[R-sig-ME] PredictSE.mer error message

[Jake Westfall]

Apparently predictSE.mer() is a function from the "AICcmodavg" package (it doesn't come with lme4). You should take this up with the maintainer of that package.

Jake

> From: yuki_himawari at hotmail.com
> To: r-sig-mixed-models at r-project.org
> Date: Wed, 13 Aug 2014 22:13:49 +0000
> Subject: [R-sig-ME] PredictSE.mer error message
> 
> Hi
> 
> I am running glmer model with binomial distribution and want to get confidence interval using fixed effect for each treatments. 
> However I get the error message I don't really understand therefore I don't know how to modify the code. 
> 
> My model is 
> 
> model_2b <- glmer(cbind(NumberCorrect,10-NumberCorrect)~ ADHDYN + EmotionType + (1 | ID),
>                   family="binomial",data = data)
> 
> and here is the summary output
> 
> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
>  Family: binomial  ( logit )
> Formula: cbind(NumberCorrect, 10 - NumberCorrect) ~ ADHDYN + EmotionType +      (1 | ID)
>    Data: data
> 
>      AIC      BIC   logLik deviance df.resid 
>   2000.5   2034.5   -992.3   1984.5      508 
> 
> Scaled residuals: 
>     Min      1Q  Median      3Q     Max 
> -3.4809 -0.7256  0.1760  0.7238  3.4354 
> 
> Random effects:
>  Groups Name        Variance Std.Dev.
>  ID     (Intercept) 0.1337   0.3656  
> Number of obs: 516, groups:  ID, 86
> 
> Fixed effects:
>              Estimate Std. Error z value Pr(>|z|)    
> (Intercept)   1.15637    0.10193  11.345  < 2e-16 ***
> ADHDYNYES    -0.27728    0.10236  -2.709  0.00675 ** 
> EmotionType2  2.04670    0.17845  11.469  < 2e-16 ***
> EmotionType3 -0.24761    0.10744  -2.305  0.02119 *  
> EmotionType4 -1.79462    0.10811 -16.600  < 2e-16 ***
> EmotionType5 -0.04793    0.10939  -0.438  0.66129    
> EmotionType6 -1.01871    0.10445  -9.753  < 2e-16 ***
> ---
> 
> And I used the code below to get CI's but got error message.
> 
> > predictSE.mer(model_2b, NEWdata,se.fit=TRUE, type="response",level=0, print.matrix=FALSE)
> Error in predictSE.mer(model_2b, NEWdata, se.fit = TRUE, type = "response",  : 
>   no slot of name "offset" for this object of class "glmerMod"
> Can you please guide me where I am getting wrong? I noted the sd of random effect is quite large- so might be worth to do the simulation. I heard about MCMCsamp but is this the way to go?
> 
> Thank you kindly
> Yuki
> 
> 
>  		 	   		  
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
 		 	   		  
	[[alternative HTML version deleted]]