Spearman correlation: How can I adjust for a 3rd variable?

Tiago Munhoz

Join Date: Mar 2015

Posts: 16
#1

Spearman correlation: How can I adjust for a 3rd variable?

09 Oct 2019, 19:52

Dear all,

Could I estimate Spearman's rho adjusting for a 3rd variable? ( "confounding variable" ) . If so, how could I do that?

Thanks in advance. Best wishes
Tags: None
Joseph Coveney

Join Date: Apr 2014

Posts: 4398
#2

09 Oct 2019, 20:07

Maybe you could use Pearson's after creating ranks?

Code:

help egen help pcorr

Last edited by Joseph Coveney; 09 Oct 2019, 20:09.
Comment
Tiago Munhoz

Join Date: Mar 2015

Posts: 16
#3

09 Oct 2019, 20:08

Originally posted by Joseph Coveney View Post

Maybe you're looking for something like this?

Code:

help pcorr

Thanks. I have tried it.
Comment

Joseph Coveney

09 Oct 2019, 20:12

Try the update

Code:

sysuse auto
egen byte pr = rank(price)
egen byte pm = rank(mpg)
correlate pr pm
spearman price mpg
pcorr pr pm trunk

Comment

Tiago Munhoz

Join Date: Mar 2015

Posts: 16
#5

11 Oct 2019, 11:46

Originally posted by Joseph Coveney View Post

Try the update

Code:

sysuse auto egen byte pr = rank(price) egen byte pm = rank(mpg) correlate pr pm spearman price mpg pcorr pr pm trunk

Thanks, Joseph.
Comment