Other Census Bureau population datasets

The US Census Bureau manages and publishes data from several different data programs. Three of these that we use to estimate the population are the decennial census (the “census”), the American Community Survey (the ACS), and the Population Estimate Program (PEP).

While these are the most used for population and basic population characteristics, the Census Bureau conducts more than 130 surveys and programs each year. See a list of all the Census Bureau programs and surveys.

Decennial census

The main goal of the decennial census is to provide counts of people for the purpose of congressional district apportionment. This data is representative of population counts as of April 1 of the census year. This data is collected from everyone (everyone who answers the census) every 10 years. The data also includes some limited demographic variables, like age groups, sex, race, Hispanic or Latino origin, household type, relationship to householder, group quarters population, housing occupancy, and housing tenure. Learn more about the 2020 census data products available.

American Community Survey (ACS)

The American Community Survey (ACS) is a nationwide survey that collects data on the social and economic characteristics of the U.S. population—education, housing, jobs, and more. ACS data is published every year, and survey results are pooled throughout the whole year. ACS estimates reflect data that have been collected over a period of time rather than for a single point in time like the decennial census. Learn more about the ACS.

Population Estimate Program (PEP)

The Census’ Population Estimate Program (PEP) estimates overall population numbers (with a few demographic details) for a limited number of geographies: the United States, its states, counties, cities, and towns, as well as for the Commonwealth of Puerto Rico and its municipios.

PEP data is not based on a survey – it is based on population data on births, deaths, and migration. We rarely use PEP measures in City Performance projects, because we are typically looking for sub-county level data (geospatial trends within SF), or because we want more detailed demographic data that is only available in the ACS. However, if you do have a need for an overall SF population estimate, PEP is a good option. Learn more about PEP.

Note that these three are used for total population counts, but there are many other useful Census datasets by topic area. For example, the Current Population Survey (CPS) tool is useful for labor force analyses.

When to use what

We most commonly use the ACS in our projects when demographic data is needed. However, it’s good to consider your options when starting your analysis. Some considerations when choosing a data source:

• What type of data you are looking for (population counts, demographic data, other social or economic variables)?

• What is your geography of interest (state, county, or sub-county data)?

• Based on when you are looking for data, which datasets are available (and most up to date) on data.census.gov?

  • If you’re planning ahead, take a look at the ACS (or decennial census) release schedule on the Census Bureau site.

• Is it important that this data source is consistent with other projects or publications? Do you need to align with other reports, particularly if the analysis is published publicly? If so, be aware that you may need to choose the dataset that includes data for all your analyses.

  • Ex: In reporting COVID-19 data publicly, the Department of Public Health wanted the resident population estimates to be consistent across webpages. Therefore, the 5-year ACS estimates were used for everything.

Flow chart showing the decision matrix for when to choose different population datasets. If you are looking for detailed population data, you'll always use the ACS. If you are just looking for the population counts, you may use the PEP, the Decennial Census (if one has been released recently) or the ACS. You'll need to consider if consistency between publications is important. If so, we almost always use the ACS.

If you are seeking detailed population information in addition to population counts, you will use the ACS. We typically use the ACS estimates in CP projects as it is regularly published with annual estimates. It’s easiest to track trends over time with the ACS data.

If you are seeking an overall count of San Francisco residents, the PEP estimates are likely best. If you are seeking a count of residents for smaller geographies and you are within 2-3 years after a census year, the decennial census is a good choice.

As mentioned above, you may value consistency in a project. So even if you could use PEP or decennial census data, using the ACS across the board for a project may simplify communications (as the total numbers would be consistent across all tables or maps).