Skip to main content

New AHS 2001 File

HUD.GOV HUDUser.gov
eList
AHS
There is a new version of the 2001 AHS microdata on the HUD USER web site
at https://www.huduser.gov/portal/datasets/ahs.html.  Here is the Census
Bureau descripton of the change:

"Based on the situation reported by a data user, we re-examined the way
that the alterations and remodeling data were processed.  The problem was
that the summary recode (RAN) for the number of jobs completed by a
household sometimes did not match the number of individual job records for
the same household.

"This occurred because slightly different criteria were used to tally a job
in the counter -- that is, the variable, RAN -- than were used to output a
job record (which would include the variables RAS, RAD and RAH).  The
differences were limited to a few job types -- those related to carpeting,
other inside and other outside jobs, and disaster-related work.  Sometimes
more job records were generated than tallied in RAN and sometimes the
reverse occurred.  However, the number of inconsistencies was not large.

"The programs have been rewritten to make the value in RAN agree with the
number of job records.  This change will affect RAN (total number of jobs
done) and RAC (total amount spent).  In addition, the number of RAS codes
of 1 (disaster loss), 51 (carpeting added over subflooring), 52 (carpeting
added over finished flooring), 64 (other major inside improvements) and 70
(other major outside improvements) may change along with the related
variables of RAD (cost of individual job) and RAH (who did the work for the
individual job)."

We at HUD have prepared a new ASCII version, a new documentation archive,
and an update package. However, at the moment only the full SAS version of
the new dataset is on the web site.  The other files should be appearing
shortly.  I will send announcements to the mailing list when that happens.

These changes affect only alterations and repairs.  If you don't work with
that part of the dataset, you may want to simply note this change and
update your files later, when you need these data.  While we don't know of
any other problems in the current dataset, experience teaches that another
update or two would not be out of the ordinary.

Our thanks to Alvaro Martin-Guerrero, of the Joint Center at Harvard, for
discovering the problems.

Dav Vandenbroucke
Economist
U.S. Dept. HUD
david_a._vandenbroucke@hud.gov
202-708-1060 ext. 5890