collin command

[anto atha]

Hi,

I am trying to download "collin" command. I tried using "findit" command, however, I couldn't find it.

Does anyone know how I can find and install "collin" command?

[Nick Cox]

It's there at http://www.ats.ucla.edu/stat/stata/a...sis/collin.ado

You could start with

 
net from http://www.ats.ucla.edu/stat/stata/ado
net cd analysis

I guess the author might want to say something like this: If it helps you, I am happy. If it's not up-to-date or as general as you wish, sorry about that. (I am not the author, and am imputing.)

[Richard Williams]

UCLA has a lot of good commands, including Collin. I hope they aren't permanently making their commands harder to find.

[Richard Williams]

-findit collin- works now. Maybe UCLA was just tweaking its pages or they were temporarily down.

[Jessica Thacker]

Dear Nick Cox , I would humbly like to ask the following questions

(a)Does collin command takes into consideration only linear collinearity?

(b)Also, I wanted to humbly ask can this command be used to find collinearity between dummy and continuous variables?

[Joro Kolev]

You can find a good discussion on why the "problem of multicollinearity is overrated, and you should not "test" for multicollinearity" here: https://www.statalist.org/forums/for...ical-variables

As to your questions

a) there is only one type of collinearity, the linear one. It is suggested by the name co-linearity.

b) yes, collinearity between a dummy and a continuous variable is like any other collinearity. So you can use -collin-.

Dear Nick Cox , I would humbly like to ask the following questions

(a)Does collin command takes into consideration only linear collinearity?

(b)Also, I wanted to humbly ask can this command be used to find collinearity between dummy and continuous variables?