Effects of self-monitoring on-task behavior, academic productivity, and academic accuracy were assessed with 6 elementary-school students with learning disabilities in their general education classroom using a mathematics task. Following baseline, the three self-monitoring conditions were introduced using a multiple schedule design during independent practice sessions. Although all three interventions yielded improvements in either arithmetic productivity, accuracy, or on-task behavior, self-monitoring academic productivity or accuracy was generally superior. Differential results were obtained across age groups: fourth graders' mathematics performance improved most when self-monitoring productivity, whereas sixth graders' performance improved most when selfmonitoring accuracy. DESCRIPTORS: learning disabled, self-monitoring Self-monitoring procedures have been applied to diverse populations and behaviors (Mace & Kratochwill, 1988). These interventions offer an acceptable alternative to more intrusive procedures (Lloyd & Landrum, 1990;Roberts & Nelson, 1981) and are useful with atypical learners to decrease direct teacher intervention and to improve academic performance. Numerous studies have shown that self-monitoring either attention or academic responding increases on-task behavior (e.g., Blick & Test, 1987;DiGangi, Maag, & Rutherford, 1991;Hallahan, Lloyd, Kosiewicz, Kauffman, & Graves, 1979;Harris, 1986). However, it remains undear which of the variables in self-monitoring best promotes academic gains (Hallahan & Sapona, 1983;Harris, 1986;Lloyd & Landrum, 1990;Snider, 1987).Advocates of self-monitoring attention stress its simple and practical nature (Lentz, 1988 that any differential effects between the two variables are too small to influence academic achievement (Hallahan & Sapona, 1983;Lloyd, Bateman, Landrum, & Hallahan, 1989;Lloyd & Landrum, 1990). Proponents of self-monitoring academic outcome, on the other hand, question the link between on-task behavior and achievement and cite research indicating that targeting productivity or accuracy is more likely to affect academic responding (e.g., Hoge & Andrews, 1987;Klein, 1979;Petersen & Swing, 1982;Ruggles & LeBlanc, 1985;Treiber & Lahey, 1983). They reason that if increases in on-task behavior are independent of achievement, self-monitoring interventions should focus on active academic responding (Graden, Thurlow, & Ysseldyke, 1983).Only five studies have examined the differential efficacy of self-monitoring attention versus academic outcome (Harris, 1986;Lloyd et al., 1989; Reid & Harris, in press;Roberts & Nelson, 1981;Rooney, Polloway, & Hallahan, 1985). All studies reported increased on-task behavior for both selfmonitoring approaches; neither form of self-monitoring emerged as dearly superior. Lloyd et al. (1989), Rooney et al. (1985), and Roberts and Nelson (1981) Leary, 1979).The present study compared the effects of selfmonitoring attention, academic productivity, and academic accuracy on mathematics performance. Content difficulty was control...