Dr Zhou Replies

Xiao-Hua Zhou,Nancy Obuchowski,Donna McClish

DOI: https://doi.org/10.1016/s1076-6332(03)00251-4

IF: 5.482

2003-01-01

Academic Radiology

Abstract:It seems that our remarks ( 1 Zhou X.H. Obuchowski N.A. McClish D.K. Statistical Methods in Diagnostic Medicine. Wiley-Interscience, New York, NY2002 Crossref Google Scholar ) concerning Beiden, Wagner, and Campbell’s (BWC) variance component analysis of multireader ROC (receiver operating characteristic) data ( 2 Beiden S.V. Wagner R.F. Campbell G. Components-of-variance models and multiple-bootstrap experiments. Acad Radiol. 2000; 7: 341-349 Abstract Full Text PDF PubMed Scopus (109) Google Scholar ) have offended some ( 3 Wagner R.F. Beiden S.V. Campbell G. An approach to multiple-reader, multiple-case (MRMC) ROC Analysis controversial - or subtle?. Acad Radiol. 2003; Google Scholar ); we apologize for this. We corresponded with these authors before writing this section of our book ( 1 Zhou X.H. Obuchowski N.A. McClish D.K. Statistical Methods in Diagnostic Medicine. Wiley-Interscience, New York, NY2002 Crossref Google Scholar ) to discuss our concerns about their methodology. We had hoped that their response would alleviate our concerns; it did not, so we proceeded with voicing our concerns in the book and meeting the editor’s fast-approaching publication deadline. It is unfortunate to both parties that these issues were not resolved before the book’s publication. We recognize the scientific merit of open-forum two-way debates of such issues, and we again apologize to BWC and users of their methodology (existing and future).