A further study of the impact of bias on accountability
When measurements are obtained for the purpose of material accountability, any deviation of averages from true value is one of great concern. The key term here is true value. Any deviation of averages from their true value will result in inventory discrepancies. It will not matter that the bias is of no statistical significance. Even small biases will over a period of time result in large inventory discrepancies. The terms `target value`, `accepted value`, and `certified value` are all very useful terms, and they certainly have their place in the statistical arena. However, the one of concern here is true value. True value, of course, is never known, but even with best efforts, deviations from it are often much larger than one has historically-believed them to be. With all of this in mind and with the aid of statistical calculations, simulation techniques, and statistical reasoning, the discussion here appears to provide reasonable estimates of the truth. The authors conclude that it is possible, in theory, to apply propagation of variance (POV) techniques when assigning limits of uncertainty about inventory differences which are based on the art of measuring. However, because of one`s inability to obtain reliable estimates of true biases, it is misleading to attempt to apply them with any stated degree of confidence. When, for example, one says that one is 95% confident, one could be in error by almost any amount. Such error depends almost totally on how well one is able to estimate these small `difficult to identify` deviations from true value. Even when biases are of significant size, large percentages of them are very often the result of randomness in the data and/or random variation in control chart center lines. Also, this study would not have seen those cases where both the center line and data have biases of the same direction and size. In such cases the combined random variation would permit both biases to go undetected.
- Research Organization:
- Oak Ridge Y-12 Plant (Y-12), Oak Ridge, TN (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC05-84OS21400
- OSTI ID:
- 505368
- Report Number(s):
- Y/DG-26634; ON: DE97007844; TRN: 97:013367
- Resource Relation:
- Other Information: PBD: Mar 1997
- Country of Publication:
- United States
- Language:
- English
Similar Records
THE EVOLUTION OF CATACLYSMIC VARIABLES AS REVEALED BY THEIR DONOR STARS
Evaluation of Uncertainties in the Anthropogenic SO2 Emissions in the USA from the OMI Point Source Catalog