Documentation: ACS 2007 -- 2009 (3-Year Estimates)
you are here: choose a survey survey document chapter
Publisher: U.S. Census Bureau
Document: ACS 2009-3yr Summary File: Technical Documentation
citation:
Social Explorer; U.S. Census Bureau; American Community Survey 2007-2009 Summary File: Technical Documentation.
ACS 2009-3yr Summary File: Technical Documentation
Chapter 2. How to Use the ACS Summary File
2.1. Locating the Summary File
There are two methods to easily find the 2005-2009 ACS 5-year Summary File: (1) through the ACS home page at www.census.gov/acs or (2) through the American FactFinder web site at factfinder.census.gov.

  1. Through the ACS home page at www.census.gov/acs:
From the American Community Survey home page, www.census.gov/acs, 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. As you can see at the top of the next page, the data are arranged by each ACS data year and estimate type (1-year, 3-year and 5-year). Click on 2005-2009 ACS 5-Year Summary File to go to the ACS Summary File.



There you will see the following screen:



This is the ACS Summary File-it is actually comprised of three folders that are explained the next chapter.

  1. Through the American FactFinder web site at factfinder.census.gov:
The Summary File Directory may also be accessed through the American FactFinder at factfinder.census.gov. Under the American Community Survey heading, select Get Data, as shown in the screenshot below.



Inside the grey box titled 2005-2009 American Community Survey 5-Year Estimates, click on the Download 5-Year Block Group Data in Summary Files, as the screenshot below shows.



That will lead to the following screen:



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, which are the first three folders shown in the screenshot on the previous page. These three directories contain the same combination of files; they are simply arranged differently to accommodate various user needs:

  • 2005-2009_ACSSF_All_In_2_Giant_Files(Experienced Users Only)
The "All in 2 Giant Files" directory contains two zip files, with one containing the files for all tables at the census tract and block group levels only and the other containing the files for all tables for all geographies except for census tracts and block groups. Each zip file contains geography, estimate, and margin of error files. These are ideal to download if users want estimates and margins of error for all geographies throughout the nation at once. These files are very large and should only be used by those that can easily process very large files. For a graphical representation of this folder and its contents, see Appendix B.1.
  • 2005-2009_ACSSF_By_State_All_Tables
The "By State All Tables" directory contains zip 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 state level file is broken into two zip files with one containing the files for all tables at the census tract and block group levels only and the other containing the files for all tables for all geographies except for census tracts and block groups. Each zip file contains geography, 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.

  • 2005-2009_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. Each state level folder is broken into two additional folders: one for the census tract and block group levels only and the other for all geographies except for census tracts and block groups. Within those folders are a geography file and zip 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 is the following:


File Name: 20095ak0001000.zip
Example Name Range or Type
2009 Reference Year ACS data year (last year of the period for multiyear periods)
5 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
1 Sequence Number 0001 to 9999
0 Place Holder Iteration value for future use

As Appendix B shows, the "All in 2 Giant Files" 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 234 files. The "All in 2 Giant Files" zipped files contain all sequence files for all states, which is about 17,000 files.

It is highly recommended that data users download from the "By State By Sequence Table Subset" folder to decrease download time, especially if (like most users) they need only a few tables for a few geographic areas. The chart below shows the average download time for each directory as a reference, based on download speed and which file is being downloaded:
Folder Estimated Time to Download One File
High-Speed Download Low-Speed Download
All geos except Tracts and Block Groups Tracts and Block Groups All geos except Tracts and Block Groups Tracts and Block Groups
All In 2 Giant Files (Experienced Users Only) 1 hour 38 minutes 10 hours 7 hours
By State All Tables Up to 5 minutes Up to 3 minutes Up to 1 hour Up to 45 minutes
State by Sequence Table Subset less than 1 minute less than 1 minute less than 1 minute less than 1 minute

As mentioned earlier, the zip 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.

Below 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. It also shows that for the state and state-level equivalents, the census tracts and block groups are in separate folders and files than the other summary levels.

Each State, DC, and Puerto Rico United States
All_Geographies_Not_Tracts_Block_Groups All_Geographies_Not_Tracts_Block_Groups
State United States
County Region
County subdivision Division
Place All metropolitan or urban statistical areas
Congressional districts (110th Congress) New England City and Town Area (NECTA)
Public Use Microdata Area (PUMA) All American Indian/Alaska Native/Hawaiian Homeland areas
School Districts Urban areas
Alaska Native Regional Corporation  
Tracts_Block_Groups_Only  
Tracts  
Block Groups  


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. 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 117 sequences for the 2005-2009 ACS 5-year Summary File.
  • Tables are grouped into sequences according to subject area, but they are not in numerical order (i.e., Table B00001 is not in sequence file 0001).
  • Tables with race iterations are grouped in the same sequence.
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 his/her 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 or SAS Dataset, lists Table IDs associated with each sequence number. This spreadsheet, formerly known as "merge_5_6", is available at www2.census.gov/acs2009_5yr/summaryfile/.

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 "0003". 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 "0003".



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 a position based format. A geography file exists for each state or state level equivalent.

Geography files are named using the following convention:



File Name: g20095ak.txt
Example Name Range or Type
g File Type e=estimate, m=margin of error, g=geography
2009 Reference Year ACS data year (last year of the period for multiyear periods)
5 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.

2.5. Layout of the Geography File
The following table provides the layout of the geography file:

Data Dictionary Reference Name Description Field Size Starting Position Geographic Summary Levels For Single-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 Reserved for future use
STATE State (FIPS Code) 2 26 040, 050, 060, 067, 070, 080, 140, 150, 155, 160, 170, 172, 230, 260, 269, 270, 280, 283, 286, 290, 311, 312, 313, 315, 316, 320, 321, 322, 323, 324, 331, 333, 336, 338, 340, 341, 345, 346, 351, 352, 353, 354, 356, 357, 358, 360, 361, 362, 363, 364, 365, 366, 500, 510, 550, 610, 612, 620, 622, 795, 950, 960, 970
COUNTY County of current residence 3 28 050, 060, 067, 070, 080, 140, 150, 155, 270, 313, 316, 322, 324, 353, 354, 357, 358, 362, 363, 365, 366, 510, 612, 622
COUSUB County Subdivision (FIPS) 5 31 060, 067, 070, 080, 354, 358, 363, 366
PLACE Place (FIPS Code) 5 36 070, 080, 155, 160, 172, 269, 312, 321, 352, 361
TRACT Census Tract 6 41 080, 140, 150
BLKGRP Block Group 1 47 150
CONCIT Consolidated City 5 48 170, 172
AIANHH American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 4 53 250, 251, 252, 254, 260, 269, 270, 280, 283, 286, 290, 550
AIANHHFP American Indian Area/Alaska Native Area/ Hawaiian Home Land (FIPS) 5 57 Reserved for future use
AIHHTLI American Indian Trust Land/ Hawaiian Home Land Indicator 1 62 252, 254, 283, 286
AITSCE American Indian Tribal Subdivision (Census) 3 63 251, 290
AITS American Indian Tribal Subdivision (FIPS) 5 66 Reserved for future use
ANRC Alaska Native Regional Corporation (FIPS) 5 71 230
CBSA Metropolitan and Micropolitan Statistical Area 5 76 310, 311, 312, 313, 314, 315, 316, 320, 321, 322, 323, 324, 332, 333, 341
CSA Combined Statistical Area 3 81 330, 331, 332, 333, 340, 341
METDIV Metropolitan Statistical Area-Metropolitan Division 5 84 314, 315, 316, 323, 324
MACC Metropolitan Area Central City 1 89 Reserved for future use
MEMI Metropolitan/Micropolitan Indicator Flag 1 90 010, 020, 030, 040, 314, 315, 316, 323, 324
NECTA New England City and Town Area 5 91 335, 336, 337, 338, 345, 346, 350, 351, 352, 353, 354, 355, 356, 357, 358, 360, 361, 362, 363, 364, 365, 366
CNECTA New England City and Town Combined Statistical Area 3 96 335, 336, 337, 338, 345, 346
NECTADIV New England City and Town Area Division 5 99 355, 356, 357, 358, 364, 365, 366
UA Urban Area 5 104 400
BLANK   5 109 Reserved for future use
CDCURR Current Congressional District *** 2 114 500, 510, 550
SLDU State Legislative District Upper 3 116 610, 612
SLDL State Legislative District Lower 3 119 620, 622
BLANK   6 122 Reserved for future use
BLANK   3 128 Reserved for future use
BLANK   5 131 Reserved for future use
SUBMCD Subminor Civil Division (FIPS) 5 136 067
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, 321, 352, 361
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 200 219 All Summary Levels
BLANK   50 419 Reserved for future use

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. Below is a screenshot of 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.

2.6. 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: e20095ak0003000.txt
Example Name Range or Type
e File Type e=estimate, m=margin of error, g=geography
2009 Reference Year ACS data year (last year of the period for multiyear periods)
5 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
3 Sequence Number 0001 to 9999
0 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 fields contain 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. 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.



2.7. Instructions for Applying Statistical Testing to 2005-2009 ACS 5-Year Data
This document provides some basic instructions for obtaining the ACS standard errors needed to do statistical tests, as well as performing the statistical testing for multiyear estimates.

In general, ACS estimates are period estimates that describe the average characteristics of the population and housing over a period of data collection. For example, the 2009 ACS 1-year estimates are averages over the period from January 1, 2009 to December 31, 2009 because this is the period of time for which sample data were collected. Similarly, multiyear estimates are averages of the characteristics over several years. For example, the 2005-2009 ACS 5-year estimates are averages over the period from January 1, 2005 to December 31, 2009. Multiyear estimates cannot be used to say what was going on in any particular year in the period, only what the average value is over the full time period.

More information regarding the 2005-2009 5-year ACS data products may be found in the Multiyear ACS Accuracy document at:
www.census.gov/acs/www/data_documentation/documentation_main/

2.7.1. Obtaining Standard Errors
Where the standard errors come from, and whether they are readily available or users have to calculate them, depends on the source of the data. If the estimate of interest is published on American FactFinder (AFF), then AFF should also be the source of the standard errors. Possible sources for the data and where to get standard errors are:

  1. ACS data from published tables on American FactFinder
All ACS estimates from tables on AFF include either the 90 percent margin of error or 90 percent confidence bounds. The margin of error is the maximum difference between the estimate and the upper and lower confidence bounds. Most tables on AFF containing single-year or multiyear ACS data, display the margin of error.

Use the margin of error to calculate the standard error (dropping the +/- from the displayed value first) as:

Standard Error = Margin of Error / Z

where Z = 1.645 for 2006 and later years as well as all multiyear estimates and Z = 1.65 for 2005 and earlier years.

If confidence bounds are provided instead (as with most ACS data from 2004 and earlier years), calculate the margin of error first before calculating the standard error:

Margin of Error = max (upper bound - estimate, estimate - lower bound)
All published ACS estimates use 1.645 (for 2006 and later years) or 1.65 (for 2005 and previous years) to calculate 90 percent margins of error and confidence bounds. Other surveys may use other values.

  1. ACS Public Use Microdata Sample (PUMS) tabulations
Using the methods described in the Accuracy of the PUMS documentation users can calculate standard errors for their tabulations using a design factor method or a replicate weight method. For example, 2005-2009 Accuracy of the PUMS documentation can be used with the 2005-2009 ACS PUMS file to calculate standard errors. This document is located at:
www.census.gov/acs/www/data_documentation/pums_documentation/

NOTE: ACS PUMS design factors should not be used to calculate standard errors of full ACS sample estimates, such as those found in data tables on AFF. In addition, Census 2000 design factors should not be used to calculate standard errors for any ACS estimate.

2.7.2. Obtaining Standard Errors for Derived Estimates
Once users have obtained standard errors for the basic estimates, there may be situations where they create derived estimates, such as percentages or differences that also require standard errors.
All methods in this section are approximations and users should be cautious in using them. They may be overestimates or underestimates of the estimates standard error, and may not match direct calculations of standard errors or calculations obtained through other methods.

  • Sum or Difference of Estimates


  • Proportions and Percents
Here a proportion is defined as a ratio where the numerator is a subset of the denominator, for example the proportion of persons 25 and over with a high school diploma or higher.
Let:



If the value under the square root sign is negative, then instead use



If P = 1 then use



  • Means and Other Ratios
If the estimate is a ratio but the numerator is not a subset of the denominator, such as persons per household or per capita income, then
  • Products
For a product of two estimates - for example if a user wants to estimate a proportions numerator by multiplying the proportion by its denominator - the standard error can be approximated as


Users may combine these procedures for complicated estimates. For example, if the desired estimate is

then SE(A+B+C) and SE(D+E) can be estimated first, and then those results used to calculate SE(P).

For examples of these formulas, please see the 2005-2009 Accuracy of the Data, available at www.census.gov/acs/www/data_documentation/documentation_main/.

2.7.3. Instructions for Statistical Testing
Once standard errors have been obtained, doing the statistical test to determine significance is not difficult. The determination of statistical significance takes into account the difference between the two estimates as well as the standard errors of both estimates.

For two estimates, A and B, with standard errors SE(A) and SE(B), let



If Z 1.645, then the difference between A and B is significant at the 90 percent confidence level. Otherwise, the difference is not significant. This means that there is less than a 10 percent chance that the difference between these two estimates would be as large or larger by random chance alone.

Users may choose to apply a confidence level different from 90 percent to their tests of statistical significance. For example, if Z 1.96, then the difference between A and B is significant at the 95 percent confidence level.

This method can be used for any types of estimates: counts, percentages, proportions, means, medians, etc. It can be used for comparing across years, or across surveys. If one of the estimates is a fixed value or comes from a source without sampling error (such as the Census 2000 SF1), use zero for the standard error for that estimate in the above equation for Z.

Using the rule of thumb of overlapping confidence intervals does not constitute a valid significance test and users are discouraged from using that method.