Research on Chi-Square Statistics in Data Analysis of Inter-laboratory Comparison

HANG Chenzhe; XU Dinghua; YUAN Zundong

doi:10.12338/j.issn.2096-9015.2020.9042

Volume 65 Issue 5

Jun. 2021

Turn off MathJax

Article Contents

Article Navigation > Metrology Science and Technology > 2021 > 65(5): 108-114

HANG Chenzhe, XU Dinghua, YUAN Zundong. Research on Chi-Square Statistics in Data Analysis of Inter-laboratory Comparison[J]. Metrology Science and Technology, 2021, 65(5): 108-114. doi: 10.12338/j.issn.2096-9015.2020.9042

Citation:

HANG Chenzhe, XU Dinghua, YUAN Zundong. Research on Chi-Square Statistics in Data Analysis of Inter-laboratory Comparison[J]. Metrology Science and Technology, 2021, 65(5): 108-114. doi: 10.12338/j.issn.2096-9015.2020.9042

Citation:

PDF( 1222 KB)

Research on Chi-Square Statistics in Data Analysis of Inter-laboratory Comparison

doi: 10.12338/j.issn.2096-9015.2020.9042

National Institute of Metrology, Beijing 100029, China

Available Online: 2021-04-21
Publish Date: 2021-06-24

Abstract

Abstract

The chi-square statistics, which can be used in the consistency test of comparison results and the estimation of reference value uncertainty, is a key statistical tool in data analysis of inter-laboratory comparison. In this study, under the condition that the comparison results obey Gaussian distribution with a common mean, a chi-square statistic containing generalized linear estimation is proposed, and the properties and distribution of the statistic are investigated. The statistic enables consistency testing and uncertainty estimation of general linear reference value estimates, and provides a statistical tool for a wider range of linear reference value estimates, which can be used for comparison data analysis or multi-laboratory fixed value measurements. As an example, against the limitation of using the traditional chi-square test for the arithmetic means based on the same uncertainty claimed by each laboratory, this method gives the chi-square statistic for the arithmetic mean under any combination of uncertainties, providing a new statistical analysis method for the extended application of this common linear reference value estimation.
- inter-laboratory comparison,
- reference value estimation,
- chi-square statistics,
- generalized model,
- random effects model,
- Monte Carlo simulation

FullText(HTML)

References(12)

References

[1]	原遵东, 王池, 林延东. 确定量值比对参考值的准则[J]. 仪器仪表学报, 2010, 31(10): 2381-2385.
[2]	张海云, 吴忠杰, 杭晨哲, 等. MCMC法确定关键比对参考值及其不确定度[J]. 计量技术, 2020(7): 3-7, 26. doi: 10.3969/j.issn.1000-0771.2020.07.01
[3]	Cox M G. The evaluation of key comparison data[J]. Metrologia, 2002, 39(6): 589-595. doi: 10.1088/0026-1394/39/6/10
[4]	Koepke A, Lafarge T, Possolo A, et al. Consensus building for interlaboratory studies, key comparisons, and meta-analysis[J]. Metrologia, 2017, 54(3): S34-S62. doi: 10.1088/1681-7575/aa6c0e
[5]	Toman B, Possolo A. Laboratory effects models for interlaboratory comparisons[J]. Accred Qual Assur, 2009, 14(10): 553-563. doi: 10.1007/s00769-009-0547-2
[6]	Elster C, Toman B. Analysis of key comparisons: estimating laboratories’ biases by a fixed effects model using Bayesian model averaging[J]. Metrologia, 2010, 47(3): 113-119. doi: 10.1088/0026-1394/47/3/001
[7]	Steele A G, Douglas R J. Extending chi-squared statistics for key comparisons in metrology[J]. Journal of Computational and Applied Mathematics, 2006(192): 51-58.
[8]	Rukhin A L. Weighted means statistics in interlaboratory studies[J]. Metrologia, 2017, 46(3): 323-331.
[9]	Toman B, Fischer J, Elster C. Alternative analyses of measurements of the Planck constant[J]. Metrologia, 2012, 49(4): 567-571. doi: 10.1088/0026-1394/49/4/567
[10]	Paule R C, Mandel J. Consensus values and weighting factors[J]. J. Res. Natl Bureau Stand, 1982, 87(5): 377-385. doi: 10.6028/jres.087.022
[11]	Graybill F A, Deal R B. Combining unbiased estimators[J]. Biometrics, 1959(15): 543-550.
[12]	BIPM. CCQM guidance note: estimation of a consensus KCRV and associated degrees of equivalence[DB/OL]. http://www.bipm.org/cc/CCQM/Allowed/19/CCQM13-22_Consensus_KCRV_v10.pdf, 2013: 23-25.