Effect of a Public Policy Change on Respondent Reporting of Intellectual Disabilities

Even minor changes in the wording of a survey question may have important ramifications for measurement quality. I report on legislatively mandated changes in terminology that affected questions in the National Health Interview Survey (NHIS) and the effects of those changes on proxy and self-reports of an important condition. I specifically examine changes from use of the term “mental retardation” to “intellectual disability” that were mandated by a law signed by President Obama in October 2010. The wording changes were quickly implemented in the NHIS starting in January 2011. My analyses focus on changes in national prevalence of measures of intellectual disability during the period 2011-2016, compared to 1997-2010, for the population as a whole, and also for demographic subgroups, using multivariate models designed to isolate potential trend changes that can be attributed to the introduction of the revised question wording. Findings evaluate the impact of the 2010 legislation on the continuing estimation of this construct. Also, I examine how respondents perceive and interpret a variety of terms including “intellectual disability” and “mental retardation,” that may underlie the pattern of findings I have previously observed in the longitudinal NHIS data. Using Amazon’s Mechanical Turk, I collect two rounds of data from a sample of US adults to address these questions. These data address potential effects of both question interpretation and perceived stigma by examining a range of clinical and non-clinical terminology commonly used to describe persons with intellectual disabilities. Employing both quantitative and qualitative methods, I examine evidence regarding two competing hypotheses; that the Rosa’s Law effects are a consequence of (1) differences in perceptions of stigma associated with the terminology employed in these questions, vs. they are a consequence of (2) differences in perceptions of the content domain of the various terms employed. Findings enable us to understand the processes driving the observed differences in reporting that were identified subsequent to the implementation of Rosa’s Law.