Employment data with a high degree of accuracy, completeness, spatial precision, temporal consistency and categorization have become the primary means by which government agencies locate economic activity for transportation planning. Unfortunately, locating a data set with all these attributes has proven elusive. A number of sources for employment data exist but each has its own idiosyncrasies. The Ohio Department of Transportation (ODOT) had a unique opportunity to compare and contrast 5 such sources at both detailed and aggregate levels to help them decide on their future employment data source(s).

Since 2000 ODOT has utilized the Quarterly Census of Employment and Wages (QCEW) as its primary source of employment data. This dataset based upon unemployment insurance regulations contains some known systematic gaps in completeness and in some cases its spatial precision below county level is suspect. In addition, due to its confidential nature, obtaining access to it has become increasingly onerous.

ODOT also obtains Bureau of Economic Analysis (BEA) data on employment. This dataset is more complete than QCEW since it also uses tax return data; however, it is only available at the county level and thus is used mainly for establishing control totals.

The Census Bureau’s Longitudinal Employer-Household Dynamics (LEHD) dataset is primarily based upon QCEW and is more readily available and provides linkages between employee residence and employment locations, however, it only provides data by area (Census block) rather than by business site, making local review and enhancement difficult and is subject to the Census Bureau’s disclosure rules.

Infogroup’s InfoUSA (and its online interactive version ReferenceUSA) is one of several privately developed sources of data using phone directories and other publicly available sources and then enhanced and verified by their staff. It was obtained specifically for this study. It contains individual employer records with data similar to that contained in QCEW.

Lastly, two versions of the Dun and Bradstreet Global Commercial Database (also sometimes referred to as Hoover’s based upon a D&B subsidiary) were previously purchased by ODOT for various purposes (largely due to the confidentiality issues connected to using QCEW). This database is assembled in a manner similar to Infogoup’s data and contains similar attributes.

In general the private data sources have better spatial resolution than QCEW and come with positional coordinates though the accuracy of these varies by source. In addition, each has its own level of completeness and there appears to be no systematic way to identify the missing employers. Finally, unlike QCEW the private sources contain duplicate records and defunct businesses, again the level varying by source.