Submitted to: American Association of Cereal Chemists Meetings
Publication Type: Abstract Only
Publication Acceptance Date: 9/26/2003
Publication Date: 10/2/2003
Citation: Delwiche, S.R. 2003. Practical statistics for collaborative studies [abstract]. American Association of Cereal Chemists Meetings. p. 102.
Technical Abstract: Collaborative, or ring, studies are frequently performed by analytical chemists as a means to document the precision of quantitative procedures. In the cereals community, this may be a gravimetric, titrimetric, amperometric, or physical procedure to measure the concentration of a naturally occurring component, an ingredient, or the response to a processing condition such as dough mixing or baking. Through application of classical statistical principles, the collaborative study produces information on the level of variability (precision) that can be expected of a procedure when it is performed by one laboratory over a reasonable timeframe (repeatability), as well as when it is performed by multiple laboratories (reproducibility). The Approved Methods Committee of the AACC has for a number of years adopted the guidelines of the internationally harmonized procedures of the AOAC International for conducting collaborative studies. This talk will focus on these guidelines to address issues of required numbers of samples, blind duplicates (or similar) and laboratories, statistical tests for outliers, analyses of variance for determining repeatability and reproducibility, and calculation of recovery. User-friendly spreadsheet program macros will be used to demonstrate collaborative study data analysis.