Sid Talks Science: What Uncertainty Means in Testing

4/20/15

Should you be concerned about the Uncertainty of Measurement?

Obtaining accurate measurements is a Science unto itself, which has always been quietly working in the background trying to describe, as well as, reduce the size and rate of errors made in science. In Science, the word Error does not carry the usual connotations of the terms mistake or blunder. Error in a scientific measurement means the inevitable uncertainty that is associated with all measurements. Errors are not mistakes; they cannot be eliminated by being more careful. The best we can hope for is to ensure that errors are as small as possible and to have a reliable estimate of how large they are. The information provided by that estimate is in fact a key component of test result analysis.

To illustrate the inevitable occurrence of uncertainties, we will look at the measurement of density of gold, to decide whether a sample is actually gold, of known density 15.5 gram/cm3, or a substitute of lesser value of known density 13.8 gram/cm3. Suppose we ask two experts to provide us with the measurement of the density. The first expert makes a quick measurement and reports his best estimate is 15 and that it almost certainly lies between 13.5 and 16.5 gram/cm3. The second expert takes a little longer to make her best estimate of 13.9 with a probable range from 13.7 to 14.1 gram/cm3 If we knew only the two best estimates, we would not be able to draw a valid conclusion, and could actually be misled, because the first expert’s result (15) seems to suggest the gold is genuine. This is where Measurement Uncertainty helps reduce uncertainty.

Each expert gives a range within which he or she is confident the density lies, and these ranges overlap.   It just so happens that the uncertainty (range) in the first experts measurement is so large that his results are of no actual use, since the densities of both the gold and the substitute lie within his range, from 13.5 to 16.5 gram/cm3 ; so no conclusion can be drawn from his measurements. On the other hand, the second expert’s measurements indicate clearly that the gold is not genuine, as the density of the suspected substitute, 13.8, lies comfortably inside her estimated range of 13.7 to 14.1 and the density of gold, 15.5, is far outside it.

As clearly illustrated above, we must conclude that not all measurements are created equally and it is important to know this is the case and what questions to ask to determine the validity of any scientific measurements.

While this conclusion may be convincing some of you, it should raise many more questions for others, like how these ranges were determined and why should we believe these experts claims? Without a brief explanation of how the uncertainty was estimated, the results are almost useless. This process is currently required for Accreditation to ISO 17025, and having it is just plain good business and good Science. For now, though, the most important point about the two experts’ measurements is this: Like most scientific measurements, they would both have been useless if they had not included reliable statements of their uncertainties.

Sydney Sudberg

Each month Sidney Sudberg, Alkemist Founder & CSO, helps to demystify the science behind testing by discussing a testing-related topic.

Categories: News & Events