• 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.

Need for a GOOD class on multi-linear regression/Stats analy

I am interested to this level in taking a RESIDENTIAL multi-linear regression or statistical analysi

  • I am willing to pay up to $200 for the classwork

    Votes: 0 0.0%
  • I am willing to pay up to $400 for the classwork

    Votes: 0 0.0%
  • I am willing to pay up to $600 for the classwork

    Votes: 0 0.0%
  • I am willing to pay up to $800 for the classwork

    Votes: 0 0.0%
  • Fer-gid-dit I prefer to stick my head in the sand and hope this goes away

    Votes: 0 0.0%
  • I am cautiously interested but need to know MUCH more about costs

    Votes: 0 0.0%

  • Total voters
    0
Status
Not open for further replies.

Farm Gal

Elite Member
Joined
Jan 14, 2002
Professional Status
Licensed Appraiser
State
Nebraska
OK folks, here is your chance to get something happening! Please indicate your interest level so that I and others can get something going on this!

Make the following assumptions:
1. The class will actually provide the BASICS for understanding and using a program to perform these analyses.

2. You will HAVE to do some pre-and post class reading in order to familiarize yourself with the terminology and processes.

3. Assume a minimum of 3 locations nationwide, East, West and Central US.

4. Probable 2 to 3 day course, which MAY affect your future ability to work effectively in this field(hint hint - I think this is important!!!

5. Need to have a laptop computer and purchase DISCOUNTED software. I am contacting the two major program makers and dangling SALES in front of them. So to get this happening we are going to have to get some solid (ok maybe wispy) numbers to dangle!!!

So PLEASE vote as often as your interface permits :wink:
 

TC

Elite Member
Gold Supporting Member
Joined
Jan 31, 2002
Professional Status
Certified Residential Appraiser
State
Pennsylvania
I believe the McKissock school already offers a class covering this topic.

TC
 

Jo Ann Meyer Stratton

Elite Member
Joined
Jan 16, 2002
Professional Status
Certified Residential Appraiser
State
Arizona
Just got off the phone with an appraiser in Tucson who is developing a class for the Appraisal Institute. Check with the AI to see when it will be available. Auston: could you email me your email address, so I can pass it on to hime? [email protected]
 

Farm Gal

Elite Member
Joined
Jan 14, 2002
Professional Status
Licensed Appraiser
State
Nebraska
I am aware of the Mckissock class and prefer not to air all my reasons for thinking it less than a reasonable solution. Getting a class approved for state ed credit with an underlying purpose of selling a propriatary program which does not meet MY needs doesn't further the trade/professional interests of the appraisal community. It might suffice to start a person thinking about the possibility of regression...

Jo Ann:
I am aware of the AI class in planning, However I think there is a tri-polar need:

First for the general education in stats (background) that is needed to understand the use of the tool(s)

Secondly: specific training in the use of the currently available statistical programs... Either from a local college or university OR distance learning OR the purveyors of the programs themselves.

and the final need: SPECIFIC training as applicable to appraisers.

I personally want the AI class to address the last of these rather than the first two: digressing into Austin's feared lynch mob is also a concern...Though I have to mention that the AI conference did not so degenerate!~

The first need (general education about stats and regression) frankly needs to be on our own, it is too cumbersome a topic to be addressed by anything less than a week-long AI course! A year long grad level college course combined the first and second for me.

The second topic would also compress into a week long highly intensive seminar, but anything less would be shortsighted.

The last would best be taught at the roundtable or conference level by folks currently useing the programs, specific to appraising.

My intent in starting this poll is to establish a firm core group of individuals that I can harry a sales person at one or both of the primary stats program marketing departments into getting the core group a cost effective entre into their program, with the intent of broadening their market if/when the core group becomes the preferred program of choice for the industry! Shhh Don't tell 'em my plan 8) !

Additional private or e-mail responses to this post are welcomed (thank you to those who have already written), and will be saved for the purpose of beating up the companies for our mutual benefit!
 

Austin

Elite Member
Joined
Jan 16, 2002
Professional Status
Certified General Appraiser
State
Virginia
To all:
I just received a phone call from Jo Ann’s friend in Tucson, Ariz. We had a long chat. He has a graduate degree in statistics and his partner has a PHD in economics. He has been teaching regression for 12 years and is constructing a course for the AI. We are pretty much on the same page. From talking to him this subject has been taught for a long time but I don’t know who he was teaching, AI students apparently. So if you want a course one is in the works. He said he would e-mail me and we could talk some more. I will let you know what is going on.
 

John SRA

Junior Member
Joined
Jan 19, 2002
The AI has had a regression class for quite a while. I attended a few years ago, and found it very useful.

The class included general background and specific examples applicable to appraising. The course material included a floppy disk with Excel overlays. If you have Excel, there is no need for additional software.

While I do not pursue it with the same zeal as Austin, we regularly use regresson methods in our analysis.

Best Wishes

JC
 

Farm Gal

Elite Member
Joined
Jan 14, 2002
Professional Status
Licensed Appraiser
State
Nebraska
John et all:

Sir I respectfully submit that simple linear regression is to what I am looking for what tinder steel and flint are to a closed glass top cookrange.

Sort of like proclaiming paired sales is a sufficient research methodology in a market where there ain't any pairs... :evil:
~~~~~~~~~~

I KNOW there is more out there and I want it N.O.W.. If I have to drum up other like-minded folk myself I intend to do so.

~~~~~~~~~~~~~~~

Multiple regression lets you consider more than one independent variable and assess the combined ability of the each predictor/factor to account for changes in the dependent variable. In simpler terms:

Like how much are patios REALLY worth and how much are fireplaces REALLY worth and how much does size within a narrow range of similar homes in the same neigborhood REALLY affect value and you can on occasion get all of this from ONE 'clean' set of data!!!

It ain't a perfect solution for wildly disparite markets nor places where there are NO similar sales, but even then it can predict with slightly better than "you" can by eye some parameters for adjustments.

And there really is an easter bunny too! OK, sarcasm aside these are powerful programs folks, I used tightly scheduled mainframe computer time 15 years ago to do this sort of analysis... well the times they are a changin, and you can do this from home (be careful to have [proper supervision] and training)

Now my take on the following is that the articles may not be everyones read, your eyes may glaze and cross, but plow through anyway, it's for your own good, Mom-me says so. The following links are to the NAIFA e-Gram articles written by our own Doug Smith, which give a nice overview of the future. Those of you new to the concepts might want to peruse them.


part 1
http://www.naifa.com/gram/2002apr/smith-apr02.html

part 2

http://www.naifa.com/gram/2002jul/smith-jul02.html

part 3

http://www.naifa.com/gram/2002oct/smith-oct02.html

I agree with Doug on some points disagree on others, but fell REAL strong that this type of program is in all our futures...
 

KD247

Senior Member
Joined
Jan 24, 2002
Professional Status
Certified Residential Appraiser
State
California
I'd be very interested in the learning more about the use of statistical tools in appraising. But, from the point of view of a statistician looking at appraising, not through the eyes of an appraiser looking at statistics.

The use of statistics typically yields two numbers, the "best value" for an experiment and also a measurement of uncertainty. Most statisticians treat these two numbers with equal importance, but appraisers rarely breach the subject of uncertainty. Why? Maybe because the data we deal with usually does not produce statistically reliable results.

Maybe we could use more instruction that teaches us which tool to pick up for a particular job. Multilinear regression analysis is a tool with a very cool name, but is it really the right tool for most appraisal work? I have a feeling that most statisticians would say "No". Not with the number of variables, or quality and quantity of data, available for most appraisals.

In some large tract neighborhoods, an AVM can produce an incredibly accurate estimate of value. But this is the exception, not the rule. In most areas, particularly those with unique properties and limited sales data, even the most experienced appraisers have trouble agreeing on a reasonable value range.

In most cases, appraisers are more like the Antiques Roadshow guys, who say "Well, I've seen these things sell at auction from $10,000-$14,000 but with the current interest in them, I'd expect that you'd get around $15,000." Are his opinions really going to be more precise or reliable if he whipped out a laptop with a sophisticated statistical program?
The only exception I've seen is when the market uses the same tools as the appraiser. When I was selling commercial real estate, buyers, sellers, and agents all used the same cash-flow analysis software and conversed in terms of specific net rates of return. There was no way that you could appraise a property in that market without using those same software tools.
 

Austin

Elite Member
Joined
Jan 16, 2002
Professional Status
Certified General Appraiser
State
Virginia
Lee Ann: Doug Smith’s articles are some of the best regression explanations I have seen but you have to consider the algorithm too. We have talked by e-mail in the past and he was on the forum this week posting. Don’t be to quick to disagree until you get some flying time under your belt. The system works perfectly and if I could show you a demo of a contrived problem with no market variance you would be amazed at the results. The reason it doesn’t work perfectly in practice is that noise creeps into the equation, like for instance: normal random variance; unknown and unknowable independent variables like the lady next door is a nudists, etc. These unknown variables corrupt the estimated value factors and that is another reason to treat them in the aggregate. Who cares what a patio contributes as long as you can accurately predict the most probable price.
By the way, the man’s name constructing the AI program is Jim Sanders.
 

Terry Russell

Senior Member
Joined
Feb 24, 2002
Professional Status
Appraiser Trainee
State
Montana
Lee Ann,

Tripolar? Well, from what I understand is that 88% of all people make up their own statistics. And that's a fact!
terry
 
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