In the past, every ten years the Census Bureau conducted our decennial census to provide information to Congress and policymakers about the country. This was done through two forms – the short form (with only 7 or 8 basic questions) and the long form (which had lots of questions on everything from income to education).
Beginning in 2010, the decennial census will ONLY be done with the short form. In place of the long form, a new tool called the American Communities Survey was fully implemented in 2005. The American Community Survey (ACS) is a nationwide survey designed to provide communities a fresh look at how they are changing.
With the ACS, the Census Bureau now collects and produces population and housing information every year, instead of every ten years. About three million housing unit addresses are sampled each year throughout the United States and Puerto Rico to produce this data.
Beginning with the 2005 ACS (and continuing every year thereafter), a variety of demographic, social, economic and housing data for each year was made available for geographic areas with a population of 65,000 or more. These are called the “1 Year Estimates” – they are based on a single year’s sampling of our local population. Because these “1 Year Estimates” were only released for communities with populations of 65,000 or more, in our region this only covered Oneida County as a whole.
In 2008, however, the ACS released its first multi-year estimates based on ACS data collected from 2005 through 2007. These are called the “3 Year Estimates”. The “3 Year Estimates” of demographic, social, economic and housing characteristics are available for geographic areas with a population of 20,000 or more. For our region, this now means data are available for Oneida County, Herkimer County, Utica, Rome and the Town of New Hartford.
This past December the Census released “5 Year Estimates”. With their release, data is now available for ALL levels of geography, regardless of population size. These data cover the sample period of 2005 to 2009.
The 3-year and 5-year estimates require slightly different thinking about the data they contain. These 3-year estimates are called “period estimates”. So when discussing the child poverty data, for example, the best way to characterize this information would be along the following lines: “The child poverty rate for Oneida County over the period 2005 to 2007 was 'X' percent.” These are NOT an average of 3 or 5 individual years; they are the average of a sample taken over 3 or 5 years.
Comparing the 3-year and 5 year estimates to other Census measures is possible, but with some caveats. Generally the following are good guidelines:
- If comparing across geographic areas (like between two towns or two counties), make sure you are comparing apples to apples. That is, compare 1-year estimates to 1-year estimates, or 3-year estimates to 3-year estimates, and 5 year to 5 year estimates. DO NOT compare one-year estimates with three-year estimates across geographies.
- If comparing data for a single area (like just one town or one county to itself) over time, this can be done but it must be done with caution. Comparing the 2000 Census for Herkimer County to the 2005-2007 period estimate, for example, in Herkimer County is acceptable, but keep in mind there can problems with differences in residency rules, reference period differences, and question wording changes.
- If you are more interested in current data, 1-year or 3 year estimates are the better source to use; if precision is more important, then use the 5-year estimates. The five year estimates represent a larger sample size over a longer period of time so they have more reliability, especially for smaller areas.
A good source for what is and isn’t comparable when using the ACS data sets is found here or by going to the Census Bureau website at: http://www.census.gov/acs/www/guidance_for_data_users/estimates/ . The site url is case sensitive so be sure to use lower and upper case letters as in the address listed for the site.
