The present study tested the efficacy of the On-Task in a Box program for increasing on-task behavior and academic accuracy of highly off-task students. Six students in 2nd and 3rd grades were identified by their classroom teacher as highly off-task. Following identification, the students participated in the On-Task in a Box intervention. Results of the study found immediate and large effects, which were maintained following discontinuation of the intervention. Collateral improvements in accuracy on math probes completed during independent seatwork were also observed. Teacher and participant responses to intervention acceptability questionnaires indicate the program was viewed positively. Implications for school-based adoption of the program are presented, and limitations and future research are discussed.