[R-sig-ME] Unable to fit glmer with family=binomial(link=identity)

Ken Beath ken at kjbeath.com.au
Wed Aug 19 23:52:16 CEST 2009


On 20/08/2009, at 1:58 AM, Doran, Harold wrote:

> I may be missing something here, but why are you using an identity  
> link with binomial outcomes? If there is a reason why, you might as  
> well just use lmer since the identity link is for a normal  
> distribution and that is the distributional assumption used for the  
> errors in that function.
>

There are good reasons, as mentioned it results in relative risks  
rather than odds ratios. It can be made to work for fixed effects  
models, the main problem is that the linear predictor must always  
remain within 0 and 1 or the algorithm fails. For random effects  
models this is impossible, as the linear predictor plus a normally  
distributed random effect must always include points outside this  
range, so it is sensibly not allowed by the software.

Ken



>
>
>> -----Original Message-----
>> From: r-sig-mixed-models-bounces at r-project.org
>> [mailto:r-sig-mixed-models-bounces at r-project.org] On Behalf
>> Of David Evans
>> Sent: Wednesday, August 19, 2009 11:03 AM
>> To: r-sig-mixed-models at r-project.org
>> Subject: [R-sig-ME] Unable to fit glmer with
>> family=binomial(link=identity)
>>
>> Fellow R-users,
>>
>> I need to estimate the "risk difference" (in epidemiological
>> terminology) for a neighbourhood-level exposure (e.g.
>> presence of parks) with a binary outcome adjusted on 5 or so
>> covariables. Study participants are clusted by neighbourhood
>> .  I was hoping to use a generalized linear mixed model with the  
>> code:
>>
>> mod  <-  glmer(outcome ~ exposure + (1 | neighbourhood),
>> family=binomial(link=identity), data=rec)
>>
>> but the identity link is not available for the binomial
>> family with glmer (or for glmmPQL).  Is there any way to do
>> this with lme4 or is there another package in R which could
>> fix my problem?
>>
>> I'd be very grateful for any help.
>>
>> David.
>>
>> --
>> David Evans
>> UMR-S 707 Inserm - Université Pierre et Marie Curie - Paris 6
>> Faculté de Médecine Saint-Antoine 27, rue Chaligny
>> 75571 Paris cedex 12
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>




More information about the R-sig-mixed-models mailing list