, Volume 9, Issue 8, pp 441-444
Date: 28 May 2004

Comments on “Cofino statistics”

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Cofino et al. have put forward a method for the analysis of data from interlaboratory studies. The method uses ideas and notation from quantum chemistry, and is quite sophisticated. A critical examination shows that it has some relationship with both kernel density estimation and robust estimation methods. The quantitative results it provides are not founded on any statistical probability model, and it is not at all clear how the variance in particular is to be interpreted. There is no clear argument for adopting Cofino statistics in preference to simple problem-free methods that are at least as effective in estimating a concensus value.

This report was prepared for the Statistical Subcommittee by Professor T. Fearn, with the assistance of Mr T. Robinson, and approved by the AMC on 25/11/03. The Statistical Subcommittee comprised: Dr. S.L.R. Ellison, Professor T. Fearn, Mr M. Gardner, Dr. F. Hollywood, Professor R.J. Howarth, Dr. P. Lowthian, Professor J.N. Miller, Dr. E.J. Newman, Professor B.D. Ripley, Professor M. Thompson (Chair), Dr. R. Wood, Dr. A. Williams and Mr J.J. Wilson (Secretary).