Sensitivity analysis is a broad method that can be applied to virtually any statistical method and therefore any SAS procedure or program. For example, you might want to find out how sensitive the fit of your regression model is to a change in the nature of one or more of the predictors. Or you might want to see how much the results of a specialized program you wrote in SAS/IML change as you change inputs.

Because of this broad applicability, the capability of SAS to do sensitivity analysis is not confined to a single procedure. Rather, the SAS language provides many components that you can use to conduct a sensitivity analysis. For example, using the random number generators and other capabilities in the DATA step and/or the experimental design construction capabilities in SAS/QC software, you can construct data sets covering a range of scenarios or assumptions about the underlying phenomenon under study. You can run the desired statistical procedure or program on each of these data sets in turn and then output the statistics whose sensitivity is of interest. Note that you can use the ODS OUTPUT statement to output any table produced by any procedure to a data set for further processing. Finally, you can use any of our statistical procedures to summarize the behavior of the relevant statistics.

Operating System and Release Information