[R] Problems using na.include

[Frank E Harrell Jr]

Greg Distiller wrote:
> Hi
> I am having problems using na.action=na.include in a nlme call. Basically I 
> am trying to run a particular model without including any covariates but my 
> dataset does include some covariates (factors) with missing values. My 
> understanding from the Pinheiro and Bates book is that one would use 
> na.action=na.include in order to treat the missing values as another factor 
> level and then one can also use naPattern to specify which rows must be 
> excluded from the calculation of the objective function.

The approach of adding new categories for missing values has been shown 
to be highly problematic, completely changing the interpretation of the 
predictor and causing biases.  See an article by Werner Vach on this in 
the Encyclopedia of Biostatistics.

> 
> I have tried this but keep getting error messages about being unable to find 
> the object or the function "na.include". I have also tried to replicate the 
> particular Phenobarbital example verbatim from the book but am getting the 
> same error message. Could it be that my version of the nlme package is 
> outdated? I also cannot find any help on na.include...
> 
> Thanks
> 
> Greg
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
> 


-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University