How Does Data Access Shape Science? Evidence from the Impact of U.S. Census’s Research Data Centers on Economics Research
Author: Abhishek Nagaraj
Author: Matteo Tranchero
Abstract: This paper studies the impact of access to confidential administrative data on the rate and direction of economics research. We estimate how the progressive geographic expansion of the U.S. Census Bureau’s Federal Statistical Research Data Centers (FSRDCs) shapes the research productivity of affected economists. We find that increased access to administrative data increases its diffusion in local universities and helps empirical researchers publish 24% more articles in top-ranked outlets and increase citation-weighted publication metrics by 40%. These effects are driven both by the direct adoption of administrative data as well as spillovers to non-adopters through increased awareness of Census databased research. Spillovers operate by leading researchers to explore a new set of topics using more robust designs and accrue only to those citing research using Census data. Our work highlights the role of data access in shaping scientific advances in the social sciences and informs the design of data infrastructures to support evidence-based policies.
Date Added: 3/31/2023, 9:41:58 AM
Seminar Notes:
Venue: CES Seminar 2023
Objective: To understand how access to high quality administrative data affects the quantity, quality, and direction of research
Importance: Scientific benefits from public data discussed broadly in context of James Webb telescope, but discussion of social and economic data mostly focuses on privacy. Tradeoff between benefits and risks
Background: Gradual expansion of Census RDC system since 1994 to 30 locations. First RDC at NBER in 1994
Data & Key Variables: FSRDC locations & opening dates
EconLit 839,000 article -> 98,000 researchers -> 15,000 economists (matched to AEA)
CES annual reports, CES working paper series, keyword search to identify those papers coming from RDC data
Methodology: Variation in access from both researcher mobility and RDC expansion. Access=ever had access, even if move away
DC always treated because data starts in 1990 and CES was created in 1982
Classify a paper as empirical or theoretical then classify people as empirical/theoretical based on majority of papers
Regression of outcome (# papers, cites to Census papers, papers in top journals)
Results: 3% of researchers in sample use an RDC, 40% had access to an RDC at some point, 23% cite a Census paper
Use of Census data and cites to Census papers more than doubles with access to RDC (for empirical researchers)
23% more publications in top journals
Treated researchers work on a more diverse set of topics
Key Table/Figure:
