On Monday, December 12, the members of the Commission for Evidence-Based Policymaking gathered for their last meeting of 2016 at the National Press Club. The meeting, which focused on key considerations for the Commission related to data management and infrastructure, featured speakers from the Census Bureau, the IRS Statistics of Income Division (SOI), Stanford University, the Bureau of Labor Statistics (BLS), the National Center for Education Statistics (NCES), and the Centers for Medicare and Medicaid Services (CMS). In order to maximize interaction time with the various speakers, , each presentation was limited to a set amount of time and each commissioner was allowed two minutes for any questions that arose. Most questions focused on specific details of the programs, though many touched on various privacy and security concerns.
Ron Jarmin, Associate Director for Economic Programs at the U.S. Census Bureau, first gave remarks on Census’s current models for acquiring data, linking data, and making data available to researchers. Examples of such programs include the Center for Administrative Records Research and Applications (CARRA), Federal Statistical Research Data Centers (FSRDCs), and the Longitudinal Employer-Household Dynamics program (LEHD). All of these programs link data from multiple sources, including federal data from agencies like the IRS, Medicare, and Medicaid, as well as state data and third-party data. In addition, Census has several joint projects with other agencies, including the Bureau of Justice Statistics, Bureau of Prisons, and the U.S. Department of Veterans Affairs.
Barry Johnson, Director of the IRS SOI Division, gave the history and mission of one of the only statistical agencies that relies wholly on administrative records rather than survey or census data. He went on to say that as a rule, research projects with SOI data must support tax administration. Finally, he described the protocol for acquiring permission, including relevant timelines, data-sharing ability, steps for linking data, and regulations for bringing in non-IRS data.
David Grusky, Professor of Sociology at Stanford University and Committee Member of the National Academy of Sciences Standing Committee on Creating the American Opportunity Study, brought a researcher’s perspective to the meeting by delivering remarks on how the process could be improved for those trying to access the data. He described how the country’s capacity for evidence could be improved in three steps: building a population frame by converting cross-sectional data into longitudinal data, which preserves and records inter-generational linkages; ramping up the population frame; and hanging administrative data off the frame by pursuing an aggressive program of data-sharing agreements.
Erica Groshen, Commissioner of the U.S. BLS, gave details on the existing partnership between federal-state cooperative programs. In exchange for BLS funding for training, IT systems, quality standards, processing, analysis, and enhancements, the states prepare and provide quarterly unemployment microdata. By working together when acquiring, transforming, reviewing, and editing data, the BLS and state MI agencies are producing the “gold standard”.
Finally, Marilyn Seastrom, Chief Statistician of NCES, and Niall Brennan, Chief Data Officer of CMS, gave a joint perspective on making federal data available to researchers. Dr. Seastrom gave an overview of protocol and processes of the Institute for Education Sciences and focused on questions from commissioners on the difficulties of compiling education statistics in the face of regulations on student-level data. Mr. Brennan, who led a push for open data from CMS, spoke on developments in the field, including agreements to release large amounts of data, creating consumer friendly interfaces, and enabling cutting-edge healthcare research by lowering barriers to entry.
The Commission’s major task is to create a report by next September detailing their recommendations on how to create evidence-building capacity in the federal government (for general topics that might be included in the report, you can read posts covering the first, second, and third meetings, as well as the public hearing). They are planning on holding at least one additional meeting on state and international models for evidence in January, as well as two additional public hearings in Chicago and the Bay area in early 2017. We will continue to follow the process… make sure to check ASA’s Science Policy Blog for more!