The degrees of variance among three groups of evaluators relative to their assessments of the teaching competencies of preservice teacher education students were studied. Subjects included groups of 23 and 32 undergraduates who were certified to teach by the teacher preparation program at Washburn University in Topeka (Kansas) in 1987 and in 1988, respectively. At the end of student teaching, the student, the cooperating teacher, and the university supervisor completed a rating scale containing the same criteria related to the teaching-learning process. A matrix of probabilities was obtained for each competency examined through the multiple comparisons provided in Fisher's levels of significant differences (LSDs). The mean values derived were not, in most cases, significantly different for the three groups of evaluators. The greatest differences, which were not large, were found between students' self-assessments and the evaluations of university supervisors for long-range and daily lesson planning and competence in assessing and interpreting student progress. Appendix A gives the means and standard deviations of the study in 25 tables. Appendix B contains 25 additional tables showing the application of Fisher's LSD matrix. Appendix C is a summary of levels of significance, and Appendix D contains the Follow-Up Performance Rating Scale for Student Teachers. (SLD)