The effectiveness of a multi-component computer based training package that consisted of competency based instructions, video modeling, and two forms of feedback was evaluated in terms of treatment integrity of two procedures across four staff. Treatment integrity in completing critical steps of discrete-trial and backward chaining procedures were measured using a multiple baseline design across participants, counterbalanced for procedures. All four participants reached 100% treatment integrity on at least one skill in a role play setting. Maintenance probes showed skills maintained at the same levels 6 weeks following training. (Contains 2 figures and 3 tables.)