>>On Fri, Sep 14, 2007 at 09:46:57AM +0100, S Ellison wrote:
>>
>>... only you probably shouldn't be doing that at all. Words like 'bias'
>>spring to mind...
>>
>> Woudn't it be better to accept the NA's and find methods that handle them as
>> genuinely missing.
>> R is usually quite good at that.
And Gabor Csardi <[EMAIL PROTECTED]> replied
>Although in some cases the proper handling of NA values is to treat
>them az zeros....
Yup. And sometimes not...
>I like this list because if you ask a question,
>they don't only solve it immediately (in five different ways), but they
>persuade you that what you're trying to do is actually
>incorrect/stupid/uninteresting or your problem just makes no sense at all.
>:)
Being a chemist, I have to confess that I can't always tell that what I'm about
to attempt is barking, trivial, uninteresting or better done a completely
different way; myself, I'd rather be warned too often than left to dig my own
pit and fall into it ...
On NA's vs zero, I usually have the reverse problem in my corner of the world;
folk will often call nondetects 'missing', which is also often a silly thing to
do; nondetect means 'I looked and it was too low to see' but NA means 'I didn't
look'. All that leaves me a bit nervous about replacing NA with 0 and vice
versa ... hence the knee-jerk. Apologies if I'm teaching egg-sucking to an
expert.
S
*******************************************************************
This email and any attachments are confidential. Any use, co...{{dropped}}
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
