[R] MCMClogit: normal intercept priors don't work

AlexanderDueDo alex at ccin.de
Wed Oct 7 18:31:40 CEST 2009


Dear R users,

I think that the question I posted yesterday was not specific enough.
Apologies for this.

The question should have been: why does MCMClogit stop working if I a) set a
normal prior for the intercept and b) remove all variables except for the
intercept as explanatory?

Example code:

simulatedCase <- rbinom(100,1,0.5)
simDf <- data.frame(CASE = simulatedCase)
posterior_m0 <<- MCMClogit(CASE ~ 1, data = simDf, b0 = 0, B0 = 1)

It should easily be possible to fit this model with the intercept part
around -0.2 or so, and, in fact, when I use flat priors with

posterior_m0 <<- MCMClogit(CASE ~ 1, data = simDf)

this works -- and -0.2 is easily allowed by a N(0,1) prior on the intercept,
so why doesn't this work any more? It this a bug, or am I simply overlooking
the most obvious error?

Best wishes,

Alexander Düdo


-- 
View this message in context: http://www.nabble.com/MCMClogit%3A-normal-intercept-priors-don%27t-work-tp25789993p25789993.html
Sent from the R help mailing list archive at Nabble.com.




More information about the R-help mailing list