Efforts to determine work and social skills related to job success have followed one of three approaches (Langford, 1982): (a) establishing correlations between personality and demographic variables and work success (Bronner, 1933;Greenspan, 1979;Madison, 1964), (b) obtaining lists from employers and supervisors of desirable employee behavior (Langford & Salzburg, 1982), and (c) placing people with handicaps in competitive jobs and offering individualized intervention and training (Levy, 1978;Mithaug, 1978;Rusch, 1979;Wehman, Hill, & Koehler, 1980). Some of the problems with each of these methods include (a) globally defined variables, (b) lack of correlations between variables and job success, (c) inconsistency in variables across employers, settings, and so forth, and (d) intervention of such specificity that educators had difficulty generalizing findings across individuals and settings. In all approaches, little guidance was offered for decisions on which skills to train, which behaviors to increase or diminish, and so forth. For example, Rusch and Mithaug's (1980) and Wehman's (1981) vocational training programs, while otherwise comprehensive and systematic, include only limited assessment of competitive job requirements, and they do not provide much direction for prioritizing training objectives. This difficulty leaves teachers and supervisors of people without handicaps without definitive guidelines for placement decisions or for training work-related skills.The larger question is, in fact, what is desirable work behavior. An initial answer to this question was supplied by Levy (1978), who attempted to gain a better understanding of behaviors of retarded and nonretarded workers on the job. His study provided a basic methodology for comparing trainees with competitive employment requirements. Using a variety of observational techniques, he identified several different on/off-task behaviors and used them to evaluate the differences among workers in sheltered and competitive settings. The results indicated that workers both with handicaps and without engaged in the same types of on-and off-task behaviors but that the frequency and duration differed. These conclusions shed light on desirable work behaviors, illustrated critical differences between employees with handicaps and those without, and suggested objectives for vocational interventions.Because Levy's study involved a small sample, further research with more subjects in different sheltered and competitive settings seemed warranted, as did expansion of the behavioral categories. We (Levy & lascoe, 1986) expanded and replicated Levy's (1978) study. We found that observational techniques could be used to measure different vocational tasks, in different work sites, and with employees of differing levels of ability and expertise. This study generated extensive descriptive data. These observations and findings are presented below.