Skip to content
CSDE News & Events

Pelletier Presents at Census Bureau’s Workshop on Advancing Research on Race, Ethnicity, and Inequality

Posted: 11/17/2023 ()

CSDE Trainee Elizabeth Pelletier (Evans School of Public Policy & Governance) presented work at the Census Bureau’s Workshop on Advancing Research on Race, Ethnicity, and Inequality last Tuesday Nov 14th. This research is co-authored with CSDE Affiliates Dr. Jennifer Romich (Social Work), CSDE Trainee Tess Abrahamson-Richards (Social Work), CSDE research scientist, Dr. Sofia G. Ayala, and Santino G. Camacho (Social Work). Administrative data arises from citizens’ interactions with government agencies, creating records that contain large observation counts over long periods of time and can yield accurate reports on earnings, transfer income, voting, residential addresses, and other factors relevant to studies of human populations. Furthermore, data derived from administrative records has the powerful potential to address one weakness of national survey data: that probability sampling does not yield sufficient sample sizes of smaller racial and ethnic populations. This paper focuses on using administrative records to capture the experiences of several populations with relatively small sizes within the US: American Indians, Alaska Natives, Native Hawaiians, Pacific Islanders, and Asian-Americans, including ethnic subgroups, the largest of which are Chinese Americans and Filipino Americans.

Authors draw on their experience developing a new integrated data resource for Washington State to present techniques for developing full population data that includes useful information on small subgroups as well as ethnoracial data for the majority of working-age adults. First, they show how integrating records across different administrative sources can create a strong approximation of a state’s population. Second, they demonstrate the capacity and limitations of current methods in creating full-population ethnoracial information, particularly the implications of standard imputation methods for developing evidence about labor market inequality. They show how applying the Bayesian Improved Surname Geocoding (BISG) imputation method yields under- and over-estimates of populations, earnings, and employment rates for different ethnoracial groups. Finally, they address the future of integrated data in Washington as their team works on a state-supported effort to build WashPop, which will draw information from state agency records to create and continually update a longitudinal, full-population data resource. They outline a framework for improving full-population ethnoracial data while practicing community-led research methods and aligning to ethics of Tribal data sovereignty and community data interests.