Multiple jobholding — holding more than one job in a reference period — is a long-standing category in labour statistics. The category was designed for a labour market in which secondary jobs were often visible as second employers on a payroll. Platform-mediated and intermittent secondary work sit awkwardly inside that frame. Published research has therefore spent as much effort on how the question is asked as on the headline rate itself.
Katz and Krueger on alternative work
Lawrence Katz and Alan Krueger compared measures of alternative work arrangements across survey instruments and periods, including analyses that used the Contingent Worker Supplement to the Current Population Survey. Their work is frequently cited for the claim that nontraditional arrangements are economically consequential and for the parallel claim that measurement choices move the estimated share. They emphasise that contingent and alternative categories are not interchangeable with “gig” or platform labels, and that trends depend on which instrument and year are compared. Later commentary has debated how far their findings travel outside the United States and outside the specific supplements they analyse.
The Contingent Worker Supplement
The United States Bureau of Labor Statistics Contingent Worker Supplement asks about contingent work and selected alternative arrangements. The supplement’s strength is a probability sample and a documented questionnaire. Its limits, which BLS documentation and external researchers both note, include intermittent fielding, question wording that may not cue digitally mediated secondary tasks, and a focus on arrangements rather than on all forms of extra earnings. When the supplement is fielded less often than monthly employment surveys, short-run movements in platform activity are not visible in the same series.
Abraham and Amaya on undercount
Katharine Abraham and Ashley Amaya, among others, have examined how survey design affects counts of informal, intermittent and electronically mediated work. Their findings support the view that standard employment questions can miss activity that respondents do not code as a “job,” especially when hours are irregular or pay arrives outside a familiar payroll channel. Redesigned probes improve capture in experimental settings, but Abraham and colleagues caution that residual undercount remains and that administrative sources answer a different question — reported receipts under tax rules — rather than a corrected version of the survey concept.
What remains unsettled
Across these strands, researchers agree that multiple-jobholding rates from traditional instruments are sensitive to wording and that some secondary digital work is under-recorded. They do not agree on a single adjustment factor that converts an official series into a platform-inclusive rate. Katz and Krueger’s alternative-arrangement estimates, BLS supplement tables and Abraham’s measurement critiques should be read as distinct claims about distinct instruments. This note records that disagreement; it does not resolve it.
Sources named in the text · September 2026