This paper presents a relatively context-free method of assessing teacher behavior. The context is removed through the use of extremely brief (10-second) clips of videotaped teacher behavior, separated into isolated nonverbal and verbal channels. This method makes it possible to trace subtle within-teacher differences in isolated visual and verbal channels, such as the face, the body, speech content, and tone of voice, as well as to compare teachers who work in disparate educational contexts. To examine a commonly held belief about the relatively high quality of Israeli preschool teachers, samples of preschool, remedial, and elementary school teachers were videotaped, and brief clips were rated on 10 scales by 15 judges. Following a principal component analysis, the 10 rating scales were reduced to three interpretable factor-based scores. The groups of teachers did not differ in their “active teaching behavior”—a composite score consisting of task-orientation, clarity, dominance, and activity/energy/ enthusiasm. The groups differed in “nondogmatism” (ratings of teacher as democratic, flexible, and warm) and “negative affect” (ratings of teacher as hostile, condescending, and tense/nervous/anxious) manifested in their behavior. For these composite variables, a clear linear trend was observed, with preschool teachers showing least dogmatism and negative affect, elementary school teachers showing most dogmatism and negative affect, and remedial teachers falling between these two groups. Dogmatic behavior was detected in both the face and the verbal channels, whereas negative affect was detected only in the teachers’ faces.