The U.S. Census Bureau announced in March that it is releasing its Synthetic Longitudinal Business Database Beta Data Product (SynLBD) for use by the general public. The SynLBD was produced by the Census Bureau in collaboration with the National Institute of Statistical Sciences (NISS), Duke University, Cornell University, the Internal Revenue Service, and the National Science Foundation. Research statistician Satkartar (Saki) Kinney led the NISS effort.
The purpose of SynLBD is to provide users access to a longitudinal business data product that can be used outside of secure Census Bureau facilities. The SynLBD contains synthesized information about establishments’ employment and payroll, as well as their birth and death years and actual industrial classification. The synthetic data are generated by fitting models to the confidential data and using these to simulate the replacement values. The goal is to preserve broad analyses in the data while protecting the confidentiality of individual establishments.
According to NISS Director Alan Karr, “Within the data confidentiality world, establishment data are more problematic than household or individual data, especially because large establishments are so easy to recognize. Although the nature of the difficulties is still emerging, longitudinal data are proving to be challenging in multiple ways, especially when there are ongoing releases of data.”
Researchers and others interested in using the SynLBD product may apply for a free user account on the Cornell University Virtual RDC. For more information, visit the SynLBD website.
Leave a Reply