The unemployment rate is an estimate built from interviews, not a monthly headcount of everyone without a job. The federal government asks a carefully selected group of households about work and job search, then uses their answers to estimate the labor force across the country. That approach makes a monthly measure practical, but it also means the figure has a margin of sampling error. Understanding the survey’s reference week and definitions helps explain what the rate captures—and what it leaves outside its count. Bureau of Labor Statistics
Why interview a sample?
A census of every household would take too long and cost too much to repeat each month, and repeated visits would burden households. Unemployment insurance records cannot substitute: they cover people who applied for benefits, while some unemployed people are ineligible, have exhausted benefits or never apply. The Current Population Survey (CPS) offers a regular national measure without either limitation. Bureau of Labor Statistics
The Census Bureau interviews about 60,000 eligible households, representing roughly 110,000 people. Its sample is selected to reflect the country’s states, urban and rural areas, and other geographic differences. A portion of households rotates out over time, while many remain in consecutive monthly samples; this overlap helps make month-to-month comparisons more reliable. Bureau of Labor Statistics
One week anchors the questions
Interviewers ask about each person’s activity during a reference week, usually the week containing the 12th of the month. The survey does not ask respondents to label themselves unemployed. Instead, answers to standardized questions determine whether someone had a job, was looking for one and was available to start. That common procedure makes responses comparable. Bureau of Labor Statistics
For example, imagine a hypothetical household with one adult who worked for pay during the reference week and another who had no job, submitted applications in the previous four weeks and could have started work. The first person is classified as employed; the second as unemployed. A person without a job who did not actively search or was unavailable is generally outside the labor force, rather than counted as unemployed. Bureau of Labor Statistics
From answers to a national estimate
Survey answers are weighted to independent population estimates from the Census Bureau. The adjustments account for characteristics including age, sex, race, Hispanic ethnicity and state of residence, so each interviewed person represents an appropriate share of the broader population. The published totals therefore describe estimated numbers of people, not just the respondents in the sample. Bureau of Labor Statistics
The unemployment rate is the number of unemployed people as a share of the civilian labor force, which consists of employed and unemployed people. Someone who is not working is not automatically in its numerator: active job search and availability usually matter. A person waiting to be recalled from temporary layoff is an exception to the active-search requirement. Bureau of Labor Statistics
What the estimate can—and cannot—say
Because the CPS interviews a sample, its result can differ from what a complete census would produce. The Bureau of Labor Statistics says there is a 90-in-100 chance that the monthly unemployment estimate is within about 300,000 of a total-census figure. That margin concerns the estimate; it does not mean the same number of individuals are misclassified, or that every part of the labor market has equal uncertainty. Bureau of Labor Statistics
In the September 2026 report, the household survey put the unemployment rate at 4.2 percent and the estimated number of unemployed people at 7.1 million. Those figures describe the survey’s definitions and reference period. The same release states that October’s employment report was scheduled for November 6, after the November 1 source update, so September is the latest reported month here. Bureau of Labor Statistics
Sources & further reading
- How the government measures unemployment ↗Bureau of Labor Statistics
- September 2026 employment report ↗Bureau of Labor Statistics
