Skip to content
Prev 11574 / 20628 Next

MCMCglmm variance estimates Poisson distribution

Hi Jarrod 
 >  is there a reason that the data frames differ in each (uc11 and uc12)?
Yes. Two different baseline conditions, cut and paste error. 

With the same data frame

summary(mcmc.c11.cf2)
 Iterations = 10001:99911
 Thinning interval  = 90
 Sample size  = 1000 

 DIC: 7489.396 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units    0.7111   0.6276   0.7912     1000

 Location effects: totflct ~ 1 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     1.211    1.160    1.263     1000 <0.001 ***
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
(Intercept) 
     4.7921
[1] 4.793908

And for glmer... 
Generalized linear mixed model fit by the Laplace approximation 
Formula: totflct ~ (1 | obs) 
   Data: uc11 
  AIC  BIC logLik deviance
 6205 6216  -3100     6201
Random effects:
 Groups Name        Variance Std.Dev.
 obs    (Intercept) 0.083672 0.28926 
Number of obs: 1607, groups: obs, 168

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  1.50622    0.02535   59.43   <2e-16 ***
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1 
__

for which exp(1.50622 + 0.083672/2) = 4.70232

I take your point re prior. Number of observations is 1607 so I thought this should be sufficient to limit the influence of the prior?

Cheers
John