Vermont Housing Data: Vermont housing data profiles

Vermont housing data profiles

American Community Survey

Background

Most users of U.S. Census Bureau data are familiar with the Decennial Census. Since 1790, the United States has fulfilled its constitutional requirement to conduct a "head count" of all residents of America every ten years. Over the past few decades, the decennial Census has consisted of two surveys:

The Census "Short Form," received by every housholds in America and from which results the basic data on numbers of people and housing units, gender and age, race and ethnicity, tenure and vacancy status, and more.
The Census "Long Form," received by a sample of households, from which results the data on income, poverty, disability status, housing condition and costs, household/family status, veteran status, ancestry, languages spoken, and more.

For several years now, the U.S. Census Bureau has been developing a new survey in order to replace the decennial Long Form. The name of this new survey is the American Community Survey (ACS). The ACS is designed to release sample-based data on a yearly basis and cover all of the data topics and geographies addressed by the decennial Long Form. Once this transition is complete, the ACS will be continuously collected, and the decennial Census will consist of the Short Form only.

What to expect with ACS data

As stated abve, the American Community Survey will replace the Census Long Form and cover approximately the same questions and geographies. However the ACS and the data produced differ from the Census Long Form data that we have grown used to. Here are some key differences:

ACS data will be released on an annual basis.
The ACS is currently a household-only survey and does not include data on the group quarters population. However the ACS will include group quarters with the 2006 ACS data released in 2007 and should continue to do so indefinitely.
Currently the best ACS sampling rate to expect in Vermont is 1 in 10 households. With the Census 2000 Long Form, Vermont households were subject to sampling rates from 1 in 6 to 1 in 2.
ACS data for many counties and towns in Vermont will be released in 3- and 5-year averages. Example: the "average median family income between 2005 and 2009." With this example, data releases that are fully independent of each other will only be available every 5 years.
ACS data is released with a margin of error, which will be expressed by a +/- range associated with each data variable. [All Census data products are subject to margins of error, but the ACS will be one of the first to publish them explicitly.]
The ACS has a 2+ month usual residence rule, in contrast to the decennial Census' 6+ month usual residence rule. This rule makes it highly likely that the ACS will include data on seasonal residents, not just primary residents. There are currently no plans on the part of the U.S. Census Bureau to distinguish seasonal and primary residents within the ACS.
Many ACS questions ask the respondent to reference the previous 12 months when answering. In contrast, equivalent Census Long Form questions asked respondents to consider the previous full calendar year.

Due to these methodological differences, direct comparisons between ACS data and past decennial Census data are discouraged.

Timeline

ACS data products are being released in a staggered manner, with full implementation intended for 2010. Here is the current timeline as it pertains to Vermont:

2006: Single year ACS data releases began for Vermont (statewide only) and for Chittenden County. The first data release was for 2005. The 2006 data released in 2007 will be the first time that the group quarters population is included.
2008: 3-year average data releases begin for every geogaphy in Vermont larger than 20,000 people. This includes Burlington and most counties. The first data release will be for 2005-2007.
2010: 5-year average data releases begin for every other geography in Vermont, including towns and down to the Census block-group level. The first data release will be for 2005-2009. The ACS will be fully implemented at this point.

For more information

For more on the American Community Survey and the implications to Vermont data users, please contact: