We discuss two major reviews of the accuracy of dichotomous (yes/no) v
iolence predictions (Monahan, 1981; Otto, 1992), which reported rates
of false positive errors that were calculated using conceptually and m
athematically dissimilar methods. We outline potential problems that r
esult from analyzing data in the form of 2 x 2 contingency tables and
offer some recommendations for future research on the prediction of vi
olence.