Background: Pain can elevate stress in people with dementia. Although salivary cortisol is used as a biomarker of stress in people with dementia, few studies have reported the feasibility of collection methods to assess salivary cortisol in nursing home residents with both dementia and chronic pain. Objective: To explore the feasibility of collecting cortisol via salivary swab as an indicator of stress in people with dementia and chronic pain. Methods: Participants (N = 43) aged ≥ 65 years and living with dementia and chronic pain were randomly assigned to the PARO (individual, nonfacilitated, 30-min sessions with the robotic seal PARO, 5 days per week for 6 weeks) or usual-care group using computer-generated random numbers. Salivary cortisol was collected in the early morning before the intervention (Week 0) and at the completion of the intervention (Week 6) for comparison. Results: There were multiple challenges associated with saliva collection and analysis, including cognitive impairment of participants, ability to obtain repeated samples with saliva volume adequate for assay, and overall cost. Ultimately, adequate saliva was collected from only 8 participants (both pre- and post-intervention) for assay and quantitative analysis. Conclusion: Considering the multiple challenges involved in obtaining valid saliva samples in this population, salivary cortisol may not be a feasible biomarker of physiological stress in people with dementia and chronic pain.