• Welcome to AppraisersForum.com, the premier online  community for the discussion of real estate appraisal. Register a free account to be able to post and unlock additional forums and features.

You Want Coefficients?

Status
Not open for further replies.

Elliott

Elite Member
Joined
Apr 23, 2002
Professional Status
Certified General Appraiser
State
Oregon
In the April 2003 issue of The Appraisal Journal, put out by AI,
Gary Grudnitski, PHD, did a sample with 2300 sales in
Las Vegas. He was hunting for the golf course influence,
but I think there were some interesting findings.

And, since I live in a market too small to ever correlate anything
with anything, I'm a reader of correlation rather than producer
of the statistics.

Fireplace 3.4% (per fireplace)

Bathrooms 2.7% (per bathroom)

Pool 8%

Storys (as in two) -13% (I found this one the most interesting-especially in a retirement community like LV, indicating a two story house sells 13% less than
a ranch with similar GLA)

Private golf course community 12.5%

Semi-private gc community 6%

Public gc community 5.7%


elliott
 

Jim Bartley

Senior Member
Joined
Jan 20, 2002
Professional Status
Certified Residential Appraiser
State
Virginia
Elliot,

Regarding the 2 story adjustment....Doesn't it make since that, all things being equal, a rancher will sell for more than a 2 story?
 

Elliott

Elite Member
Joined
Apr 23, 2002
Professional Status
Certified General Appraiser
State
Oregon
Jim,
Of course. I usually end up comparing two story houses
with two story houses. But I don't think I've ever made
an adjustment in design between the ranches and a two
story. I was surprised by how large the negative correlation was.

Can you imagine coming in low on a property and having made
the 10% adjustment to a two story house and getting the
call from the realtor??

elliott
 

Restrain

Elite Member
Joined
Jan 22, 2002
Professional Status
Certified General Appraiser
State
Florida
In regards to your two-story adjustment, I've seen that here in upper-end developments where the homes are owned by older people, but the adjustment is for the Master Bedroom/Bath being upstairs vs. downstairs. Otherwise, I've not found a differential. And it's only for markets with older occupancy.

Roger
 

Caterina Platt

Senior Member
Joined
Jan 17, 2002
Professional Status
Certified Residential Appraiser
State
New Mexico
Elliot,

That is interesting. Surprising though. Roger and others who are well versed in these high enders....does a fireplace really add anywhere near $17,000 and a bath $13,500 to a half million dollar dwelling???

Yowsa!!!
 

Austin

Elite Member
Joined
Jan 16, 2002
Professional Status
Certified General Appraiser
State
Virginia
I hope you guys don’t read those numbers and start making adjustments based on them, because if you do you are in for a rude awakening. For example, that two story number said 13% lower per square foot not per house. That sounds about right but the point of concern is that you can’t pay much attention to the coefficient factors. You can come up with the perfect regression equation that will predict the price of every sale in the regression perfectly indicating a perfect solution and have every coefficient number totally meaningless and incorrect. The reason is that there about 5 value influencing independent variables that determine price and if you enter 9 independent variables in the regression it will take the slack in the regression and allocate it out among the 9 variables. It doesn’t leave any factor out of the deal. If you happen to pick the correct 5 independent variables it will take the normal random variance, influence from outliner sales, and distort the value coefficients.
I took a regression class once and the instructor was teaching students to use these coefficients to make adjustments, which prompted me to write an article that was in RV magazine in the spring of 1998, warning of this. You can’t do that for reasons just cited. You guys are at the danger stage of using regression so be careful. If you put to many or two few independent variables in your test statistics the coefficients will look good but every coefficient can be totally mislabeled. There is another little problem that is a fly in the ointment called multicollinearity of variables. This means that if the independent variables are correlated to each other, and they all are correlated to each other to different degrees, that the coefficients will be distorted. If the correlation is high enough the regression can’t even calculate an answer. Look that word up in your statistics book and try to understand it because that is the problem that has prevented regression from working in the past. I found a webb site once where math professors and graduate students were discussing multicollinearity and there were 1,400 different views posted. I read a lot of them and strangely the only one I agreed with was a professor from my old college. That is the fly in the soup that separates the men/women from the boys/girls.
 

EDWARD BERRY

Senior Member
Joined
Jan 15, 2002
Professional Status
Certified General Appraiser
State
Arkansas
Could the 1 story vs 1.5 (or 2) have anything to do with the COST)

In our market the upper level will never sell at the same price as main levels.(60 to 80%)

Lower levels about 50%.

We THINK it has to do with original COSTS.

We are going to find out with MR "classes"

You all bought all the used textbooks up so it is goinf to cost me more to do the "class" BUT IT WILL BE WORTH IT.

ed in arkansas
 

George Hatch

Elite Member
Gold Supporting Member
Joined
Jan 15, 2002
Professional Status
Certified General Appraiser
State
California
Austin,

I found a webb site once where math professors and graduate students were discussing multicollinearity and there were 1,400 different views posted.   I read a lot of them and strangely the only one I agreed with was a professor from my old college.

I find it very surprising that mathmaticians would disagree to this extent in relation to a statistical question. Actually, it sounds worse than appraisers arguing over the development and application of adjustments, which by the way, you never see.

So exactly what were they arguing about? Certainly not the effects of multicollinearity. Was it how to parse the factors to find the 'right' ones?


George Hatch
 

Austin

Elite Member
Joined
Jan 16, 2002
Professional Status
Certified General Appraiser
State
Virginia
George:
The questions about multicollinearity were: What is it? How do you deal with it? I am not stretching things, there were over 1,400 opinions of what is it and what to do about it. It is a very complicated subject. It is not just a math question, it is a question of logic or applying the correct algorithm to engineer your way through the math. The solution involves the correct sequence of attacking the problem. Does the the phrase "correct sequence of adjustments" ring any bells in your head? Hint hint hint!!!!!
 

Jim Bartley

Senior Member
Joined
Jan 20, 2002
Professional Status
Certified Residential Appraiser
State
Virginia
I have found that in some neighborhoods, a 3%-7% adjustment for design is warranted in comparing a ranch to a 2 story. I make my sq ft adjustment as usual after adjusting under "Design" for ranch vs other. But every neighborhood is different. I think MR analysis is fine so far as looking at trends and identifying the real value related features we should be adjusting for, but it all comes down to a judgement by the appraiser. Rocket science it ain't.
 
Status
Not open for further replies.
Find a Real Estate Appraiser - Enter Zip Code

Copyright © 2000-, AppraisersForum.com, All Rights Reserved
AppraisersForum.com is proudly hosted by the folks at
AppraiserSites.com
Top

AdBlock Detected

We get it, advertisements are annoying!

Sure, ad-blocking software does a great job at blocking ads, but it also blocks useful features of our website. For the best site experience please disable your AdBlocker.

I've Disabled AdBlock
No Thanks