Research on religious attributions has been limited by a preoccupation with disentangling “religious” from “naturalistic” attributions and a failure to capture the attributions that people make in response to meaningful events. Thirty years of research has shown that even under optimum conditions religious attributions are rare compared to naturalistic ones. This research draws on unique archival materials comprising letters written to the Panacea Society containing self‐reported effects of a spiritual healing treatment based on water‐taking practices. This analysis examined attributions over time among a sample of letter writers (N = 19) from the 1920s using the Leeds Attributional Coding System to examine patterns of attributions that correspondents made in response to improvement and worsening of health outcomes. In line with previous research, religious attributions were more common for positive outcomes than negative outcomes. Contrary to previous research, religious attributions accounted for the majority of attributions made compared with nonreligious attributions. We discuss the implications for future research in studying attributions in real‐life meaningful settings and in expanding the repertoire of attributions to include religious ritual and community.