The information presented below came from mine and Elaine Murakamis
notes of a meeting we attended concerning the American Community
Survey. While many issues were discussed, the items presented below
stood out in our minds and may be of interest to the transportation
community. The agenda is attached.
On April 14th the Census Bureau hosted an invited meeting for the
transportation community to discuss the American Community Survey.
There were approximately 31 people in attendance, 18 from the Census
Bureau and 13 from the transportation community. The meeting took place
at the Department of Commerce Building in Washington D.C.
ACS questionnaire items
The Census Bureaus goal is to have the ACS from 2008 to 2012
approximate a 2010 long form. Thus, the ACS questionnaire items must be
locked down over this period. Any recommendations for changes to the
questionnaire need to be tested in the 2006 Content Test. Any cognitive
tests would need to occur in 2005. Any recommendations for additions
must be submitted to the OMB interagency group on ACS, and must include
statutory justification. Congress wants the questionnaire to be as
short and with the least burden as possible.
(As background: DOT submitted justification for the current version of
ACS, which retains all the journey-to-work questions and household
vehicle availability question. The DOT report can be found at
The CB plans to issue a Federal Register Notice within the next 30 to 45
days, with a list of tabulations that they plan to publish from the ACS.
These would be standard products from the ACS, as opposed to CTPP
tabulations which are considered a special tabulation, paid for by the
transportation community. Standard products come out quickly and the
tables are not subject to rounding.
We informed the CB at this meeting that we would like them to consider
adding workplace tabulation and flow tabulation to their standard
tabulation, which is currently limited to residence geography. We
provided a DRAFT of tables we would like considered, as well as a few
changes to the current list of standard tabulations available on the ACS
webpage. Appended at the end of our meeting notes is a list of the
new tables we would like considered. We reiterated that users of
workplace and flow tabulations were not limited to transportation
planners, but included emergency/evacuation planners, environmental
planners, employment/labor analysts, and the marketing community.
Using a Federal Register Notice for public participation in table design
differs considerably from previous Census Bureau activities to design
the SF3 tables for 2000, and STF3 tables in 1990.
The Census Bureau said they plan to implement at least 4 different rates
of Non-Response Follow-Up (NRFU) in the implementation of ACS starting
around January 2005. The ACS follow-up currently targets one-in-three
housing units, that is, if the Census Bureau does not receive a
mail-back form, 1 of every 3 non-respondents is contacted, and the other
2 of 3 are left as missing. Weighting is used to bring the responses to
Because the Census Bureau cannot increase the total number of samples to
be followed up without incurring additional costs, they plan to base
their NRFU on historic trends in mail-back returns. That is, those
census tracts with the lowest mail-back returns will have the highest
proportion of NRFU. The goal is to have results with approximately the
same statistical quality over space.
Mail-back rate (NRFU rate)
Better than Average (Less than 1:3)
The Census Bureau plans to use the 2000 Public Use Microdata Areas
(PUMAs) as a standard tabulation geography for ACS. This means that
large jurisdictions with population over 250,000 will have a close
approximation of subareas within their jurisdiction using 3-year
averages, rather than waiting for 5-year averages. There are no plans
to change or redefine PUMAs until 2010. However, because the
transportation data community was often not involved in the definition
of the 2000 PUMAs, an earlier redefinition might make sense for the
User Defined Statistical Areas
We asked if the CB would allow for a new layer of geography that was
smaller than a jurisdiction, for jurisdictions with population larger
than 250,000. While the CB acknowledged the notion of User Defined
Statistical Areas, staff quickly moved the focus to the use of PUMAs.
Tables Proposed for Consideration as Standard Tables in ACS
Residence Based Tables
--Household Size (4) by Vehicles Available (5)
--Household Size (4) by Number of Workers in Household (4)
--Number of Workers in Household (4) by Vehicles Available (5)
--Means of Trans. to Work for Workers 16+ (5) by Vehicles Available (5)
--Means of Trans. to Work for Workers 16 and over (13)
--Travel Time to Work for Workers 16+ (12)
--Vehicles Available (5) by Means of Trans. to Work for Workers 16+ (6)
--Household Income (16)
Journey to Work Flow Tables
--Number of Workers
--Means of Trans. to Work for Workers 16+ Who did not Work at Home (4)