Skip to main content
All CollectionsGlossaryC
Class of Worker (CoW)
Class of Worker (CoW)
Updated over a week ago

Class of worker categorizes jobs according to the type of employment of the worker. This variable identifies whether the respondent is a salaried employee of a business, or is self-employed (Canada)/proprietor (UK).


For various reasons, Lightcast further splits each of these categories in two, resulting in four classes of worker in Lightcast Data.

  • QCEW Employees: The Bureau of Labor Statistics’ Quarterly Census of Employment and Wages (QCEW) dataset is the best source for job counts data in the United States. This quarterly near-census of workers is a byproduct of unemployment insurance reporting, which businesses are required to file monthly. QCEW covers 95% of the positions held by employees in the United States.

  • Non-QCEW Employees: The remaining 5% of employment not covered by QCEW occurs marginally in specific industries and is accounted for in other government datasets. Lightcast collects employment data for these industries and puts it in the Non-QCEW Employees Class of Worker. In other words, these jobs are held by employees of businesses, but for various reasons they are not covered by unemployment insurance and therefore aren’t counted in QCEW.

  • Self-Employed: This Class of Worker includes job counts for work we typically think of as constituting self-employment. The data comes from the Census’ American Community Survey, and counts respondents who list self-employment as their primary source of income.

  • Extended Proprietors: This Class of Worker contains miscellaneous job counts recorded by the BEA that exceed counts reported in American Community Survey data. Many of these jobs are incidental self-employment that does not constitute a primary source of income (e.g. selling handmade goods on Etsy). It’s important to note that, although the goal of this Class of Worker is to account for miscellaneous income from labor, it inherently contains miscellaneous income from capital as well (since the BEA looks at profits rather than earnings).

You can use any combination of these categories, but we recommend several groupings particularly.

  • To match the BLS’s QCEW dataset most closely, use the QCEW Employees class of worker by itself.

  • For a complete picture of the employed workforce, use the QCEW and non-QCEW classes together.

  • To capture the entire employed workforce, plus self-employed persons, use the QCEW Employees, non-QCEW Employees, and Self-Employed classes in conjunction. This is the default Class of Worker setting in Analyst, and generally fits most use cases.

  • The gig economy can be approximated using the Extended Proprietors Class of Worker; however, it is critically important to keep in mind that some income and “jobs” from capital will likely be included, due to the nature of the BEA’s data. This will be an approximation of the gig economy only, likely with jobs and earnings higher than they are in actuality because of the inclusion of some “extra” jobs and income.

Did this answer your question?