R help - Oct 2005 - measurement error model - "simple" linear regression

Dear friends, I found the thread on this subject this summer but 
wonder whether it has been taken any further. I have an important 
medical problem where X is computed from a three independent and 
complicated measurements (exchangeable sodium and potassium and total 
body water - i.e. X = (Nae+Ke)/TBW ) and Y is serum sodium 
concentration (all data in Edelman, JCI 1958). I have the individual 
data in the paper (but have not yet had time to enter it) and Edelman 
et al in fact reported measurement errors for most individual items. 
Edelman and coworkers made a linear regression of Y on X and now I 
would like to ascertain the importance of the fact that X was 
measured with error for the utility of the formula in clinical 
practice (since I believe it is being misused by ignoring this fact - 
and more). I found in Neter, Wasserman and Kutner 3. ed a general 
discussion of the problem - but would very much like a more recent 
presentation if available and application in R.

Best wishes

Troels Ring
Aalborg, Denmark

Dear Troels,

you might try McArdle, Limnol. Oceanogr., 48(3), 2003, 1363-1366 for
some reading, and the pls package (newly updated) for tools.

(I apologize if this information replicates previous postings, but
Googling R-help McArdle 2005 draws a blank so I think I'm on safe
ground!)

Good luck!

Andrew

On Sun, Oct 16, 2005 at 10:40:17PM +0200, Troels Ring
wrote:
> Dear friends, I found the thread on this subject this summer but 
> wonder whether it has been taken any further. I have an important 
> medical problem where X is computed from a three independent and 
> complicated measurements (exchangeable sodium and potassium and total 
> body water - i.e. X = (Nae+Ke)/TBW ) and Y is serum sodium 
> concentration (all data in Edelman, JCI 1958). I have the individual 
> data in the paper (but have not yet had time to enter it) and Edelman 
> et al in fact reported measurement errors for most individual items. 
> Edelman and coworkers made a linear regression of Y on X and now I 
> would like to ascertain the importance of the fact that X was 
> measured with error for the utility of the formula in clinical 
> practice (since I believe it is being misused by ignoring this fact - 
> and more). I found in Neter, Wasserman and Kutner 3. ed a general 
> discussion of the problem - but would very much like a more recent 
> presentation if available and application in R.
> 
> Best wishes
> 
> Troels Ring
> Aalborg, Denmark
> 
> ______________________________________________
> 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
-- 
Andrew Robinson
Senior Lecturer in Statistics                       Tel: +61-3-8344-9763
Department of Mathematics and Statistics            Fax: +61-3-8344-4599
University of Melbourne, VIC 3010 Australia
Email: a.robinson at ms.unimelb.edu.au    Website: http://www.ms.unimelb.edu.au