Skip to main content
v2026.11,986 entries · CC-BY 4.0

Current Population Survey: Data and Research Use

Choose CPS monthly files, ASEC or supplements, check the population and weights, and document a reproducible research analysis.

Written and maintained by CASRAI Editorial Board

Published

On this page

The Current Population Survey (CPS) is a U.S. household survey that supplies labor-force statistics and supplemental social and economic data. Researchers can use published tables or public-use microdata, but must distinguish the basic monthly survey from the Annual Social and Economic Supplement (ASEC). This guide connects file selection, reference periods and reporting, with a fictional worked example showing why different labor-market rates need different denominators.

What it is

The CPS is both an ongoing statistical survey and a source of research data. The U.S. Census Bureau identifies it as the primary source of U.S. labor-force statistics. Its public data products let researchers investigate population characteristics without conducting a new household survey.

That makes CPS analysis a form of secondary data analysis: the research team develops a question using information collected for an established statistical program. The first decision is which product answers that question. A published estimate, a custom table and an analysis of individual records are different outputs, even when each carries the CPS name.

The Census Bureau separates the regular survey from supplemental collections. Its About page describes supplements covering subjects such as school enrollment, health insurance and volunteering, with schedules that vary by sponsor. The plural phrase “current population surveys” should therefore lead you to identify the particular collection, rather than assume all CPS variables appear in every monthly file.

Who runs it and what it covers

The CPS is jointly sponsored by the Census Bureau and the Bureau of Labor Statistics (BLS); Census administers the survey. The documented design covers households in the 50 states and the District of Columbia. It uses probability sampling rather than enumerating every resident.

Census describes a rotating design: a household participates for four consecutive months, leaves for eight months, then returns for four more. Rotation provides overlap between collections while bringing households into and out of the sample. This matters when planning repeated observations: adjacent monthly files are not simply independent sets of new people. The distinction between a repeated snapshot and a panel survey analysis should be explicit in the study design.

Population definitions also require care. Census describes interview eligibility beginning at age 15, whereas BLS generally publishes labor-force statistics for the civilian noninstitutional population aged 16 and older. A file can therefore contain records outside the population for a particular estimate. The survey excludes institutional populations such as people in prisons and nursing homes; one household member generally supplies information for other eligible members.

For reporting, write the universe in words before selecting records. “People in the data file,” “adults,” “employed people” and “people in the labor force” are not interchangeable descriptions. Make the eligibility statement specific to the outcome, rather than relying on a generic description copied from a dataset landing page.

How to search or use Current Population Survey data

Start with the Census CPS Data page and choose between Data Tables, Data Tools and Datasets. These are the official product labels. The following workflow combines that navigation with the Census public-use guide’s sequence for identifying files, variables, weights and comparison estimates.

  1. Define the output. Decide whether the project needs an existing published statistic, an interactive tabulation or a reproducible microdata analysis. Census describes Data Tools as a route to analysis without downloading microdata; Datasets is the route to the underlying public-use files and documentation.
  2. Select the collection. Under Datasets, choose Basic Monthly CPS for regular monthly files, Annual Social and Economic Supplements for ASEC, or CPS Supplement and Replicate Weights for another supplement. Base that decision on the measured topic, not merely a year appearing in a filename.
  3. Select the period. Monthly files are organized by year and month; other supplements are organized by supplement name and collection year. Record the chosen collection separately from the period the questions describe. The downloadable data table below distinguishes these product families and their documentation needs.
  4. Read the matching dictionary. Identify the outcome, eligible universe and valid responses before interpreting a column. The public-use guide explains that dictionaries specify variable positions and definitions. Preserve the documentation alongside the file so that an import can be checked against the intended layout.
  5. Choose the appropriate weight. Census documents different weights for different calculations. Identify the weight that belongs to the outcome and population before estimating totals or rates. A weight suitable for one analysis should not be carried into another solely because the variable is already in the script.
  6. Calculate and compare. Reproduce a relevant official estimate where possible, checking that the population, period and adjustment basis agree. The Census guide recommends this comparison and explains that confidentiality protections can produce small differences in public-use subgroup estimates.

This sequence gives a researcher a documented route from a question to an estimate. It also creates a useful stopping point: if the chosen collection does not contain the required outcome, change the data source before building the model. Finding a similarly named variable in another collection is a reason to revisit the research design.

Strengths and limits

The CPS combines labor-force measurement with demographic information and supplemental subjects. Researchers can work from the agency’s published results or formulate their own analyses from public files. The Census Data page provides these routes together, which helps connect a study’s custom result to its official statistical context.

Its limits begin with the question being measured. BLS distinguishes usual weekly earnings in the monthly survey from annual income and earnings in ASEC. It also distinguishes seasonally adjusted series from unadjusted observations. A difference between two figures can therefore reflect a different measure or adjustment, rather than a disagreement about the same quantity.

Public-use files also involve disclosure protection. Census explains that identifiers are removed and some information is altered or restricted to protect respondents. Small subgroup estimates can differ from official results; very finely divided groups can also be unreliable. Avoid treating every mismatch as proof of a software error or every computable subgroup as a publishable result.

A practical analysis record should separate three decisions: what the variable measures, how the selected observations represent the target population, and how uncertainty is assessed. Software support for survey analysis in R can help implement a design, but choosing commands does not settle those substantive decisions. State the actual choices made in the study rather than naming a package as a substitute for a methods description.

Worked example: three rates from one fictional population

Illustrative composite: the following weighted totals are fictional and are not CPS estimates for any place or date. Suppose an analysis represents 90,000 employed people, 10,000 unemployed people and 60,000 people outside the labor force. The labor force is 100,000 and the total population represented is 160,000.

BLS defines the unemployment rate as the unemployed share of the labor force, the employment-population ratio as the employed share of the population, and participation as the labor-force share of the population. Applying those definitions gives:

  • Unemployment rate: 10,000 ÷ 100,000 × 100 = 10%.
  • Employment-population ratio: 90,000 ÷ 160,000 × 100 = 56.25%.
  • Labor-force participation rate: 100,000 ÷ 160,000 × 100 = 62.5%.

Dividing 10,000 by 160,000 instead produces 6.25%. That arithmetic is valid, but it answers a different question: the unemployed share of the represented population. Calling it the unemployment rate would change the definition of the statistic.

For a research paper, this example suggests a useful check before interpreting any association: write both the numerator and denominator beside the outcome label. If two analyses use the same numerator with different denominators, their estimates should not be presented as direct replications. The distinction remains even when both calculations use the same underlying records.

Alternatives to the Current Population Survey

Other Census programs serve different research questions. The agency’s population-program directory distinguishes community information, resident counts and longitudinal economic data. The table summarizes those documented purposes; it is not a ranking of survey quality.

Source Documented purpose Research distinction
American Community Survey (ACS) Recurring demographic, social and economic information for communities. Consider its community information when that is the object of the study, rather than substituting a monthly labor-force measure.
Survey of Income and Program Participation (SIPP) Income, program participation and changing household circumstances through repeated interviews. Its longitudinal design supports questions about economic change over time.
Decennial census A count of residents used for congressional apportionment and allocating federal resources. A population count serves a different purpose from the CPS labor-force sample.

The Census SIPP overview describes following people over several years and collecting monthly information about changing circumstances. That is a concrete design distinction when the research question concerns a sequence of experiences. Specify the observation structure needed before choosing a familiar survey name.

Publishing a reproducible CPS analysis

A publication should identify the CPS product precisely enough for another researcher to locate the same inputs. The dataset menu and revision archive provide the starting evidence; the analysis record supplies the decisions made after download. The following reporting checklist is an editorial synthesis for organizing that record.

  • Identify the collection and release. Name basic monthly CPS, ASEC or the particular supplement. Keep the file identifier, collection period and revision information rather than citing only the program homepage.
  • State the reference period. Separate when interviews occurred from when the outcome occurred. For example, Census’s 2026 CPS ASEC report listings describe income, poverty and insurance outcomes for 2025. A publication year alone does not identify the outcome year.
  • Describe the analytical population. Record eligibility restrictions, exclusions and the unit being analyzed. Give the reader enough detail to distinguish a population estimate from a count of records left after filtering.
  • Record the estimation decisions. Preserve outcome definitions, weight choices, transformations and the uncertainty method actually used. Keep these decisions with the script so that a later rerun does not depend on recollection.
  • Explain the comparison target. Name the official table or statistic used to check the analysis and document meaningful differences in definitions. Do not silently alter the analytical population merely to force agreement.

The Census Dataset Revision Archive retains prior and corrected datasets. When updating a manuscript, keep the original file identity and the replacement file identity separate. Rerun the documented analysis against the intended release and record which results changed. This creates a clear explanation for a revised figure instead of leaving readers to infer whether the data, the eligibility rules or the calculation changed.

Use a specific data citation alongside the methods description and retained documentation. The citation identifies the source; the methods explain what the team did with it. Together they make it possible to evaluate the result and distinguish a changed data release from a changed analysis.

Sources

Data: CPS data-product and documentation crosswalk

Original compilation connecting Census CPS download categories with the measurement distinctions described by BLS and Census. It compares file families and documentation needs, not observed population estimates. Sources were checked on 2026-11-14.

Coverage: United States: 50 states and District of Columbia.

CPS data-product and documentation crosswalk
File familyDocumented subject coveragePeriod distinctionOfficial dataset menu labelDocumentation sources
Basic monthly CPSLabor-force status and demographic characteristicsMonthly reference week; distinguish from interview weekBasic Monthly CPSCPS Datasets; BLS Concepts and Definitions
Annual Social and Economic Supplement (ASEC)Income, poverty, health insurance and household/family characteristicsCollection year can differ from outcome year: 2026 ASEC reports describe 2025 outcomesAnnual Social and Economic SupplementsCPS Datasets; Census CPS program page
Other CPS supplementsMeasures specific to the named supplemental collectionIdentify supplement name and collection year; frequency varies by sponsorCPS Supplement and Replicate WeightsCPS Datasets; About the CPS

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about Current Population Survey: Data and Research Use

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →