Persons with dementia have difficulty communicating, which has led to the underdetection and undertreatment of pain in this population. In this paper, we compare pain assessments using 3 different measurement types: self-report, informant rating, and observational assessments. Participants were 153 nursing home residents from 4 nursing homes. Four different self-report assessments were compared with 3 observational scales and 2 informant-based assessments, which included 3 summary rating scales. Only 60% of participants were able to complete at least 1 of the self-report assessments. The strongest correlations between assessments were found among instruments of the same type, that is, among the different self-report measures, among the different observational measures, between the 2 informant rating questionnaires, and among the 3 informant rating summary questions. Our results show that the different formats of pain assessments resulted in somewhat overlapping results, yet the differences were much larger than those obtained from different assessments within the same format type. The results reraise the questions of what pain is in this population, and whether the new assessments aimed to uncover such pain have targeted the correct construct. Results suggest that a multimethod type assessment may be needed to effectively manage pain.