Students' evaluations of a lesson were made on the same videotaped lesson after students were informed that the lesson was difficult or that it was easy, and after they were given a hard or easy content test. A two-way factorial design was employed with 92 undergraduate students, who were randomly assigned to conditions. After all students listened to a 10-minute lesson of an abstract relational concept, half were briefly told that the concept was used for training professional mathematicians, and half that it was used to teach math in the primary grades. In a completely crossed design, half took a difficult objective test over the content, and half took an easier version of the same test. Immediately afterwards, all filled out a 10-item teaching evaluation on the lesson. The findings suggest that students tend to give lower evaluations to teachers who give easy tests. (Author/SW)