Researchers should investigate statistical models that can help counselors decide how to treat individual clients. This research investigated the questions, "Given an effect size (ES) from a counseling outcome study, what is the probability that a client would have a negative response to the treatment, and what is the probability that the client would receive a minimally acceptable benefit?" Statistical models of 144 experimental situations were evaluated. ES, sample size, the variance of clients' treatment reactions and the distribution of these reactions were varied. With respect to the probabilities of a negative response to treatment, it was found that if the counselor is willing to assume that all clients respond identically to a treatment, then the counselor need only attend to the sample sizes and the effect size. If the effect size is a least 0.50 or the sample size is at least 15 in each group, the probability is less than 0.10 that the true effect size is negative. This percentage increases as the effect size decreases and the standard deviation of the individual responses to treatment increases, regardless of the sample size. Similar patterns hold for determining the probability of effect sizes exceeding a minimally acceptable benefit. Two tables provide statistical summaries. Contains eight references. (RJM)