Wouldn't it be nice if we could find the perfect travel behavior
dataset? A lot of us are suffering from the ACS small sample sizes that
result in large margins of error, especially when trying to look at
tract geography. The ACS results are not perfect either and rely on
self-reported addresses for the workplace. People who work in
construction may report their work location as the place from where they
get their check, and not at the actual construction site. Similarly,
contract employees may report an address for their home office, even if
they work at another location on a daily basis.
One of the advantages of the LEHD OTM is that the number of primary job
records is very large, resulting in a very large number of small area
flows. Understanding the source of the work location and home location
In the LEHD OTM, the quality of the workplace coding varies widely from
state-to-state. While the workplace is reported using administrative
records (values are from the Quarterly Census of Employment and Wages
and Multiple Worksite Reports), there have been many instances of many
workers coded to a headquarters office. Even in states that "require"
multiple worksite reporting, there is often lack of staff resources at
State ESD offices to work on improving this list.
For example, in the December 2010 LEHD Webinar
, Robert McHaney
reported that for a project in downtown Austin, two blocks were
identified as having "very high" employment densities. These 2
locations were found to be a post office regional office, and a school
district headquarters office. After editing those 2 blocks, the
employment in the area was reduced from over 60,000 to about 33,000. As
far as I know, there was no attempt at moving those 27,000 workers to
another location, as the project was only concerned about downtown
Austin. I also don't know how they subtracted the home ends for those
27,000 workers from the flow tabulation, or perhaps the project did not
need this information.
Also, previous research has shown problems with state government
employment most likely to be coded to a state capitol address, rather
than coded to a local office. So, depending on the area in which you
are working, this may be more or less of a problem.
The LEHD OTM data synthesis process includes assigning "workers" (job
holders) to an employer location. If the employer has multiple work
sites, then the worker is linked to a specific worksite based on an
assignment algorithm using data from Minnesota, as Minnesota requires
employers to link workers to a specific location. If there are no
entries in the multiple worksite report, then the worker is linked to
the single address in the QCEW.
The home address of the worker generally is taken from their IRS record.
This can be problematic for people like college students who have jobs
at school, but continue to use their parent's address for IRS forms, and
similarly, some snowbirds may take part-time jobs in their wintering
location, and keep their summer address for the IRS.
In addition, remember that the QCEW includes "covered employment" that
is jobs covered by unemployment insurance. Therefore, self-employed and
some other categories of workers are not included. This is generally
estimated at about 10% of all workers.
So, no matter which dataset you use, try to understand the source of the
data and how the data have been (possibly) synthesized, and
check/validate the results before you use them!
FHWA Office of Planning
[mailto:email@example.com] On Behalf Of Ross Friedman
Sent: Tuesday, May 03, 2011 1:56 PM
Subject: Re: [CTPP] looking for journey to work data newer than 2000
LEHD (Longitudinal Employment Household Dynamics) is a synthetic data
set that is provided at the block level. However, noise is added to it
so it is recommended that one aggregate the data up to TAZ or Tract
level. There is an origin-destination file which gives an approximate
route of workers. The data is accessible on the Census website at:
(OntheMap Data) Currently, years 2002
through 2009 are available on this website. Hopefully, this will help
with what you are looking for.
Ross Marc Friedman
>> Nancy Reger <Nreger(a)morpc.org>
05/03/2011 3:34 PM >>>
Does anyone know a source for journey to work data
for small geographies
(any geography really) post 2000 census?
Nancy Reger, AICP
Deputy Director, Transportation
Mid Ohio Regional Planning Commission
111 Liberty St, Ste 100, Columbus OH 43215
P: (614) 233-4154 /E: nreger(a)morpc.org
ctpp-news mailing list