Little consensus exists regarding the indications for and effectiveness of many back pain treatments. This clinical uncertainty arises because most back pain research has been flawed by poor methodology. The authors discuss strategies for improving the quality of back pain research on treatment efficacy. Design features, including randomized treatment allocation, independent outcome assessors, comprehensive outcome measures, appropriate statistical analyses, and close patient follow-up can increase study validity. Complete descriptions of enrollment criteria, patient characteristics, and clinical interventions can increase the generalizability of results. Although large scale trials often involve university centers, community-based researchers can collaborate on randomized trials or conduct valuable cohort studies.