Subcommittee on Census Data For Transportation Planning
Below is a reproduction of the letter sent to the Census Bureau on behalf of the USDOT by Ashish Sen, Director of the Bureau of Transportation Statistics, USDOT .  Signed paper copies are available by contacting the CTPP Working Group at

June 16, 2000

Kenneth Prewitt
U.S. Census Bureau
Federal Office Building 3,  Room 2049
4700 Silver Hill Road
Suitland, Maryland  20233

Dear Mr. Prewitt:

I was surprised to learn that the Census Bureau was planning major changes to the PUMS coding and aggregation.  While we understand that some changes may be necessary, we would like to see the Bureau work more closely with the transportation planning community in developing any changes.

There are 350 Metropolitan Planning Organizations (MPOs)  {49 U.S.C. Section 5303(c)} and 51 states including the District of Columbia that are required to prepare regional transportation plans under Federal Law.  These plans rely on forecasted travel that is derived from a series of models, which rely in part on decennial Census data.  As a result, there is a very large planning infrastructure built around the census data products.

Aside from the "larger" transportation planning communities' uses and concerns, there is one project that may be impacted that is of particular interest to DOT.  Currently, the DOT and EPA are working on a $40 million effort to replace the traditional four-step travel demand system used by MPOs.  The new system is based on a microsimulation methodology called TRANSIMS.  Being a microsimulation, it relies on microdata and the designers have assumed the availability of "unchanged" Census 2000 PUMS data.  Given the magnitude of this project, it is important that the responsible DOT and Census Bureau staff come together and discuss the current proposal, evaluate any impacts on this project and assure that we are not working at cross purposes.

Specific to the PUMS proposal we heard about at a May 22nd meeting, we would prefer not to see "across the board" rounding of travel time or departure times.  Although, we understand that some rounding at the tail end of the distributions may be necessary, we do not see how exact data reporting would adversely effect confidentiality especially where the sample is very rich.

Regarding micro area definitions, we were happy to see that the Census Bureau is proposing to keep 100,000 as the population threshold for PUMAs in state-level PUMS.  Keeping the geographic unit as small as possible is very important to the transportation community, especially in evaluating travel mode choice, and household acquisition of vehicles.  However, in terms of defining PUMAs, we would like to work with your staff to bring the transportation community closer to process of defining PUMAs, specifically in urban areas.

Overall, I believe that DOT and Census Bureau have had an excellent working relationship.  I also believe that it is important that our respective staff keep this cooperative spirit alive.  To this end, we look forward to furthering our discussions on PUMS and facilitating a dialogue between the Census Bureau and the "larger" transportation planning community on this matter.

My staff contact on this project is Ed Christopher (202-366-0412).

Ashish Sen

        cc: William G Barron, Deputy Director

Back to Subcommittee HomePage
Comments? Send to Ed Christopher (subcommittee chair)