BACKGROUND: Quality educators are a core component of successful residency training. A structured, consistent, validated evaluation of clinical educators is important to improve teaching aptitude, further faculty development, and improve patient care. STUDY OBJECTIVES: The authors sought to identify specific domains of instructional quality and to develop a composite instrument for assessing instructional quality. METHODS: The study setting is a 3-year residency program. Residents rated the quality of faculty member instruction using an 18-item survey twice over a 2-year period (2004-2005). Each survey item used a 9-point scale. Factor analysis employing a Varimax rotation identified domains of instructional performance. Cronbach's alpha was used to assess the internal consistency of the identified domains. RESULTS: There were 29 faculty members evaluated. Using 2004 data, five domains of instructional quality were identified that explained 92.5% of the variation in survey responses (chi(2) = 2.33, P = 0.11). These were: Competency and Professionalism (30% of variation), Commitment to Knowledge and Instruction (23%), Inclusion and Interaction (17%), Patient Focus (13%), and Openness to Ideas (9%). Competency and Professionalism included appropriate care, effective patient communication, use of new techniques, and ethical principles. Commitment to Knowledge and Instruction included research, mentoring, feedback, and availability. Inclusion and Interaction included procedural participation and bedside teaching. Patient Focus included compassion, effective care, and sensitivity to diverse populations. Openness to Ideas included enthusiasm and receptivity of new ideas. These five domains were consistent in the 2005 data (Cronbach's alpha 0.68-0.75). CONCLUSIONS: A five-domain instrument consistently accounted for variations in faculty teaching performance as rated by resident physicians. This instrument may be useful for standardized assessment of instructional quality.