RE: AHS: Do You Need Degree Days?

From: American Housing Survey (AHS) ListServ <ahs@huduser.gov>

I have analyzed the public use AHS files using the climate zone data, especially when looking at AC/Heating equipment in relation to other variables.
Results in this article were based in part on a negative statistical relationship between housing costs and lack of built-in heating equipment that prevailed in some, but not all, climate zones.
(I apologize for the poor quality of the scanned-in pdf file; I lost the original electronic version in a hard drive crash).

PAUL EMRATH
Vice President
Survey and Housing Policy Research

National Association of Home Builders
1201 15th Street, NW
Washington, DC 20005
F 202 266 8333 T 800 368 5242 x8449
pemrath@nahb.com

Everything you need to know at www.nahb.org
International Builders’ Show / January 12-15, 2011 / Orlando, FL

On Thu, Aug 26, 2010 at 10:12 AM, AHS wrote:

From: American Housing Survey (AHS) ListServ <ahs@huduser.gov>

The original goal of having degree days in AHS data was to enable and encourage analyses of housing characteristics in different climates, and the relationship among energy use, clmate, and housing characteristics

The categories of degree days, ranging from very cold to hot, show the rough climate zone for each unit.

The national books show several items by climate zone, such as utility costs. Books can continue to do so even if zone is not on the public use files.

Climate can certainly be helpful for studying how utility use is related to housing characteristics, controlling for climate, or the range of energy use for people within the same climate and similar housing. This kind of study will be important in encouraging people to sue less energy.

Simple data on utility spending by climate zone, I believe were and probably still are available from DOE.

Data on degree days in each market are available from NOAA, in much more detail than from AHS, as others have noted on this list.

The old reason to have climate zone rather than states on the public use files was that the sample design usually covered only one or a few areas per state, not representative of the whole state. How well will the 2015 sample design represent each state?

Paul Emrath is right that blurring is a good approach to confidentiality. HUD's and the Census Bureau's research shows that revealing areas of 100,000 population does NOT protect confidentiality, and proper blurring does (e.g. the 2001 book: Confidentiality, disclosure, and data access : theory and practical application for statistical agencies by Pat Doyle et al.)

It would be useful to hear from people who have analyzed the public use files, using the climate zone data, if anyone has.

Paul Burke
------------------------------------

On Tue, Aug 24, 2010 at 10:31 AM, AHS wrote:

From: American Housing Survey (AHS) ListServ <ahs@huduser.gov>

Not having DEGREE would make it a lot more difficult to evaluate a unit’s heating equipment (e.g., use of room heaters or portable electric heaters as main heating equipment may be a sign of an inadequate unit and possibly a health/safety concern, depending on the climate).

It would also make an energy efficiency module close to useless, at least from the standpoint of a public use data set.

Wouldn’t it be possible to retain the DEGREE variable, but blur the data on the public use file to preserve confidentiality?

PAUL EMRATH
Vice President
Survey and Housing Policy Research

National Association of Home Builders
1201 15th Street, NW
Washington, DC 20005
F 202 266 8333 T 800 368 5242 x8449
pemrath@nahb.com

Everything you need to know at www.nahb.org
International Builders’ Show / January 12-15, 2011 / Orlando, FL

On Mon, Aug 23, 2010 at 2:26 PM, AHS wrote:

From: American Housing Survey (AHS) ListServ <ahs@huduser.gov>

If you drop this variable, How are we supposed to get this information. Will it be available at another government site?

On Fri, Aug 20, 2010 at 2:36 PM, AHS wrote:

From: American Housing Survey (AHS) ListServ <ahs@huduser.gov>

The American Housing Survey microdata files include a variable, DEGREE, which indicates the heating or cooling degree days of the housing unit's local area. The codebook defines the variable this way:

DEGREE =
Average heating/cooling degree days
1 Coldest: 7,001+ heating degree days and < 2,000 cooling degree days
2 Cold: 5,500-7,000 heating degree days and < 2,000 cooling degree days
3 Cool: 4,000-5,499 heating degree days and < 2,000 cooling degree days
4 Mild: < 4,000 heating degree days and < 2,000 cooling degree days
5 Mixed: 2,000-3,999 heating degree days and 2,000+ cooling degree days
6 Hot: < 2,000 heating degree days and 2,000+ cooling degree days

HUD is considering dropping this variable when we redesign the survey and draw a new sample for the 2015 AHS. The reason is that having DEGREE on the file interferes with our ability to identify certain kinds of geography, such as states and Census divisions. Degree is set at the county level. There are certain Census divisions (and, even more so, states) where the combination of one of the degree codes and the division code identifies a section of the division whose population is less than 100,000. Under the Census Bureau's confidentiality rules, such small areas may not be identifiable on a public use file. Thus, we will not be able identify Census divisions on our public use file. As for states, there will certainly be states that are too small to identify no matter what. However, the presence of the DEGREE variable makes the job even more difficult.

Thus, HUD is proposing that we drop DEGREE from the survey beginning in 2015, so that we can identify Census divisions and at least some of the larger states. If any data users would be seriously inconvenienced by the loss of DEGREE, we would like to hear from you.

Dav Vandenbroucke
Senior Economist
U.S. Dept. HUD
david.a.vandenbroucke@hud.gov
202-402-5890

I disclaim any disclaimers.

--
Laura Montllor, AIA

Executive Director
HomeFreeHome

516.883.0403
http://homefreehome.org/

66 Hillview Avenue
Port Washington, NY 11050