American Community Survey data provides the most in-depth picture of Seattle’s population and housing. ACS estimates need to be interpreted carefully and can be complicated to use so you should educate yourself on the basics of the ACS. The Census Bureau’s Compass Products can help you effectively use the ACS data while navigating the challenges involved.
It is very important to recognize that, as a sample survey, the American Community Survey provides information on the characteristics of the population – and is not meant to count the population. You should keep in mind that ACS estimates carry larger margins of error than decennial census sample estimates. This is especially true for small areas and population groups.
Estimates from the ACS represent the characteristics of the population and housing over a period of time: 1 year, 3 years, or 5 years, depending on the estimate series. (For example, the 5-year ACS estimates provide a weighted average of ACS responses collected over 60 months.) Because ACS estimates are for a whole period of time, you cannot use ACS estimates to determine what is happening at any particular day, month, or year. This is crucial to keep in mind for periods with a great deal of change.
It is important to select the estimate series that best suits your needs. Depending on the desired geography, there may be up to three different estimates to choose from:
The U.S. Bureau provides an excellent presentation “Understanding Multi-Year Estimates from the American Community Survey” that gives guidance on how to choose an estimate series and how to use the data appropriately.
The first 5-year ACS data release is for 2005-2009 and provides geographic coverage all the way down to the block group level. It is important for you to know that the 2005-2009 ACS data are controlled to population and housing estimates that are based on Census 2000 counts.
The 2006-2010 5-year ACS data is controlled to Census 2010 population and housing counts. The 2006-2010 data is intended to replace the decennial census long-form data, although important differences exist. See below for information about comparing ACS data to the decennial census.
All estimates that are based on samples, such as the ACS and the decennial census long-form estimates, include some uncertainty or “sampling error.” The margin of error is a measure of sampling error that helps to tell us the range within which the real value most likely falls.
Knowing the size of the margin of error is important because it helps you understand how reliable the estimate is: the bigger the margin of error, the less certain the estimate. In general, estimates from the ACS have larger margins of error than estimates from the decennial census long form because of the lower sampling rates in the ACS.
The Census Bureau lists margins of error alongside the ACS estimates. These published ACS margin of errors are for the 90 percent confidence interval, meaning we can be 90 percent confident that the true number falls somewhere within the range or “confidence interval,” described by the published estimate plus or minus the margin of error.
Most of the time it’s very important for you to include the margins of errors when using ACS estimates because margins of errors can be particularly high for small population groups and for small geographic areas.
If a margin of error is too high for a suitable level of accuracy, you may need to:
If you need to derive new margins of errors because the desired answer combines geographies or data categories, this requires a formula that involves more than just adding up the margins of errors of the estimates. The following are two good resources for understanding more about combining margins of errors:
The strength of the American Community Survey is in estimating characteristic distributions. The U.S. Census Bureau recommends users compare derived measures such as percents, means, medians, and rates rather than estimates of population totals.
Whether you are comparing ACS data across different time periods or against the decennial census, it is important to be aware of the differences in methodology, time periods, and geographies.
When comparing ACS data for different time periods:
When comparing ACS data to the decennial census be aware of:
For more information about comparing ACS with decennial census data, margins of error, or estimate series differences, consult the following:
Additional information about the ACS is available from: