The current study used data from a clinical trial to identify variables that are associated with and/or mediate the beneficial effects of four psychological chronic pain treatments: one teaching patients self-hypnosis to reduce pain intensity (HYP), one teaching self-hypnosis to change thoughts about pain (hypnotic cognitive therapy, or HYP-CT), one teaching cognitive restructuring skills to change thoughts about pain (cognitive therapy, or CT), and one providing education about pain (ED; included as an active control condition). Of 17 possible mechanism variables examined, and with alpha not corrected for multiple comparisons, significant between-group differences were observed for three. Two of these (changes in beliefs about control over pain and number of days of skill practice) were supported as mediators of the beneficial effects of HYP, CT, or HYP-CT, relative to ED. Six mechanism variables evidenced significant pre- to post-treatment changes in the sample as a whole, without showing significant between-group differences. Pre- to post-treatment changes in all six were associated with improvements in pain interference, pain intensity, or both. In addition, participant ratings of therapeutic alliance at post-treatment were associated significantly with improvements in both pain intensity and pain interference in the sample as a whole. Thus, of the 17 possible mediators examined, there were relatively few that serve as mediators for the beneficial effects of specific treatments; a larger number of variables predicted treatment outcome overall. The extent to which these variables are treatment mediators (i.e., are responsible for, rather than merely associated with, treatment-related improvements) will require further research.