U.S. flag

An official website of the United States government

Skip Header


Sample Size Definitions

1. How large is the sample?

The full implementation of the ACS, which began in 2005, sampled approximately 2.9 million housing unit addresses annually stateside. The PRCS sampled approximately 36,000 housing unit addresses each year in Puerto Rico. The 2011 ACS sampled approximately 3.3 million housing unit addresses - this corresponds to an increase in the targeted annual sample size of 3.54 million addresses that began with the June 2011 ACS sample. This increase in the targeted annual sample size has continued since then, resulting in a current annual ACS sample of approximately 3.54 million housing unit addresses.

The full implementation of the ACS and PRCS Group Quarters (GQ) data collection began in 2006. Sample sizes approximated 200,000 persons through 2016 and around 170,000 persons starting in 2017.  The counts through 2016 included sample persons in GQs that were out-of-scope – these sample persons were removed from subsequent sample counts.

The ACS sampled between 740,000 and 900,000 housing unit addresses annually between 2000 and 2004.

2. What is sampling error?

The ACS estimates are based on data from a sample of housing units and people in the population, not the full population. For this reason, ACS estimates have a degree of uncertainty associated with them, called sampling error. In general, the larger the sample, the smaller the level of sampling error.

3. Why is it important to measure sampling error?

The estimates produced by the ACS are not exact because they are based on a sample. The sampling error measures the degree of uncertainty associated with the estimates. If the degree of uncertainty is too large, then users should be cautious in how the estimates are used.

4. How does the ACS measure sampling error?

The ACS calculates standard errors for each estimate produced and publishes the 90 percent confidence level margins of error (the Census Bureau standard). You can be 90 percent confident that the interval within the margin of error from the estimate includes the true value. See Accuracy of the Data for more details on how margins of error and confidence intervals are calculated and interpreted.

5. What other numbers provide important information on the reliability of ACS estimates?

Note: Sample and interview counts were rounded for 2018 to help reduce disclosure risk. In 2019, noise injection replaced rounding as our method to reduce disclosure risk for interview counts. As part of that review, we determined that the sample counts did not require protection. For more information, see this user note.

Housing Units Initial Addresses Selected - The number of addresses in each state and for the nation that were selected for the ACS sample for a particular year. Each year's sample is systematically divided into 12 monthly samples for ACS interviewing. This initial number includes addresses later determined to be commercial or nonexistent, as well as housing units that are not interviewed due to subsampling for personal visit follow-up, refusals, or other reasons. It excludes addresses that were deleted from the sample, e.g., addresses that were part of a sample reduction.

Housing Units Final Interviews - The final number of interviews across all modes of data collection for the ACS in a given year for the nation and by state. This number includes occupied and vacant housing units that were interviewed by the internet, mail, or personal visit methods between January 1 - December 31. It excludes addresses determined to be nonexistent or commercial, addresses not selected in the subsample for personal visit follow-up, and addresses that are not interviewed due to refusals or other reasons.

Internet interviewing began in January 2013, giving the ACS four interview modes for data collection.  Computer-assisted telephone interviewing then ended in September 2017, resulting in the ACS having three modes of data collection again. 

Group Quarters People Initial Sample Selected - The number of people living in in-scope GQs that were selected for ACS interviewing in a given year for the nation and by state. Each year's sample is systematically divided into 12 monthly samples for ACS interviewing. This initial number includes people not interviewed due to the group quarter refusing entry, the person refusing to respond, the person being unavailable at the time of the interview, or other reasons. It excludes potential sample cases where the GQ was determined to be out-of-scope or nonexistent as well as any sample reductions where there was no opportunity for a person to be contacted.

Refinements were made to the methodology for computing initial sample beginning with the 2017 sample.  One of these refinements was to omit people in out-of-scope GQs from the initial sample size.  The effect of these refinements is a smaller sample size.

Group Quarters People Actual Final Interviews - The final number of actual person interviews for the ACS for those living in group quarters in a given year for the nation and by state.

Group Quarters People Synthetic Final Interviews - The final number of synthetic person interviews for the ACS in a given year, for the nation and by state. For more information, see the User Note on synthetic interviews.

6. Are these numbers used to calculate the survey response rate?

These numbers by themselves are not used to measure the response to the request to participate in the ACS. See Survey Response Rates - Definitions for more information.

7. How does the ACS calculate these five measures of sample size?

Housing Units Initial Addresses Selected (state x, year y) =

The number of housing unit addresses selected for sample from the Master Address File, for state x in year y

Housing Units Final Interviews (state x, year y) =

The number of all housing unit interviews successfully conducted by internet, mail, or personal visit between January 1 and December 31 of year y, in state x, adjusted for disclosure avoidance

Group Quarters People Initial Sample Selected (state x, year y) =

The number of expected sample persons across sample GQs in state x, year y

Group Quarters People Actual Final Interviews (state x, year y) =

The number of all actual person interviews between January 1 and December 31 of year y, for state x, adjusted for disclosure avoidance

Group Quarters People Synthetic Final Interviews (state x, year y) =

The number of all synthetic person interviews between January 1 and December 31 of year y, in state x, adjusted for disclosure avoidance

Page Last Revised - August 22, 2023
Is this page helpful?
Thumbs Up Image Yes Thumbs Down Image No
NO THANKS
255 characters maximum 255 characters maximum reached
Thank you for your feedback.
Comments or suggestions?

Top

Back to Header