Special Issues
of Interest to Users of Census Data
Special Issues
Census
Geography
[Go to More information on
geography]
It is important to understand the geography of the Census because different items of information are available for different units of geography. Simply knowing the "normal" geography of an area isn't always sufficient since the Census Bureau defines some geographical areas using different boundaries and different terminology. The Census Bureau does attempt to make its geographical areas conform to what an area is really like, however, and they use local input whenever possible.
Terminology is perhaps the biggest stumbling block to understanding Census geography. Here are the basic Census geographical terms and their definitions, moving from the smallest area to the largest:
| Census Block (or, simply, Block). Blocks, in urban
areas especially, conform to what is usually meant by a city block, i.e.,
a small area bounded on all sides by streets. Other types of boundary lines
may be used when necessary, such as railroad tracks, streams, city and county
lines, etc. In rural or remote areas, however, Blocks may be large (many
square miles) in area and irregular in shape. But for Census data reporting
purposes, one kind of block is the same as another. Blocks vary widely in
population, and there are 10 million Blocks nationwide. Each one is given
a four-digit identification number within its Census Tract (see at right).
Block Group. As its name would imply, this is a cluster of Census Blocks, all of which share the same first digit of their four-digit identification number. Block Groups ideally contain about 1500 people, with ranges generally between 600 and 3000. |
Census Tract. Tracts, larger than Block
Groups but still relatively small areas, are, ideally, fairly stable county
subdivisions. They are usually delineated with local input. The Census Bureau
likes for Census Tracts to have populations between 2500 and 8000, and as
a consequence they vary considerably in size. The 2000 Census is the first
one for which the entire country has been divided into Census Tracts; previously,
Tract data was not available, as such, for mostly rural areas.
Census Designated Place (CDP). Settled concentrations of population that are identifiable by name but not legally incorporated are given this designation, usually named in consultation with the residents. This is an important definition in highly rural states or areas. |
|
| Zip Code Tabulation Area (ZCTA). These areas approximate
Postal Service 3- and 5-digit Zip Code areas, with some adjustments so as
not to truncate Census Blocks. In rural areas, Zip Codes dedicated entirely
to post office boxes are included. This is a new geographical category
for the Census Bureau, acknowledging that Zip Codes are more easily recognizable
as geographical divisions than are some of the Census Bureau's areas, and
that many researchers and businesses are used to looking for data by Zip
Code. Census.gov has a page
of information and links to technical data on ZCTAs.
|
Metropolitan Area (MA). The Office of Management and
Budget defines Metropolitan Areas for the purposes of collecting and reporting
statistical data. An Area will include a large city or cities, typically,
with their surrounding suburbs. The key point is that the suburbs are closely
integrated, socially and economically, with the central city. Terminology
has varied over time on this topic, with the Areas referred to as Standard
Metropolitan Areas (SMA) from 1949-1959, Standard Metropolitan Statistical
Areas (SMSA) from 1959-1982, Metropolitan Statistical Areas (MSA) from 1983-1989,
and, currently, simply Metropolitan Areas. The alphabet soup gets thicker
with the following sub-divisions of MAs (consult
"About
Metropolitan Areas" for more details and explanation):
States/Counties. These geographical areas are not modified by the Census Bureau, but retain their normal definitions and boundaries.
|
|
Generally speaking, the smaller the geographical area, the fewer items of data are available. The smallest areas for which the important income, education and other social data items are available, is the Census Tract. At the Block level, the smallest geographical area for which the Census reports data, the items available are largely those on the 100% form (see below): age, sex, and race. One of the reasons for this is that the very large number of Blocks would increase the cost of Census data dissemination tremendously were other data to be reported at this level. Another concern, however, is confidentiality. A fundamental Census principle is that they will not display any data that can be used to identify a particular person, household or other basic counting unit. There might very well be so few people in an income class, for example, in a certain block, that reporting the numbers would actually give information on a particular household.
| There are a great many sources of more detailed, expert
information on Census geography, at Census.gov. Here are some recommended
sites:
Geographic Areas Reference Manual. This massive work covers all aspects of Census geography in detail, and explains the criteria and procedures the Census Bureau uses to define and modify the geographical concepts. Reference Resources for Understanding Census Bureau Geography. This extensive set of links to Census.gov sources covers Census geographical heirarchies, regions and divisions, codes, cartographic resources, and Census reference files. Geographic Terms and Concepts. Detailed definitions of an exhaustive list of Census terms. Coming from Census.gov, these are the official definitions. Geographic Changes for Census 2000+ Glossary. This Census.gov page combines the definitions listed in the "Terms and Concepts" page with a series of questions/answers on changes from previous Censuses. Geography FAQ (Frequently Asked Questions). Includes information on maps, coding, and mapping software. A FAQ specifically focusing on the TIGER mapping system is also available. GODORT's Government Information Technology Committee (GITCO) has a section of their Census 2000 Toolkit which deals with Census Geography. The Toolkit includes links to Census.gov and other sites with tutorials, reference information and technical tips. |
Comparability
between Censuses: the importance of definitions and methodology
The two main reasons for difficulty in comparing statistics from one Census to another are:
When you keep in mind that all Census data is derived from
the questions asked of the population during the Census-taking process, it
is obvious that the inclusion or exclusion of a particular question will
have a great effect on the data available for that particular year. For example,
the 2000 Census is the first one which asked a question about grandparents
as primary caretakers of grandchildren. Therefore, this 2000 data cannot
be compared with data from earlier years, at least not with Census data.
(Other organizations may have information on this subject from surveys or
other means.) Conversely, the 2000 Census did not ask about the source of
a household's water, a question which had appeared on previous Census
questionnaires. Therefore data on this topic from a previous Census cannot
be compared with 2000 data, since there is no 2000 data (from the Census,
at least) on this topic. Which questions are asked (or not) is largely a
matter of Congressional mandate or executive agency requirements, making
the seemingly arcane and pedestrian subject of Census methodology subject
to all the whims of political maneuvering and varying public opinion. If
you are interested in the reason behind the questions asked in the 2000 Census,
consult this page:
 Uses for Questions on the
Census 2000 Forms.
To complicate matters further, other factors than simply the presence or absence of a question can affect comparability. For example, questions may be switched from the 100% section of the Census questionnaire to the sample section, and vice versa. You will recall that some basic questions (age, race, sex) are asked of all residents, while approximately one-sixth of the households in the country are asked a set of much more detailed questions (income, education level, etc.). The Census Bureau will report data for all of these items, of course, but in comparing one set of data to another, users should be sensitive to whether the data came from a 100% or a sample question. |
In 2000 the question on marital status (as well
as some others on housing topics) which had been on the 100% questionnaire
for many Censuses, was dropped to the sample questionnaire. For this reason,
comparisons of marital status data from 1990 and 2000, for example, must
be made carefully. The numbers can be compared, certainly, especially for
some applications not requiring great precision, but note should be made
that the sources of the data are different for the two years. (For more
information the problems with shifting questions, see
parts one
and five
of Daniel Cornwall's tutorial.)
Differences in the questions themselves are not the only comparability problem stemming from the actual Census questionnaires. The answers may also be different from Census to Census. Some Census questions leave space for open-ended answers, but for most of them, respondents must choose one (or more) answers from several options. Since the responses are constrained in this way, the data will be similarly constrained, and these constraints might have differed over time. For example, the answers to a question about disability were greatly expanded in the 2000 Census, in order to gather more detailed information about the nature of the disabled population. The total numbers of disabled people as reported in 2000 should be comparable to total numbers from previous Censuses. But any details on the nature or extent of disabilities will not be. (For more details on this issue, see part two of Daniel Cornwall's tutorial.)
|
| Comparing Statistics on
Race
The very complicated issue of comparability of racial data also arises from differences in the way Census questions are answered. And since this area is such an important one, I'll go into some detail discussing it. Major changes were instituted for the 2000 Census in the collection of racial data; therefore the most pressing problems in comparability occur when 2000 data is compared with that from earlier years. But even when the 2000 Census is not involved, modifications in definitions and questionnaire structure from Census to Census may cause problems with other years' data as well. Prior to 2000, there were complaints from many different sectors of the population that the Census Bureau's collection of racial data was too restrictive, since respondents had to choose one and only one racial category for their answer. The "Tiger Woods'" in the population, those who consider themselves to be multi-racial, if you will, were forced to choose one or the other, but not both. With the 2000 Census, respondents were able to mark more than one category from a list of major racial categories: American Indian, Asian, Black, Native Hawaiian, White, and other. Given the possible permutations of this list, there end up being 63 different racial categories for which data is being reported. Since in previous years, there were only 6 (or however many categories were offered as answers to the question) it is obvious that racial data from 2000 gives a much better picture of the diversity of the country. But the data from 2000 cannot be directly compared with that from any earlier Census. Another major departure for Census 2000 lies in the categories offered as choices for Asian- and Polynesian-Americans. Prior to 2000, there was only an "Asian and Pacific Islanders" category. In 2000 these major categories were broken into two. Looking at the statistics for sub-categories of these groups (for example: Chinese, Japanese, Filipino, Hawaiian, Samoan, etc.) might be a way to find comparisons with earlier data. For more information on how 2000 Census racial data can be interpreted, as well as more detail on comparability problems, see these Census.gov web pages: Parts three and four of Daniel Cornwall's Census tutorial also have more explanations and a detailed case study of data interpretation (in part four). |
Comparing Statistics on
Hispanics
The biggest problem in finding and using data on the Hispanic population comes from confusion between the concepts of race and ethnicity. Quoting from Census.gov's Overview of Race and Hispanic Origin, "The federal government considers race and Hispanic origin to be two separate and distinct concepts." Hispanics can be of any race. Unfortunately, reporters and others in the media don't always understand this distinction, and one often sees new items comparing Hispanics with African Americans, for example, adding to the confusion surrounding this issue. In addition to this fundamental confusion, however, Census collection of data on Hispanics has varied so much over the last 60 years that comparisions are practically impossible between any two Census' data on the topic. Data has ranged from "mother tongue" to a comparison of names with a list of Spanish surnames, to "Spanish origin" to "Spanish ancestry," etc. Census.gov's The Hispanic Population gives a summary of these changes, and their Working Paper number 56 (titled "Historical Census Statistics on Population Totals By Race, 1790 to 1990, and By Hispanic Origin, 1970 to 1990" and available in HTML and PDF versions) goes into much more detail. It is important to note that 1970 was the first Census to ask specifically about Hispanic origin and in that Census it was asked only on the sample quesionnaires. One final, somewhat arcane, point may affect comparability of data on Hispanics (as well as that of the racial categories) between the 2000 and, at least, the 1990 Census. In 2000, the Hispanic origin question was asked before the race question, whereas in 1990 it followed the race question. The order in which the questions are read might affect how they are perceived by respondents and, therefore, how people decide to answer them. Since the Census Bureau lets each respondent determine for him-or herself how to answer these questions, respondent perceptions (and their understanding of what is being asked) are important. For more detailed information on the problems and issues surrounding data on Hispanics, see the following:
. |
Comparability
problems due to geographical changes
The problems that arise for comparability due to changing geographical boundaries aren't as intellectually challenging as those due to methodology as discussed above, but they can be very important nonetheless. A general discussion of the broad kinds of changes (new geographical entities, etc.) made for Census 2000, in addition to a geographical glossary, can be found at this Census.gov page: "Geographic Changes for Census 2000".
As you can imagine, in the ten years between Censuses, new cities will attain Metropolitan Area status (although the changing definitions applied to urban areas account for some of the differences, too; see "Historical Metropolitan Area Definitions" for links to definitions back to 1950). And the boundaries of MAs will shift considerably over time as well, making comparability of data for some applications difficult.
Census Tracts are also variable over time, since shifts of population concentrations may necessitate new boundaries. The Census Bureau has produced relationship files (for previous Censuses they are called comparability files) showing how 1990 and 2000 Census Tracts relate. Description of the files, which are in fixed length ASCII format, along with links for downloading and technical documentation, can be found at "Census Tract Relationship Files." Even such supposedly stable geographic entities as counties can be changed over time. See "Significant Changes to Counties and County Equivalent Entities: 1970-present" for detailed lists of changes.
A New Problem/Opportunity: The American Community Survey (ACS)
Beginning with the 2010 Census, however, some of these problems with comparability will not be issues, because there will be NO long form questions asked. This data will still be available, through the new American Community Survey. A few years ago the Census Bureau began the survey in a limited fashion, but in 2005 it is scheduled for full implementation. Through it, all areas of the U.S. will be surveyed for data previously included on the decennial Census long questionnaires (income, education, etc.) on a continuing basis. This new system is designed to 1) make such data available in a more timely manner; and 2) streamline the decennial Census, making it, again, simply a counting of persons.
Regarding the timeliness issue, in Census cycles past, by 2005 the data reported from the long form questionnaires would be 5 years out of date. (And before it would be taken again, it would be 10 years out of date.) Now, however, with the ACS, there is data for San Antonio and other Texas cities, with the Census Bureau imprimatur and freely available to all, for 2003. The categories of data available for each locality (states, cities and counties) are General Demogrphic, Social Characteristics, Economic Characteristics, and Housing Characteristics, plus a handy "Narrative Profile" that includes graphs.
This more timely gathering and release of data is certainly a benefit to researchers and policy planners. The problems will arise in the area of comparability, since the figures making up ACS are gathered from samples much smaller than the previous long questionnaire data gathered during the decennial Censuses.
The ACS page offers a great deal of information on the survey methodology, sampling procedures, etc., that should be carefully studied by users of the data. Look in particular at the "Survey Basics" section and also "Using the Data."
Other
guides to Census information
| Census.gov offers a great many resources
to help users get the most out of their pages. Some of the major sites are
listed below. In addition to these general resources, the Census Bureau
has a collection of web-based tutorials and training materials that can be
found on their
"Census Bureau
Training" or
"Customer
Information" pages. Several libraries and other organizations have also
compiled helpful guides, tutorials and link collections; some major resources
of this type are listed below. For more resources, see our
Web Subject Index under
"Census/Demographics."
CensusScope. This site, provided by the Social Science Data Analysis Network (SSDAN) at the Univ. of Michigan, offers many maps, charts and other analytical products, which are developed using Census 2000 data but are not available at Census.gov. FactFinder Help at Census.gov. From any of the pages of the American Factfinder, click on "Help" at the upper right and a Javascript window will open containing extensive instructions and explanations on using this highly interactive service. The initial help window is context-sensitive; that is, it will display information about the main screen you are currently viewing. But there is a "Table of Contents" tab that will give you access to all the help topics from any point. In addition, Census.gov offers a collection of AFF Tutorials with step-by-step instructions in many detailed areas. This American FactFinder Guide from the University of Wisconsin has step-by-step instructions answering a series of "how can I...." questions.
Measuring America:
The Decennial Census from 1790 to 2000.
History and
Organization.
GODORT Frequently Used Sites: Census. Listing of important sites with brief annotations. Another resource from GODORT is their Government Information Technology Committee's (GITCO) Census 2000 Toolbox. It contains tips on working with different file formats, links to historical and other information, tutorials and reference information. Grace York's American Factfinder Tutorial. This tutorial, developed by a librarian at the University of Michigan, is available as web pages or in a PowerPoint version and gives step-by-step instructions for finding information in American FactFinder. Note that some of the local resources mentioned in this tutorial are not available at the Blume Library. Public Use Presentation Library. This collection of PowerPoint presentations are offered by Census.gov. They are designed to acquaint users with Census products and services and can be used and copied freely. SUNY Buffalo: Census Information. Category listing with annotations. University of Colorado Demographics and Statistics page. Has good annotations of Census.gov pages and other resources. University of North Carolina at Greensboro Data from the Census Bureau. Has a particularly good list of business and economics links. |
Back to Census Guide Table of Contents
