>> Date: Tue, 5 Jan 1999 15:16:39 +0200
>> Subject: Deeming Regression
>> From: Ralph Katzenell <[log in to unmask]>
>> To: "'ACB Discussion Group'" <[log in to unmask]>
>> Reply-to: Ralph Katzenell <[log in to unmask]>
>
>> Happy New Year, All
>>
>> Would someone be kind enough to point me to a resource which allows Deeming
>> Regression analysis.? The stuff I have been able to find just skirts around
>> the issue. Help, please !!
>>
>> And why is discussion of this procedure not appearing in many modern Clin.
>> Chem. Text books?
>>
>> Ralph Katzenell
>> Customer Support
>> Orsys Ltd
>>
>>
>There is a good comparison between Deming's regression and the
>classical regression in:
>
>Wakker, P.J.M., Hellendoorn, H.B.A., et al.
>APPLICATION OF STATISTICS IN CLINICAL CHEMISTRY
>A CRITICAL EVALUATION OF REGRESSION LINES
>Clinica Chimica Acta 64, 173-184 (1975)
>Dr.H.Mueller
>Institut fuer Laboratoriumsmedizin
There also is a short but more recently written page about Deming
regression in CLINICAL INVESTIGATION AND STATISTICS IN LABORATORY MEDICINE
by Richard Jones and Brian Payne, ACB Venture Publications, London, 1997.
Software for method evaluation is available that may include Deming
regression capability:
ANALYSE-IT
Analyse-It Software, Ltd.
P.O. Box 77
Leeds, LS12 5 XA, UK
EP Suite
MarChem Associates, Inc.
325 College Road
Concord, MA 01742
EP Evaluator
David G. Rhoads Associates
504 Meetinghouse Lane
Kennett Square, PA 19348
CB Stat
Kristian Linnet
Vibavej 3
DK-8240 Risskov
Denmark
Deming regression is becoming reasonably common on routine scientific
calculators, as well.
For what it's worth, I don't think a discussion of the Deming regression
procedure belongs in routine clinical chemistry textbooks. Method
evaluation concepts should be covered in these texts, but least squares
linear regression
is sufficient for the great majority of evaluation experiments encountered
in clinical chemistry. The effect of the failure to know the "x" values
without error (which the Deming regression overcomes) is negligible when
the range of the data is sufficiently large.
Best regards,
Dave Koch, Ph.D.
Associate Professor, Pathology & Laboratory Medicine
University of Wisconsin
Madison, WI
608-263-1507
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|