Critique of the AIAG cross-tabulation procedure for attribute gauge R&R study

Emese Vágó, S. Kemény

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The automotive industry action group’s manual for attribute measurement system analysis has two major deficiencies: 1. the improper use of cross-tabulation and Cohen’s kappa when operators are compared in case of repeated ratings, and 2. calculation of the probabilities of wrong decisions with the Bayes theorem, however the conditions of its application are not fulfilled. Besides the AIAG manual we reviewed the approaches to attribute measurement system analysis proposed in the literature and used by some of the most popular statistical software. The conclusion is that all the methods referred to are statistically flawed. They do not properly consider the multiple ratings. The need has arisen for a different approach.

Original languageEnglish
Pages (from-to)75-93
Number of pages19
JournalInternational Journal of Quality Engineering and Technology
Volume2
Issue number1
DOIs
Publication statusPublished - 2011

Fingerprint

Gages
Systems analysis
Automotive industry

Keywords

  • AIAG
  • attribute
  • Automotive Industry Action Group
  • Bayes
  • kappa
  • measurement system
  • R&R
  • repeatability
  • reproducibility

ASJC Scopus subject areas

  • Safety, Risk, Reliability and Quality

Cite this

Critique of the AIAG cross-tabulation procedure for attribute gauge R&R study. / Vágó, Emese; Kemény, S.

In: International Journal of Quality Engineering and Technology, Vol. 2, No. 1, 2011, p. 75-93.

Research output: Contribution to journalArticle

@article{2967bb439df44bc69138dbbb1c7390b5,
title = "Critique of the AIAG cross-tabulation procedure for attribute gauge R&R study",
abstract = "The automotive industry action group’s manual for attribute measurement system analysis has two major deficiencies: 1. the improper use of cross-tabulation and Cohen’s kappa when operators are compared in case of repeated ratings, and 2. calculation of the probabilities of wrong decisions with the Bayes theorem, however the conditions of its application are not fulfilled. Besides the AIAG manual we reviewed the approaches to attribute measurement system analysis proposed in the literature and used by some of the most popular statistical software. The conclusion is that all the methods referred to are statistically flawed. They do not properly consider the multiple ratings. The need has arisen for a different approach.",
keywords = "AIAG, attribute, Automotive Industry Action Group, Bayes, kappa, measurement system, R&R, repeatability, reproducibility",
author = "Emese V{\'a}g{\'o} and S. Kem{\'e}ny",
year = "2011",
doi = "10.1504/IJQET.2011.038724",
language = "English",
volume = "2",
pages = "75--93",
journal = "International Journal of Quality Engineering and Technology",
issn = "1757-2177",
publisher = "Inderscience Publishers",
number = "1",

}

TY - JOUR

T1 - Critique of the AIAG cross-tabulation procedure for attribute gauge R&R study

AU - Vágó, Emese

AU - Kemény, S.

PY - 2011

Y1 - 2011

N2 - The automotive industry action group’s manual for attribute measurement system analysis has two major deficiencies: 1. the improper use of cross-tabulation and Cohen’s kappa when operators are compared in case of repeated ratings, and 2. calculation of the probabilities of wrong decisions with the Bayes theorem, however the conditions of its application are not fulfilled. Besides the AIAG manual we reviewed the approaches to attribute measurement system analysis proposed in the literature and used by some of the most popular statistical software. The conclusion is that all the methods referred to are statistically flawed. They do not properly consider the multiple ratings. The need has arisen for a different approach.

AB - The automotive industry action group’s manual for attribute measurement system analysis has two major deficiencies: 1. the improper use of cross-tabulation and Cohen’s kappa when operators are compared in case of repeated ratings, and 2. calculation of the probabilities of wrong decisions with the Bayes theorem, however the conditions of its application are not fulfilled. Besides the AIAG manual we reviewed the approaches to attribute measurement system analysis proposed in the literature and used by some of the most popular statistical software. The conclusion is that all the methods referred to are statistically flawed. They do not properly consider the multiple ratings. The need has arisen for a different approach.

KW - AIAG

KW - attribute

KW - Automotive Industry Action Group

KW - Bayes

KW - kappa

KW - measurement system

KW - R&R

KW - repeatability

KW - reproducibility

UR - http://www.scopus.com/inward/record.url?scp=84871015694&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871015694&partnerID=8YFLogxK

U2 - 10.1504/IJQET.2011.038724

DO - 10.1504/IJQET.2011.038724

M3 - Article

AN - SCOPUS:84871015694

VL - 2

SP - 75

EP - 93

JO - International Journal of Quality Engineering and Technology

JF - International Journal of Quality Engineering and Technology

SN - 1757-2177

IS - 1

ER -