Reading the monthly jobs report: payrolls, the unemployment rate and revisions

What the monthly Employment Situation release measures, why its two surveys can disagree, and how revisions and sampling error change what a headline number means.

Empty desks and chairs in an office in morning light.

Once a month, news reports lead with two numbers from one government release: how many jobs the economy added and the unemployment rate. This guide explains where they come from and what they count, so you can tell when a monthly change is clear and when it is within the margin of error or likely to be revised. It does not interpret any particular month.

What the jobs report is and who publishes it

The report’s official name is The Employment Situation. It is published by the U.S. Bureau of Labor Statistics (BLS). Each release covers the previous month, and BLS posts upcoming dates on its release schedule (BLS, checked November 16, 2026). Every date on the current schedule is a Friday, at 8:30 a.m. Eastern time.

The latest release, with its summary, tables, technical note and FAQ, is at the Employment Situation table of contents (BLS, checked November 16, 2026).

Two surveys, two different questions

The release combines two surveys, described in its technical note (BLS, checked November 16, 2026). The establishment survey, formally the Current Employment Statistics survey, asks employers. Each month BLS collects payroll records from about 119,000 businesses and government agencies, covering roughly 622,000 worksites. It produces the payroll jobs number, plus hours and earnings by industry. Its results appear in the “B” tables.

The household survey, formally the Current Population Survey, asks people. It samples about 60,000 eligible households and is conducted by the Census Bureau for BLS. It produces the unemployment rate, labor force participation and demographic breakdowns. Its results appear in the “A” tables.

Both refer to a single week or pay period: generally the calendar week containing the 12th of the month for households, and the pay period including the 12th for employers.

Why the two can point different ways

The surveys count different things. According to the technical note, the household survey includes self-employed people whose businesses are unincorporated, agricultural workers, unpaid family workers and private household workers; the establishment survey excludes them. The household survey counts a person with two jobs once; the establishment survey counts each job on each payroll. People on unpaid leave count as employed in the household survey but not on payrolls.

Sample size matters too. BLS’s Employment Situation FAQ (checked November 16, 2026) says an over-the-month employment change of about 122,000 is statistically significant in the establishment survey, while the threshold in the household survey is about 650,000. So in a given month the two can move in different directions without either being wrong.

Two clipboards with blank paper side by side on a table.

How the unemployment rate is defined

The household survey sorts each person 16 and over into one of three groups. In the technical note’s words, people count as unemployed if “they had no employment during the reference week; they were available for work at that time; and they made specific active efforts to find employment” in the four weeks ending with that week. People laid off and expecting to be recalled count as unemployed even if they are not looking.

The employed and unemployed together form the civilian labor force; everyone else is outside it. That includes people who want a job but, as the release summary explains, are not counted as unemployed because they were not actively looking in the previous four weeks or were unavailable to take a job (BLS, checked November 16, 2026). The unemployment rate is the number unemployed as a percent of the labor force. The labor force participation rate is the labor force as a percent of the population.

The unemployment count has nothing to do with unemployment insurance: the technical note says the figures “in no way depend upon the eligibility for or receipt of” benefits.

The alternative measures, U-1 to U-6

BLS also publishes six measures of labor underutilization in Table A-15 (checked November 16, 2026). U-3 is the official unemployment rate. The narrower U-1 and U-2 cover people unemployed 15 weeks or longer, and people who lost jobs or finished temporary jobs. The broader U-4 to U-6 add discouraged workers, then everyone marginally attached to the labor force, and finally, in U-6, people working part time for economic reasons who want full-time work.

Why the payroll numbers get revised

The headline payroll figure is preliminary. The FAQ says BLS revises each monthly estimate twice, in the next two monthly releases, to include additional survey responses and recalculated seasonal adjustment factors. The technical note says an estimate is considered final only after those two revisions, “when nearly all sample reports have been received.” Each release reports these revisions near the end of its establishment section.

Once a year, BLS also re-anchors the survey to nearly complete employment counts from unemployment insurance tax records. The technical note calls the March difference between the survey estimate and those counts a benchmark revision. Over the prior 10 years the absolute average was 0.2 percent of total nonfarm employment, with revisions ranging from -0.4 percent to 0.3 percent.

Sampling error and the confidence interval

Both surveys are samples, so their estimates carry sampling error. BLS analyses are generally conducted at a 90 percent level of confidence. The technical note states that “the confidence interval for the monthly change in total nonfarm employment from the establishment survey is on the order of plus or minus 122,000.”

In BLS’s own example, if payrolls rose by 50,000, the 90 percent interval would run from -72,000 to +172,000. Because that range includes values below zero, BLS says “we could not say with confidence” that employment rose. A reported rise of 250,000 would put the whole interval above zero.

For the household survey, the interval for the monthly change in the unemployment rate is about plus or minus 0.3 percentage point at an unemployment rate of around 6 percent, so a move of a tenth of a point falls well inside it.

Average hourly earnings

The establishment survey also reports average hourly earnings for all employees on private nonfarm payrolls, in Table B-3, and for production and nonsupervisory employees in Table B-8 (BLS, checked November 16, 2026). Because these are private-sector payroll figures, they leave out government workers and the unincorporated self-employed. As a general point about averages, not a BLS statement, the figure can shift when the mix of jobs changes, not only when individual pay changes. The latest two months in the table are marked preliminary.

A short checklist for reading a release

  1. Note which month the release covers and its publication date, at the top of the summary.
  2. Read the payroll change and the unemployment rate as coming from different surveys.
  3. Compare the payroll change with the plus or minus 122,000 interval before treating it as a clear rise or fall.
  4. Check the revisions to the prior two months.
  5. Look at labor force participation and Table A-15 alongside the unemployment rate.
  6. Treat the latest two months of payroll and earnings data as preliminary.

Common questions

Does the unemployment rate count only people receiving unemployment benefits?

No. The FAQ says the household survey has no requirement or question relating to unemployment insurance. People without jobs who are available and actively looking count either way.

Why can payrolls rise while the household survey shows fewer people employed?

The surveys cover different groups, count multiple jobholders differently and have very different margins of error. BLS says both have strengths and limitations.

How much can a monthly payroll figure change after it is first reported?

It is revised twice in the following two months and again, indirectly, in the annual benchmark.

Sources checked November 16, 2026: BLS — Employment Situation table of contents; BLS — Employment Situation Summary; BLS — Employment Situation Technical Note; BLS — Employment Situation FAQ; BLS — Table A-15, Alternative measures of labor underutilization; BLS — Table B-3, Average hourly and weekly earnings; BLS — Schedule of Releases for the Employment Situation.