The major purpose of this investigation was to develop a program for teaching a specific counselor verbal skill and to examine the effect of simulation methods within the program. The program consisted of model, practice, and test phases. High and low fidelity methods were used within each phase. Sixty‐four counselor trainees were randomly assigned to the 16 treatment groups. The low fidelity composite treatment was the least effective. Results are discussed in terms of training and future researc