Administrative Data Censuses in US States
Main Article Content
Abstract
This paper describes work underway to conduct administrative data censuses in several US states. We are helping states identify sources that, when combined, reflect their entire population, and introduce approaches to efficiently and routinely integrate their data.
The national 2020 Census of Population and Housing faces numerous challenges. Operationally, there are concerns about self-response, inadequate technology integration and testing, hacking and data manipulation of this first online census, and the application of differential privacy to published results. This project’s main aim is to support population benchmarking in states to ensure correct counts affecting federal funding and political representation. This capacity-building has spillover effects for states: it can support their program administration, emergency preparedness, and policymaking.
We describe our plan and results to date. We have developed governance approaches and templates and proposed schema to de-duplicate individuals across data sources. We established a secure project workspace designed for cross-institution collaboration to design approaches for states to clean and standardize identifiers, to link across and de-duplicate files, to determine or impute relationships, and to resolve discrepancies observed between sources on residence and demographic characteristics.
We have assessed operational readiness in the states and gathered the concerns and recommendations of state integrated data system owners, demographers, advocacy groups, and program administrators. This stakeholder engagement was critical to design workflows and planned products. We need to report on the coverage, completeness, and overlap of data sources, and of the linked data. Next we will work with states to determine what population measures they need from this new capacity: what demographic characteristics and levels of geography are desired for tables and maps.
The project aims to build trust in the data and process from the ground up. We will develop training materials and offer technical assistance to help other states adopt this population benchmarking capacity.
This paper describes work underway to conduct administrative data censuses in several US states. We are helping states identify sources that, when combined, reflect their entire population, and introduce approaches to efficiently and routinely integrate their data.
The national 2020 Census of Population and Housing faces numerous challenges. Operationally, there are concerns about self-response, inadequate technology integration and testing, hacking and data manipulation of this first online census, and the application of differential privacy to published results. This project’s main aim is to support population benchmarking in states to ensure correct counts affecting federal funding and political representation. This capacity-building has spillover effects for states: it can support their program administration, emergency preparedness, and policymaking.
We describe our plan and results to date. We have developed governance approaches and templates and proposed schema to de-duplicate individuals across data sources. We established a secure project workspace designed for cross-institution collaboration to design approaches for states to clean and standardize identifiers, to link across and de-duplicate files, to determine or impute relationships, and to resolve discrepancies observed between sources on residence and demographic characteristics.
We have assessed operational readiness in the states and gathered the concerns and recommendations of state integrated data system owners, demographers, advocacy groups, and program administrators. This stakeholder engagement was critical to design workflows and planned products. We need to report on the coverage, completeness, and overlap of data sources, and of the linked data. Next we will work with states to determine what population measures they need from this new capacity: what demographic characteristics and levels of geography are desired for tables and maps.
The project aims to build trust in the data and process from the ground up. We will develop training materials and offer technical assistance to help other states adopt this population benchmarking capacity.