The U.S. Census Bureau runs two programs that people often confuse: the decennial census, a head count taken every ten years, and the American Community Survey, an ongoing survey that produces new estimates every year. This article explains what each one does, how they differ and what to keep in mind when you see a local statistic that comes from either.
What is the decennial census?
According to the Census Bureau, the decennial census counts the people living in the United States every 10 years. It is conducted in years ending in zero, and Census Day is April 1. The count is a constitutional requirement: the Census Bureau cites Article I, Sections 2 and 9 of the Constitution, and Section 2 calls for an enumeration “within every subsequent term of ten years.”
Congress later organized the legal details in Title 13 of the U.S. Code, which it codified in 1954. Among other things, the Bureau says Title 13 requires it to tell Congress which subjects the census will cover at least three years before the census, and the exact wording of the questions at least two years before.
The first census took place in 1790, during George Washington’s first term. According to the Bureau, it was led by Thomas Jefferson and counted 3,929,625 people, and 105 House seats were apportioned among 15 states.
Why does the count matter for representation?
The Bureau describes apportionment, the division of the 435 seats in the U.S. House of Representatives among the 50 states, as the original legal purpose of the census. The House reached 435 seats in 1913, according to the Bureau. Because seats are divided according to population counts, a state’s population relative to other states’ affects how many representatives it has.
A second use follows from the first. After each census, state officials redraw the boundaries of congressional and state legislative districts to reflect population shifts. That is why census results are closely tied to the political maps voters see. California’s own process for creating law is covered in our explainer on how the California Legislature makes law, and the rules for putting measures before voters are described in how California ballot initiatives qualify.
This article describes how the process works and does not take a position on any particular map or count.
How is the American Community Survey different?
The census is a once-a-decade count. The American Community Survey, or ACS, is what the Bureau calls “an ongoing survey that provides vital information on a yearly basis.” It covers subjects the head count does not emphasize, such as jobs and occupations, educational attainment, veterans, and whether people own or rent their homes.
The Bureau also says ACS data help determine how federal funds are distributed each year. That is one reason local governments, nonprofits and researchers pay close attention to it. The Bureau’s page does not state the ACS sample size, so this article does not give one; readers who need it should check the Bureau’s methodology documentation.
A customized version, the Puerto Rico Community Survey, provides data each year for communities in Puerto Rico.
A quick comparison
| Feature | Decennial census | American Community Survey |
|---|---|---|
| How often | Every 10 years | Every year |
| What it is | A count of the population | An ongoing survey that produces estimates |
| Main stated uses | Apportionment of House seats and redistricting | Information on jobs, education, veterans, housing tenure and more; helps determine federal fund distribution |
| Reference date or period | Census Day, April 1, in years ending in zero | Data collected over a 12-month or 60-month period, depending on the product |
The distinction matters because a survey produces estimates, while a full count aims to reach every household. Estimates are useful, but they carry uncertainty that a reader should not ignore.
What do 1-year and 5-year estimates mean?
The Bureau publishes ACS results in two main forms, and the choice between them is a tradeoff. According to the Bureau’s guidance:
| Feature | 1-year estimates | 5-year estimates |
|---|---|---|
| Collection period | 12 months (for example, January 1 to December 31, 2025) | 60 months (for example, 2021 through 2025) |
| Geographic coverage | Areas with populations of 65,000 or more | All areas |
| Sample size | Smallest | Largest |
| Reliability | Less reliable | Most reliable |
| Currency | Most current | Least current |
| Best for | Large populations, when currency matters more than precision | Small populations and smaller geographies such as tracts, when precision matters more than currency |
The Bureau’s guidance describes the 1-year estimates as less reliable and the 5-year estimates as more reliable, reflecting their sample sizes. It does not give numeric margins of error on that page. As of this writing, the Bureau lists the 2025 ACS 1-year estimates and 2025 ACS 1-year supplemental estimates as available on its ACS page.
A worked example of the tradeoff
Consider two hypothetical readers, both looking for a statistic on renters. The first wants to know about a large urban area with more than 65,000 residents and wants the most recent picture. The 1-year estimate fits that need, because it covers a single 12-month period and the area meets the population threshold. The second wants figures for a small neighborhood-sized area. A 1-year estimate is not published for areas below the threshold, so the 5-year estimate is the product available, and it averages data collected over 60 months. The illustration is hypothetical and uses no real data; it simply shows how the choice follows from the size of the place and the question being asked.
Which should a San Diego reader use?
The answer depends on the geography. The 65,000 threshold the Bureau states for 1-year estimates means larger jurisdictions can often be found in 1-year tables, while smaller communities, individual tracts and other small areas rely on 5-year products. Regional agencies often draw on census data in their planning, as described in our explainer on what SANDAG does and how regional planning works. Readers should check the Bureau’s own tables for the specific place they care about rather than assume a figure applies.
Is the information collected private?
The Bureau says census records are kept confidential for 72 years. Its page cites Title 13 as the statute that codified census law and set the notification deadlines, but does not itself say that Title 13 sets the confidentiality rule, so readers who want the legal basis for that protection should consult the statute or the Bureau’s own explanation of its confidentiality obligations.
In practical terms, this long hold on individual records is why genealogists can eventually use old census records while current-era responses are not released to the public. The Bureau’s published statistics are summaries, not lists of individuals.
How should you read a census-based statistic?
A few habits help when you come across a number attributed to the census or the ACS:
- Check the product. Is it from the decennial census or the ACS? Counts and estimates are not interchangeable.
- Check the period. A 1-year estimate covers 12 months; a 5-year estimate covers 60. Two figures with different periods may not be comparable.
- Check the geography. Small areas generally appear only in 5-year products, according to the Bureau.
- Look for uncertainty. Survey-based estimates come with margins of error. The more precise the claim, the more it is worth confirming.
- Date the number. Population and housing change, so a figure is only as current as its reference period.
Other federal statistics follow similar logic. Our explainer on how the U.S. unemployment rate is measured covers another widely quoted number that comes from a survey, and the same caution about estimates applies.
Two programs, two jobs
The decennial census answers a constitutional question: how many people live in the country and in each state, so House seats can be divided and districts redrawn. The American Community Survey answers a more everyday one: what are communities like right now, and how are they changing? For a reader in San Diego, the first shapes representation and the second supplies much of the detail behind local planning and reporting. The most reliable way to use either is to go to the Census Bureau’s pages for the current release, confirm the product, period and geography, and treat survey results as estimates.
This article explains how federal statistical programs work and is general information, not an endorsement of any policy, map or count.
Sources: U.S. Census Bureau
