A survey was conducted of 55 regular classroom teachers who referred students to special education in three Connecticut school systems to determine their predictions of the results of psycho-educational assessment and their perceptions of the instructional relevance of standardized psychological and educational test results. The findings of this study indicate that teachers could accurately predict the standardized test performance of students referred for special education using the same standard score and percentile rank formats employed by school psychologists and other diagnosticians. It is, therefore, argued that psychological and educational reports that focus primarily or exclusively on test scores may do little more than confirm teachers' impressions of their students. This investigation also found that teachers are ambivalent about the instructional value of psycho-educational assessment. Overall, these results are used to argue for greater teacher input into special education identification and placement procedures. These would include more in-depth involvement of specialists in the pre-referral stage of the eligibility process and a stronger focus on assessment domains that are more likely to yield new and useful information for teachers, such as affective development, behavioral adjustment, motivation, information processing, and data-based intervention strategies. Data tables are attached. (Contains 47 references.) (Author/CR)