Stepwise Regression

[Derreck Shaw]

I need help with stepwise regression. I opened the stepwise estimation, but I am clueless after that. I would like a step by step explanation of the process.

[Clyde Schechter]

The section in the [R] manual is, I think a clear explanation and has several worked examples.

But better still, I refer you to http://www.stata.com/support/faqs/st...ems/index.html for a really good explanation of why you probably shouldn't be doing it in the first place.

[Nick Cox]

I agree with Clyde. In addition to the FAQ cited see

Harrell, F. 2001. Regression modeling strategies. New York: Springer.

which contains detailed discussion of stepwise methods, but the executive summary is "Don't".

[Euslaner]

Before you do this, consider that stepwise is very atheoretical and has statistical issues:

http://www.stata.com/support/faqs/st...sion-problems/

http://www.nesug.org/proceedings/nesug07/sa/sa07.pdf

http://www.danielezrajohnson.com/stepwise.pdf

Seems that we all agree.

[Carlo Lazzaro]

In the same fashion, "don't do it" recommendations about stepwise regression are well summarized in: Kirkwood BR, Sterne JAC. Essential Medica Statistics, 2nd edition. Malden, Ma. Blackwell Science, 2003: 340-342.

Kind regards,
Carlo