OBJECTIVE: To compare the perceptions of students and clinical instructors regarding helpful clinical instructor characteristics. DESIGN AND SETTING: We developed a questionnaire containing helpful clinical instructor characteristics for facilitating student learning from a review of the medical and allied health clinical education literature. Respondents rated clinical instructor characteristics from 1 (among the least helpful) to 10 (among the most helpful). Respondents also identified the overall 10 most helpful and 10 least helpful characteristics. SUBJECTS: A total of 206 undergraduate students and 46 clinical instructors in the National Athletic Trainers' Association District 4 athletic training education programs accredited by the Commission on Accreditation of Allied Health Education Programs responded to the survey. MEASUREMENTS: We computed individual-item and subgroup mean scores for students, clinical instructors, and combined students and instructors. Pearson product moment correlations were computed to evaluate the level of agreement between students and instructors. Correlations were also computed to evaluate the level of agreement between the open-ended responses and the Likert-scale responses. RESULTS: Agreement was high between the students' and the clinical instructors' ratings of individual items. Agreement was also high between individual-item means and the directed, open-ended 10 most helpful and 10 least helpful clinical instructor characteristics. Modeling professional behavior was considered the most helpful subgroup of clinical instructor characteristics. Integration of knowledge and research into clinical education was considered the least helpful subgroup of clinical instructor characteristics. CONCLUSIONS: Clinical instructors should model professional behavior to best facilitate student learning. Integration of research into clinical education may need more emphasis.