The purpose of this research was to compare group-based and internet-based physical activity interventions in terms of desirability, participant characteristics, exercise self-efficacy, and barrier self-efficacy. Pretest questionnaires were completed prior to voluntary enrollment into either of the ten-week physical activity interventions. Both interventions were based on Social Cognitive Theory and the Transtheoretical Model. Interventions were followed with posttest questionnaires. Results demonstrated that the internet intervention attracted more participants, but only the group-based participants showed significant increases in exercise and barrier self-efficacy. At pretest, participants who selected the internet intervention were significantly lower in life and job satisfaction than those who selected the group intervention. Results suggest that traditional group-based exercise interventions are helpful for improving cognitions associated with exercise behavior change (e.g., exercise self-efficacy) and that the internet intervention may help employees who fall into an “unhappy employee ”typology.