Interpretation of lme results with intercorrelation between fixed factors

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Interpretation of lme results with intercorrelation between fixed factors

Vpapatha37
I�m running a mixed model analysis with 2 fixed factors that are intercorrelated using the lme function and I�m having difficulties in interpreting the results.
As I�m quite novice I�ll try to use a very simple example.

My model is lme(Y~A*B). A has 3 levels (1, 2 and 3) and B has 2 levels (I and II).
My results are something like this:



P statistic

B:II

P=0.01

A:2

P=0.01

A:3

P=0.09

II*2

P=0.61

II*3

P=0.031


My question is as follows. I understand that R keeps a level of each factor and reports any statistical differences to it. So in this example it reports that II is different than I and 2 against 3. However when it comes to the intercorrelation, what does it report? Does it compare II*2 and II*3 to II*1 and if so what happens with I*2 and I*3? Or does it compare II*2 to I*2 and II*3 to I*3 and if so what happens to I*1 and II*1?

Thank you


Vasillis


        [[alternative HTML version deleted]]


______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Interpretation of lme results with intercorrelation between fixed factors

David Winsemius

> On Mar 10, 2017, at 3:22 AM, Vasillis Papathanasiou <[hidden email]> wrote:
>
> I�m running a mixed model analysis with 2 fixed factors that are intercorrelated using the lme function and I�m having difficulties in interpreting the results.
> As I�m quite novice I�ll try to use a very simple example.

I'm guessing this is perhaps the `lme` function in pkg:nlme. It's not really an appropriate question for rhelp in any event. See the Posting Guide (link below).

>
> My model is lme(Y~A*B). A has 3 levels (1, 2 and 3) and B has 2 levels (I and II).
> My results are something like this:
>

Better to include actual output. I think most statisticians would advise focussing on model comparisons rather than p-values of individual coefficients.

The experts hang out at:

https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

... and they tend to be more welcoming for "interpretation" questions.

Or post to CrossValidated.com where this would be on topic. That group generally tries to avoid purely R questions, but your question is so devoid of actual R (no code and no actual output) that it would probably not be closed for that reason.

Best;
David.

>


>
> P statistic
>
> B:II
>
> P=0.01
>
> A:2
>
> P=0.01
>
> A:3
>
> P=0.09
>
> II*2
>
> P=0.61
>
> II*3
>
> P=0.031
>
>
> My question is as follows. I understand that R keeps a level of each factor and reports any statistical differences to it. So in this example it reports that II is different than I and 2 against 3. However when it comes to the intercorrelation, what does it report? Does it compare II*2 and II*3 to II*1 and if so what happens with I*2 and I*3? Or does it compare II*2 to I*2 and II*3 to I*3 and if so what happens to I*1 and II*1?
>
> Thank you
>
>
> Vasillis
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius
Alameda, CA, USA

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Loading...