[R] imputation of sub-threshold values [off-topic]

Thomas Petzoldt thpe at hhbio.wasser.tu-dresden.de
Tue Mar 16 09:06:55 CET 2004

Thomas Lumley wrote:


> If you can persuade the people measuring the values to give you the 
> numbers (assuming they are just below `limit of detection' rather 
> than genuine non-detects) you will reduce the need for imputation.
> This is often the most powerful technique -- analytical chemists are
> trained not to give out numbers less than some multiple of the
> measurement error, but they can often be persuaded that statisticians
> can be trusted with these numbers.


Dear Thomas and dear R people,

thank you very much for these words. Similar discussions are frequently 
observed with people of the "applied side". I repeatedly point out, that 
scientific data should include also values lower than the "measurement 
limit", but in most cases they do are not willing, sometimes because of 
legal reasons and in other cases simply, because someone told them, that 
this is would be an unsound behaviour. Are there citable references, I 
can use as antidote?

Thomas P.

More information about the R-help mailing list