Challenging behaviors displayed in both general education classrooms and self-contained classrooms by students with emotional behavioral disorders (EBD) can be a difficult challenge for novice and experienced teachers (Kennedy & Jolivette, 2008; Turnbull, Turnbull, & Wehmeyer, 2010). Furthermore, students with EBD demonstrate broad deficits in academic areas such as reading and math (Lane, Wehby, Little, & Cooley, 2005). An instructional strategy that has been found to maximize the effectiveness of classroom instruction by increasing student participation and academic outcomes is Numbered Heads Together (NHT) (Maheady, Michielli-Pendl, Harper, & Mallette, 2006; Haydon, Maheady, & Hunter, 2010). NHT requires little preparation time while ensuring active student participation and positive academic and behavioral outcomes. To date, there have been four NHT studies conducted (Haydon et al., 2010; Maheady et al., 2006; Maheady, Mallette, Harper, & Sacca, 1991; Maheady, Michielli-Pendl, Mallette, & Harper, 2002). In the Haydon et al. (2010) study, an individual reward system was used in the incentive condition. The authors investigated the use of providing the Numbered Heads Together + Incentives (NHT+I) condition as an alternative to NHT without rewards. Results of the study indicated improvement in social and academic behavior in favor of the NHT+I strategy. The present study systematically replicates and extends the Haydon et al. (2010) study by (a) working with students identified with EBD in a self-contained classroom, (b) investigating the use of a preference assessment in the NHT+I condition, and (c) working in a new content domain (math). The study involved four student participants and took place over a 2-month period during math instruction in a self-contained classroom for students with EBD, which was located within an urban middle school (Grades 6-8) in a Midwestern city. An alternating treatment design (Alberto & Troutman, 2009) was used to compare the effectiveness of the two NHT conditions and the baseline condition. The purpose of this study was to further investigate the NHT intervention with and without incentives on the percentage of on-task behavior and quiz scores for 4 students with EBD. In terms of on-task behavior, there was an increase of an overall mean score of 17.4% during the NHT+I condition over the NHT condition. In terms of overall quiz scores, all 4 students scored a higher mean percentage on daily quizzes during the NHT+I condition in comparison with the NHT condition. The findings give some evidence that the use of preference assessments may supplement effective instructional practices in self-contained classrooms for students with EBD (DeLeon & Iwata, 1996; Fisher et al., 1992). Finally, the high rate of student on-task behavior is noteworthy, as research indicates that students with EBD exhibit high percentages of off-task behavior in classrooms (Wehby, Symons, & Shores, 1995).