Skip to main content

AHS: Updated 2005 Public Use File

HUD.GOV HUDUser.gov
eList
AHS

An updated version of the AHS 2005 public use file has been posted on the HUD USER web site, https://www.huduser.gov/portal/datasets/ahs.html . Here is an extract from the release notes:

UNIT SQUARE FEET AND LOT SIZE
Because of a processing error, the topcoding for UNITSF and
LOT was done incorrectly. In the previous version of the
dataset, some records had values of 4300 for UNITSF, and some
had 10421. The correct topcoded value is 10099. Similarly, for
some records LOT was erroneously given the value 999987
instead of the correct topcode, 983834. The new file has the
correct values.

ADDITIONAL RECORDS FOR NONINTERVIEW CASES
The AHS is a longitudinal survey which can be used to trace
changes in the housing stock over time. Because of this, our
practice is to include records for housing units that are no
longer in the sample, either because the units have ceased to
exist or because of a change in the sampling strategy. The
previous version of the dataset did not include records for
some "Type C" noninterview cases. Type C cases are housing
units that were permanently deleted from the stock (for
example, by demolition). In addition to these, the previous
dataset omitted records for the manufactured (mobile) home
units that were dropped from the sample when the Census Bureau
drew a new sample of such units from the 2000 census. These
omitted records have been restored, so that users who are
interested in longitudinal analysis can trace what happened to
them. The new version of the dataset includes 9,439 additional
records. The records for the dropped manufactured housing
units have NOINT=38 “Type C - Unit eliminated in subsampling.”

Note that all Type C cases have zero weights. Thus, most
tabulations performed using the previous version of the
dataset should be unaffected by the addition of these cases.

CHANGES IN INCOME LIMITS AND FAIR MARKET RENTS
The inclusion of additional noninterview cases led to small
changes in some values for the income limit and fair market
rent variables. Specifically, some noninterviewed cases had
been inappropriately included in the calculation of minimum
and maximum income limits and fair market rents (FMRA, FMRB,
L50A, L50B). Also, a small number of interviewed cases outside
of identified metropolitan areas (that is, cases with SMSA =
9999) had been inappropriately excluded from the calculations
of minimum, maximum, and mean income limits and fair market
rents (FMR, FMRA, FMRB, L50, L50A, L50B). A total of 4,623
interviewed cases (STATUS = 1,2,3), all with SMSA=9999, change
compared to the last PUF. The FMR change is always less than
2.5 percent, and averages 0.7 percent. The changes to L50
(very low income threshold) are always less than 1 percent.

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