[R-sig-ME] MCMCglmm update

[Jarrod Hadfield]

Hi,

I've update MCMCglmm (2.02) and it should appear on CRAN soon. The  
main changes are:

1) After many questions regarding multivariate binary and multivariate  
ordinal models I've now implemented a sampler for correlation  
matrices. Simply  use "cor" instead of "us".

2) I found a bug regarding a cholesky update. Potentially it can  
affect analyses that use parameter expansion and analyses that have  
conditionally independent blocks of missing data. The effects are  
likely to be subtle; I only noticed it during an analysis of 2 data  
points both of which were missing, and even then the discrepancy  
between the sampled posterior and that expected (the prior) was small.

3) The companion paper to MCMCglmm has now been published in The  
journal of Statistical Software. If you use MCMCglmm I would be  
grateful if you could cite this paper.

If anyone could provide feedback on the behaviour of the correlation  
sampler I'd be grateful.

Thanks,

Jarrod


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