[R-sig-ME] Interpreting a MCMCglmm model with a bivariate response variable

Jarrod Hadfield j.hadfield at ed.ac.uk
Wed Sep 23 16:25:02 CEST 2015


Hi Iker,

You need to follow the advice given to your previous post. With  
unconstrained residual variance the model is largely generating  
nonsense. Use `corg' instead of `us'. Also, depending on what the  
outcomes are you almost certainly need to have `trait' in the fixed  
effect specification.

Jarrod




   Quoting Iker Vaquero Alba <karraspito at yahoo.es> on Wed, 23 Sep 2015  
13:28:09 +0000 (UTC):

>
>    Hello all,
>    I am implementing a model with a multiple (bivariate) response  
> variable using MCMCglmm. Both response variables and all the  
> explanatory variables are categorical variables, with between 2 and  
> 6 levels. The model is as follows:
>   
> testmodel1<-MCMCglmm(cbind(natapshort,nataplong)~gender+age+religion+sexor+selfattr+partnerattr+gender:age+gender:religion+gender:sexor+gender:selfattr+gender:partnerattr+age:religion+age:sexor+age:selfattr+age:partnerattr+religion:sexor+religion:selfattr+religion:partnerattr+sexor:selfattr+sexor:partnerattr+selfattr:partnerattr,random=NULL,rcov=~us(trait):units,family=c("threshold","threshold"),data=extphen,nitt=100000,singular.ok=TRUE)
>    And this is the summary of the model after all the iterations:
>   
> summary(testmodel1)  Iterations =3001:99991 Thinninginterval  =  
> 10 Sample size  = 9700   DIC:   R-structure:   
> ~us(trait):units                            post.mean l-95% CI u-95%  
> CI eff.sampnatapshort:natapshort.units    114108   992.1   213000     
> 7.105nataplong:natapshort.units      33245   310.8    66300     
> 4.656natapshort:nataplong.units      33245   310.8    66300     
> 4.656nataplong:nataplong.units       82964   671.5   155869     
> 2.218  Location effects:cbind(natapshort, nataplong) ~ gender + age  
> + religion + sexor + selfattr +partnerattr + gender:age +  
> gender:religion + gender:sexor + gender:selfattr +gender:partnerattr  
> + age:religion + age:sexor + age:selfattr + age:partnerattr+  
> religion:sexor + religion:selfattr + religion:partnerattr +  
> sexor:selfattr +sexor:partnerattr + selfattr:partnerattr  
>                        post.mean   l-95% CI  u-95% CI eff.samp    
> pMCMC   (Intercept)           8.934e+02 -6.995e+02 2.596e+03    
> 275.73 0.22660   genderM              -8.936e+01 -7.437e+0
> 2 5.066e+02  4236.67 0.75794   genderO              -1.292e+03  
> -1.934e+05 1.765e+05  9700.00 0.99052    
> age                  -3.493e+02 -1.170e+03 3.615e+02   505.92  
> 0.31918   religionY            -7.361e+02 -1.598e+03 1.481e+01     
> 33.71 0.03402 * sexorHOM             -1.235e+03  
> -1.808e+05 1.679e+05  9700.00 0.99814   sexorOT               
>   2.193e+03 -1.589e+05  1.687e+05 10583.09 0.97814    
> selfattr             -2.367e+02 -7.424e+02 1.706e+02   314.82  
> 0.24948   partnerattr           1.391e+02 -2.667e+02 6.147e+02    
> 966.40 0.49546   genderM:age           2.696e+01 -1.313e+02   
> 1.748e+02  3786.10 0.69670   genderO:age          -1.055e+04  
> -8.325e+04 7.163e+04  1722.63 0.78474    
> genderM:religionY    -1.295e+02 -3.725e+02 8.194e+01   200.08  
> 0.20495   genderO:religionY    -1.016e+04 -1.589e+05 1.505e+05   
> 8731.86 0.89052   genderM:sexorHOM     -2.245e+02  
> -5.713e+02 4.443e+01   105.67 0.10495    
> genderO:sexorHOM     -8.104e+03 -1.620e+05 1.385e+05  5318.22  
> 0.90474   genderM:sexorOT      -1.52
> 0e+02 -5.124e+02 1.856e+02   423.76 0.33402    
> genderO:sexorOT       2.628e+03 -1.654e+05 1.658e+05  9700.00  
> 0.97670   genderM:selfattr      9.029e+01 -3.152e+01 2.334e+02    
> 119.78 0.12907   genderO:selfattr      6.281e+03  
> -6.511e+04 8.524e+04  3504.67 0.88412    
> genderM:partnerattr  -7.284e+01 -2.160e+02 6.729e+01   263.29  
> 0.25052   genderO:partnerattr   2.536e+02 -5.113e+02 1.121e+03   
> 1291.76 0.49113   age:religionY         8.732e+01  
> -1.283e+02 3.457e+02   727.46 0.42289    
> age:sexorHOM          2.809e+02 -8.592e+04 8.847e+04  9700.00  
> 0.99711   age:sexorOT          -1.246e+03 -8.447e+04 7.941e+04   
> 9370.57 0.97526   age:selfattr          1.195e+02  
> -6.636e+01 3.452e+02   212.35 0.19567    
> age:partnerattr      -8.598e+00 -1.963e+02 1.714e+02  9700.00  
> 0.92227   religionY:sexorHOM    8.506e+01 -2.392e+02 4.612e+02   
> 2059.92 0.59959   religionY:sexorOT     1.420e+01  
> -5.170e+02 5.464e+02  9700.00 0.96268    
> religionY:selfattr    2.782e+01 -1.198e+02 1.833e+02  3520.80  
> 0.68701   religionY:part
> nerattr 1.407e+02  1.423e+01  2.886e+02   22.99 0.00928  
> **sexorHOM:selfattr     1.160e+02 -1.141e+02 3.707e+02   394.74  
> 0.28495   sexorOT:selfattr      1.006e+02 -8.528e+01 3.050e+02    
> 305.50 0.24577   sexorHOM:partnerattr  1.231e+02  
> -1.246e+02 3.990e+02   415.43 0.31072    
> sexorOT:partnerattr  -1.401e+00 -2.007e+02  1.956e+02 9700.00  
> 0.99237   selfattr:partnerattr  5.483e+00 -6.017e+01 7.207e+01   
> 3007.45 0.85464   ---Signif. codes:  0?***? 0.001 ?**? 0.01 ?*? 0.05  
> ?.? 0.1 ? ? 1  Cutpoints:                           post.mean l-95%  
> CI u-95% CI eff.sampcutpoint.traitnatapshort.1     235.2   62.34     
> 376.2    8.822cutpoint.traitnatapshort.2     633.4  202.16     
> 944.0    3.578cutpoint.traitnatapshort.3   1139.5   364.35    
> 1683.7   5.832cutpoint.traitnataplong.1      293.4   54.85     
> 433.7    5.203cutpoint.traitnataplong.2      651.8  223.70     
> 961.6    2.604cutpoint.traitnataplong.3    1023.1   344.82    
> 1483.0   2.353
>   
> So, my question is: in that summary, where are the effect sizes, are  
> they the "post. mean" column? And have they been transformed in some  
> way? Because obviously, for response variables that can only take  
> values 1,2,3,4 or 5, I would expect to see those as the effect size.
> Also, is there any way of knowing to what extent are those results  
> due to each specific response variable, and the degree of covariance  
> between both? Is it possible to get all that information from that  
> summary output I have copied above?
>    Thank you very much.   Iker
>
> __________________________________________________________________
>
>    Iker Vaquero-Alba
>    Visiting Postdoctoral Research Associate
>    Laboratory of Evolutionary Ecology of Adaptations
>    Joseph Banks Laboratories
>    School of Life Sciences
>    University of Lincoln   Brayford Campus, Lincoln
>    LN6 7DL
>    United Kingdom
>
>    https://eric.exeter.ac.uk/repository/handle/10036/3381
>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>



-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.



More information about the R-sig-mixed-models mailing list