Objectives. In the first randomized controlled trial of yoga on UK prisoners, we previously showed that yoga practice was associated with improved mental wellbeing and cognition. Here, we aimed to assess how class attendance, self-practice, and demographic factors were related to outcome amongst prisoners enrolled in the 10-week yoga intervention. Methods. The data of 55 participants (52 male, 3 female) who completed a 10-week yoga course were analysed. Changes in pre- and postyoga measures of affect, perceived stress, and psychological symptoms were entered into linear regression analyses with bias-corrected and accelerated bootstrap confidence intervals. Class attendance, self-practice, demographic variables, and baseline psychometric variables were included as regressors. Results. Participants who attended more yoga classes and those who engaged in frequent (5 times or more) self-practice reported significantly greater decreases in perceived stress. Decreases in negative affect were also significantly related to high frequency self-practice and greater class attendance at a near-significant level. Age was positively correlated with yoga class attendance, and higher levels of education were associated with greater decreases in negative affect. Conclusions. Our results suggest that there may be progressive beneficial effects of yoga within prison populations and point to subpopulations who may benefit the most from this practice.