AkLA Census Tutorial Part 5: 100% DATA VS. SAMPLE DATA

My thanks to all of you who've hung on so far. We are now more than half way through our tutorial on Census 2000. Please pat yourself on the back for getting this far.

100% DATA VS. SAMPLE DATA

If you look closely at census reports, you will see that some of them say that they're based on "100% data", like the Census Briefs at [http://www.census.gov/population/www/cen2000/briefs.html], while others state they are based on "sample data", like 1990 Census Summary Tape File 3 available at [http://homer.ssd.census.gov/cdrom/lookup NOTE: no longer available on Census Bureau website. "The 1990 Census Lookup data access tool is no longer available, due to resource limitations. Please use American FactFinder to access STF 1 and STF 3 data." 9/2010]. So what's the difference? "100% data" is the data taken from questions sent to every Census respondent (i.e. all of the questions on the "Short Form," or (if these refer to the same questions) the first six questions of the "Long Form"). Sample data is based on questions 7-53 of the "Long Form." Nationwide, one out of six people the "Long Form," but that doesn't tell the whole story. To ensure a large enough sample was taken, the Census varied the sampling ratio of the "Long Form" by population size. According to the Encyclopedia of the US Census, p. 277, areas were sampled as follows:

Population                              Sampling Rate
Governmental Unit*
Fewer than 2100         		One in two
2100-31000                      One in four
Census Tract
5200 or more                    One in Eight
Any other area                  One in Six

*Governmental Unit - A government is an organized entity which, in addition 
to having governmental character, has sufficient discretion in the management of its own affairs 
to distinguish it as separate from the administrative structure of any 
other governmental unit. To have governmental character, an entity must 
have existence as an organized entity and responsibility to the public. 
(Source: Decennial Management Division Glossary, 
http://www.census.gov/dmd/www/glossary.html)

Alaska had 124 incorporated cities with populations less than 1000 back in 1997. According to Census Bureau criteria, all of them would have sampled at the one-in-two ratio for the "Long Form."

AVAILABLE TOPICS

100% Data:
Household relationship
Sex
Age
Hispanic or Latino origin
Race
Tenure (Own or Rent)
Vacancy Characteristics

Sample Data:
Population
Ancestry
Disability
Education: Enrollment and Attainment
Grandparents as Caregivers
Income
Industry, Occupation, and Class of Worker
Labor Force Status
Language Spoken at Home
Marital Status
Place of Birth, Citizenship, and Year of Entry
Place of Work and Journey to Work
Residence 5 Years Ago (Migration)
Veteran Status
Work Status Last Year

Housing:
Bedrooms
Farm Residence
House Heating Fuel
Kitchen Facilities
Plumbing Facilities
Rent
Rooms
Selected Monthly Owner Costs
(Shelter Costs):
Utilities and Fuels
Mortgage Costs
Taxes
Insurance
Condominium Fee
Mobile Home Costs
Telephone Service Available
Tenure (Owner/Renter)
Units in Structure
Value
Vehicles Available
Year Moved into Unit
Year Structure Built

I'm giving you ALL of the available topics so you can see what Census 2000 data can and cannot do for your patrons. If someone needs Internet usage by town from the 2000 Census. They're out of luck. It's just not there. You can give them telephone service by Census tract, but not computer ownership or Internet use. Telephones are tracked, computers are not. If they feel they need that level of detail in the future, ask them to write their Member of Congress.

Actually, the Census Bureau *does* do a little tracking of computer ownership and Internet usage in their Current Population Surveys. You can see what's available by going to [http://www.census.gov/population/www/socdemo/computer.html]. However, you will see that the available reports are not broken down geographically, plus the most current data was gathered in 1997.

I used the example of Internet usage to emphasize the importance of checking your patron's request against what's available. What's available is really a two-part question. We answered the first part this week what topics did Census 2000 gather information on? Next week we will search for the answer to part two of that question when will this data be out? While you may have heard "Census 2000" information came out in April, some products won't be coming out till 2003, and many popular items won't be around until the summer of 2002.

See you next week,

Daniel Cornwall, Member Alaska Library Association Government Documents Roundtable.


If you find any part of this tutorial useful, you may use it in training materials with proper attribution, and if you drop me a line at dan_cornwall@eed.state.ak.us telling me how you're going to use it. Thanks!


Census Tutorial Home


This Page was last updated June 27, 2001. (Links checked September, 2010)