits containing the middle 99.73% of a normal process. SPECLIMITS=(LSL,USL,k) requests the estimates of the parameters PTR(LSL,USL,k) and Cp(LSL,USL,k) to be displayed.


specifies that certain additional ratios of variance components should also be computed and displayed, such as proportion of total variance due to the process. These ratios are listed in Table 97.4.

For more information see the section Computational Methods.


specifies an unsigned integer used to start the pseudo-random number generator. If you do not specify a seed or if you specify zero, the seed is generated from reading the time of day from the computer clock. You can use a SAS date as a seed. The random number generation is used in the computaion of generalized confidence limits; see the section Confidence Limits.