A study is described in which students anxious about speeches were treated by variations of stress inoculation training. The purpose of the experiemnt was to evaluate the contribution of an educational phase to the training. Three treatment groups received either the educational phase only, the skills phase only, or both. These were compared with a no-treatment control group. The education-only group was the only group to improve significantly on self-reported anxiety measured before giving a public speech. The education-only and combination group imporved on self-reported self-efficacy as a speaker. All groups improved on behavioral rating of anxiety. Discussion focuses on the apparent potency of using an educational model in the treatment of anxiety. Future research is suggested for delineating the conditions under which education is a powerful ingredient.