People, jobs and positions
The household survey measures people’s labor-force status. The payroll survey measures jobs at establishments. A person with multiple jobs may be represented differently in the two systems. Unemployment uses a defined labor-force denominator, so a low rate does not mean everyone who wants work has the work they want.
Sources: U.S. Bureau of Labor Statistics
Stocks and flows
Openings are measured at the end of the month. Hires, quits and layoffs cover activity during the month. Stocks and flows should not be subtracted or divided without a defensible interpretation. In particular, openings divided by unemployed people is not a personal offer probability: geography, occupation and hiring requirements prevent one-to-one matching.
Sources: U.S. Bureau of Labor Statistics
Dates carry meaning
An observation month says when the measured activity occurred. A release date says when the publisher made the estimate available. A retrieval date says when Opening obtained it. These dates should not be substituted for one another. The dashboard preserves retrieval vintages and labels them as such; it does not call the retrieval date a BLS release date.
A missing observation is missing data. It is never a zero. Lines on a graph must not imply invented monthly estimates when a source has a gap. Revised history can change the apparent path even when the latest point is unchanged.
Levels, changes and uncertainty
A payroll level and monthly payroll growth are different quantities. A percentage-point change in unemployment is different from a percentage change in unemployment. The dashboard displays the source units. Calculate monthly changes only between consecutive calendar months, never across an unavailable observation.
A number printed to a decimal place is not necessarily precise to a decimal place. Surveys have sampling and nonsampling error, and many monthly estimates are revised. Small moves deserve caution; a chart should not turn every change in the last digit into a story.
Use and reproduce the data
Choose an indicator and period, inspect its definition, then download the source-unit values. Retain the retrieval timestamp with the file. The download link includes that vintage, so later revisions do not change the data you cited. Opening’s archive begins with this publication; it cannot reconstruct vintages from before the archive existed.
Credit the U.S. Bureau of Labor Statistics as the producer and Opening Research for the presentation or analysis. State any transformations. The data dashboard is a public reference; it does not expose private career records or provide evidence about any one person’s prospects.
Scope and limits
This is an interpretive reference, not an exhaustive statistical manual. Consult the linked BLS program documentation before conducting inference or comparing incompatible populations.
Sources
- U.S. Bureau of Labor Statistics. How the Government Measures Unemployment. Accessed September 9, 2026.
- U.S. Bureau of Labor Statistics. JOLTS data definitions. Accessed September 9, 2026.
Cite this article
Opening Research. “How to read a labor market without flattening it into one number” September 9, 2026. https://opening.works/research/reading-labor-market-data. Include your access date. Cite the original study when using its estimates.