Schema-based instruction (SBI) was used to teach nine middle school students with high-incidence disabilities (HID) to solve proportional reasoning word problems. A multiple-baseline-across-groups design was used to determine student understanding of the SBI process. The design consisted of three phases including baseline, SBI process, and maintenance. Training of the SBI process with worked problems occurred between baseline and SBI process. Students' performance was measured using a five-category rubric, under two conditions: solving problems with and without calculator use. Results indicated that all groups improved problem-solving performance and maintained improvement 4 weeks after instruction. Implications for revising the rubric's mastery criteria and future research are described. [For a related dissertation, see ED559601.]