Until the year 2010, the U.S. conducted a decennial census consisting of a short form, completed by everyone, and a longer, more extensive form, completed by certain households, so that information about many variables was available for almost all geographies for the decennial year. As a result, we had a lot of information for the decennial year and very little for the years in between. With the advent of the American Community Survey (ACS), that has changed.
What is the ACS? It is an ongoing nationwide survey that replaces the long form. It does not actually “count” the population but it does give information about the same variables that were available from the decennial census averaged into either 1-year, 3-year or 5-year estimates (periods of time vs. a point in time). So, this means we have more information about the years in between the census, but less detail about the decennial year itself.
It will take some time to adjust to this new way of looking at census data and it helps to keep these important tips in mind:
- Given the differences between the ACS and the decennial census, comparing data from the two sources is not recommended. The only data that can be compared is the short form data from 2010 to the previous decennial censuses.
- ACS data can be compared to ACS data. Best practice is to compare only 1-year estimates with other 1-year estimates, 3-year estimates with other 3-year estimates and 5-year estimates with other 5-year estimates and the time period should not overlap. For example, comparing data from 2005-2007 with 2006-2008 is not recommended but it is ok to compare 2005-2007 with 2008-2010.
- Due to the nature of survey data and the sample sizes, data for the smallest geographies may only be available for the 5-year estimates.
- Label your ACS data correctly: “2005-2007 ACS data” vs. either “2005,” “2006” or “”2007.”
- Most important: pay attention to “Sampling Errors,” especially to “Margin of Error,” which is presented with the data.
- Need help? Contact a librarian. email@example.com