An Analysis of Non-response Bias in the World Trade Center Health Registry
Murphy, J. J., Brackbill, R. M., Sapp, J. L., Thalji, L., & Pulliam, P. A. (2005, August). An Analysis of Non-response Bias in the World Trade Center Health Registry. Presented at Joint Statistical Meetings, Minneapolis, MN.
The World Trade Center Health Registry is designed to assess the health effects of the disaster of September 11, 2001. Baseline data collection was conducted by RTI International for the New York City Department of Health and Mental Hygiene and ATSDR. Over 70,000 30-minute interviews were conducted, but some nonresponse bias may be present in the baseline data. It is important to understand the direction and magnitude of this bias as it may impact conclusions drawn from the data. Since we know little about the nonresponders, we rely on responders’ data to predict the size and direction of bias. We assume that the relationship between demographic and health outcomes is equivalent between the responders and nonresponders; those who initial cooperators are less like nonresponders on the health measures than the initial nonresponders who eventually responded; and those who self-identified are less like the nonresponders on health outcomes than those who were reached by the survey because they were included on a list source. We provide ranges of the potential bias in the survey estimates due to differential nonresponse to inform interpretation of survey results.