why p-value is change with the scale of variable

[chenyi...@gmail.com]

hello,

I’m estimate a logit model, but after I multiply all the variables by 10, the p-value changed from 0 to 1. What are the reasons for this phenomenon?

Thank you for your help!

[Michel Bierlaire]

Your model is not identified. It seems that the B_ coefficient have no impact on the likelihood.

> On 12 Sep 2023, at 04:59, chenyi...@gmail.com <chenyi...@gmail.com> wrote:
>
> hello,

> I’m estimate a logit model, but after I multiply all the variables by 10, the p-value changed from 0 to 1. What are the reasons for this phenomenon?<微信图片_20230912104538.jpg><微信图片_20230912104522.jpg>

> Thank you for your help!
>

> --
> You received this message because you are subscribed to the Google Groups "Biogeme" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to biogeme+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/biogeme/ddf444b5-03fd-472d-8fd7-df552610feb0n%40googlegroups.com.
> <微信图片_20230912104522.jpg><微信图片_20230912104538.jpg>

Michel Bierlaire
Transport and Mobility Laboratory
School of Architecture, Civil and Environmental Engineering
EPFL - Ecole Polytechnique Fédérale de Lausanne
http://transp-or.epfl.ch
http://people.epfl.ch/michel.bierlaire