The Guidance for Data Users page on the American Community Survey (ACS) website is a valuable resource for assistance in using ACS estimates. Here are some suggestions for avoiding common mistakes when working with ACS statistics:
- The ACS was designed to provide estimates of the characteristics of the population, not to provide counts of the population in different geographic areas or population subgroups. If you are looking for population totals, we recommend using the Decennial Census or Population Estimates Program. When an estimate from the Population Estimates Program (or PEP) is available, such as the total population or number of males/females in a county, the PEP data is the official value and is preferred.
- Use caution in comparing ACS data with data from the Decennial Census or other sources. Every survey uses different methods, which could affect the comparability of the numbers.
- Be careful in drawing conclusions about small differences between two estimates because they may not be statistically different.
- Data users need to be careful not to interpret annual fluctuations in the data as long-term trends.
- Data users should not interpret or refer to 5-year period estimates as estimates of the middle year or last year in the series. Learn more about this in the Period Estimates in the American Community Survey.
- The Census Bureau discourages direct comparisons between estimates for overlapping periods. Instead, compare nonoverlapping estimates. This means we discourage you from comparing the 2017-2021 ACS 5-year estimates to the 2018-2022 ACS 5-year estimates because four of the years overlap (2018, 2019, 2020, and 2021).
- When comparing estimates for different areas, use the same period length for each estimate. This means you should not compare a 1-year estimate to a 5-year estimate.
More suggestions can be found in the handbook Understanding and Using ACS Data: What All Data Users Need to Know.