"The amicus brief asserted that therapists cannot predict future violence, citing studies tending to show that therapists overpredict violence and are more often wrong than correct in such predictions. The studies cited were studies of people who had been committed because of alleged violent tendencies and later released. There are, however, no comparable studies demonstrating the same level of unreliability regarding predictions made about people who are in the community or have ready access to it...

[The] respondents proved to be rather more confident about their ability to predict future violence than the arguments of the Tarasoff critics suggested. When asked to indicate the firmest prediction they would be willing to make about the possibility that an outpatient of theirs might physically harm another, only 5% of [the] respondents felt that there was "no way to predict" such behavior, and over three-quarters felt that they could make a prediction ranging from "probable" to "certain." The finding that therapists are quite confident in predicting future violence by outpatients does not entirely meet the criticism that such prediction is idiosyncratic or that there are, in fact, no agreed upon criteria for predicting violence." [Daniel J. Givelbe, William J. Bowers and Carolyn L. Blitch, "Tarasoff, Myth And Reality: an Empirical Study of Private Law in Action," 1984 Wis. L. Rev. 443, 462-463].