Documentation: ACS 2010 -- 2012 (3-Year Estimates)
you are here: choose a survey survey document chapter
Publisher: U.S. Census Bureau
Document: ACS 2012-3yr Summary File: Technical Documentation
citation:
Social Explorer; U.S. Census Bureau; American Community Survey 2012 3yr Summary File: Technical Documentation.
ACS 2012-3yr Summary File: Technical Documentation
2. How to Use the ACS Summary File
2.1. Locating the Summary File
The 2010-2012 ACS 3-year Summary File is accessible from the American Community Survey home page. From the ACS home page, http://www.census.gov/acs/www/, click on the Data and Documentation tab, select the option for Summary File, as shown below:


That will take you to the ACS Summary File page. Click on 2012 ACS 3-Year Summary File to go to the ACS Summary File ftp site:



This is the ACS Summary File-it is actually comprised of three folders that are explained the next chapter.
2.2. Summary File Organization
The Summary File is organized in three folders as shown in the above screenshot. These three directories contain the same combination of files; they are simply arranged differently to accommodate various user needs:
  • 2010-2012_ACSSF_All_In_l_Giant_File(Experienced-Users Only)
The "All in 1 Giant File" directory contains a zipped file, which includes geography, estimate, and margin of error files. This zipped file is ideal to download if users want estimates and margins of error for all geographies throughout the nation at once. The file is very large and should only be used by those that can easily process a very large file. For a graphical representation of this folder and its contents, see Appendix B.1.
  • 2010-2012_ACSSF_By_State_All_Tables
The "By State All Tables" directory contains zipped files for each state or state level equivalent-each of the 50 states, District of Columbia, Puerto Rico, as well as cross-state geographies such as metropolitan areas. Each zipped file contains a geography file and multiple estimate and margin of error files. Downloading from these state level folders is ideal if users want all the tables for a state level geography or cross-state geographies. For a graphical representation of this folder and its contents, see Appendix B.2.
  • 2010-2012_ACSSF_By_State_By_Sequence_Table_Subset
The "By State By Sequence Table Subset" directory contains folders for each state or state level equivalent-each of the 50 states, District of Columbia, Puerto Rico, as well as cross-state geographies such as metropolitan areas. Within those folders are a geography file and zipped files containing the estimate and margin of error files, one per "sequence" (sequences are explained in Chapter 2.3). Downloading from these folders is ideal if users only want a few tables for a state level geography or cross-state geographies. For a graphical representation of this folder and its contents, see Appendix B.3.

The naming convention used for the zipped files in this directory is the following:



File Name: 20123ak0001000.zip
Example Name Range or Type
2012 Reference Year ACS data year (last year of the period for multiyear periods)
3 Period Covered 1=1-year, 3=3-year, 5=5-year
ak State Level US or abbreviations for state, District of Columbia, and Puerto Rico
0001 Sequence Number 0001 to 9999
000 Place Holder Iteration value for future use

Note: Iteration ID is always "000" for the standard 1-Year, 3-Year, and 5-Year products.

As Appendix B shows, the "All in 1 Giant File" and the "By State All Tables" folders contain the same tables as the "By State By Sequence Table Subset" folder. The difference is in the organization. The "By State All Tables" zipped files contain all of the sequence files for the given state, so each zipped file contains 354 files. The "All in 1 Giant File" zipped file contain all sequence files for all states, which is thousands of files.

As mentioned earlier, the zipped files are divided by state or state-level equivalents. Those state- level equivalents include the District of Columbia and Puerto Rico. There is also a level called "United States," which is for summary levels that can cross state boundaries, such as the Nation, and all Regions, Divisions, Metropolitan Statistical Areas and Tribal Reservations. The United States level does not contain tables for geographies that are always entirely within a state, such as counties and places; for those tables, go to the folder or files for that state.
The following is a table that gives examples of the types of summary levels are in the state and state-level equivalent folders and files and those that are in the United States folders and files.
Each State, DC, and Puerto Rico United States
State United States
County Region
County subdivision Division
Place Metropolitan or urban statistical areas
Congressional districts (110th Congress) New England City and Town Area (NECTA)
Public Use Microdata Area (PUMA) American Indian/Alaska Native/Hawaiian Home Land areas
School Districts Urban areas
Alaska Native Regional Corporation  


2.3. Sequence Numbers
Detailed Tables for similar subject areas are grouped together in "sequences". A sequence number is an assigned number to a grouping of ACS tables. Table sequencing now follows these new rules:

1) Tables are sorted numerically by the "root" of their Table ID, where the "root" is defined as the numeric section after the first letter and before any additional letters, so for example the root of B06004APR is "06004". For tables with the same root, additionally sort them in the following order:

Non-iterated, non-collapsed, non-PR version (e.g., Table B06003)

Iterated, non-collapsed, non-PR versions (e.g., Tables B06004A, B06004B…B06004I)

Non-iterated, collapsed, non-PR version (e.g., Table C06001)

Iterated, collapsed, non-PR version (e.g., Tables C08505A, C08505B… C08505I)

Non-iterated, non-collapsed, PR version (e.g., Table B06003PR)

Iterated, non-collapsed, PR versions (e.g., Tables B06004APR, B06004BPR…B06004IPR)

Non-iterated, collapsed, PR version (e.g., Table C06001PR)

Iterated, collapsed, PR version (e.g., Table C06001APR)

2) With tables sorted in this order, start with the first table and assign it to the first sequence. For each subsequent table, if the table has either a new "subject", a new "geography type", or would cause the number of cells in the sequence to exceed 245, then start a new sequence. "Subject" is described using the second and third characters in the Table ID, so for example the subject of B06004APR is "06". "Geography type" can be one of three things: Place of Residence geography type, Place of Work geography type, or Residence 1 Year Ago geography type see Appendix E.

3) If a table does not fit in one sequence, then put the first 245 cells of it in one sequence, and the rest in the next. If a table does not fit in two sequences, then put the first 245 cells of it in one sequence, the next 245 cells of it in the next sequence, and the rest in a third sequence.

The rules governing how many tables can be assigned the same sequence number depend on the following:

  • There are no more than 256 columns per sequence, so the data can be read into a spreadsheet.
  • There are 178 sequences for the 2010-2012 ACS 3-year Summary File.
  • Tables are grouped numerically by the "root" of their Table ID, (i.e., Table B00001 is in sequence file 0001).
  • The sequencing algorithm groups' tables sequentially by TBLID, which means that tables with race iterations are often grouped in the same sequence, but this is not always the case.
It is critical to know the sequence number associated with a Detailed Table (Table ID) for two reasons. First, one needs it in order to access the correct estimates and margins of error files for the desired table. Second, the field start position for the estimates or margins of error of a certain Detailed Table depends on its sequence number.

The Sequence Number and Detailed Table Number Lookup file, available as an Excel spreadsheet, text file, and SAS dataset, lists Table IDs associated with each sequence number. This spreadsheet, formerly known as "merge_5_6", is available at www2.census.gov/acs2012_3yr/summaryfile/. The name of the Excel spreadsheet is "Sequence_Number_and_Table_Number_Lookup.xls," the text file is called "Sequence_Number_and_Table_Number_Lookup.txt," and the SAS dataset is called "SequenceNumberTableNumberLookup.sas7bdat."

For example, to find the sequence number associated with the table B08406, a user must open and look for that Table ID in the Sequence Number and Table Number Lookup file. Shown below is a screenshot of this file opened to where the "tblid" is B08406. The next column in the file, "seq", shows that this Table ID is associated with the sequence number "0043". In order to access the estimate and margin of error file for Table B08406, a user will need to download the estimate and margin of error files labeled with the sequence number "0043".



2.4. Geography File
There is a geography file that comes with the estimate and margin of error files. This file begins with a "g" and is an ASCII file using either a position based format or comma delimited format. A geography file exists for each state or state level equivalent.

Geography files are named using the following convention:



File Name: g20123ak.txt
Example Name Range or Type
g File Type g=geography
2012 Reference Year ACS data year (last year of the period for multiyear periods)
3 Period Covered 1=1-year, 3=3-year, 5=5-year
ak State Level US or abbreviations for state, District of Columbia, and Puerto Rico


The geography files contain geographic information for an ACS tabulated area, including the name of the area. One variable on the file, called LOGRECNO, is the logical record number and is used to link the level of geography to the estimate and margin of error files. An example of how to use LOGRECNO is discusses in Chapter 2.5.
The following table provides the layout of the geography file:


Variable Name Description Field Size Starting Position Geographic Summary Levels For 1-Year Tables
RECORD CODES        
FILEID Always equal to ACS Summary File identification 6 1 All Summary Levels
STUSAB State Postal Abbreviation 2 7 All Summary Levels
SUMLEVEL Summary Level 3 9 All Summary Levels
COMPONENT Geographic Component 2 12 All Summary Levels
LOGRECNO Logical Record Number 7 14 All Summary Levels
GEOGRAPHIC AREA CODES        
US US 1 21 10
REGION Census Region 1 22 20
DIVISION Census Division 1 23 30
STATECE State (Census Code) 2 24 Blank for all summary levels
STATE State (FIPS Code) 2 26 040, 050, 060, 160, 230, 312, 352, 500, 795, 950, 960, 970
COUNTY County of current residence 3 28 050, 060
COUSUB County Subdivision (FIPS) 5 31 60
PLACE Place (FIPS Code) 5 36 160, 312, 352
TRACT Census Tract 6 41 Not applicable for this release
BLKGRP Block Group 1 47 Not applicable for this release
CONCIT Consolidated City 5 48 Not applicable for this release
AIANHH American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 4 53 250
AIANHHFP American Indian Area/Alaska Native Area/ Hawaiian Home Land (FIPS) 5 57 Not applicable for this release
AIHHTLI American Indian Trust Land/ Hawaiian Home Land Indicator 1 62 Not applicable for this release
AITSCE American Indian Tribal Subdivision (Census) 3 63 Not applicable for this release
AITS American Indian Tribal Subdivision (FIPS) 5 66 Not applicable for this release
ANRC Alaska Native Regional Corporation (FIPS) 5 71 230
CBSA Metropolitan and Micropolitan Statistical Area 5 76 310, 312, 314
CSA Combined Statistical Area 3 81 330
METDIV Metropolitan Statistical Area-Metropolitan Division 5 84 314
MACC Metropolitan Area Central City 1 89 Blank for all summary levels
MEMI Metropolitan/Micropolitan Indicator Flag 1 90 010, 020, 030, 040, 314
NECTA New England City and Town Area 5 91 335, 350, 352
CNECTA New England City and Town Combined Statistical Area 3 96 335
NECTADIV New England City and Town Area Division 5 99 355
UA Urban Area 5 104 400
BLANK   5 109 Reserved for future use
CDCURR Current Congressional District *** 2 114 500
SLDU State Legislative District Upper 3 116 Not applicable for the release
SLDL State Legislative District Lower 3 119 Not applicable for the release
BLANK   6 122 Reserved for future use
BLANK   3 128 Reserved for future use
ZCTA5 5-digit ZIP Code Tabulation Area 5 131 Not applicable for the release
SUBMCD Subminor Civil Division (FIPS) 5 136 Not applicable for the release
SDELM State-School District (Elementary) 5 141 950
SDSEC State-School District (Secondary) 5 146 960
SDUNI State-School District (Unified) 5 151 970
UR Urban/Rural 1 156 010, 020, 030, 040
PCI Principal City Indicator 1 157 010, 020, 030, 040, 312, 352
BLANK   6 158 Reserved for future use
BLANK   5 164 Reserved for future use
PUMA5 Public Use Microdata Area - 5% File 5 169 795
BLANK   5 174 Reserved for future use
GEOID Geographic Identifier 40 179 All Summary Levels
NAME Area Name 1000 219 All Summary Levels
BTTR Tribal Tract 6 1219 Not applicable for the release
BTBG Tribal Block Group 1 1225 Not applicable for the release
BLANK   43 1226 Reserved for future use


This year, we created an Excel template for the geography file named "2012_SFGeoFileTemplate.xls". The template provides users with two rows containing the variable names and their descriptions (as displayed in the above table) for each column in the geography file. It is meant to be used with the comma delimited version of the geography file. The template is available at http://www2.census.gov/acs2012_3yr/summaryfile/UserTools/ in the zipped "2012_SummaryFileTemplates" folder. Here is a screenshot of the Excel file:



Each state, the District of Columbia, Puerto Rico and the set of cross-state geographies, have one geography file associated with them regardless of how the Summary File is accessed. For example, the following screenshot shows the beginning of the state geography file for Maryland. In the screenshot, the logical record numbers corresponding with the state of Maryland, Allegany County, and Anne Arundel County are circled. The logical record number for the state of Maryland is "0000001", for Allegany County it is "0000012", and for Anne Arundel County it is "0000013".


Excess spaces in the pictured geography file have been removed for illustrative purposes.

2.5. Estimate and Margin of Error Files
Each of the three Summary File directories include zipped files containing estimate files (file names beginning with an "e") and margins of error files (file names beginning with an "m"). The estimate files contain published ACS estimates and the margin of error files contain published ACS margins of error for their respective estimates. Here is the naming convention used for those files:



File Name: e20123ak0003000.txt
Example Name Range or Type
e File Type e=estimate, m=margin of error
2012 Reference Year ACS data year (last year of the period for multiyear periods)
3 Period Covered 1=1-year, 3=3-year, 5=5-year
ak State Level US or abbreviations for state, District of Columbia and Puerto Rico
0003 Sequence Number 0001 to 9999
000 Reserved for future use Iteration value for future use


The estimates and margins of error for Detailed Tables are grouped together in by sequence numbers, as discussed in Chapter 2.3. There is an estimate and margin of error file for each sequence number.

The format of the estimate and margin of error files are identical; they are strings of comma- delimited ASCII text. Each row represents a different geographic area and the first six fieldscontain metadata such as the geographic area and the sequence number. Following those fields are the estimates or margins of error for the Detailed Tables. Starting and ending positions of the fields associated with each Detailed Table can be found using the Sequence Number and Table Number Lookup file, which is discussed in Chapter 2.3. The estimates or margins of error for one Detailed Table span several fields within a row.

Here is the record layout of the estimates and the margin of error files:

Field Name Description Field Size
FILEID File Identification 6 Characters
FILETYPE File Type 6 Characters
STUSAB State/U.S. - Abbreviation (USPS) 2 Characters
CHARITER Character Iteration 3 Characters
SEQUENCE Sequence Number 4 Characters
LOGRECNO Logical Record Number 7 Characters
Field # 7 and up Estimates Various


Going back to the example from Chapter 2.3, we know that table B08406 corresponds to sequence "0003". Additionally, the Sequence Number and Table Number Lookup file (as shown earlier) tells us that table B08406 begins at position seven and contains 51 cells.

In order to get estimates for Maryland; Allegany County, MD; and Anne Arundel County, MD one must recall the logical record numbers associated with each geography. In Chapter 2.4, we identified these to be "0000001", "0000012", and "0000013", respectively. The logical record number, LOGRECNO, must be used to merge the geography information to the estimate and margin of error files.

The example below shows the estimate file for sequence "0003" and all geographies except census tracts and block groups for the state of Maryland using the 2009 ACS 1-year Summary File. For the 2010 ACS 1-year Summary File, the dots "." in the below screenshot will be replaced by empty cells as documented in Appendix C.3. Note that each row has a uniquely assigned logical record number, called LOGRECNO, which links the estimate to a specific geographic area. The pictured example has the logical record numbers corresponding to Maryland, Allegany County, and Anne Arundel County circled. Estimates for table B08406 at these geographic levels can be found within their respective rows at field seven and continuing for 50 additional fields.




2.6. How to Join the ACS Summary File to the TIGER/Line Shapefiles
TIGER/Line Shapefiles allow data users to directly link geographic areas to data from the American Community Survey and other surveys. The TIGER/Line Shapefiles are designed for use with geographic information system (GIS) software. Learn more about TIGER/Line Shapefiles at http://www.census.gov/geo/maps-data/data/tiger-line.html.

The variable GEOID joins the ACS Summary File to the TIGER/Line Shapefiles. For the ACS Summary File, GEOID is located in column AW of the geography file. It is not found in the estimates or margins of error files. (As discussed in previous chapters, the variable LOGRECNO is needed to join together the parts that make up the Summary File: the geography, estimates, and margins of error files). GEOID's corresponding variable in the 2012 TIGER/Line Shapefiles is GEOID10.

We will walk through an example of joining these files using Kent County, Delaware. In the ACS Summary File, the GEOID is 05000US10001. In the TIGER/Line Shapefiles, the GEOID10 is 10001. (GEOID is a concatenation of all the codes associated with a given geographic area, such as the state FIPS code, county FIPS code, etc. The exact concatenation varies by geographic area. In this example, 10=state FIPS code and 001=county FIPS code.)

The ACS Summary File GEOID contains the necessary information to connect to the TIGER/Line Shapefiles, but it needs to be modified in order to exactly match up. Notice that GEOID, 05000US10001, contains the GEOID10 string, 10001.

In order to create an exact match of GEOID and GEOID10, it is necessary to remove all of the characters before and including the letter "S" in the ACS Summary File. By removing these characters, the new GEOID in the ACS Summary File exactly matches the field GEOID10 in the TIGER/Line Shapefiles.

The following is an example of how to modify the ACS Summary File's GEOID in Excel 2007 so it can be joined with TIGER/Line Shapefiles:

1) Open the ACS Summary File comma delimited geography file in Excel. This example uses Delaware's geography file (g20121de.csv) available at www2.census.gov/acs2012_1yr/summaryfile with the column headers from the geography file template copied into Delaware's geography file. Learn more about the geography file template in Chapter 2.4.

2) Insert 2 blank columns to the right of the column "GEOID." Your modified GEOID will eventually go into the second column. (Note: Columns F through AV in the diagrams following are hidden for illustrative purposes.)



3) Next, select the column "GEOID."




4) Select the "Data" tab from the top menu, then select "Text to Columns."

The "Convert Text to Columns Wizard" box should pop up.




5) In the "Convert Text to Columns Wizard," select "Delimited" under "Choose the file type that best describes your data:" then click "Next."




6) Check "Other" as the delimiter and type the letter "S" into the box. Click "Next."




7) In the "Data preview" window, click on the top of both columns in "Data preview" and select "Text" under "Column data format." In "Destination," select the two blank columns that you created in Step 1. Click "Finish."



8) Column AY should now contain the modified GEOID that corresponds to GEOID10 in the TIGER/Line Shapefiles. The second screenshot shows the TIGER/Line Shapefile for Kent County, Delaware.






9) The ACS Summary File and the TIGER/Line Shapefile should now be ready to be joined using GIS software. Visit "Working with TIGER/Line Shapefiles" at https://www.census.gov/geo/education/howtos.html to learn more about how to access and use the TIGER/Line Shapefiles.