Census tracts are subdivisions of counties that generally have stable boundaries that normally follow visible features. Census tracts are delineated by a local committee of census data users for presenting data. Census tracts are designed to be units with similar population characteristics, economic status, and living conditions, and average about 4,000 inhabitants Urban communities can be comprised of several census tracts, whereas a single rural census tract can contain several small communities.

U.S. Census Bureau (2005). Census tracts and block numbering areas. Geographic Areas Reference Manual (10-2 – 10-3). [Retrieved] 1/10/2011 [from] http://www.census.gov/geo/www/garm.html

Census-designated places are locally recognized communities with population clusters which lack separate municipal government, but otherwise physically resemble incorporated places. CDPs are created to provide data for settled concentrations of population that are identifiable by name by residents but are not legal entities. However, in the Oregon Communities Reporter Tool, CDPs are treated the same as Census-designated municipalities for the purposes of comparison of communities.

Census Designated Place (CDP) Program for the 2010 Census – Final Criteria, Federal Register 73:30 (February 13, 2008) pgs. 8269 – 8273, [Retrieved] 1/10/2011 [from] http://www.census.gov/geo/www/psap2010/cdp_criteria.html

Prior to the 2010 database update of the Oregon Communities Reporter, the years selected for inclusion in the database corresponded to the decennial census years 1990 and 2000. Additional non-census data were included in 2005, but the census was not performed in this year. In December 2010, the U.S. Census Bureau released the first annual installment of 5-year spanned averages from the American Community Survey (ACS). The ACS has replaced the decennial census for the collection of most demographic, social, housing, and household economic data nationwide. From 2010 onward, the Oregon Communities Reporter will be updated annually with the releases of the ACS and other non-Census Bureau data.

Since the ACS data refer to a 5-year average, it is impossible to select a specific year within those displayed for data from the U.S. Census Bureau. For example, starting in 2010, users will still be able look up Median Household Income, but will have to select the spanned average from 2005 – 2009 instead of a specific year in that range. On the other hand, data from sources other than the U.S. Census Bureau will continue to be selectable by specific years.

The American Community Survey (ACS) is an ongoing nationwide statistical survey conducted by the U.S. Census Bureau. The ACS provides communities with reliable and timely demographic, social, economic, and housing data every year. This represents a remarkable shift in the way the U.S. Census Bureau disseminates and collects data from a large sample of the U.S. population. Historically, the U.S. Census Bureau has collected detailed questions about socioeconomic and housing characteristics of the American people through the “long form” of the 10-year or decennial census. Starting in 2010, this information will come only from the ACS.

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What General Users Need to Know, U.S. Government Printing Office, Washington D.C., 2008.

Prior to 2010, the U.S. Census Bureau collected socioeconomic and detailed housing data once every ten years using a sample of about 18 million households that completed the “long-form” of the census. In contrast, with the American Community Survey (ACS), the Census Bureau samples nearly 3 million addresses each year (approximately 250,000 surveys per month). Since the annual ACS sample is smaller than that used for the past decennial censuses, the Bureau needs to combine ACS data spanning up to 5 years to produce reliable numbers for small counties, neighborhoods, and other local areas. While decennial census data will continue to be available once every ten years, starting in 2010 the decennial census will only provide basic population information. For this reason, the U.S. Census Bureau will now release 5-year detailed socioeconomic and housing estimates from the ACS for all geographical units in the nation on an annual basis.

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What General Users Need to Know, U.S. Government Printing Office, Washington D.C., 2008.

Because ACS estimates provide information about the characteristics of the community over an entire time frame, multiyear estimates contrast with “point-in-time” estimates we may be used to, such as those from the decennial census. The ACS data collection is spread evenly across the entire reference period so as not to over-represent any particular month or year within the period. Therefore, we cannot interpret these multiyear estimates by any single year, but must attribute these values to the full span of years from which they were collected (e.g. Median Household Income for 2005 – 2009).

When users report data from the ACS, they should clarify that the data represent a multiyear time-span and have a margin of error. For example, the estimated percentage of Oregonians with a high school education or greater (from the 2005–2009 ACS) was 88.3 percent, with a margin of error of 0.2 percent. The best way to report this estimate would be to state, “On average, between 2005 and 2009, approximately 88.3 percent of adult Oregonians (age 25+) had a high school or greater education level, plus or minus 0.2 percent.”

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What Rural Users Need to Know, U.S. Government Printing Office, Washington D.C., 2009.

Yes, the U.S. Census Bureau has created several guides for understanding and using data from the ACS. You can find them at: http://www.census.gov/acs/www/guidance_for_data_users/handbooks/

Caution is needed when using ACS multiyear estimates for estimating year-to-year change in a particular characteristic. This is because four years of the data in a 5-year estimate overlap with the data in the next year’s 5-year estimate. Thus, as shown in the figure below, when comparing 2005–2009 estimates with 2006–2010 estimates, the differences in the multiyear estimates are driven by differences in the non-overlapping years (in this case, the difference between 2005 and 2010).

A data user interested in comparing overlapping multiyear estimates will not be able to isolate the differences due to any particular year using these two successive 5-year estimates. While the interpretation of difference between overlapping multiyear estimates is difficult, these comparisons can be made with caution. Users who are interested in comparing overlapping multiyear estimates should refer to Appendix 4 in the U.S. Census Bureau guide for Understanding and Using American Community Survey Data (Rural Users) for more information.

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What Rural Users Need to Know, U.S. Government Printing Office, Washington D.C., 2009

Because ACS estimates provide information about the characteristics of the community over an entire time frame, the specific single year value for a community characteristic cannot be determined from 5-year estimates.

Margin of error is a measure of the accuracy of a sample at a given confidence level. The American Community Survey is administered to a random sample of the nation's population every month. The monthly values are combined and annual estimates are calculated. The margin of error describes how close the estimate likely is to the true population value for all Americans.

For example, the 2009 ACS Data Profile for Oregon shows that on average 731,554 married-couple families resided in the state between 2005 and 2009. The 90 percent margin of error for this statistic was calculated to be give or take 5,327 families (or about seven tenths of a percent). That is, we can be 90 percent sure that the true number of married-couple families that resided in Oregon on average between 2005 and 2009 lies somewhere within this calculated interval. The U.S. Census Bureau uses the 90 percent confidence level as its standard. If we wanted to be more certain, say 95 percent, then the margin of error would be larger and the upper and lower bounds around the estimate would be larger.

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What General Users Need to Know, U.S. Government Printing Office, Washington D.C., 2008

Stokes, Lynne: Tom Belin (2004). What is Margin of Error. What is a Survey? (63-67). Survey Research Methods Section, American Statistical Association. [Retrieved] 1/6/2011 [from] http://www.amstat.org/sections/srms/pamphlet.pdf

Often users want to compare the characteristics of one area to those of another area. Whenever you want to make a comparison between two different geographic areas you need to take the type of estimate into account. It is important that comparisons be made within the same estimate type. That is, 1-year estimates should only be compared with other 1-year estimates, 3-year estimates should only be compared with other 3-year estimates, and 5-year estimates should only be compared with other 5-year estimates. Because 5-year estimates cover all communities in the state, the Oregon Communities Reporter Tool only lists these estimates, making it possible to compare all communities in the 2009 database year (2005 – 2009 ACS estimates). However, since the data for 2000 and 1990 were collected by the decennial census, these values were derived using different method and they are not directly comparable to data collected by the ACS.

U.S. Census Bureau, A Compass for Understanding and Using American Community Survey Data: What Rural Users Need to Know, U.S. Government Printing Office, Washington D.C., 2009

To prevent “double-counting” geographic sub-units, the Oregon Communities Reporter Tool does not allow users to combine dissimilar units. This is because when a user selects a county, all of the smaller geographic units within the county are automatically included in the calculation of community characteristics, including municipalities (census designated places) and census tracts. If a user were able to combine a municipality with a county, for example, the municipality that has already been included in the county data could be added again, causing the report of inaccurate data.

The likely reason is that the data simply is not collected and stored at the municipal (city) level. Another possibility is that the data are available for a few cities instead of all of Oregon. Where possible, we have avoided “patchy” data sources, instead choosing comprehensive data sets for the whole state.

Information found in the “Other Data” section are displayed in single years, while Census data from the American Community Survey are displayed in the span of years from which they were collected. Selecting data from a single year is fundamentally different from choosing a multiyear span and these sections of the report have been separated for ease of use.

Under each community characteristic in the Oregon Communities Reporter is a “Details” link that will open a window that contains the source of the data and the formula used to derive the characteristic.

Click on the graph icon on the right side of the community report for the characteristic you are interested in. A graph, table, and the source of the data for all years in the Oregon Communities Reporter database will pop-up. You are authorized to use and reproduce the displayed output.

If the community you are looking at either went from about 4,000 people to a much smaller number, or declined to about 4,000 then it is likely that this community went through a change in geography.

Sometimes unincorporated towns that were allocated to census tracts one year get recognized by the Census Bureau in a later year and become Census Designated Places. In those instances, the town’s population will appear to go from about 4,000 people (the population of the census tract) to a population size that corresponds to the much smaller boundaries of the Census Designated Place. The opposite can also occur. Sometimes towns that are recognized as Census Designated Places (CDP) one year are not recognized as such the following year. In those cases, the population of this town will shift from the population residing within the boundaries of the CDP to the population residing with the census tract that contains the town (which is typically about 4,000 people).

Vincent Adams
Coordinator, Rural Communities Explorer
Office 216 Ballard Extension
Extension Family & Community Health
Oregon State University
Corvallis, Oregon 97331-5106
Phone 541-737-1412
vince.adams@oregonstate.edu