Using methods intended to minimize experimenter and expectation effects, this field-based study tested the difference in exercise session attendance between formerly sedentary men and women (N = 46) whose scores on the Profile of Mood States (34) scales of Depression, Tension, Fatigue, and Vigor demonstrated moderate-to-high improvements over 10 weeks, and those whose responses did not. The self-selected cardiovascular exercise amounts approximated 30 min at a moderate intensity (60 to 70% maximal heart rate). Participants with moderate-to-high reductions (> 0.5 SD reduction based on normative values) in Depression, Tension, and Fatigue scores had significantly (p <.013) higher attendance rates over the 10-week study than participants with less reduction. The effect sizes were large (ds = 0.87 to 1.64). No such attendance difference was found in participants with moderate-to-high increases in Vigor scores. The correlation between changes in VO2max over 10 weeks and attendance was not significant (r44 =.13) ruling out a possibility that exercise session attendance was related to greater gains in fitness. Limitations and the need for replication were emphasized. Applied uses of the findings were suggested, including the possibility of promoting early exercise adherence through monitoring of individuals' exercise-related psychological changes. © 2004 by Human Kinetics Publishers and the European College of Sport Science.