R help - Oct 2012 - Interpreting and visualising lme results

Dear R users,
I have used the following function (in blue) aiming to find the linear
regression between MOE and XLA and nesting my data by Species. I have obtained
the following results (in green).
model4<-lme(MOE~XLA, random = ~ XLA|Species,
method="ML")summary(model4)
Linear mixed-effects model fit by maximum likelihood Data: NULL         AIC    
BIC   logLik  -1.040187 8.78533 6.520094
Random effects: Formula: ~XLA | Species Structure: General positive-definite,
Log-Cholesky parametrization            StdDev       Corr  (Intercept)
1.944574e-01 (Intr)XLA         6.134158e-06 -0.884Residual    1.636428e-01    
  
Fixed effects: MOE ~ XLA                 Value  Std.Error DF   t-value
p-value(Intercept) 3.0558697 0.15075939 32 20.269847  0.0000XLA        
0.0000005 0.00000335 32  0.150811  0.8811 Correlation:     (Intr)XLA -0.861
Standardized Within-Group Residuals:       Min         Q1        Med         Q3
       Max -1.8354171 -0.4704322  0.1414749  0.5500273  1.5950338 
Number of Observations: 38Number of Groups: 5 
I have read that large correlation values such as,Correlation:     (Intr)XLA
-0.861"reflect an ill-conditioned model", in addition XLA does not
have an effect on the model p=0.88. These results are not logic when I look at
my data and therefore I think I am missing something in the model? It would be
very helpful if someone has some tips on this? In addition, I was wondering if
somebody knows what is the best way to visualise this kind of data (nested
data)?
Thank you very much for any help and time.


	[[alternative HTML version deleted]]

On Oct 25, 2012, at 10:32 PM, Santini Silvana wrote:
> Dear R users,
> I have used the following function (in blue)
No, we do not do "in blue" here. This is a monochrome mailing list.
> aiming to find the linear regression between MOE and XLA and nesting my
data by Species. I have obtained the following results (in green).
> model4<-lme(MOE~XLA, random = ~ XLA|Species,
method="ML")summary(model4)
> Linear mixed-effects model fit by maximum likelihood Data: NULL         AIC
BIC   logLik  -1.040187 8.78533 6.520094
> Random effects: Formula: ~XLA | Species Structure: General
positive-definite, Log-Cholesky parametrization            StdDev       Corr 
(Intercept) 1.944574e-01 (Intr)XLA         6.134158e-06 -0.884Residual   
1.636428e-01
> Fixed effects: MOE ~ XLA                 Value  Std.Error DF   t-value
p-value(Intercept) 3.0558697 0.15075939 32 20.269847  0.0000XLA        
0.0000005 0.00000335 32  0.150811  0.8811 Correlation:     (Intr)XLA -0.861
> Standardized Within-Group Residuals:       Min         Q1        Med       
Q3        Max -1.8354171 -0.4704322  0.1414749  0.5500273  1.5950338
> Number of Observations: 38Number of Groups: 5 
> I have read that large correlation values such as,Correlation:    
(Intr)XLA -0.861"reflect an ill-conditioned model", in addition XLA
does not have an effect on the model p=0.88. These results are not logic when I
look at my data and therefore I think I am missing something in the model? It
would be very helpful if someone has some tips on this? In addition, I was
wondering if somebody knows what is the best way to visualise this kind of data
(nested data)?
> Thank you very much for any help and time.
> 
> 
> 	[[alternative HTML version deleted]]
We also do not do HTML. This message is mangled.

-- 

David Winsemius, MD
Alameda, CA, USA

On 10/26/2012 04:32 PM, Santini Silvana wrote:
> Dear R users,
> I have used the following function (in blue) aiming to find the linear
regression between MOE and XLA and nesting my data by Species. I have obtained
the following results (in green).
> model4<-lme(MOE~XLA, random = ~ XLA|Species,
method="ML")summary(model4)
> Linear mixed-effects model fit by maximum likelihood Data: NULL         AIC
BIC   logLik  -1.040187 8.78533 6.520094
> Random effects: Formula: ~XLA | Species Structure: General
positive-definite, Log-Cholesky parametrization            StdDev       Corr 
(Intercept) 1.944574e-01 (Intr)XLA         6.134158e-06 -0.884Residual   
1.636428e-01
> Fixed effects: MOE ~ XLA                 Value  Std.Error DF   t-value
p-value(Intercept) 3.0558697 0.15075939 32 20.269847  0.0000XLA        
0.0000005 0.00000335 32  0.150811  0.8811 Correlation:     (Intr)XLA -0.861
> Standardized Within-Group Residuals:       Min         Q1        Med       
Q3        Max -1.8354171 -0.4704322  0.1414749  0.5500273  1.5950338
> Number of Observations: 38Number of Groups: 5
> I have read that large correlation values such as,Correlation:    
(Intr)XLA -0.861"reflect an ill-conditioned model", in addition XLA
does not have an effect on the model p=0.88. These results are not logic when I
look at my data and therefore I think I am missing something in the model? It
would be very helpful if someone has some tips on this? In addition, I was
wondering if somebody knows what is the best way to visualise this kind of data
(nested data)?
Hi Santini,
I am currently illustrating the results of nested analyses using the 
barNest function from the plotrix package. The illustrations display 
nested frequencies, proportions or location parameters, but convey the 
fairly complex relationships in a way understandable to most readers.

Jim