[R] Computation of pD in R2WinBUGS

[Nanye Long]

Hi all,

About how R2WinBUGS computes pD (effective number of parameters), in
the manual it says that pD = var(deviance)/2.
However, I run some data set and find that the reported pD not equal
to var(deviance)/2, but much smaller than that.
Is var(deviance)/2 the actual formular used in R2WinBUGS, or did I
miss anything? Thanks for your answer.

Nanye Long