[R-sig-ME] glmer: Downdated X'X is not positive definite [related to nAGQ]

[Ross Boylan]

On 2/27/2013 3:31 PM, Ross Boylan wrote:
> On 2/24/2013 6:20 PM, Ben Bolker wrote:
>> Ross Boylan <ross at ...> writes:
>>
>>   [snip]
>>
>>> Ben Bolker wrote
>>>> It is indeed a little surprising (to me) that you're getting
>>>> downdating problems with an intercept-only model.
>>>>     Hard to say without a closer look at the data.
>>>>     I know it won't solve your problem, but does fitting the model
>>>> _without_ the singleton clusters work?
>>    I didn't mean without the clusters, I meant dropping the singleton
>> clusters from the data ...
>>
> Oops.  Good question.  No, it didn't help.  I had trouble getting the 
> development version, detailed in a separate message.
> Ross
I discovered that omitting the nAGQ argument is sufficient to avoid the 
error.  I also tried nAGQ=5 and removing the missing data in advance; 
neither eliminated the downdating error.

> r2 <- glmer(sexActs~(1|id), data=sexpartner, family=poisson())

> r2 <- glmer(sexActs~(1|id), data=sexpartner, family=poisson(), nAGQ=5)

Error in mer_finalize(ans) : Downdated X'X is not positive definite, 1.

The error message sounds as if it is just a function of the input 
covariates, which obviously don't change when I change nAGQ.  My 
interpretation of X in the error message may be wrong.

This is with the version of lme4 from CRAN.

Ross