Current Population Survey (CPS) Modernization

Modernizing the Current Population Survey (CPS)

The Bureau of Labor Statistics (BLS) and the Census Bureau are modernizing the Current Population Survey (CPS) to address declining response rates, manage rising collection costs, and strengthen data quality. A key part of this effort is the development of an Internet Self‑Response (ISR) option that will allow eligible sampled households to respond online.

CPS will continue to collect data by personal visit and telephone interview, but the new web option will give respondents greater flexibility. Expanding collection modes is essential because reaching respondents through traditional methods has become more difficult and more costly.

Introducing a web response option into a complex survey like the CPS poses challenges, including managing potential impacts on the survey data caused by using a new collection mode. To ensure a smooth transition, BLS and the Census Bureau are conducting extensive testing and evaluation of the self‑administered web questionnaire.

The modernization initiative will help ensure that CPS continues to deliver accurate and reliable labor force statistics for the nation.

Internet Self-Response (ISR) testing timeline

The CPS modernization effort has been a multi-year development and research project. The project accelerated in 2022–2024, when BLS and the Census Bureau began work to develop and test a new Internet Self-Response (ISR) instrument. Staff converted the existing questionnaire for web use and began building the supporting enterprise systems needed for digital data collection. This phase also included extensive expert review, cognitive testing, and online testing to refine content and the user experience. 

In 2025, the first field test contacted roughly 16,000 households across three regions and evaluated ISR feasibility by using new data collection tools, digital incentives, and alternative questionnaires. The sample included 8,000 newly selected households and another 8,000 households who had recently completed their eight monthly interviews in the CPS. These recently-completed households were contacted again for this test and asked to provide one month of data using the ISR mode. This 2025 field test was conducted outside of the normal collection period for the CPS. (Data collected in the test were not used in official CPS estimates.) Results showed ISR was workable, with web response rates around 21%, though improvements were needed in field-collection logistics, contact strategies, and questionnaire clarity.

In 2026, modernization activities included the continued evaluation of the 2025 field test results and conducting a second field test involving about 10,000 households. This sample included 5,000 newly selected households, 4,000 households who participated in the 2025 field test, and another 1,000 households who had recently completed their eight monthly interviews in the CPS. This 2026 field test was conducted outside of the normal collection period for the CPS. (Data collected in the test were not used in official CPS estimates.) This phase tested updated questions, added help text, piloted texting as a contact strategy, and evaluated longitudinal cases and revised earnings questions. By the end of 2026, ISR and other system development will be substantially complete, while work continues on refining ISR logistics and assessing mode effects.

In 2027, BLS and Census plan to analyze findings from the 2026 field test, assess industry/occupation coding and nonresponse patterns, and conduct a field test of the modernized Annual Social and Economic (ASEC) Supplement between January and April.

Next steps

Future field tests will more closely simulate CPS collection by occurring during the typical collection period. This “parallel testing” of the ISR instrument will begin in summer 2027. (Data collected in the test will not be used in official CPS estimates.) The parallel design will enable the study of ISR mode effects on key labor force estimates before the new collection mode is phased in to production. The parallel testing will differ from prior field testing in several key ways:

• Parallel testing will be resourced at levels comparable to the official CPS per sample case. This allows for a more direct comparison.
• The parallel test will use the same reference period and collection period as the official CPS, which was not logistically possible for initial field testing.

The parallel testing will have two phases, starting with a limited regional test of a small number of states. After the limited regional parallel test, BLS proposes to expand to a nationwide parallel test. (Data collected in this test would not be used in official CPS estimates.) During this national testing, BLS would produce experimental estimates using the new parallel sample with a mixed-mode design including ISR. The earliest months of results would be for internal evaluation only, but shortly thereafter, BLS would plan to release experimental results publicly. This process would allow BLS to identify and (when needed) mitigate mode effects introduced by the addition of ISR. This nationwide parallel testing would be the final step before transitioning the official survey to include ISR.

Additional resources

Last modified date: October 27, 2026