STUDY DESIGN: Randomized Clinical Trial. OBJECTIVE: To identify a subgroup of patients with low back pain who are likely to respond favorably to an intervention including mechanical traction. SUMMARY OF BACKGROUND DATA: Previous research has failed to find evidence supporting traction for patients with low back pain. Previous studies have used heterogeneous samples, although clinical experts tend to recommend traction for a more limited subgroup of patients with low back pain. METHODS: Sixty-four subjects (mean age 41.1 year, 56.3% female) with low back and leg pain and signs of nerve root compression were randomized to receive a 6-week extension-oriented intervention with or without mechanical traction during the first 2 weeks. Between-group comparisons were conducted for changes in pain, disability, and fear-avoidance beliefs. Baseline variables were explored for potential as subgrouping criteria defining a subgroup of subjects likely to benefit from traction. RESULTS: The group receiving traction showed greater improvements in disability (adjusted mean difference in Oswestry change 7.2 points) and fear-avoidance beliefs (adjusted mean difference in FABQPA change 2.6 points) after 2 weeks. There were no between-group differences after 6 weeks. Two baseline variables were associated with greater improvements with traction treatment; peripheralization with extension movements and a crossed straight leg raise. CONCLUSION: A subgroup of patients likely to benefit from mechanical traction may exist. The results of this study suggest this subgroup is characterized by the presence of leg symptoms, signs of nerve root compression, and either peripheralization with extension movements or a crossed straight leg raise. Further research is needed to validate this finding.