Statistical Aspects of Linkage Analysis in Interlaboratory Studies

Published

November 9, 2005

Author(s)

W Strawderman, Andrew L. Rukhin

Abstract

This paper investigates issues that arise in statistical inference in interlaboratory studies known as Key Comparisons when one has to link several comparisons to or through existing studies. A new approach to the analysis of such a data is proposed using a Gaussian distributions model often employed in meta-analysis. We develop conditions for the set of sufficient statistics to be complete and characterize unique uniformly minimum variance unbiased estimators of the contrast parametric functions. New procedures are derived for estimating these functions with estimates of their uncertainty. In particular, the dependence of these procedures on reported uncertainties based on statistical estimates of standard deviations (Type A) and on scientific judgment (Type B) of participating laboratories is examined.These estimates lead to associated confidence intervals for the laboratories (or studies) contrasts. Several examples demonstrate statistical inference for contrasts based on linkage through the pilot laboratories. Monte Carlo simulation results on performance of approximate confidence intervals are also reported.

Citation

Journal of American Statistical Association

Pub Type

Journals

Download Paper

Local Download

Keywords

Behrens-Fisher distribution, common mean problem, completeness, confidence intervals, contrasts, Graybill-Deal estimator, key comparisons, sufficient statistics, unbiasedness

Citation

Strawderman, W. and Rukhin, A. (2005), Statistical Aspects of Linkage Analysis in Interlaboratory Studies, Journal of American Statistical Association, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=151793 (Accessed January 7, 2025)