SERVICES FOR SINGLE-RACE BRIDGED POPULATION ESTIMATES BY IMMIGRATION STATUS

点击翻译按钮获取中文摘要

The study of racial disparities in cancer incidence and mortality among immigrants requires population data being stratified by immigration status and single-race categories. The American Community Survey (ACS) is the only authoritative data for estimating populations, as immigration status information is not collected in the decennial censuses. However, the ACS data cannot be directly used to estimate such populations because the race data is collected in the multiple-race format in compliance with the 1997 OMB recommendations for collecting racial data in federal statistics systems. Prior effort of converting individual level multiple-race data in the ACS to single-race format exists. The University of Minnesota’s Integrated Public Use Microdata Series (IPUMS) program previously derived one single-race response for every individual in the ACS who reported to have multiple races using regression-based bridging probabilities. This method was previously implemented in the ACS sample data to derive a single-race variable (i.e. RACESING), which is released in University of Minnesota’s IPUMS data processing and release program. Despite the merit of this method, the data previously used may not reflect contemporary racial dynamics, as they are based on the pooled 1997-2000 National Health Interview Survey (NHIS) data. Because of concerns over bias that this temporal mismatch may cause to the estimation of racial distribution and challenges of integrating the single-race variable into the modernized IPUMS data processing infrastructure, starting from 2015, the IPUMS program have stopped producing and publishing this bridged single-race variable in ACS samples. The Division of Cancer Control and Population Sciences (DCCPS), Surveillance Research Program (SRP) requires 1) updated population estimates by single-race/ethnicity and immigration status for 2005-2014; 2) population estimates for 2015-2017 calculated using individual level data from the American Community Survey (ACS); and 3) population estimates for 2018-2019 calculated using individual level data from the American Community Survey (ACS). The deliverables provided under the subsequent purchase order will provide the denominator data required for the NCI to release cancer rates by single race and ethnicity among immigrants, and easy-to-access ACS samples that the NCI will be able to use to generate population estimates by alternative specifications.