Hi Everyone:

I asked Andy Pickard from the Hampton Roads MPO and Kristen Rohanna from SANDAG to look at 2006 ACS, because their regions have substantial Group Quarters populations (military).  For single year ACS reporting, the threshold of 65,000 population must be reached before tabulation. 


What did they find?   I am attaching 3 documents:  one from Andy, one from Kristen and one from me (US Total)


To estimate the workers in Group Quarters by Means of Transportation to Work, you can use B08006 Sex by Means of Transportation to work (all workers) and subtract B08141 Means of Transportation to work by Vehicles Available (workers in households since Vehicles Available is a household variable).   (Please keep in mind that the 2006 ACS is the FIRST year that Group Quarters were surveyed and the sample is small.)


Based on previous data, we would expect to see MORE walk to work for workers in group quarters.  However, many respondents in Group Quarters are showing up as “worked at home”.  My guess is that this is an artifact of different question wording on the Group Quarters ACS questionnaire Q26 that says “worked AT THIS ADDRESS” rather than “worked AT HOME”.   So, the category “worked at home” is “worked at home” (for households) plus “worked at this address” (for group quarters).   


For the Census 2000 Military form Q21 used the SAME PHRASE “worked at home” as the housing unit questionnaire.



So, let’s say I live and work at Fort Lewis in Pierce County, Washington.  My address would be Fort Lewis and I live in a barracks on the Fort and WALK to a different building at Fort Lewis.  I would likely select  “WORKED AT THIS ADDRESS” rather than “walk.”    


If you have done other work using the 2006 ACS, please share your results with the CTPP listserv.  Thanks!


Elaine Murakami

FHWA Office of Planning 



I did some analysis of the 2006 ACS data, with and without Group Quarters. Like Andy, we are finding an increase of those who walk to work with the inclusion of GQs. Additionally, “drove alone” is much higher when you don’t include GQs. The change between 2005 and 2006 is significant for those that “drove alone” to work, UNTIL you use the household workers in 2006; then, the change is not significant.


I’ve attached a spreadsheet with my calculations. The “analysis” worksheet explains what I found….

I hope this helps! Please let me know if you need anything else.


Kristen Rohanna

Associate Research Analyst

San Diego Association of Governments (SANDAG)


From: Andrew PICKARD [mailto:apickard@hrpdcva.gov]
Sent: Wednesday, October 31, 2007 2:00 PM
To: Murakami, Elaine
Cc: Srinivasan, Nanda <FHWA>; Rob CASE; kroh@sandag.org; Paul Agnello
Subject: Re: 2006 ACS data with group quarters


Elaine -

Attached is a pdf w/ some comparisons for you.  The first page I just created and is a quick look at our MSA and the city of Norfolk, which has ~10% of its population in group quarters [I would have included a summary for Williamsburg as well (which has about 5k gq pop of 12k total pop in 2000 Census) but they aren't included in ACS reports yet].  The second page is an analysis I presented to a CTPP/ACS/NHTS workshop that VDOT, Ed, and Nanda put together last May in northern Va. 


A couple general observations:

 - drove alone increased about 2% region-wide and possibly up to ~10% in Norfolk w/out gq pop

 - bike or walk decreased over 1% region-wide w/out gq pop

 - travel time to work increased by 0.3 minutes w/out gq pop

 - the number of walkers about cut in half w/out gq pop

 - pub transit decreased about 5k w/out gq pop

 - work at home is much higher in 2006 ACS region-wide and for Norfolk (not sure why this might be.  any ideas?)