Hello!
I tried searching for Centered clusterwise vs. Clusterbased centering.
In GAMLJ Clusterbased centering I believe was within-subjects clustering
in GAMLJ3 it got replaced by Centered clusterwise?
Just trying to clarify this..
Thanks so much!!
Centered clusterwise vs. Clusterbased centering
-
- Posts: 47
- Joined: Fri Feb 09, 2018 3:41 pm
- mcfanda@gmail.com
- Posts: 530
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Centered clusterwise vs. Clusterbased centering
Do yoy have a repeated measure design?
- mcfanda@gmail.com
- Posts: 530
- Joined: Thu Mar 23, 2017 9:24 pm
-
- Posts: 47
- Joined: Fri Feb 09, 2018 3:41 pm
Re: Centered clusterwise vs. Clusterbased centering
Yes, it is a repeated measures design:
Translation: Spanish (Casa) --> English (Home) vs. English (Home) --> Spanish (Casa).. Lexical decision task and Controls.
The Language Dominance covariate (continuous variable) ranges from -30 (Spanish dominant) to +30 English dominant.
So, it would seem that I would need a "between-subjects" centered covariate
If I were to use the Language Dominance as a factor, I would have three levels: Spanish Dominant, English Dominant, Balanced Bilingual.
R
Translation: Spanish (Casa) --> English (Home) vs. English (Home) --> Spanish (Casa).. Lexical decision task and Controls.
The Language Dominance covariate (continuous variable) ranges from -30 (Spanish dominant) to +30 English dominant.
So, it would seem that I would need a "between-subjects" centered covariate
If I were to use the Language Dominance as a factor, I would have three levels: Spanish Dominant, English Dominant, Balanced Bilingual.
R
- mcfanda@gmail.com
- Posts: 530
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Centered clusterwise vs. Clusterbased centering
If language dominance is a variable that varies across participants, you need to center it using the sample mean (the default method), otherwise it becomes a column of zeros
-
- Posts: 47
- Joined: Fri Feb 09, 2018 3:41 pm
Re: Centered clusterwise vs. Clusterbased centering
Thank you so much!! I did the Between-groups centering which is basically what you are suggesting.
Muchas gracias!
Muchas gracias!