Forward Stepwise Variable Selection

[Caleb Hall-Paterson]

I had a quick question regarding forward stepwise selection and was hoping someone may be able to confirm whether my code is right or not?

stepwise, pr(.051) pe(.05) forward: logistic

I want to screen my regression for variables that meet my 0.05 sig level.

Thanks in advance

[Carlo Lazzaro]

Caleb:
before embarking yourself in this risky methodological journey, I would recommedn you to take a look at: https://www.stata.com/support/faqs/s...sion-problems/

[Marcos Almeida]

As Carlo point out, stepwise selection is a controversial strategy. That said, to check your code, you just need to type - help stepwise - in the command window.

[Caleb Hall-Paterson]

Thank you both, I understand the issues with using stepwise but am basing my research's methodology on a prior paper which used it.

Following on from this, I am trying to detect whether endogeneity is an issue with my model. Everything I have read on this points me toward using

 estat endogenous

although when I enter the command the reply is simply 'estat endogenous not valid'.

My model is centred upon a Logistic regression and so was wondering whether you could point me in the right direction?

Kind regards,
Caleb

[Carlo Lazzaro]

Caleb:
it may be that -sptepwise- does not allow -estat endogenous- in post estimating.