Now that the Census Bureau is busy releasing the 2006 ACS data, I
thought it would be timely to post a few links on using the ACS data.
It is VERY IMPORTANT to understand that the biggest difference between
the 2006 and 2005 ACS is that GROUP QUARTERS population was ADDED in the
2006 sample. Areas with large Group Quarters population will see the
greatest differences between 2006 and 2005 results. Workers who live in
Group Quarters are more likely to walk to work, so you may see shifts in
distributions of means of transportation to work in these areas. The
population threshold for reporting 1-year ACS data is still 65,000
persons, based on place of residence. The 2005 and 2006 ACS standard
tabulation include tabulations for PLACE OF WORK called "for workplace
geography". Also, the "key" to finding tables on "journey to
work" is
"08", for example: B08###, or C08###, or S08##.
Cynthia Taeuber's book, "American Community Survey data in Community
Planning" (Trafford Publishing, 2006 website:
trafford.com/06-2809 ) is an easy-to-understand document that covers
basic information about ACS, but most importantly, understanding
sampling error and confidence intervals. Note: The Census Bureau is
now using the term "Margin of Error" to reflect sampling error instead
of using confidence intervals (the estimate with an upper and lower
bound).
You can order it directly from
Trafford.com, which is a "print on
demand" publishing house, but I just found it on Amazon.
(priced at $26.37 and listed as "in stock" )
Also, Cynthia's material completed for Brookings Institute "for
journalists" is on-line at:
http://www.brookings.edu/metro/umi/events/20061115_ACSreferenceguide.pdf
This document on the Census Bureau's webpage discussing the 2005 ACS is
also useful.
http://www.census.gov/acs/www/UseData/advance_copy_user_guide.pdf
The FHWA page also includes some material on using ACS data, which was
developed for earlier releases of ACS, but are still relevant.
http://www.fhwa.dot.gov/planning/census/acs.htm
Because we prepared profiles sheets from the 2005 ACS data, with
comparison to Census 2000, we are currently NOT planning to issue NEW
profiles using the 2006 data. We are currently focusing on PLANS for
the NEXT CTPP, using the first 3 year accumulation of ACS data (2005,
2006, and 2007), and conducting research on 1) improving workplace
imputation for ungeocoded survey responses, and 2) alternative
approaches to disclosure avoidance, to avoid the data suppression
problems that arose in the CTPP2000.
If you have ideas for the 3-year CTPP product, especially NEW or
DIFFERENT Tables, please let me know!
Elaine Murakami
FHWA Office of Planning
206-220-4460