Utility values to estimate quality-adjusted life years (QALYs) for use in cost-utility analyses are usually elicited from members of the general population. Public attitudes and understanding of dementia in particular may mean that values elicited from the general population may differ from patients and carers for dementia health states. This study examines how the population impacts utility values elicited for dementia health states using interviewer-administered time tradeoff valuation of health states defined by the dementia-specific preference-based measures DEMQOL-U (patient-report) and DEMQOL-Proxy-U (carer-report). Eight DEMQOL-U states were valued by 78 members of the UK general population and 71 patients with dementia of mild severity. Eight DEMQOL-Proxy-U states were valued by 77 members of the UK general population and 71 carers of patients with dementia of mild severity. Random-effects generalized least squares regression estimated the impact of population, dementia health state, and respondent sociodemographic characteristics on elicited values, finding that values for dementia health states differed by population and that the difference varied across dementia health states. Patients with dementia and carers of patients with dementia gave systematically lower values than members of the general population that were not due to differences in the sociodemographic characteristics of the populations. Our results suggest that the population used to produce dementia health state values could impact the results of cost-utility analyses and potentially affect resource allocation decisions; yet, currently, only general population values are available for usage.