What Census P60 median earnings measure

Each year the Census Bureau publishes a report in its P60 series on income in the previous calendar year. Alongside household income it reports median earnings, and those medians feed three of the trends pages here: annual earnings by education, median earnings by age and sex and the female-to-male earnings ratio. This page sets out what those figures measure.

The survey behind the report

The numbers come from the Annual Social and Economic Supplement to the Current Population Survey, the CPS ASEC, which asks about the previous calendar year. The report titled “Income in the United States: 2024,” for example, is built from the 2025 ASEC. Respondents answer for every household member aged 15 and over.

The income concept

The Census figures use “money income”: cash received before taxes and before deductions such as Social Security contributions, union dues or Medicare premiums. It does not include noncash benefits like health insurance or food assistance, and it excludes capital gains. Within money income, earnings are the part received for work: wage or salary income, which the definitions say includes wages, salary, armed forces pay, commissions, tips, piece-rate payments and cash bonuses; net income from nonfarm self-employment; and net income from farm self-employment. Earnings are thus broader than an OEWS wage in two ways, taking in bonuses and the self-employed, and narrower in one, counting only cash.

Full-time, year-round

The headline earnings medians are usually for full-time, year-round workers. The definition has two parts. A full-time worker is one who worked 35 hours or more a week in a majority of the weeks worked during the year; a year-round worker is one who worked 50 weeks or more. A person who meets both is full-time, year-round. The restriction removes most students, seasonal workers and people who entered or left the workforce during the year, and it is the main reason the Census medians exceed medians computed over all workers with earnings. The report also publishes medians for all workers, which are lower and are influenced by hours as much as by rates of pay.

What the report presents

The report compares the current year with the previous one, stating changes in medians and whether they are statistically significant. Its historical tables run back decades and are expressed in inflation-adjusted dollars so that years can be compared; the current-dollar figures are in the detailed tables. For full-time, year-round workers it gives median earnings for men and for women and the ratio between them, which gender pay gap statistics discusses. Medians by educational attainment are published in the companion detailed tables and are the source for earnings by education.

How these differ from OEWS

An OEWS annual wage is an hourly rate multiplied by 2,080 hours, reported by employers for a May reference month, for a detailed occupation in a specific area. A Census median is a calendar year’s actual earnings as the household reports them, for people, not jobs, across all occupations or by broad characteristic. Someone who worked two jobs, took unpaid leave, earned a bonus or ran a business during the year has a Census earnings figure that reflects all of that and an OEWS wage that reflects none of it. The Census report states whether a year-to-year change is statistically significant; OEWS publishes relative standard errors from its sample design.

Reading them on this site

The trends pages print the Census figures with the report year and the income year, since the two differ by one. A Census median for workers aged 25 to 34 or for women with a bachelor’s degree should be read as what such people earned over a year, not as what any occupation pays. For the latter, the occupation pages are the source, and the state pages add geography the national Census medians do not have. Neither series is a substitute for the other; they answer different questions about the same workforce.