Existing evidence on the relationship between religious involvement and health indicates that organizational religious involvement, such as attendance at services, is associated with better health. Findings concerning other dimensions of religious involvement, such as prayer, are inconsistent and analyses often neglect the potential influence of other correlated dimensions. Using cross-sectional data from 22 diverse European countries in the European Social Survey, including 18,129 men and 21,205 women, three dimensions of religious involvement (frequency of attendance at religious services; frequency of private prayer; self-assessment as a religious person) were studied. When analyzed separately, less frequent attendance was associated with poor health in men and women. Associations were weaker with less frequent prayer and lower religiousness. In models with all dimensions together, the association with attendance was strengthened and prayer became significantly inversely associated with health. The frequency of attendance at religious services and private prayer had opposite associations with self-rated health, resulting in negative confounding. These results are consistent with social contact being important in any health benefits from religious involvement and highlight the importance of using multidimensional measures.