Frequently asked questions about earnings data from the Current Population Survey

​

Back to CPS Earnings main page

1. What earnings data are available from the CPS?

The Current Population Survey (CPS) publishes earnings data for wage and salary workers (which excludes the self-employed). These estimates reflect earnings before taxes and deductions and include overtime, commissions, and tips usually received. For multiple jobholders, earnings are reported for a worker’s main job, that is, the one at which they work the most hours.

BLS publishes median usual weekly earnings for full-time workers—those who usually work 35 hours or more per week at their main job. Earnings estimates are available by characteristics such as sex, age, educational attainment, race and Hispanic or Latino ethnicity, and occupation. Additionally, median usual weekly earnings of part-time wage and salary workers are also available.

Earnings data are based on information collected from one-fourth of CPS households each month and are published as quarterly and annual averages. Because the CPS is a sample survey, estimates are subject to sampling error, which can be larger for smaller population groups. The data also may be affected by reporting errors or nonresponse.

For guidance on where and how to retrieve earnings data presented in the Usual Weekly Earnings news release, see question 3 below.

2. Why does CPS use medians for earnings data?

The Current Population Survey (CPS) uses median earnings to reflect the typical worker’s earnings. The median represents the point where half of workers earn more and half earn less. Unlike means, medians are less affected by a small number of very high or very low earnings values and are more stable over time. This makes the median a useful measure for comparing earnings across groups, especially when the distribution of earnings is skewed or contains outliers.

For example, imagine you and a few coworkers standing in a room. You all work for the same company, hold the same job title, and earn identical salaries. In this case, the mean and median earnings for the group would be exactly the same. However, if someone from a different company with extremely high earnings joined the group, the median earnings value for the group in the room would remain unchanged, but the mean would become substantially higher.

3. How does BLS calculate median earnings estimates from the Current Population Survey?

The BLS procedure for estimating the median of a weekly earnings distribution places each reported or calculated weekly earnings value into a $50-wide interval that is centered around a multiple of $50. Similarly, for hourly earnings, medians are calculated based on earnings distributions using $0.50-wide intervals that are centered around multiples of $0.50. In both cases, the median is calculated through the linear interpolation of the interval in which the median lies.

Changes over time in the medians for specific groups may not necessarily be consistent with the movements estimated for the overall median boundary. The most common reasons for this possible anomaly are as follows:

•          There could be a change in the relative sizes of the subgroups. For example, consider two age groups: those ages 16 to 24 and those age 25 and older. The median earnings of both groups may rise over time. However, if the lower-earning 16-to-24-year-old age group accounts for a greatly increased share of the total over time, the overall median for those age 16 and older could fall.

•          Large shifts in the distribution of reported earnings—that is, how the amounts workers earn are spread across a range of dollars—can affect how the median changes over time. For example, many respondents may report rounded earnings, such as $700 or $800, creating clusters in the data. When the median falls within one of these clusters, it may remain there longer even as overall earnings rise. Once the median moves past such a cluster, however, it can shift more quickly to the next commonly reported amount. Although BLS estimation procedures help reduce these irregular movements, users should exercise caution when interpreting short‑term changes in medians or in ratios based on medians.

4. How can I retrieve earnings data presented in the Usual Weekly Earnings news release?

Use the earnings table data retrieval tool to access historical data from the Usual Weekly Earnings of Wage and Salary Workers news release. Data users will be able to extract historical data by characteristics such as sex, age, educational attainment, race and Hispanic or Latino ethnicity, and occupation. For some series, additional historical data may be available on an annual average basis that is not available on a quarterly basis. To get annual averages, you must check the box at the top of the output page.

Data users can access previous editions of the Usual Weekly Earnings of Wage and Salary Workers news release in our Archived News Releases. Although archived news releases are available, please note that estimates in the news releases may have been revised after publication. See the data retrieval tool to access the latest data.

5. Why are the earnings of people who are Black, Asian, or Hispanic or Latino compared to the earnings of people who are White?

Earnings comparisons in the Usual Weekly Earnings of Wage and Salary Workers news release use people who are White—rather than another race or ethnicity group—as the reference group because doing so provides more stable comparisons across demographic groups and over longer time periods. While people who are Asian on average earn more than the other groups, data for people who are Asian are based on smaller sample sizes which can result in greater variability in quarterly average estimates. 

6. Why does the measure of women’s earnings as a percentage of men’s published by BLS differ from the measure published by the U.S. Census Bureau?

Two measures of women's earnings as a percentage of men's are widely used in the discussion of men's and women's earnings. Both come from the Current Population Survey (CPS), but they are slightly different in timing and which workers are included.

The BLS measure of women's earnings as a percentage of men's is based on the median usual weekly earnings of full-time wage and salary workers.

The usual weekly earnings data from BLS reflect earnings before taxes and other deductions and include overtime pay, commissions, or tips usually received. The weekly earnings estimates are developed from questions asked of one-fourth of the CPS sample each month. Full-time workers are those who usually work 35 or more hours per week. All self-employed workers, both incorporated and unincorporated, are excluded from these earnings estimates.

The BLS measure is available back to 1979 as a quarterly or annual average.

The Census Bureau measure of women's earnings as a percentage of men's is based on the median annual earnings of year-round, full-time workers. (See Census historical income table P-40.)

The annual earnings estimates from the Census Bureau include wages, salaries, and self-employment income received during the full calendar year. The annual earnings estimates are developed from questions in the Annual Social and Economic (ASEC) supplement to the CPS, conducted principally in March of each year and asked of the full survey sample. Year-round, full-time workers are those who worked 50 or more weeks during the previous calendar year and 35 or more hours per week during a majority of the weeks worked.

The Census Bureau measure is available back to 1960.

Neither the BLS nor the Census Bureau measure is a controlled comparison of men's and women's earnings; that is, these measures do not control for factors such as occupation, education, and work experience that can be significant in explaining earnings differences. Hence, neither measure should be construed as a comparison of workers with comparable jobs or other characteristics.

7. Does the CPS tabulate earnings data at the state level?

Yes. Earnings for men and women at the state level are available as annual averages in our online database back to 2012. BLS does not tabulate state earnings estimates for any other demographic or job characteristics.

To retrieve the relevant Series ID codes for state-level earnings data, use this link to an Excel file for data from the BLS online database. 

8. Can CPS data show whether men and women are paid the same for the same job?

No. The Current Population Survey (CPS)—the source of BLS occupational earnings measures for men and women—gathers enough information from survey respondents to identify an occupational classification. These classifications, however, are not sufficiently precise with respect to specific job skills, responsibilities, and specialization to assume that everyone grouped in a particular classification is doing the same job or comparable work. A number of the occupational classifications are in fact miscellaneous by definition.

In addition, BLS men's and women's occupational earnings measures are on a broad level and do not control for many factors that can be important in explaining differences in earnings.

In sum, BLS occupational earnings data from the CPS reflect the earnings of men and women whose jobs are included in the same occupational classification, but they should not be construed as reflecting workers with necessarily the same or comparable job, or otherwise comparable characteristics.

9. Does the Usual Weekly Earnings news release contain all the CPS earnings data?

No. Additional earnings data can be found on the CPS website. Annual average earnings statistics are presented in online tables located under the heading “weekly earnings” (tables 37 to 39). The One-Screen Data Search tool allows users to access weekly and hourly earnings data from the CPS. Also, all data users are encouraged to review the resources available in the earnings section of the CPS website.

Last Modified Date: July 30, 2026