A standard-setting experiment was conducted in which one group of participants was instructed to refrain from discussing their recommended test standards and another group was requested to report the reasons underlying their recommendations. The procedures used were designed to induce convergence of recommendations, but did not require consensus. A structured judgment procedure was used to obtain recommendations on appropriate standards of performance on the Reading and Social Studies subtests of the National Teacher Examinations. Recommended standards of performance were provided for applicants to North Carolina's teacher education programs, by public school teachers and college and university faculty. Each participant was asked to complete the subtest, then to estimate, for each subtest item, the probability that a minimally competent examinee would be able to answer the item correctly. Using the Delphi Technique, participants received feedback about the ratings and then were asked to reconsider their initial recommendations. One group was asked not to discuss the feedback; in the other, discussion was encouraged. The reduction in variance from the first judgment session to the second was substantially greater for the discussion group. (BW)