OBJECTIVES:To investigate kinship differences in the caregiver stress process by developing multiple mediation models for two distinct caregiver subgroups (spouses and adult children of older adults living with dementia). The effect of four potential mediating variables (mastery, self-efficacy, satisfaction with social support, positive caregiving appraisals) on the relationship between perceived burden and depression was evaluated. METHOD:Family caregivers of a person living with dementia were recruited through national dementia and carer organisations. Participants completed a paper-based or electronic version of the study survey. A bias-corrected, accelerated bootstrapping method was used to test the effect of the four proposed mediating variables on the relationship between perceived burden and depression in each caregiver subgroup (spouses and adult children). RESULTS:Perceived burden was directly and significantly related to depression for both spouse caregivers and adult child caregivers. The mediation models explained approximately 50% of the variance in the burden-to-depression pathway for both caregiver subgroups. Mastery and social support (but not self-efficacy, nor positive caregiving appraisals) were found to individually significantly influence the impact of perceived burden on depression in spouse caregivers. All four proposed mediators failed to reach statistical significance as individual mediators of the burden-to-depression pathway in adult child caregivers. CONCLUSION:These findings demonstrate differences in the dementia caregiver experience according to kinship, and that certain mediating variables are more relevant for some subgroups of caregivers than others. Implications for the design of psychosocial interventions are discussed.