U.S. Department of Commerce

TechTalk

You are here: Census.govTechTalk › Accessing 1990 Census of Population and Housing Data Files Directly
Skip top of page navigation

Accessing 1990 Census of Population and Housing Data Files Directly

Keep the following in mind when accessing the data files in dBase III tm format directly (without using the DOS software provided on the cd-rom) with database management software (DBMS).

The procedure below is a quick guide to this process. Additional notes follow.

  • The Subject Locator section of the file documentation can be used to get table numbers for desired tables.
  • The tables.dbf file or the chart below (from the "How to use this file" section of the file documentation) can be used to find out which data file contains a specific table by the its' table number.
  • Open the Summary Level Sequence chart section of the file documentation for area (record) selection criteria
  • Open file number "00" (file containing a complete list of geographic identifiers) to generate area lists (such as a list of county names and codes) if necessary
  • Open data file containing desired tables and query
  • The field LOGRECNU is a unique key field that can be used to link each data file back to the geographic file to obtain more geographic identifier information for specific areas.

What summary level code should you use?

There may be cases where two or more summary levels look alike. Look at the following example from 1990 Summary Tape File 3A:

SUMMARY LEVEL SEQUENCE CHARTS
Summary level area Summary level code
State-County-County Subdivision-Place/Remainder-Census Tract/Block Numbering Area 080
State-County-Census Tract/Block Numbering Area 140

What's the difference? Since a Census Tract or Block Numbering Area can cut across place or county subdivision boundaries, code 080 refers to Census Tract or Block Numbering Area parts. Part of a census tract may be in a place or county subdivision and the other part may be outside of it. Data for a census tract, in this case, will be contained in multiple records (one for each part). The forward slashes "/" mean OR.

Reading from right to left, if any area type to the right of a dash can cut across the boundaries of an area type to the left of the dash, the summary level (area type for which statistics are compiled) will include area parts.

What about code 140? Since a Census Tract or Block Numbering Area is always entirely contained within a county and, therefore, never crosses county boundaries, summary level code 140 only includes whole Census Tracts or Block Numbering Areas.

It is necessary to include the geographic identifier field for each element in the complete summary level description to uniquely identify the area. These fields would be the state FIPS code, the county FIPS code and the census tract code for summary level "140". Summary level "080" would include the fields in Summary Level "140" plus the county subdivision FIPS code and the place FIPS code.


Why do there appear to be "duplicate" records? If you are using the DOS software on the disc for lookups, you will not have to deal with this issue.

Let's again take the example of 1990 Summary Tape File 3A. Due to restrictions on the number of fields allowed in a dBase (tm) record, the database was divided into thirty five separate tables (or data files). The first table (segment STF300XX.DBF) contains a three hundred character geographical segment containing sixty seven fields (see data dictionary section). The other thirty four files contain only eight of these fields.

The variable LOGRECNU is a unique key which relates the thirty five files. This doesn't happen often, but there may be cases where the LOGRECNU number is different, but there are at least two records that have the other seven fields in common, especially for smaller areas. Does this mean that there are duplicates? No. If you refer back to the first segment (STF300XX.DBF), at least one field will be different. The fields GEOCOMP and URBANRUR are the most common examples of this.


Selecting tables

Programmers can use the metadata file TABLES.DBF to automate the process of opening up the correct data file (or segment) and extracting data from it. You can download [ZIP, 49 KB] an example from the 1990 Summary Tape File 3A CD-ROM. This is not applicable to the magnetic tape files. Programming is necessary for extracts that cannot be performed simply with the software on disc (see Accessing dBASE (tm) files).

You can also use the keyword feature of the lookup software (called by GO.BAT) on the CD-ROM to get a list tables including a keyword.

There is a section in the technical documentation titled Using the File with related information. An example from the 1990 Summary Tape File 3 appears below.

Using the File (1990 Summary Tape File 3)

File Segments

The file is segmented into 35 dBase III (.DBF) files, designated STF300ss.DBF through STF334ss.DBF where ss is the two-character State abbreviation. The STF300 segment contains the full 67 field identifiction section. The identification field names are shown in the Data Dictionary chapter of the technical documentation. Segments STF301 through STF334 each contain eight identification fields repeated from the STF300 segment. They are shown below.

Identification Fields Common To All Segments

SUMLEV ..... Summary Level
STATEFP .... State (FIPS)
CNTY ....... County (FIPS)
COUSUBFP ... County Subdivision (FIPS)
PLACEFP .... Place (FIPS)
TRACTBNA ... Census Tract/Block Numbering Area
BLCKGR ..... Block Group
LOGRECNU ... Logical Record Number

The segments are divided as shown below.

Data Tables In Each Segment

STF300 Field Identification Section
STF301 P1 - P13
STF302 P14A - P14C
STF303 P14D - P14F
STF304 P14G - P14I
STF305 P14J - P17
STF306 P18 - P26
STF307 P27 - P32
STF308 P33 - P36
STF309 P37 - P51
STF310 P52 - P60
STF311 P61 - P65
STF312 P66 - P70
STF313 P71 - P76
STF314 P77 - P83
STF315 P84 - P86
STF316 P87A
STF317 P87B
STF318 P87C
STF319 P87D
STF320 P87E
STF321 P88 - P107A
STF322 P108 - P118
STF323 P119 - P121
STF324 P122 - P123
STF325 P124A - P124B
STF326 P125 - P143
STF327 P144 - P170, H1 - H9
STF328 H10 - H21
STF329 H22 - H33
STF330 H34 - H44
STF331 H45 - H52A
STF332 H53 - H59
STF333 H60 - H81
STF334 H82 - H92

Field Names In Numeric Data Tables

Fields in numeric data tables are named according to a convention which identifies the tables and the sequence of the data item within the table. The 63 data items in P88, for example, are identified as P0880001 through P0880063. The one data item in table P107A is identified as P107A001.

Components Of The Field Name

Character 1 ..... P or H
Character 2-4 ... Table number; right justified with leading zeroes
Character 5 ..... Sub-table letter; zero if not applicable
Character 6-8 ... Item number; right justified with leading zeroes

Record Identification

Each record on a summary tape file is uniquely identified by a combination of geographic (for example, State, county, and county subdivision), summary level, and geographic component codes. For example, if you have a Maryland STF 3A file and want to look at the summary record for Prince George's County, three identification fields must be checked: the summary level code for county records (050), the total record for geographic components (00), and the county code for Prince George's County (033). If all three fields are not checked, multiple records will be found. See the Summary Level Sequence Charts and Geographic Area Component Codes section below for information on identifying summary level and geographic component codes.


Source: U.S. Census Bureau | Administrative and Customer Services Division | (301) 763-7710 |  Last Revised: May 10, 2013