Skip to content

Zero cells in contrast matrix problem

1 message · Ben Bolker

#
I don't see your data -- I see a little tiny subset, but that's not
really enough for a reproducible example.

This is the example given in the URL I sent:

cmod_blme_L2 <- bglmer(predation~ttt+(1|block),data=newdat,
                       family=binomial,
                       fixef.prior = normal(cov = diag(9,4)))

trial<-bglmer(Correct ~ Syntax.Semantics+(1|Part.name),
          data =trialglm,
         family = binomial,
        fixef.prior = normal(cov=diag(9,8)))

The last line specifies an 8x8 matrix (because you have 8 fixed effect
parameters) with a value of 9 on the diagonal, meaning the priors for
the fixed effects are independent and each is Normal with a sd of
sqrt(9)=3.


On Thu, May 28, 2015 at 3:25 PM, Francesco Romano
<francescobryanromano at gmail.com> wrote: