1970 Block Boundaries

Through this page, NHGIS provides a shapefile depicting 1970 census block boundaries for selected areas throughout the U.S. The current version (v2.0) includes all or part of 38 metropolitan areas. We will continually extend the coverage in future releases.

For the current version, we have completed only our first stage of edits, which ensures that nearly all 1970 blocks are represented in the completed areas. However, the files still contain certain types of inaccuracies and extraneous polygons, which we are now working to correct with a second stage of edits for the next release (v3.0), targeted for late 2026 or early 2027.

At this time, we provide 1970 block boundary data only through this web page. In the future, we will provide the boundary data through the NHGIS Data Finder, where we already provide 1970 block summary data and both summary and boundary data for 1990 and later census blocks. Summary and boundary data for 1980 blocks are available through the 1980 Block Boundaries page.

• Project Overview
  • Sources & Workflow
  • Two Stages of Work & Limitations of the Released First-Stage Output
• Technical Details
  • Basics
  • Geographic Coverage
  • Getting Census Data for 1970 Blocks
  • Duplicate IDs in the Census Data
• Download
• Feedback and Questions
• Citation and Use
• Credits

Project Overview

Sources & Workflow

To produce 1970 block boundaries, NHGIS staff manually edit the NHGIS 1980 block boundaries to correspond with the published paper maps and summary data for 1970 census blocks.

There is no other general source of 1970 block boundary data. The U.S. Census Bureau's 1992 TIGER/Line Files include both 1990 and 1980 block boundaries, but the 1980 blocks in these files are incomplete and often inaccurate. NHGIS has been working to edit the 1980 block boundaries to produce a more accurate and complete version. The latest version, along with documentation explaining our 1980 editing process, is available through the 1980 Block Boundaries page.

The current version of our 1970 block shapefile includes only the areas we have completed. We plan to release more versions including more areas at regular intervals until the project has ended.

NHGIS validates the edited boundary data against block-level summary data from the 1970 Summary Tape Files (Third Count). For a small portion of the blocks that appear in the summary data, we are unable to find corresponding blocks on the paper maps, so we cannot provide complete block definitions for all reported statistics.

The 1970 block shapefile comes with a supplementary list of the discrepancies between the 1970 blocks identified in the boundary file and in the summary data.

Back to top

Two Stages of Work & Limitations of the Available Data

In our first stage of 1970 block boundary editing, we modeled the process on our 1980 block boundary editing, which focuses on cases where a block identified in the census summary data does not have a corresponding polygon in the boundary data. We worked our way through lists of these cases, and in each case, we inspected block maps to determine how to edit the boundary data to correctly represent the missing block.

Following this approach, editors give relatively little attention to areas where there are no missing polygons. This is appropriate for the 1980 work because in that case, we begin with a dataset (the 1992 TIGER/Line version of 1980 blocks) that contains few boundaries that are not in fact 1980 block boundaries. The main need then is only to ensure that the missing blocks are correctly represented.

For the 1970 block boundaries, however, we begin with 1980 block boundaries, so many of the base polygons initially represent blocks that existed in 1980 but not in 1970. To achieve full correspondence with 1970 block maps, these "extraneous" 1980 block polygons need to be deleted. There are also many cases where the block codes or boundaries changed between 1970 and 1980, making it necessary to edit the 1980 block codes or boundaries to achieve full correspondence with 1970 block maps.

In our first stage of 1970 block editing, we made many of these types of edits (deleting extraneous 1980 blocks and editing 1980 block boundaries and codes), but we did not systematically seek out all cases where they may be needed because our focus was on addressing cases of "missing polygons."

The current released version of the 1970 block boundaries (v2.0) consists only of these "first-stage" edits, which means we have ensured that, in the completed areas, nearly all of the 1970 blocks that have summary data also have boundary data (wherever we were able to identify them on block maps).

We have already begun a second stage of editing where we will review all of the areas completed in the first stage to identify and execute other types of required edits, ensuring that the boundaries correspond with the original 1970 block maps as thoroughly as we can achieve. We expect to complete this work and release an improved version of 1970 blocks (v3.0) in late 2026 or early 2027.

Back to top

Technical Details

Basics

• The boundaries are provided in a shapefile within a ZIP archive
• There is a single nationwide shapefile that includes all block boundaries available at this time.
  • We also plan to provide separate shapefiles for individual states after we've created 1970 boundaries for more areas
• The shapefile's attribute table includes the standard NHGIS GISJOIN identifiers, which correspond to GISJOIN identifiers in data tables from NHGIS dataset 1970_Cnt3 (1970 Census: Count 3 - 100% Data [Blocks & Tracts: Restricted to Areas with Block Statistics])
• The attribute table also includes separate codes identifying the state, county, and census tract where the block is located
• The ZIP archive includes two supplemental documentation files:
  • block_boundaries_1970_codebook.txt - A human-readable text codebook describing the contents and source of the shapefile
  • US_block_1970_data_missing_polygon.csv - A CSV (comma-separated-values) file listing blocks that have a data record in the summary tables but no polygon in the shapefile (because we were unable to find the block on the paper maps)

Back to top

Geographic Coverage

In 1970, the Census Bureau's block data program covered only urbanized areas, some adjacent areas, and select other areas that contracted for census block data in the 1970 census.

We aim to produce 1970 block shapefiles for all areas that have block data in the 1970 Third Count Summary Tape File (1970_Cnt3) dataset and that are within 1980 SMSAs (Standard Metropolitan Statistical Areas). The reason for this scope is that the NHGIS 1980 block boundaries, which form the basis for the 1970 block boundaries, are limited to 1980 SMSAs.

At this time, our 1970 block shapefile provides at least partial coverage of the blocks in 38 1970 SMSAs, with complete coverage for 35 SMSAs and partial coverage for 3 SMSAs. In this case, "complete coverage" means that we have fully reviewed all published 1970 block maps covering any part of the SMSA, and the shapefile includes boundaries for any block identified in the summary data that we could also find on the block maps. These "complete cases" therefore still omit some blocks that appear only in the summary data and not on the maps.

In SMSAs with “partial coverage,” we have block boundaries for the urban core but omit some outlying part(s). We will complete these areas in later releases.

We detail current coverage by 1970 SMSA and by county in this spreadsheet:

• 1970 Block Shapefile Coverage (Excel spreadsheet)

Click below for a list of currently covered 1970 SMSAs ordered by 1970 population:

Back to top

Getting Census Data for 1970 Blocks

The Census Bureau provided summary data for 1970 blocks in the 1970 "Third Count" Summary Tape File, as explained in the 1970 Census Users' Guide. The corresponding NHGIS dataset is labeled 1970_Cnt3: Count 3 - 100% Data [Blocks & Tracts: Restricted to Areas with Block Statistics].

Most of the 83 tables in this dataset cover housing subjects, in many cases broken down separately for White and Black households. There are six tables that provide population counts broken down by sex, age, race, marital status, household relationship, and household type.

Users can learn more about these tables and download them through the NHGIS Data Finder. To find the tables, you can select the Geographic Levels filter for Blocks, the Years filter for 1970, and/or the Datasets filter for 1970_Cnt3.

To merge downloaded tables with the shapefile, you can join on the GISJOIN identifier, which NHGIS provides in both the table data and the shapefile. Note, however, that there are some duplicate identifiers in the table data, as explained in the next section.

Back to top

Duplicate IDs in the Census Data

Of the 1.36 million block records in NHGIS dataset 1970_Cnt3, a small fraction (1,234 blocks, or 0.09%) share a GISJOIN ID with another record. The likely explanation for these duplicates is that these records do not represent the "whole collection blocks" that appear on the 1970 block maps, but instead, they represent "tabulation blocks," which are subdivisions of the collection blocks, split by the boundaries of places, minor civil divisions, or other census reporting areas. This distinction is much more impactful in 1980 block data, so we provide a full explanation on the 1980 Block Boundaries page.

In the case of 1970 block data, unlike the 1980 data, the NHGIS data files do not contain codes for the geographic areas that might distinguish the different parts of the split collection blocks. Rather, the data files contain only state, county, census tract, and block identifiers, which are identical across the duplicate records. We plan to investigate this further and add the distinguishing codes if possible. We also plan to merge the duplicate records to produce summary data with a one-to-one correspondence with the 1970 block boundary data, as we've done with the 1980 block data. But because the number of duplicate cases is so much smaller in 1970, users should be able to proceed with using the available 1970 block summary data, unaltered, for nearly all areas that are covered in the current version of the shapefile.

In the "List by County" sheet in the 1970 Block Shapefile Coverage spreadsheet, we identify how many duplicate block IDs occur in each county. Among the counties covered in version 1.0 of the 1970 block boundaries, only two have duplicate IDs in their block summary data: Cook County, IL, has 1 duplicate ID, and Middlesex County, MA, has 3 duplicate IDs.

Back to top

Download

The 1970 block boundaries are provided in this single nationwide shapefile:

• U.S. 1970 Census Block Boundary Shapefile

Back to top

Feedback and Questions

You may direct any comments or questions about these files to [email protected].

Back to top

Citation and Use

Use of the NHGIS 1980 Block Boundary Files is subject to the same conditions as for all NHGIS data. See Citation and Use of NHGIS Data.

Back to top

Credits

Our production of 1970 Block Boundary Files is principally supported by funding from the National Science Foundation (2316650).

Back to top