Coaching may increase students' scores on an achievement test without increasing, or increasing only temporarily, skills in the area being tested. This study deals with the immediate and longer term effects of coaching on the CEEB English Composition Test. Students at each of six cooperating high schools were randomly assigned to coached and non-coached conditions. Following the coaching sessions, all students took the English Composition Test. The immediate effects varied from essentially no mean difference in test scores between coached and non-coached students at one school to a mean difference of 73 points favoring the coached students at another school. In general the coaching increment was not confined to the specific item types coached for. The coached groups achieved their superiority by getting fewer items wrong. When coached and non-coached students were retested ten months later, it was found that the coaching increment had disappeared. Estimates of growth and practice effects from test-retest data of the non-coached students varied from a mean of 15 points at one school to a mean of 94 points at another. It was concluded that the candidate who receives coaching may indeed have an advantage over the candidate who does not. But the variability from school to school both of coaching effects and of score gains due to "growth" indicate that the specifics of the teaching situation are quite important.