|
Abstract:
|
Despite the widespread concern that low response rates may produce less representative samples and contribute to larger non-response biases, several recent studies have failed to show widespread or significant discrepancies of estimated statistics in RDD surveys. Built on the literature, this study explores potential influences on relational estimates of a survey by excluding respondents who required more rigorous call effort, such as more call attempts and refusal conversion. The study will show how the exclusion would affect (1) percentage estimates and (2) estimated relationships between outcomes and covariates. This study is an important addition to the literature as RDD survey data are often used to examine and estimate how various question items are associated with each other and to understand psychological, behavioral, and socio-demographic differences.
|
