Statistics vs Excel

Hi guys,

My first post! I dont know if this is an excel question or a statistics question, I have a model, subject to X amount of variables. I want to know which of these variables sway my model the most. Are we talking of regression or correlation?

Through a Data Table or scenarios I can see which give me greater IRRs subject to certain variable changes BUT when I start doing more complex scenarios its not so easy. How would I go about getting that info?

Most definitely a statistics question before an Excel question.

Unfortunately, I don't have a strong enough background in statistics to explain it to you, but I can infer enough to maybe get you started.

The LINEST function can output supplementary statistics from its regression of the data. I'm not enough of a statistician to tell you how, but I believe that somewhere in the interpretation of those numbers is information about how much information each coefficient contributes to y. In the help file for the LINEST function (at least in my version of Excel) is an example (#5) that shows how to perform a simple T test to determine if a given coefficient is or is not significant in predicting y. This test is also described in my old intro to stats textbook. If your scenario is a regression of some kind, I expect someone with a good knowledge of statistics could use the information from the LINEST function to infer which variables are important. (I also wouldn't be surprised if such a person also had as a motto, "Friends don't let friends use Excel for statistics." )

HTH

Will look into it! thanks!