Pacific D
Vantage Point: A Likelihood-based Perspective on Reconciling Potentially Conflicting Self-report and Proxy Data (307815)
Thomas Belin, University of California, Los Angeles*Crystal Shaw, UCLA
Keywords: Proxy responses, likelihood inference, dementia, clinical trial
Many settings give rise to measurement opportunities where more than one individual could provide relevant information, including when the data being sought amounts to self-reported information from one individual, with another individual being a proxy respondent. The accuracy of the information provided by different respondents can vary, not always in a way favoring self-reported data. Specific examples include conflicting reports from a crime scene or reports of unmet needs or functional ability for those experiencing dementia or other cognitive impairment. Using the context of a clinical trial as an overarching framework, we consider a likelihood-based approach for reconciling information that might conflict between self-report and proxy respondents.