This project aims to contribute to a better understanding of the nature of gig work and how it affects the lives of workers. The project will contribute to Harnessing the Data Revolution by using machine learning methods to identify main and secondary employment in gig work generally and electronic-platform-mediated gig work in particular. The gig economy is of particular interest as new technologies facilitate such work more efficiently than ever before and are likely to only increase in the future. In addition, such work arrangements are particularly hard to measure as they may not be primary employment, may not be captured in tax data or administrative records, and may not be accurately reported in standard survey questions on work. The project will employ a convergence of economics- and information-science-based approaches to analyze existing, but not published, survey data. It will identify the relevant content in the existing internal data, explore how to leverage it to improve measurement of gig work, and determine what questions should be posed in future surveys to better understand the changing nature of gig work, gig workers, and gig work technologies. To do this, the project will use natural language processing to leverage narrative responses on industry and occupation as well as employer names in the 1996-2021 Panel Study of Income Dynamics (PSID). Such an effort will enable the production of a longitudinal dataset extending back over 25 years and use the dataset to begin to examine how the nature of gig work has changed with the introduction of electronic platforms and how those changes have affected individuals? wellbeing. The resulting dataset will be made available publicly and on the virtual enclave. The effort will inform the evaluation of the new gig work questions included in the 2021 PSID and aid in the development of new survey questions in ongoing data collection, on the PSID and beyond.
National Science Foundation
10/01/2021 to 09/30/2023