Skip to main content

RE: Revised 1999 AHS File

HUD.GOV HUDUser.gov
eList
Does this imply that there is also a problem with the 2001 national AHS
microdata file?  It seems that if, say, FRSTOC was incorrectly coded in 1999
and the same household was still in the unit in 2001, there's a chance the
program would perpetuate the 1999 error into the 2001 data. I, and many
others, look forward to any insight you can offer.

---------------------------------------------------------------------------
Here is an explanation of the change from the Census Bureau:

SUMMARY OF CHANGES TO 1999 DATA

THE PROBLEM:

A processing error was discovered which affected the 1999 data.  In every
1999 case where we had a valid entry in the 1997 survey we over wrote the
1999 answer with the 1997 answer for the variables WHNGET, FRSTOC, CUSHOM,
PREOCC, LPRICE, CPRICE, DWNPAY AND FRSTHO.  We were ONLY supposed to take
the 1997 answer if we had the same housing unit and household in 1999.
These restrictions were not applied.  The 1997 answer was taken even if we
were dealing with a different housing unit and/or household.  The new micro
data public use file (PUF) corrects this problem.

RESULTS OF CORRECTION:

The data for the following owner occupied variables (all from the RET
module) were purposely changed as a result of this correction.

WHNGET  FRSTOCC  CUSHOM  PREOCC  LPRICE CPRICE DWNPAY FRSTHO

The changes in the data for these variables were significant, especially
the changes to WHNGET.  Any previous analysis done using these variables
should be redone.

The data for the following owner occupied variables also changed.  Changes
in these variables where relatively minor.  Previous analysis using these
variables would most likely not change.  All are from the mortgage module.
The changes in these variables were the result of edits in the mortgage
module which depended on the data in the above mentioned purposely changed
RET variables (in most cases the variable WHNGET).  These variables
include.

PMT   PMT2  AMMORT  AMMRT2  MATBUY  MATBY2  NEWMOR  NEWOR2  TERM TERM2
YRMOR  YRMOR2   AMRTZ   AMRTZ2  INTW  INTW2  INTF  INTF2

The following is a list of publication (from the 1999 H-150 publication)
owner occupied items and recodes that change as a result of variable
changes mentioned above.  Beside each item is listed the variables causing
the differences in the data along with an assessment of the severity of the
change.  Minor means that any previous analysis using these data should not
be affected using the new data.  Significant means that the differences
appear to be statistically significantly and that  overall analysis using
these data may be affected.  Major means that any previously done analysis
using the old data will be affected should be redone using the new data.

Monthly Housing Costs-due to PMT and PMT2 changes---Minor

Monthly Housing Costs as Percent of Current Income-due to PMT and PMT2
changes--Minor

Previous Occupancy-due to changes in FRSTOC-Minor

Monthly Payment for Principal and Interest-due to changes in PMT and
PMT2-Minor

Year Unit Acquired-due to changes in WHNGET-Major

First Time Owners-due to changes in FRSTHO-Significant

Purchase Price-due to changes in CUSHOM PREOCC CPRICE LPRICE-Significant

Major Source of Down Payment-due to changes in CUSHOM PREOCC
DWNPY-Significant

Mortgage Origination-due to changes in MATBUY - Minor

Year Primary Mortgage Originated-due to changes in MATBUY, YRMOR and
WHNGET-Major

Term of Primary Mortgage at Origination or Assumption-due to changes in
TERM-Minor

Remaining Years Mortgaged-due to changes in MATBUY, TERM, WHNGET,
YRMOR-Minor

Current Interest Rate-All the changed mortgage variables except for PMT and
PMT2 go into this recode-Significant

Total Outstanding Principal Amount-All the changed mortgage variables
except for PMT and PMT2 go into this recode-Significant

Current Total Loan as Percent of Value-All the changed mortgage variables
except for PMT and PMT2 go into this recode-Minor

There are a total of 67 pages in the H-150 publication that have errors on
them.  As you can see from the discussion above, most of these errors are
minor.

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

We are still discussing what changes to make in the reports.  We will
probably NOT release a new paper publication.  We almost certainly WILL
change the tables on the Census Bureau web site.

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



- - - - - - - - - - - - - - - - - - - - - - - - -
This message was forwarded to you by the Listserv ahs@huduser.gov because
you had expressed an interest.
To stop receiving these messages send an email to ahs@huduser.gov
with "unsubscribe" as the SUBJECT of your message.
To reply to all members of the list address your reply to ahs@huduser.gov.
Message archives are at https://www.huduser.gov/portal/elist/archive.html
For help send an email to helpdesk@huduser.gov