ABSTRACT: Psychosocial treatments for chronic pain produce favorable outcomes. However, we still do not know precisely by what mechanisms or techniques these outcomes are wrought. In secondary analyses of a 10-week group intervention study comparing the effects of literacy-adapted cognitive behavioral therapy (CBT) with literacy-adapted pain education (EDU) among patients with chronic pain, low-socioeconomic status, and low literacy, the Learning About My Pain trial, we examined whether pain catastrophizing was a mechanism specific to CBT. Participants (N = 168) completed mechanism and outcome measures weekly for the 10 weeks of group treatment. Analyses revealed that (1) pain catastrophizing was reduced similary across CBT and EDU; (2) lagged analyses indicated that previous week reductions in pain catastrophizing predicted next week reductions in pain intensity and pain interference; (3) cross-lagged analyses indicated that previous week reductions in pain intensity and interference predicted next week reductions in pain catastrophizing; and (4) the relationships between pain catastrophizing and pain intensity and interference were moderated by session progression such that these links were strong and significant in the first third of treatment, but weakened over time and became nonsignificant by the last third of treatment. Results suggest the existence of reciprocal influences whereby cognitive changes may produce outcome improvements and vice versa. At the same time, results from analyses of changes in slopes between pain catastrophizing and outcomes indicated that CBT and EDU were successful in decoupling pain catastrophizing and subsequent pain intensity and interference as treatment progressed. Results provide further insights into how psychosocial treatments for chronic pain may work.