Uncertainties
From CheLabWiki
…from CheLabWiki, an online resource for chemical-engineering laboratories located at www.chelabwiki.org; Site Revision #356; 6 January 2009.
Community Portal → Data Analysis → Uncertainties
Although errors cannot be known, it is possible to quantify uncertainties; by an uncertainty we mean a range of values that is likely to contain the (unknown) exact value. This range is consistent with the measured data, the experimental protocol, and our understanding of how the world works. Following the ANSI guide,[1] we divide uncertainties into two kinds:
- Type A uncertainties are those in which we use repeated sampling to quantify the distribution of values. We then obtain the uncertainty from a statistical analyses of the repeated measurements.
- Type B uncertainties are those in which a distribution is assumed, based on experience, other experiments, reference works, etc. Hence, we do not use a statistical analysis to obtain these uncertainties.
We caution that type A uncertainties are not necessarily confined to statistical errors nor are type B uncertainties necessarily confined to systematic errors. Uncertainties are independent of errors; we do not attempt to quantify errors, which are unquantifiable, but we do quantify uncertainties
Steps to Estimate an Uncertainty
The steps to obtain an estimate of an uncertainty are as follows:
- Use repeated measurements to determine the mean value and to estimate the type A uncertainty.
- Use your knowledge of the apparatus, calibration curves, effects of environmental variables, etc. to estimate the type B uncertainty.
- Combine your values for the type A and type B uncertainties to obtain the total uncertainty.
- If you want to assign a confidence level, apply a coverage factor to the total uncertainty to obtain an expanded uncertainty.
- Follow accepted standards when reporting the mean and its uncertainty.
References
- ↑ U. S. Guide to the Expression of Uncertainty in Measurement, ANSI/NCSL Z540-2-1997, American National Standards Institute, NCSL International, Boulder, CO, 1997.
