Re: Excel multiple regression equations...Help please!

[Gerry Dizinno]

You seem like a very enthusiastic and nice person - so, for what it's worth from me: PLEASE, don't do this analysis yourself. While you may have some "feel" for what multiple regression does, or is used for, it also appears that you don't really have much expertise. The amount of information you need to learn is much more than is reasonably provided by members of a list such as edstat-l. There are assumptions, for example, that need to be tested. There are, also, issues with using Excel as your statistical analysis tool. Please, get yourself a consultant - describe your research to her/him, explain the nature of your questions, what you need to be able to understand and what, if any, decisions you  need to make. Describe your audience, as well. Then, make sure this consultant can explain to you - in language you understand - what she/he believes to be the most appropriate type of analysis and how it can help you. Then, have the consultant conduct a thorough analysis and report the results back to you in language you can use to communicate to the audience of interest.

While there are short answers to the question(s) you posed, I don't think it would be responsible to answer them.

I hope I haven't insulted you - that really isn't my intention - but we get a number of similar questions on this list. I pretty much stay out of the fray, but it seems like you - and your research goals - would be much better served if you worked with a professional statistician rather than trying to do it yourself. Of course, if you can take the time to educate yourself thoroughly, by all means do it! It seems, however, that you have a much shorter time frame in mind for a solution to your problem.

Cheers,
gerry



At 10:22 AM 5/29/02 -0700, GreekEagle wrote:

Hi everybody!
- I m writing an essay  using the Excel multiple linear
regression.This essay is kind of formal and i have to write down the
multiple linear regression's equations that i used..My problem is that
i cant find the equations or the method that Excel actually uses for
multiple linear regression.The Excel "help" and some relevant web
pages just refer to the equations of simple linear regression of 2(X
and Y) variables and least squares..And those which refer to multiple
regression are giving some general tips and information but none makes
100% clear the method or the equations that Excel actually uses for
multiple linear regression and coefficients caculation.... I know the
basic method for multiple regression using data matrixes giving the
coefficients (Y=Xb+E---> b=( X'X)(-1)(X'Y) etc.) but i cant tell if
Excel is using it with no differences,correlations etc.Can you help me
please?Anything......

-My final regression functions giving the Y values, have 5 or 6 (X)
different variables with 23 observations for each variable.My data are
strongly linear with determination (R Square) over 0,95.However the
t-Student values( for 95% precision)  for the variables' coefficents
are below the crtical "Student" value in many cases.I would like most
of the X variables to be statistically important in my Y functions.I
know it is impossible that all the X variables are statistically
important (with t-Student) but i would like them to be as more as
possible...I tried to improve that so I have tried to make some
combinations among the X variables different from those i originally
wanted to..i tried to increase the variables' observations with some
more data ...i transformed my data to logarithms and i tested
regression again but it didnt help much.Any ideas or tips for how i
can do it?Even a new data process method is acceptable!

Thanks for your help!Cheers!
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                   .
=================================================================

Gerry

Gerard A. Dizinno, Ph.D.
Assistant to the President for Planning & Institutional Research, and
Acting Supervisor, Information Technology Services
St. Mary's University
One Camino Santa Maria
San Antonio, TX 78228-8507
Phone: 210-436-3601
Fax: 210-431-4245
Email: [EMAIL PROTECTED]