“…Further studies were identified from references cited in studies found through this search process. In order to obtain unpublished BMT evaluations, known BMT researchers were contacted, as were the American In total, 171 BMT studies were located, of which 46 were later excluded: 26 because they lacked sufficient information for effect size calculations (e.g., Goldstein & Sorcher, 1974); 6 because they included neither pretests nor a relevant comparison group (e.g., Johnson & Marakas, 2000); 6 because training effects were measured exclusively through trainees' own self-ratings (e.g., Nunns & Bluen, 1992); 4 because, although they were described by their authors as based on behavior modeling, one or more of the four key components of behavior modeling was not included (e.g., Royster, 1981); 3 because they were implemented with children (e.g., Renaud & Stolovitch, 1988); and 1 because a pure, BMT-only condition could not be isolated (Davis & Mount, 1984).…”