This study investigated the effects of observer expectations about a speaker's nonfluency level on nonfluency counts made during a taped speech and on post‐performance evaluations of nonfluency, anxiety, central idea, organization, language, delivery, and general effectiveness. The influence of task ambiguity and medium of presentation on expectancy effects was also explored. Results indicate: (a) in a high‐ambiguity condition, observers who expected a fluent speaker counted fewer nonfluencies in his speech than observers who expected a nonfluent speaker; (b) fluent‐expectation observers rated the speaker more positively on the seven evaluative measures; (c) low task ambiguity eliminated expectancy effects on nonfluency counts and ratings of organization but not on the other six evaluative measures; and (d) auditory and auditory‐visual presentations of the speech did not produce significant diff