Review Book

data sets..." starts a large section where 1) new parameters outside of Appendix J guidelines are created which 2) pool statistics across levels and across matrices, which is not allowed by AOAC guidelines. The "False Positive Rate" and "False Negative Rate" parameters are unnecessary, as all of the information is given in the results section as to the number of discordant pairs in each matrix. Pooling is unwise here as it opens up the possibility of a Simpson's Paradox ‐ type situation where the conclusions drawn from the pooled data may be inconsistent with those from the unpooled analysis. In any case, it is unlikely these estimates are homogeneous across all 11 levels, so pooling is not appropriate. The reference to these 2 new parameters should be removed from the manuscript. They are both unnecessary and inappropriate. EDITORIAL EVALUATION CRITERIA Comments: none RECOMMENDATION: Manuscript needs the indicated revisions to be statistically acceptable and correspond with the study protocol and/or guidelines.

21. Please select the type of method being reviewed: Reviewer 1 AOAC First Action status Reviewer 2 AOAC First Action status Reviewer 3 AOAC First Action status Reviewer 4 AOAC First Action status Reviewer 5 AOAC First Action status Reviewer 6 AOAC First Action status

22. Do you recommend that the ERP adopt this method as an AOAC Official Methods of Analysis First Action Method? If no, please explain your scientific reasoning. Reviewer 1 YES Reviewer 2 YES Reviewer 3 NO The method should be in general use for 2 years as a First Action method, before being elevated to a Final Action method. Reviewer 4 YES Reviewer 5 YES Reviewer 6 (implied Yes based on review form) 23. Is there any additional information that the ERP should consider in order to recommend the method for Final Action status? If YES, please note your recommendations. Reviewer 1 NO Reviewer 2 YES Solicited feedback from users Reviewer 3 YES