"Every 10 years, the United States Census Bureau unleashes a swarm of numbers that affect everything from the political power of the states to the municipal budgets of cities and the clout of ethnic groups. And every 10 years, anyone who knows anything about how these precise-sounding numbers are arrived at knows there is no way they can be absolutely true." ~ Janny Scott
With the many changes taking place in federal information provision, some (many?) of the titles mentioned here may not be available in the future or may be available only in a digital version on a Census Bureau server. URLs are likely to change from time to time.
When lay people think of federal government statistics, they
probably think of the Census. It is the single
most
important
statistical service provided by the Federal Government. The best place
to begin is with the Census
Bureau
website. You will find very good information links, including
educational materials for schools. Census data is easily available,
particularly via the link to American
Factfinder which contains substantial and easy to use
information,
especially about the most recent census. In order to stay current, do subscribe to the Census Product Update available
free via email. It is issued every other week. Each issue includes:
"headline news," "fact of the day," "hot links," and a "hot tip"
"An enumeration of inhabitants" [The actual Enumeration shall
be
made within three years after the first
meeting of the Congress of the United States, and within every
subsequent term of ten years [ Decennial Census], in such manner as
they
shall by law
direct] to constitute Congress is the only
Constitutional requirement. Nothing else is required. The Census was
originally seen as a means to allocate political representation in the
House efficiently and fairly. The ten year insures that the House is
periodically rebalanced to reflect population change.
Congress has traditionally been reluctant to do this because change
threatens incumbents. At first, Congress added new seats as population
increased. When the size of the House became cumbersome, they tended to
ignore the problem or focus on counting people likely to be similar to
those in power and their constituents. Thus, counting rural and small
town populations were favored because they were less threatening to the
status quo. This changed only when the Supreme Court ruled that
malapportioned state legislatures were unconstitutional [377
U.S. 533 Reynolds
v. Sims]
Today, the Census results usually require that Congressional, state, and municipal legislative boundaries be redrawn. With each Census some states gain population and some states lose at least one congressional seat so there is substantial political interest in creating Census procedures that maximize the probability that particular interests will be protected. The Redistricting Resource Center of the National Committee for an Effective Congress provides current, detailed information on redistricting resulting from the 2000 Census. Forty-three states have more than one Congressional district.
Census results are also used to allocate Federal monies to states and cities. Billions of federal dollars are allocated based on Census numbers. States with a population loss will not only loss representation in the House, but will also lose Federal funding. Considerable money is at stake.
Business and industry use Census data for marketing purposes and planning. Government uses it for planning and budgeting. Many individuals and organizations depend on Census data.
The Census should locate everyone in their homes, count them,
determine male and female populations, and the ages of the household
members. Some conservatives would like to return to this simple
counting; believing that the additional questions are an invasion of
privacy.
The first Census process law was signed in 1790 by George
Washington. Federal marshals collected information door-to-door: name
of household and number of occupants. African-Americans were counted as
3/5 of a person. Native Americans were not counted. The process took 18
months and it cost $45,000 to count nearly four million people.
Historical
censuses are increasingly available on the web. Click on the
"publications" link. An
index to
earlier Censuses of Population and Housing is found here. Measuring
America:
the Decennial Censuses from 1790 to 2000 is the standard
historical
survey [about 150 pages]. Ohio State University Libraries provide an index
to decennial census publications through 1997.
The population and housing Census takes place in the year ending in zero. Individual states may take their own Census at other times and often do so because they need more current data for planning or resource allocation.
The first Census was conducted in the same year with these states included:
There was particular interest in number of men of military age. These were the questions:
Much of the 1790 Census data was destroyed by the British in the War of 1812. Until 1840, the Census focused on households. Different federal marshals collected data in their own way. Note that questions and the wording of questions vary from year to year so trend analysis may be a challenge.
In 1810 [third Census], Congress added a question about the
state of
manufacturing
establishments so the beginning of the economic census.
In 1830 [fifth Census], the first printed forms were used.
By 1840 [sixth Census], the number of questions asked had increased to more than 70 as questions on agriculture, mining, and fishing were added. As government activity increases, it needs more information to formulate policy. Commercial and industrial interests also want more information to help in planning and marketing. However, many felt that there were too many questions and a violation of privacy. In 1849, a Census Board was established and a de facto Census Director was included. Department of the Interior was not responsible for the Census.
The 1850 [seventh Census] Census was a major landmark because all persons in the residence were recorded by name. Social questions -- churches, crime, poverty, and taxation -- were added.
In 1860, native Americans ["civilized Indians"] were
separately counted.
In 1868, the ratification of the 14th Amendment eliminated the
3/5
counting rule for African Americans. In 1870 [ninth Census], maps and
charts were used in the reports. Chinese were also counted for the
first
time.
In 1880 [tenth Census], the Census Office was
officially established in the
Department of
Interior. Enumerators replaced the U.S. Marshals as data collectors.
In 1890, punch card machines were used to tabulate data. All
native Americans were counted.
In 1902, the Census Bureau was made a permanent federal agency
in
the Commerce Department. In 1913,
the Census Bureau moved to the Department of Commerce.
In 1910, Congress voted to cap the size of the House so
reapportionment became necessary. At about this time, Congress began to
use Census information to allocate grants to state government.
In 1930, questions about unemployment and income were added.
In 1940, the first sample Census was developed (the long form). It was too costly to ask many questions of all. Thus, there were now two sets of Census data: (1) a few questions (population and housing) for all and (2) many questions (population plus other aspects of family life and housing) for a smaller sample. About five percent of families received the long form. Short form data is now found in Summary Files 1 and 2 while long form data is found in Summary Files 3 and 4.
In 1950, UNIVAC, the first non-defense computer helped to
tabulate
data, creating the
possibility of machine readable data, analysis, and interpretation.
In 1960, census forms were mailed to urban households
[self-enumeration] replacing most visits.
In 1962, the Supreme Court ruled in Reynolds v. Sims that a malproportioned legislature could be constitutionally challenged.This led to the one person, one vote decisions. No longer could a rural minority dominate the suburban and urban majority. The importance of accurate head counts became even more important. In the 1960s, electronic scanning of Census responses became common.
In the 1970s, most Census forms were distributed by mail rather than having the information gathered in face to face interviews. Metropolitan areas, believing that they had been under counted, sued the Census Bureau over alleged under counts. In 1996 the Supreme Court ruled that the decision to adjust the count was left to the discretion of the Secretary of Commerce. The Secretary was unwilling to change the results. The problem is that the Bureau lacked the funding to do a better job.
In the 1980s, TIGER
files were developed to insure that geographical
references were accurate. As the geographic database became available,
accurate maps portraying Census findings became more widely available.
Customized maps could be created with the TIGER databases and map
creation software.
The 1990 Census was troubled by a decline in participation and
under counting problems. The response rate was 65 percent.
The 2000 Census cost about six billion dollars and employed
about
860,000 people. A national advertising campaign was used to encourage
participation.
Census data may be "schedule" or "summary." Schedule is
individual
data, responses to the questions, and is available from NARA.
Seventy-two years must elapse after this date for individual responses
to become public. Corporate data is held for thirty years. The 1930
Decennial Population Census scheduled information was
released in
2002. This data is especially useful for genealogical research. The Family
History Centers
of the LDS Church provide access to schedule data and are
open to
the general public.
Summary consists of aggregated statistical data and
is
available from the Census Bureau. This has been released to the public
since the first census. A good introduction to historical Census data
is U.S. Bureau of the Census's Bureau of
the
Census Catalog of Publications: 1790-1972. Government
Printing
Office; Washington D.C., 1974. (Docs. Ref. C 56.222/2-2:1790-1972). The
InterUniversity Consortium for Political and Social Research [JCPSR]
provides access to selected data from 1790 forward [UT is a member].
The Census must count all of the residents in every state. To
do
this, the Census must deal with legal/administrative units such as
states and counties as well as statistical geographic units such as
tracts or metropolitan areas. You may wish to examine the Geographic
Area
Reference Manual and visit the U.S. Census Bureau
Geography
site. In particular,
Census 2000
Geographic Definitions are useful. Geographic
Terms and Concepts is also helpful.
There is a geographical hierarchy in most census publications:
Another hierarchy would include:
The country might also be divided by the 435 Congressional districts or the 29,469 ZIP code areas. State legislative districts are included for those states providing this information. Ten states did not. However, these boundaries do not reflect post Census redistricting.
All geographic units have a unique code and the several Census variables are linked to that code. To retrieve and interpret data, researchers must use Census maps to determine how particular reporting areas are defined, especially blocks, block groups, and tracts. The U.S. Census Bureau Maps and Cartographic Resources site will answer many questions.
If you know the geographic area of interest, go directly to "data sets" in Factfinder. If not, begin with "reference maps."
These digital files automate the mapping activities needed to support the decennial Census. TIGER was originally developed for the 1990 Census. It required creating and maintaining a digital geographic data base of all the areas covered by the Census including outlying areas. All Census maps are produced from this data base. The ability to assign an individual address or code to particular geographic features is the key. GIS software uses Census data to create county level maps displaying major geographical features and then interacts with other Census CD-ROM products to map demographic data. For example, using TIGER files, you can create a county map illustrating the average educational attainment in all its Census tracts. TIGER file CD-ROMs may be used with most GIS software.
Using these files, individuals with mapping software may print individualized maps showing demographic or other data by place. Maps may show data at the block level. TIGER files have been distributed to depository libraries on CDs. Most states fit on one disk.
Maps are necessary for virtually all uses of small area Census data. Among the most useful maps are:
Landview is the product of a joint venture involving the EPA,
the
National Oceanic and Atmospheric Administration, the U.S. Geological
Survey, the Nuclear Regulatory Commission, the Department of
Transportation, the Federal Emergency Management Agency, and the Census
Bureau. Landview uses MARPLOT (Mapping Application for Response,
Planning, and Local Operational Task) to make maps quickly and easily.
With an emphasis on counties, Landview combines information from
several other data sets with those from the 2000 Census of Population
and Housing. Products include thematic mapping, printed maps and
reports. There is a tutorial on the CD-ROM and more information
may be
found at this Census site .
Landview
6 is in
the public domain and is available in a two CD set. The Landview
database retrieves Census 2000 demographic and housing data, EPA
Envirofacts, and USGS GNIS information. MARPLOT creates maps using this
data.
Enumerators are ordinary people who
know how to
read and write. They have not always been well trained or
conscientious. They should live in the area where they collect data
and should know the community well. They are paid to go door to door to
get the head count and housing information. Often, they have been given
little instruction on how to cover their territory. Under
counting
has
always been a problem. While under counting in urban areas receives
considerable publicity, under counting is also a problem in rural areas
where several families may live on the same property in an isolated
location. Historical images of Census Bureau activity are found in the Census
Bureau
PhotoZone.
Respondents have not always understood the questions and
Census
takers may "interpret" questions to gain a reasonable response. Bias
may result or the enumerator may simply record the response that seems
closest to the response. Today, most enumerators have been replaced by
mailed forms and they are now used for follow ups or used to "catch"
difficult
to count groups like the homeless or illegal migrants.
In 2010, the enumerators will enter data directly into hand
held
computers. They will also use GPS software to capture location
information.
The individual data (data that would identify a particular person or family) in the Census must remain confidential for 72 years. This is done to encourage truthful answers and accurate information. It can be frustrating for genealogists who find this information useful in compiling family histories. The latest individual Census available to the public is from the 1930 Census. The Census Bureau has a long history of successful confidentiality and has traditionally been more trusted than other statistical agencies.
Many current problems result from political groups attempting to change the process to better meet their goals. Typically, the majority party in Congress tries to influence the process so that people likely to vote for them are likely to be counted. Intuitively, politicians are less interested in counting those likely to vote for the other party. This was at the root of the dispute over the 2000 Census. Republicans fought the use of sampling because it is likely to count more of the disadvantaged. Democrats advocated the use of sampling for the same reason. Academics and researchers favor sampling because it is more accurate than the traditional head count. In a 5-4 decision, the Supreme Court said [in Department of Commerce v. United States House, 525 U.S. 326] that a head count must be used to reapportion the seats in Congress, but that sampling could be used for other purposes. Republicans vowed to permit no sampling to allocate Federal funds (about 200 billion $) or to redraw district boundaries.
Apportionment and redistricting is found in Title 13 of the U.S. Code. The apportionment counts must be delivered to the President within nine months of the census date. The President reports to the Clerk of the House who informs each state governor of the number of representatives allowed. The state legislature is responsible for defining legislative district boundaries. Under Article 1, Section 2 of the Constitution, representatives and taxes shall be apportioned to the states according to their numbers. This is the primary purpose of the census. A somewhat complicated "Method of Equal Proportions" is used to assign representatives to states.
Nearly all federal agencies use census data. For example, education information is used by Commerce, Education, Justice, Labor, Health and Human Services, Urban Development, and the National Science Foundation.
Congress, in order to save money and to decrease federal control, is attempting to move more data gathering, analysis, and interpretation responsibility to state and local government or trade associations. Often, these lack the knowledge and the funding to do what is needed.
The historical model for census-taking is based upon small town America where a housewife went door to door gathering information in a neighborhood that she knew well and with respondents who knew her. It does not work well in isolated rural areas or central cities, especially where there are homeless or immigrants. Too, people are more suspicious of government and those who ask questions today. The mail questionnaire that most recipients receive is a different experience and may not yield the same results.
Response rates are declining over time, even with the threat
of
legal penalty. The Bureau is concerned with initiatives that will
maintain or increase the response rate as well as insuring that
responses are valid and reliable. Under counts have resulted in
lawsuits by several metropolitan governments. The 1990 Census was the
first since 1940 where accuracy did not improve. About 16 million
people were miscounted with about 8.5 million missed. Those not counted
were likely to be African-Americans, Hispanics, and recent immigrants.
One of the problems has been a mediocre and sometimes incomplete
address list. For the 2000 Census, the Bureau used the Post Office
address list and Post Office employees to validate addresses and the
response rate improved.
The number of questions (length) is also a problem. The
response
rate declines with length. Short form responses are better than long
form responses. Congress often adds new questions on matters of
interest while reducing Census Bureau resources and encouraging fewer
questions in general.
Insuring that data is comparable from one Census to another is a continuing concern as questions are added, deleted, and rewritten. Trend analysis depends on comparable responses.
The costs of data gathering, analysis, and interpretation continue to increase. The Census has not been adequately funded and is not likely to be in the future. The cost in 1998 constant dollars to count one housing unit continues to increase over time:
The 1990 Census cost 65% more than the 1980 Census. Sampling would be more cost-efficient and also yield better results, but there is considerable political opposition.
As is true of many Federal agencies, the Census Bureau lacks
adequate, current information technology. For example, in 1990, Census
staff typed into computers all the answers from all the questionnaires
because Congress would not provide the money for optical scanning
devices.
1. This is the first census to rely on web-based reporting for most data sharing. American Fact Finder was created as THE portal for census data. Data is presented in an easy to retrieve and use format. SF1 - SF4 datasets are included. Users may create individualized tables and maps. Factfinder features include:
Economic data is found in:
Tutorials are available if you click on "help" in the
menu bar. Look for tutorials in the pop-up box. Grace York at Michigan
has prepared an excellent AFF
tutorial. When you go to the AFF page, click on data sets in
the
left column. Then check "data on American Factfinder" near the button
center of the left hand column on the AFF home page. The AFF
Glossary will tell you how data on a particular subject is
described in Census products.
Since analysis and data products continue, you will want to
keep up
via news
releases
and the Census
Product
Update Service.
The 100 percentage data questions were:
While most Census 2000 products have been released, a few are
forthcoming.
The 22nd Census of Population and Housing includes information from the short form (100 percent) and the long form (sample of the population). The short form asked questions about:
The long form added these questions on population variables:
The long form added these questions on housing variables:
This is what most people think of when they think of the Census. The Census is constitutionally required, but the number of questions has grown substantially over the years. The Census has been taken every 10 years since 1790. The intention is to count the number of inhabitants and the characteristics of the population as a whole. Less data is available for places of less than 10,000 people. Some questions are answered by the population and some by a sample.
This is part of the decennial Census and there is also a
national
housing survey taken every two years. Includes information on the
housing inventory, financing, and property characteristics.
A series of reports, usually annual, on various demographic topics based upon sample data. These are used to update the decennial censuses, but typically only for the nation rather than states or counties or other places. In the future, these will be available only in digital format. An excellent place to begin is Sherry Engle's "Current Population Reports and Other Population Reports Index" which includes an excellent index and links to PDF files. Series:
Population and Household Economic Topics at the Census Bureau cover about 40 topical areas with information from the census and CPR.
One CD-ROM holds the contents of 3-4 high density computer tapes or about 1500 floppy disks. Beginning with the 1990 Census, data previously available only on computer tape became available in CD-ROM. CD-ROMs made analysis easier than ever since there was no longer the need to mount tapes on main frames. Census data was no longer restricted to a small number of researchers. Content was more widely available in depository libraries since few depository libraries selected the tapes. Nearly all libraries selected some of the CDs.
Usually, the DVD versions contain more data than the tape ones. There was a dramatic increase in the use of Census data by state and local planning agencies when the CD format became available. Originally formatted for use by dBASE III software on IBM PC. Many libraries were ill-equipped to deal with data set use because of hardware and software problems as well as lack of computer and statistical knowledge.Web availability has replaced most of the digital disk sets. The CDs have been replaced by DVDs for the larger data sets. The DVDs are easier and quicker to use for more complex queries and downloads.
2.1 Public Law 94-171 Population Counts These are preliminary Census counts designed for use in legislative redistricting. They include population, racial composition, age and housing units. Data is provided for each state, county, county subdivision, city, village, Census tract, block group, city block, 1990 election ward and 1990 election precinct.
2.2 Equal Employment Opportunity Special File
In 1960, the Census Bureau distributed its first magnetic tape
of
statistical data. This allowed data to be maintained at the Census
tract level. By 1970, more extensive subject summary tabulations and
even greater geographic detail were available because tape storage
density increased. Some comprehensive data sets were available only
on tape which could be purchased from the Bureau or borrowed from the
Inter-University Consortium for Political and Social Research or ICPSR,
a data archive at Michigan [UT a member].
Summary tape files provide statistics with greater subject detail than the other reports. They also include statistics for some types of areas such as blocks and block groups not included in other reports. Four summary tape files are or will be available. The University of Michigan govdocs site provides detailed instructions for access and use. These are available via the Internet and CD-ROM.
Nearly all of the 100% data is now available in CD-ROM. All the SF data tape files will be available from ICPSR [UT is a member of ICPSR].
SSTFs are the source for the subject reports and provide
greater
subject detail than the STFs. Data is normally available only for
relatively large areas. An example might be the foreign born population
of the U.S.
Contain data from samples of long-form housing unit questions
for
large geographic areas with identifying information removed. These
allow customized cross-tabulations of most items on the Census
questionnaire. There is a 5 % sample file and a 1% sample file. Data is
limited to areas of at least 100,000 persons. [not yet released]
Details
about Census 2000
may be found here. Because of the cost of the Census process and
problems with the response rate, the Bureau has adopted a new model. It
will be interesting to see if it is politically feasible.
The Survey will be more economical, yield a better return and
provide
federal agencies and others with much more current information. The
Survey is part of the decennial Census so response is legally required.
So far, Congressional funding, especially from the Senate, is
inadequate to support ACS. If this continues, which is likely, ACS will
be abandoned, and the 2010 Census will once again include the long
form.
As true of the 2000 Census, relatively few print reports will be issued and distributed. Data bases allow users to create individualized reports or answer particular queries. Many data files will be issued on purchased DVDs. The few printed reports will be issued later by the GPO. The various divisions of the Bureau issue working papers on a variety of subjects each year. For example, the Population Division issued Historical Census Statistics on the Foreign-Born Population of the United States: 1850 - 2000 in February, 2006.
In answering census reference questions, you need to know:
The Basic Facts feature allows you to locate and manipulate date from pre-made tables and maps for frequently asked questions. It's easy to compare information for similar areas such as city. Keyword search shows what source types are available for a topic by searching all data files. Geography search lists all the geography files for counties, places and cities, but does NOT include tracts or blocks. Click on "map it' for a map of the area selected. Street address search lists all 2000 Census files including that address. Select an area, and "map it" to see what reference and thematic maps are available. Note that geographic areas are arranged from the largest to the smallest.
Data Sets provide access to all available tables for the 1990 and 2000 Censuses, American Community Surveys from 1996, the 1997 Economic Census, and Population Estimates from 2000. This is the most comprehensive entry point. Summary File 1 includes complete geographic detail to the block level. Summary file 2 includes subjects for up to 249 racial or ethnic groups. Summary file 3 includes social, economic and housing data to the block group. Within "data sets," you have access to detailed tables, geographic comparison tables, quick tables, thematic maps, reference maps, and custom maps. Tables may be saved as tab or comma delimited for manipulation in Excel.
Factfinder creates reference maps and thematic maps. Reference maps are useful for "where" questions since they display boundaries. They provide state, county, city, Census tract, block groups, block, school district, ZIP code, and Congressional district boundaries. You may customize your map in various ways using the legend. For example, you may decide to eliminate all boundaries but one or two. Thematic maps display statistical data by place. They allow comparison of topics or variables by place using colors or patterns. Thematic maps answer how many or what percent questions. Before you use the map features, review the "Creating and Using Maps" tutorial found in the FactFinder help section. Then select "Maps and Geography" on the home page. Note the importance of the "zoom" feature for viewing smaller areas such as blocks.
Population Finder provides easy access
to population data and trends. Includes data from the 1990 and 2000
Censuses as well as population estimates since then.
The various Census data collection efforts result in an immense amount of data. The Census Bureau is the most prolific data publisher in the US. It also gathers, analyzes, and interprets some foreign and international data.
Many users are not comfortable with data and statistics. The
paper
census products often required librarian help in use and understanding.
While the Census Bureau websites are typically well designed, the lack
of mediation may create problems for some users. However, availability
on the web is likely to substantially increase the number of people who
use census information.
Census data influences much national and local decision-making.
Operated by the Church of Jesus Christ Latter Day Saints
(Mormons),
the Centers for genealogical research are usually open to people of all
faiths. Among other resources, they normally provide access to a
complete selection of Census records. All are at the Salt Lake City
location, but these may be borrowed for a reasonable fee from the local
Family History Center. Some are available via the web. There is a
Family History Center in West Knoxville and it is well worth a visit.
Regional centers (one in Atlanta) provide access to a complete selection of Census records. Reminder that the 1930 Census is the most recent one available at this time because of privacy concerns.
The
Tennessee Demographic Profile is a 503 page PDF file.
Includes
information on age, sex, race and on housing, households, and families.
There is a separate table for each Tennessee county and city.
The best first source is at the University of Michigan
Document
Center
Site. It is comprehensive with helpful annotations and good navigation
to a wide variety of useful sites. The place for you to begin is Grace
York's presentation on the American
Factfinder and Census 2000. The exercises are helpful for the
serious student.
Census
Watch is a good
source of information for current information on the census.
Consortium for International Earth Science Information Network or CIESIN
The Geospatial and Statistical Data Center, University of Virginia
Quick Links to Census 2000 by the University of California at San Diego Social Sciences Data Collection.
Sources of
Zip Code Data
is the best place to go if looking for statistical data by Zip code. A
large number of useful sources are listed. An unusually useful site.
The USGenWEb Census Project uses volunteers to enter original census data from the 1790 - 1920 census into a free database. This is a noble project, but one wonders about the quality control.
The United States Historical Census Data Browser at the University of Virginia includes state and country data from 1790 through 1970.
Census Basics (C3.238). An illustrated report containing basic reference material and assistance sources.
Census ABC's--Applications in Business and Community
Census Catalog and Guide (C 3.1.163/3: (1985-).
Census Product Update found here and available via email.
200 Years of U.S. Census Taking: Population and Housing Questions, 1790-1990. (HA 37 .U52 B68 1989) Includes copies of all questionnaires from 1790 - 1990. If you know which questions were asked, you will know what data will be available. The questions [through 2000] are available from the IPUMS website. IPUMS is the Integrated Public Use Microdata Series (IPUMS-98), created at the University of Minnesota in October 1997.
This is probably the best place to begin any search for Census
statistics. All Census materials are indexed and abstracted as well as
statistical data created by other executive, legislative, and judicial
sources. Both print and digital sources are covered. ASI appears in two
parts: (1) an index and (2)abstracts. Access points include subject,
issuing agency, type of data breakdown, title, and report number. An
accession number leads the user from the index to the abstract. ASI is
available in paper, on CD-ROM and in LexisNexis Statistical. Cumulative
indexes cover 1974
- 1996. ASI also provides a fiche collection of all the statistical
data found in the index. Fiche are issued monthly. A good user guide
[pdf] is available on the website. ASI is included in LexisNexis
Statistical [available at UT].
Print reports are the most convenient and easily used source
of
data for most users. However, most of these have been discontinued.
Large PDF files are difficult to read on the web, but are available
24/7. More sophisticated users
appreciate digital files that allow manipulation and analysis as well
as downloading into spread sheets.
Census Statistical Briefs presented short
summaries on
various
demographic topics from the P-Series of Current Population
Reports. Topics includes child support, elderly, foreign born, ethnic
populations among others. This title was changed into Census Briefs
which are not
found here under subject.
This is the standard one volume collection of most used/useful U.S. statistics, including many based upon the various censuses. The first volume was issued in 1878. It does include some statistics on foreign countries and non-governmental topics. While most data comes from federal government sources, private and some unpublished sources are also used. Most data is on the national level. A typical volume will include more than 1400 tables plus charts and graphs.
In working with this source, it is important to encourage users to verify the units used for data and to use the footnotes to identify the source of the data. Often, the source document can provide more useful information. Thus, the Abstract serves as an indexing tool to locate more detailed data or time-series data.
The various sections of the Abstract
are available in PDF format on the web which is sometimes clumsy to
use. Because
of copyright problems, the web version does not include all of the
non-government content found in the print version. Still, the most
frequently requested tables, state rankings, and profiles are here.
Earlier
editions are available as PDF files. Not all editions are
available
now.
This is the historical supplement to the Statistical Abstract. It covers more than 12,500 series of statistics for population, migration, prices, national income, social statistics, health and medical care, environment, communications, and international trade. Historical Statistics is especially good for time series. The pre-1970 volumes of the Statistical Abstract have additional and MORE detailed historical data tables than those that appear here. Besides the print edition, Cambridge University Press has provided a CD-ROM edition which is available in many academic libraries. this version includes Lotus 1-2-1 files and tab-delimited text files for all tables which are Downloadable. The most recent volume of this title is privately produced and is no longer a depository item.
There is a commercial version which combines data from Historical Statistics and the Statistical Abstract: Kurian, George T. Datapedia of the United States 1790-2003: America Year by Year. 3d ed. Lanham, Md.: Bernan, 2004. Kurian has repackaged data on "several" hundred indicators of social, economic, political, and cultural development under twenty-two subject areas. Graphs supplement data tables. Projections go through 2050.
The United States Historical Census Data Browser describes the people and the economy of the U.S. for each state and county from 1790 to 1960. The site is easily used and provides a wealth of data.
The 2000 Census Atlas was to be produced with a commercial partner, but I have not found it. Mapping Census 2000: The Geography of U.S. Diversity shows a variety of interesting data. Some Census map products may be found via the Geography Division Map Gallery.
The Data Book is issued every 5 years as a supplement to the Statistical Abstracts. It covers about 11,000 cities, all 3,141 counties, and about 11,000 communities with a population of 2500 or more. Some information is available on states and regions. Data provided includes:
This compilation is good for cross-sectional statistics.
The Commercial version is City and County Extra: Annual Metro, City, and County Data Book by Bernan Press (Ref HA 203 .C68). It is also an annual and contains the much of the same information for states, counties, metropolitan areas, and cities with populations above 25,000. Subjects include population, housing, education, income and poverty, crime, manufacturing, trade, services and government. Maps, rankings, and notes complement data tables. Clear presentation. Some of the data is more current than in the original.
Places, Towns, and Townships by Bernan Press is another value added compilation with demographic, social and economic data. This is especially useful for data on smaller communities (populations less than 25,000).
This CD-ROM includes 5,000 data items on 3,142 counties from
the County
and City Data Book and from the Metropolitan Area
Data Book.
The emphasis is on time-series data, but this work is also good for
comparative data. USA Counties provides deeper
analysis and
more variables that the City and County Data Book.
Search
results may be saved as plain text, HTML, or database files. Findings
may be represented
graphically using the integrated ESRI MapObjects LT mapping
and GIS software. Intellectual access is somewhat primitive with menu
driven software, no index, and no boolean searching. The web version
has more limited options.
This Data Book is issued every 4 years. A Metropolitan area is the same as a Metropolitan Statistical Area (MSA). Data is included on a wide variety of social and economic topics for these areas, including central cities and regions. Data comes from the Census and from vital statistics. This is a good source of cross-sectional data.
Designed to meet the needs of individuals and organizations with specific needs to find Census data not available in published form. Decreasing interest as more data is available in the CD-ROM format.
The SDC is the lead agency in each state for Census information, usually in a state government agency. The Center receives free data tapes (or CD-ROMs) for their area and sometimes surrounding states. Tapes may be copied for use and data extracts may be created, often in cooperation with other state organizations. The Center is also supposed to provide feedback on how to better format data and prepare data for use. It should lead in local review of Census content and designation of geographic areas.
In Tennessee, there are two SDCs: (1) the Tennessee Department of Economic and Community Planning in Nashville and the Center for Business and Economic Research, College of Business Administration, UT. The latter appears to offer more services.
This is similar to the SDC program, but is designed to reach the business/industry audience. Intended to stimulate the use of Census information. Less than half the states are involved. Some question as to effectiveness.
Similar to the BIDC, but focused on
smaller nonprofit organizations, especially those involved with social
programs. Typically, these organizations have not used Census data
because of lack of skill and knowledge. Vanderbilt
University Library hosts a CIC.
This is the Census Bureau¹s on-line information service, 1984-, which is available via Compuserv and Dialog.
Done every 5 years in years ending in 2 and 7. Traditionally, these have been supplemented by monthly, quarterly and annual surveys. Budget pressures have caused several of these supplementary products to be reduced or eliminated.
Most of these censuses provide detailed information from three
perspectives: (1) geographic area series provide detailed information
for the U.S., states, and sub state areas (ZIP code statistics for many
sectors); (2) industry series provide mostly national information for
individual industries/products; and (3) the subject series provides
national and some state and metro area information on special subjects.
The most detailed information is at the national level. ZIP codes have
the least information. Because finance, insurance, and real estate and
some transportation, communication, and utilities components were not
included in earlier years, time series data can be limited.
The 2002 Economic Census was conducted in April and data
products were released from 2004 through 2006. The Overview
of the
Economic Statistical Program is a helpful
introduction/overview.
Besides the data on each five year Census, monthly, quarterly, and
annual reports are issued on particular topics.
Major statistics included are:
With the Agriculture and Government Censuses, the Economic censuses account for 98% of all U.S. economic activity. Data gathered here is used to create gross domestic product, input/output measures, production and price indexes. The data gathered is heavily used by state and federal governments as well as trade associations and individual firms. Private firms and not-for-profit organizations use this data for a variety of planning and evaluative purposes, including competitive intelligence. Note that the data is only as good as what is provided by the firms.
Began with the 1810 decennial Census which included questions on manufacturing. In 1840, minerals data was added. After the permanent Census Bureau was established in 1902, a Census of Manufactures was to be taken every 5 years beginning with 1905. In 1930, the first Business Census was taken for 1929. Service industries were added in 1933. The 1954 Census was the first to be fully integrated re: consistent time periods, concepts, definitions, classifications, and reporting units. Transportation was included in 1963 and construction in 1967. Minority owned business was added in 1972 and women-owned business in 1977. Communications and utilities were added to transportation in 1992. In the same year, the Census of Finance, Insurance and Real Estate was added.
The Census Bureau has said that with the 1997 Economic Census reports, the different sectors will no longer be treated separately so that the Census of Retail Trade and its supplementary material will simply be part of the Economic Census.
A variety of economic surveys, usually at the national level, are undertaken on an annual, quarterly, and monthly basis by several agencies and departments. The Census takes place every five years, including much more geographic and industry/trade particulars. The Economic Census does not include geography below the state level.
The North American Industry Classification System (NAICS) replaced the Standard Industrial Classification (SIC) in economic censuses and all the publications associated with them. The best web site on NAICS is at the Georgetown University Library. The NAICS, developed with Mexico and Canada, classifies economic activity using a scheme of two to five digits with the U.S. version including some six digit activities. Since industries, business, and services are classed this scheme, you will need to become familiar with it. With retrospective data classed according to SIC, you will also need to retain familiarity with that classification scheme. The Census Bureau has created tables making it easier to move from one classification scheme to another.
The SIC was developed in the 1930s, was revised every ten years or so, and focused on manufacturing or goods-producing. NAICS includes manufacturing, but also includes a substantial focus on services or service producing. An "establishment" is a facility at a single physical location. A "company" consists of one or more establishments under common ownership or control. NAICS includes a definition for each industry, trade, or service, an alphabetic index, and correspondence tables in each Manual.
Bernan Press publishes several reference works that focus on
Federal
government information. Besides original works, they are known for
their value-added editions of reference books based on Census
information, such as Education Statistics of the United States,
the County and City Extra: Annual Metro, City, and County Data
Book:
1998 or Housing Statistics of the United States.
They also
have continued to update and publish reference works based on Census
information that have been discontinued by the GPO. As the Census
Bureau moves away from hard copy, Bernan is likely to publish more
titles that libraries have traditionally received from GPO, such as the
Handbook of North American Industry.
This website receives more than 50,000 inquiries per day. Increasingly, data and reports appear here and not in print form. For users who have web access, this substantially reduces the time lag between collection and public release. For the Bureau, it reduces printing and distribution costs.
When a dataset is placed on the Internet, summary fact sheets about that data set are released in digital form and sometimes in print.
There is no persuasive archival plan for publications and data sets removed from the server. It has been suggested that CD-ROM will be used for archival purposes and to publish large collections of similar data. However, we can't read some computer tapes from 20 years ago, whereas we can still read books from 20 years ago.
Not all users will be comfortable in finding and using Census data on the web. Dealing with complex tables may increase the pressure on the state data centers to provide guidance.
