Anxious responses to pain may lead to avoidance of behavior expected to produce pain. McCracken et al. (1992) developed the Pain Anxiety Symptoms Scale (PASS) to assess anxiety related specifically to pain. Efforts to validate the scale, however, have been confined mostly to examining associations between the PASS and other self-report instruments. This study tested whether PASS scores were related to behavioral performance variables recorded by therapists during a physical capacity evaluation. Participants were 98 male patients with persistent pain referred to two industrial rehabilitation centers. PASS scores were correlated negatively with amount of weight lifted and carried, and results of hierarchical regressions showed that PASS scores accounted for additional variance in these variables when measures of trait anxiety, depression and pain severity were controlled. However, we did not replicate the findings of McCracken et al. (1992) that PASS scores accounted for variance in self-reported disability with trait anxiety, depression or pain severity controlled. Results extend the validity of the PASS and are consistent with models of fear of pain: patients with high PASS scores may avoid potentially painful physical exertion to reduce their fear.