OMB Meeting Book - January 8, 2015

31

DESCRIPTION OF AOAC STATISTICS COMMITTEE GENERATED DOCUMENTS

ALTERNATIVE APROAC H ES FOR MULTI‐LAB STUD Y DOCUMNTS.    Alternative approaches approved by Committee on Statistics and by OMB.  1. *tr322‐SAIS‐XXXV‐Reproducibility‐from‐PT‐data.pdf: Discussion on how  to obtain reproducibility from proficiency test data, and the issues  involved. 

2

……………………………… 

2. tr333‐SAIS‐XLII‐guidelines‐for‐use‐of‐PT‐data.pdf: Recommended use  of proficiency test data for estimating repeatability and reproducibility. ………………………………  8 3. *tr324‐SAIS‐XXXVII‐Incremental‐collaborative‐studies.pdf: Proposed 

incremental collaborative studies to find repeatability and  reproducibility via a sequential series of experiments. 

11

……………………………...

4. tr326‐SAIS‐XXXIX‐Min‐degr‐freed‐for‐random‐factor‐estimation.pdf:  The relationship of number of replicates or number of collaborators to  precision of standard deviation for repeatability or reproducibility.  5. tr323‐SAIS‐XXXVI‐When‐robust‐statistics‐make‐sense.pdf: Caveats and  recommendations on the use of so‐called ‘robust’ statistics in  accreditation studies.  TRADITIONAL PATHWAY MULTI‐LAB STUDY DOCUMENTS AND INSTRUMENTS.      Traditional study approach spreadsheet remaining as part of acceptable approaches.  6. JAOAC 2006 89.3.797_Foster_Lee1.pdf: Journal article by Foster and  Lee on the number of collaborators needed to estimate accurately a  relative standard deviation for reproducibility.  7. LCFMPNCalculator.exe: This program analyzes serial dilution assay data  to obtain a most probable number estimate of concentration with  confidence interval.  8. AOAC_BlindDup_v2‐1.xls: This workbook analyzes multi‐collaborative  study data with up to 4 replicates and reports all necessary statistics  concerning repeatability and reproducibility for quantitative studies.  9. AOAC‐binary‐v2‐3.xls: This workbook analyzes multi‐collaborative  study data with arbitrary number of replicates and reports all  necessary statistics, including POD, repeatability and reproducibility  with confidence intervals, for binary (presence/absence) qualitative  studies. 

19

……………………………… 

21

……………………………… 

29

……………………………… 

Separate Program

Separate Program

Separate Program