This study addressed the effectiveness and efficiency issues in a self-instructional Iearning activity in counselor training. The study specifically examined three media (written, audio, and video) of model and of self-generated performance feedback presentation. There were no differences in trainee production of counselor tacting response leads because of media of model or media of performance feedback presentation. Results, however, did indicate that learning is significantly enhanced when a written performance feedback component is added to a written modeling component. The same did not hold true for audio and video presentations. The implications of these cost-and labor-efficient instructional packages are discussed.In several recent issues of Counselor Education and Supervision, specific articles have outlined approaches for developing and implementing self-evaluation or self-instructional modules for counselor trainees (Cormier & Cormier, 1976;Hector, Elson, & Yager, 1977; Fuhrmann, 1978). Although each author has reverified the importance of such programs, there is a dearth of empirical research investigating the impact of self-instructional programs and the most effective and efficient methods of presenting such programs.Cormier and Cormier (1976) suggest three components of a selfinstructional learning activity designed to assist trainees in the acquisition of discrete interviewing concepts and skills. The specific components are 1) modeled examples, 2) practice, and 3) immediate feedback on practice. These were the steps employed by Robinson (1978) who compared instructions, modeling, model plus self-generated performance feedback and model plus expert performance feedback groups. She found that the addition of feedback to the modeled examples had more impact than models presented alone. Her results also indicated that self-generated performance feedback (a self-instructional learning activity) was as effective as expert performance feedback (a learning activity requiring supervisor input). The equivalence of self and expert feedback, without the modeling phase, was also supported by a study conducted by Robinson, Kurpius, & Froehle (1978). These two empirical studies confirm the effectiveness of an instructional module using a self feedback condition. In the both of these studies, the self-generated performance feedback was facilitated via a correctness cue presented on videotape. The models were also presented on videotape in the Robinson (1978) study. While the majority of counselor training research has used videotaped training materials, few researchers have compared different media for presenting instructional packages. Stone (1975) presented low fidelity models (instructional manuals and audiotapes) and high fidelity models (videotapes and in vivo), but he collapsed across the types of high and low fidelity models, which did not allow for a true comparison of video, audio or written models. Robinson, Forehle, and Kurpius (1979) compared video and written models and found no differences. A study ...