how to test the random factor effect in lme

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

how to test the random factor effect in lme

Xiang Gao-2
Hi

I am working on a Nested one-way ANOVA. I don't know how to implement
R code to test the significance of the random factor

My R code so far can only test the fixed factor :

anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
            numDF denDF   F-value p-value
(Intercept)     1    12 1841.7845  <.0001
Area              1     4    4.9846  0.0894


Here is my data and my hand calculation.

> PCBdata
   Area Sites PCB
1     A     1  18
2     A     1  16
3     A     1  16
4     A     2  19
5     A     2  20
6     A     2  19
7     A     3  18
8     A     3  18
9     A     3  20
10    B     4  21
11    B     4  20
12    B     4  18
13    B     5  19
14    B     5  20
15    B     5  21
16    B     6  19
17    B     6  23
18    B     6  21

By hand calculation, the result should be:
Source SS DF MS
Areas      18.00  1    18.00
Sites        14.44  4    3.61
Error        20.67  12  1.72
Total        53.11   17   ---


MSareas/MSsites = 4.99 --- matching the R output
MSsites/MSE = 2.10
Conclusion is that Neither of Areas nor Sites make differences.


My R code so far can only test the fixed effect :

anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
            numDF denDF   F-value p-value
(Intercept)     1    12 1841.7845  <.0001
Area              1     4    4.9846  0.0894



--
Xiang Gao, Ph.D.
Department of Biology
University of North Texas

______________________________________________
[hidden email] mailing list
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.
Reply | Threaded
Open this post in threaded view
|

Re: how to test the random factor effect in lme

David Winsemius

On Feb 14, 2012, at 5:36 PM, Xiang Gao wrote:

> Hi
>
> I am working on a Nested one-way ANOVA. I don't know how to implement
> R code to test the significance of the random factor

Have you read what the unofficial Mixed Model FAQ says about testing  
for significance on random effects?

http://glmm.wikidot.com/faq

>
> My R code so far can only test the fixed factor :

That may be the intent of the authors. They may want to make it  
sufficiently  difficult so that an adequate barrier prevents the  
unwary from taking some "easy way out". You probably need to describe  
your study (assuming this is not an assigned homework exercise) in  
sufficient scientific detail and do so on the mixed-models mailing list.

--
David.

>
> anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
>            numDF denDF   F-value p-value
> (Intercept)     1    12 1841.7845  <.0001
> Area              1     4    4.9846  0.0894
>
>
> Here is my data and my hand calculation.
>
>> PCBdata
>   Area Sites PCB
> 1     A     1  18
> 2     A     1  16
> 3     A     1  16
> 4     A     2  19
> 5     A     2  20
> 6     A     2  19
> 7     A     3  18
> 8     A     3  18
> 9     A     3  20
> 10    B     4  21
> 11    B     4  20
> 12    B     4  18
> 13    B     5  19
> 14    B     5  20
> 15    B     5  21
> 16    B     6  19
> 17    B     6  23
> 18    B     6  21
>
> By hand calculation, the result should be:
> Source SS DF MS
> Areas      18.00  1    18.00
> Sites        14.44  4    3.61
> Error        20.67  12  1.72
> Total        53.11   17   ---
>
>
> MSareas/MSsites = 4.99 --- matching the R output
> MSsites/MSE = 2.10
> Conclusion is that Neither of Areas nor Sites make differences.
>
>
> My R code so far can only test the fixed effect :
>
> anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
>            numDF denDF   F-value p-value
> (Intercept)     1    12 1841.7845  <.0001
> Area              1     4    4.9846  0.0894
>
>
>
> --
> Xiang Gao, Ph.D.
> Department of Biology
> University of North Texas
>
> ______________________________________________
> [hidden email] mailing list
> 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, MD
West Hartford, CT

______________________________________________
[hidden email] mailing list
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.
Reply | Threaded
Open this post in threaded view
|

Re: how to test the random factor effect in lme

glsnow
In reply to this post by Xiang Gao-2
This post https://stat.ethz.ch/pipermail/r-sig-mixed-models/2009q1/001819.html
may help you understand why the standard p-values in some cases are
not the right thing to do and what one alternative is.

On Tue, Feb 14, 2012 at 3:36 PM, Xiang Gao <[hidden email]> wrote:

> Hi
>
> I am working on a Nested one-way ANOVA. I don't know how to implement
> R code to test the significance of the random factor
>
> My R code so far can only test the fixed factor :
>
> anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
>            numDF denDF   F-value p-value
> (Intercept)     1    12 1841.7845  <.0001
> Area              1     4    4.9846  0.0894
>
>
> Here is my data and my hand calculation.
>
>> PCBdata
>   Area Sites PCB
> 1     A     1  18
> 2     A     1  16
> 3     A     1  16
> 4     A     2  19
> 5     A     2  20
> 6     A     2  19
> 7     A     3  18
> 8     A     3  18
> 9     A     3  20
> 10    B     4  21
> 11    B     4  20
> 12    B     4  18
> 13    B     5  19
> 14    B     5  20
> 15    B     5  21
> 16    B     6  19
> 17    B     6  23
> 18    B     6  21
>
> By hand calculation, the result should be:
> Source  SS      DF      MS
> Areas      18.00  1    18.00
> Sites        14.44  4    3.61
> Error        20.67  12  1.72
> Total           53.11   17   ---
>
>
> MSareas/MSsites = 4.99 --- matching the R output
> MSsites/MSE = 2.10
> Conclusion is that Neither of Areas nor Sites make differences.
>
>
> My R code so far can only test the fixed effect :
>
> anova(lme(PCB~Area,random=~1|Sites, data = PCBdata))
>            numDF denDF   F-value p-value
> (Intercept)     1    12 1841.7845  <.0001
> Area              1     4    4.9846  0.0894
>
>
>
> --
> Xiang Gao, Ph.D.
> Department of Biology
> University of North Texas
>
> ______________________________________________
> [hidden email] mailing list
> 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.



--
Gregory (Greg) L. Snow Ph.D.
[hidden email]

______________________________________________
[hidden email] mailing list
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.