This study investigates father involvement among a sample of ethnically diverse, low-income men participating in a randomized controlled trial of a 14-hour relationship education program that teaches skills and principles for healthy relationships. Utilizing data from 137 fathers, we examined contextual, individual, and coparental relationship pretest correlates of father involvement and found the strongest predictors were income, religiosity, ethnicity, and parenting alliance. Pre-post analyses on a subsample of 112 fathers revealed that workshop participants increased father involvement compared to controls and men whose partner attended workshops alone. Among workshop attendees, increased father involvement was significantly predicted by increased parental alliance. Implications for improving fathering by targeting the couple relationship are discussed.