Comparison of Bootstrap Confidence Interval Methods in a Real Application: The Kappa Statistic in Industry

Authors

Krik O. Monson, Brigham Young University
Dr. Bruce G. Schaalje, Brigham Young University

Keywords

bootstrap confidence interval methods, kappa statistic

College

Physical and Mathematical Sciences

Department

Statistics

Abstract

The kappa statistic is defined as the proportion of agreement between two instruments after chance agreement is removed from consideration (Cohen 1960). It is very useful when assessing the similarities of two instruments or procedures. Industries often use an automated system for testing their products to ensure a level of quality before packaging their goods. Since these systems parts must be periodically replaced, it is necessary to determine if the new components have the same level of accuracy or performance as the old ones. The kappa statistic would be very useful in this situation if a more reliable method for determining confidence intervals was available.

Recommended Citation

Monson, Krik O. and Schaalje, Dr. Bruce G. (2013) "Comparison of Bootstrap Confidence Interval Methods in a Real Application: The Kappa Statistic in Industry," Journal of Undergraduate Research: Vol. 2013: Iss. 1, Article 2812.
Available at: https://scholarsarchive.byu.edu/jur/vol2013/iss1/2812