Research has demonstrated the usefulness of the Structured Interview of Reported Symptoms (SIRS) for the accurate classification of those feigning mental illness. In this study we examined the eifects of coaching (i.e., providing a general description of bona fide psychiatric disorders with implicit guidelines on how to minimize detection) on subjects' ability to fake a serious mental disorder. Two important findings emerged: (a) Coached subjects were able to modify their responses on all SIRS scales designed to assess malingering, and (b) despite these changes, the SIRS maintained its discriminability (i.e., 96.7% of inpatients and 84.6% of simulators). The implications of these findings for other measures, such as the Minnesota Multiphasic Personality Inventory, are discussed.