Abstract The use of a programmed generalization strategy in general education classrooms was evaluated as to the results and the practicality of the process. Results indicate that the intervention, i.e., social skills training and self-monitoring with teacher matching, led to generalization of improved student behavior (meeting their teachers expectations) in up to six different class settings. This behavior was also maintained over time. This evaluation adds to the existing literature suggesting this programmed generalization strategy is a viable means to increase generalization of learned behaviors in numerous class settings with different teachers and classmates. Implications for research and practice are discussed. Introduction Several risk factors have been associated with students who are at-risk for violence, delinquency, school failure, and drug, alcohol, and tobacco abuse. They include poorly developed social interaction and resistance skills, poor academic skills, and dysfunctional families (Callahan, Young, West, & Mason, 1994: Schinke, Botvin, & Orlandi, 1991). The focus of successful prevention programs is to reduce or eliminate these risk factors by developing social competence, teaching self-management and problem-solving skills, remediating academic deficits, and strengthening families. The success of these programs depends in part on the extent to which students can use these social and self-management skills in settings other than those in which they were learned. Research has consistently shown that a variety of social skills can be taught to students at-risk and/or to students with disabilities using a structured learning approach encompassing: identifying the skills, modeling, role-playing, and performance feedback (Kiburz, Miller, & Morrow, 1984; Mathur & Rutherford, 1991; Schloss, Schloss, Wood, & Kiehl, 1986). The problem is that these skills have not always generalized to non-training situations (Fox & McEvoy, 1993; Gresham, 1981). Stokes and Baer (1977) define generalization as the occurrence of a response targeted in a training condition also occurring in different, non-training conditions (i.e., across subjects, settings, people, behaviors, and/or time) without the scheduling of the same events in those conditions as were scheduled in the training conditions." In past years, researchers in special education have acknowledged that, in order to reliably produce generalization effects, some type of programming strategy is necessary (Fox & McEvoy, 1993; Marholin & Siegel, 1978; Mathur & Rutherford, 1994; Mathur & Rutherford, 1991; Schloss et. al., 1986; Stokes & Baer, 1977; Stokes & Osnes, 1989). Although previous studies have produced moderate generalization effects, the results have been limited in terms of the number of settings where generalization occurred. The experiments of Lonneker, Brady, McPherson, & Hawkins (1994), Mathur, & Rutherford (1994), Sasso, Melloy, & Kavale (1990), and Smith, Nelson, Young, & West (1992) showed generalization of treatment gains to one setting. Clees' (1994) experiment had generalization occurring in two additional settings. Rhode, Morgan, & Young (1983) had generalization to different classes, however, only one generalization class per student (the students were in the same class with the same teacher all day other than the special education class where the initial training occurred). These previous experimental findings provide support for student self-management to promote generalized use of pro-social, cooperative behaviors in other settings. However, since secondary (middle schools, junior high schools, and high schools) students typically have several different classes per day, frequently all with different teachers, there is a need for strategies that promote the generalization of socially appropriate behaviors in all classes. …