Despite increasing interest in the relationship between religious involvement and health outcomes for cancer patients, research has been limited by the lack of appropriate measures. Few of the many instruments available are well suited to cancer patients. The current study examined the psychometric properties of one recently developed measure, the Duke Religious Index (DRI), which assesses several aspects of religious involvement. The DRI was evaluated in two distinct samples: 104 cancer patients receiving treatment at a bone marrow transplantation program and 175 gynecology clinic patients. The instrument demonstrated good internal consistency (coefficient alphas 0.87-0.94). Moderate to high correlations with other measures of religiosity provided support for convergent validity. Modest relationships with other measures commonly used in psychosocial oncology (e.g., optimism, social support, purpose in life) indicated that the instrument provides unique information (all rs's < 0. 42). Small relationships with social desirability response bias, negative affect, and relationship cohesion further supported the divergent validity of the instrument (all rs's < 0.22). The DRI was significantly associated with demographic characteristics but not with medical variables. Findings support the value of the DRI for use in oncology settings. PMID: 10739356