Multiple Regression Modeling Challenges

Multiple regression has been touted in various ways by companies promising that their regression is better than other canned regression programs.  At the other extreme, some curricula claim multivariate regression as an ‘advanced’ appraisal method.  Either extreme may misstate its true capability and usefulness.

Regression in and of itself is just a math formula.  It minimizes the deviations of data points.  Usually this is the square root of the squared deviations.  The issue is not the mathematics.  It is the modeling.

This 7-hr. seminar considers:

  1. When is multiple regression (MR) a reasonable analysis method?

  2. What are the three uses of MR?

  3. How to prepare data for MR?

  4. Why is transforming some data very important?

  5. What type of regression should be used:  simple, two-predictor, multiple?


This class will not resolve all.  MR modeling issues can be perplexing.  But this overview and checklist of challenges will help provide appropriate skepticism, while serving confidence when appropriate.