RESEARCH REPORT BEGUILED BY BAYES - A REANALYSIS OF MOSSMAN AND HARTSESTIMATES OF MALINGERING

Bayes' theorem has been successfully applied to clinical decision maki ng in circumstances where the prevalence rate is both stable and well established. An attempt by Mossman and Hart (1996) to apply Bayes' the orem to malingering is critically examined. When marked fluctuations i n the base rates of malingering (both within and across settings) are considered, this model appears to have little clinical utility. For th e sake of illustration, Mossman and Hart data were reanalyzed, taking into account the pronounced variation in base rates. Within a 90% conf idence level, the estimates of malingering for many scores extended th e full range from 0% to 100%. (C) 1998 John Wiley & Sons, Ltd.