How to address too many missing
Posted on
Xin Lin
Joined: 06/10/2019
Forums
Dear Dr. Cheung and colleagues,
We are trying to investigate the predictors of an outcome using longitudinal studies. Yet, there are some studies with only one or two predictors.
Is there an algorithm that allows us to handle missing correlation or remove those correlation matrices containing only one correlation coefficients?
Thanks for your time and patience.
Dear Xin Lin,
Dear Xin Lin,
If you want to drop studies with many missing values, you may use is.na() to check whether there are missing values and drop them using standard R functions.
Best,
Mike
Log in or register to post comments
In reply to Dear Xin Lin, by Mike Cheung
Dear Mike,
Dear Mike,
Thank you for your reply.
Are there any citations support the removal of correlation matrices containing too much NA in MASEM?
Best,
Xin
Log in or register to post comments
Dear Xin,
Dear Xin,
No, I don't think so.
Personally, I will try to keep as many data as possible.
Best,
Mike
Log in or register to post comments
In reply to Dear Xin, by Mike Cheung
Dear Mike,
Dear Mike,
I see. Thanks for your explanation！
Best,
Xin
Log in or register to post comments