AHS Digest

Here are the AHS Digest messages:


= = = = = = = = = = = =

From: David_A._Vandenbroucke@HUD.GOV on 12/11/2002 3:30:00 PM
Subject: AHS 2001 Public Use File Update

There is a new public use dataset for 2001 national AHS available for
download from the HUD USER web site, at:

https://www.huduser.gov/portal/datasets/ahs.html

The Census Bureau and HUD released a new dataset in order to add a
variable, WGT90GEO.  No existing variables have been changed.  The new
variable is an alternative weight for AHS records.  This new weight is
adjusted for 1990 metropolitan geography.  The standard weight variable,
WEIGHT, is adjusted for 1980 geography, in order to be consistent with
previous surveys.  Both weights sum to national control totals as of 2001.
Thus, totals for all units will match when computed with both sets of
weights, although occupied unit totals do not match.

We recommend that the 1990 geography-based weight be used for matching
numbers from the public use file (PUF) with numbers in the publication at
the US and Census region level. For historical comparisons, use the 1980
geography-based weights, as these are comparable to previous public
use files.  Use the standard (1980 geography-based) weights for current
comparisons of geographic areas,  because only 1980 geography is identified
on the 2001 PUF.

For more details about how weights are computed, see Appendix B of the
publication.  Also, Table V in Appendix C compares some key tabulations
using the two sets of weights.

The web site current has the new full dataset in SAS and ASCII formats.  It
also has a new documentation file which gives basic tabulations using both
sets of weights.  We will soon publish an update package, so that those of
you that already have the dataset can save yourselves the trouble of
downloading the whole thing again.

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




= = = = = = = = = = = =

From: "Jill Meron" <Jmeron@nmhc.org> on 12/13/2002 12:48:09 PM
Subject: New weights

Has anyone successfully replicated table 1A-1 for units in structure of
all housing units?
I have a discrepancy between the published results and my tabulations
using the new weights.
Total housing units 119,117 is fine as are total occupied units,
116,038, and owner occupied and renter occupied, 72,265 and 33,996
units.
My problem arises when tabulating by units in structure for those units
in buildings with two or more apartments.
These are the differences I recorded tabulated for all housing units:
2 to 4 units in structure - I get 9,276 instead of 9,354
5 to 9 - 5,633 instead of 5,682
10-19 - 5,309 instead of 5,367
20-49 - 3,901 instead of 3,898
50+ - 4,266 instead of 4,084
The totals sum to the published numbers but the breakdowns differ.
Does anyone have any ideas?
Thanks,

Jill Meron
Research Analyst
National Multi Housing Council
202 974 2343


- - - - - - - - - - - -

To stop receiving the digest, send a message with DIGEST OFF in the subject.