| 1990 Census: An Overview of Resources at Columbia University |
|
Over time, the government has played a more active role in the lives of its citizens; each increase in government activity has led to an increase in the amount of information it has needed to formulate policy. To gather more information, it has been necessary to draw a sample of the population, as it would be too expensive to poll every single person. With the use of sampling techniques, analysts have been able to study in great detail a smaller portion of the population from which they can infer characteristics of the population in general. At the same time, the need for an exact count of the population has remained. The result has been that there are two sets of Census data--a broad look at the whole population on a narrow range of questions and a closer look at a sample of the population on a deep range of questions. As the government collects more information from people, two complicating issues have arisen: the need to protect the confidentiality of respondents and the infeasibility of every single individual providing the depth of information needed to set policy.
For purposes of counting the population, the Census Bureau polls 100% of the households in the United States with a brief questionnaire asking a limited number of questions about every household member and a number of questions about the household itself. For learning about the lives of the people, the Census Bureau sends a longer questionnaire to two samples of the households (a 5% sample and a second 1% sample of all households), which asks very detailed questions about the members of the household and of the household itself.
Since the original purpose of the census was to guarantee proportionate representation to the citizens of the United States, the census must initially count all of the residents of every state. The territory of the United States is organized hierarchically for the purpose of subdividing the population in order to obtain the most accurate count of all people in their place of residence. The basic organization is like this:
National (United States)
This scheme does not show all of the various ways in which census geography is presented in its publications and other data products. The United States can be divided in three manners: regions, divisions, or the states. The sum of the populations of each of these three groupings is equal to the sum for the entire United States. Similarly, besides county subdivisions, states can also be divided into Congressional Districts (the original purpose of the Census) and into ZIP Code areas. Congressional Districts, like counties, cover 100% of a state's territory and account for 100% of a state's population. Zip Code areas, on the other hand, do not account for 100% of a state's territory. (Whereas unoccupied areas have census and congressional competence, mail is not delivered where there are no addresses or where no one lives.)
Within each state there are also concentrations of people in some areas, while others are sparsely populated. States are divided to assure a level of administrative burden sharing within for the provision of education, infrastructure, and other community services. Frequently, the concentrations of the populace and the administrative divisions have not exactly coincided. The Census Bureau has divised a number of mechanisms to assure everyone is counted, but the population can be measured according to a variety of geographic criteria.
General Outline of 1990 Census Geography
Summary No. of
Level Units Geographic Description
010 1 United States
020 4 Regions
030 9 Divisions
040 51 States, plus District of Columbia
050 3,141 Counties
060 35,298 County sub-divisions
070 23,788 Places/remainders
080 61,231 Census Tracts/Block Numbering Areas
(43,691 Tracts/3,423 BNAs)
090 226,399 Block Groups
100 6,961,148 Blocks
State Subdivision Variations
3,141 Counties
435 Congressional Districts
29,469 Zip Code Areas (There are about 8,000 zip code areas that
are only commercial.)
Covering less than 100% of the population and geography
Places (crossing county divisions)
Indian reservations, trusts, etc.
Return to Contents
The 100% population data are released in two formats, printed books and machine readable form. The 100% count publications include the CP-1 (General Population Characteristics) and the CH-1 (General Housing Characteristics). There are two sets of machine readable formats: the PL 94-171 data summarizing all questions down to the tract level--for the purposes of reapportionment--and the Summary Tape Files 1 and 2, which provide comparative summaries for the 100% population counts.
The sample data are also released in printed and machine readable form. Sample publications include CP-2 (Social and Economic Characterisitcs) and CH-2 (Detailed Housing Characteristics). There are two sets of machine readable formats: the Summary Tape Files 3 and 4, which provide statistics for the samples of the population, and the Public Use Microdata Series (PUMS) which consist of the actual sample data with most geographical identifiers removed. The 1990 Census has released three PUMS versions--the full 5% (1 in 20) sample (consuming about 25 reels of computer tape or 7 CD/ROM disks), a (1 in 100) 1% sample (about 5 reels of computer tape or 2 CD/ROM disks), and a .1% (1 in 1,000) sample (one tape).
The 1990 Census of Population and Housing reports are available in the Document Service Center's stacks located adjacent to the DSC in Room 327 of Lehman Library. As reports are released, they are processed and placed in the stacks. The following is a list of these reports together with their cataloging numbers. (An index of Census terms is at the end of this guide.)
CP-1 GeneralPopulation Characteristics (C3.223/6:990 CP-1-#)
CP-2 Social and Economic Characteristics (C3.223/7:990 CP-2-#)
CH-1 General Housing CharacteristicsE (C3.224/3:990 CH-1-#)
CH-2 Detailed Housing Characterisitics (C3.224/3:990 CH-2-#)
CPH-1 Summary Population and Housing Characteristics (C3.223/18:990 CPH-1-#.)
CPH-2 Population and Housing Unit Counts (C3.223/5:990 CPH-2-#)
CPH-3 Population and Housing Characteristics for Census Tracts and Block Numbering Areas (C3.223/11:990 CPH-3-#)
CPH-4 Population and Housing Characteristics for Congressional Districts of the 103rd Congress, (C3.223/20:990 CPH-4-#)
CPH-5 Summary Social, Economic, and Housing Characteristics (C3.223/23:990 CPH-5-#)
CPH-6 Summary Social, Economic, and Housing Characteristics--outlying areas (C3.223/23:990 CPH-6-#)
The Census Bureau has released a large portion of the 1990 Census on CD-ROM as well as on computer tape. The Columbia Electronic Data Service (EDS), located in Room 201 of Lehman Library, has all CD products released by the Census Bureau and tape of other Census materials pertinent to the New York area. The CD-ROMs were created from the original computer tapes and have made access to census data easier than ever before. Data can be browsed using Go, a menu-driven software developed by the Census Bureau, or special subsets of data can be created using Extract, a more complex software.
STF 1 Summary Tape File 1--100% data of population and housing
Technical documentation is on reserve in EDS.
STF 2 Summary Tape File 2--100% data with more subject detail.
STF 3 Summary Tape File 3--sample data of population and housing.
Technical documenation is on reserve in EDS.
STF 4 Summary Tape File 4--sample data with more subject detail.
PL 94-171 Public Law 94-171 Redistricting Data File
EEO Census/Equal Employment Opportunity Special File
PUMS Public Use Microdata Samples
Technical documentation is on reserve in EDS.
TIGER TIGER/Line 1992