, Volume 9, Issue 8, pp 439-440

“Decisions” vs “conclusions”

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access
This is an excerpt from the content

It seems to have escaped many people who are involved in measurement, that ultimately accepting a measurement result (which consists of a quantity value with an associated measurement uncertainty) is a matter of decision rather than an automatic acceptance by the analyst of numbers.

Let us have a closer look at this.

In a typical traditional approach, one may have gathered a lot of evidence through earlier measurements to be confident that a result will lie in a certain range, expresses by a confidence interval with which even a quantified probability can be associated. This eventually leads to stating a result as a value with a confidence interval of, say, 95% probability.

Such a statement is said to imply that if another series of 20 measurements is carried out under similar circumstances, one of them would yield a result which lies outside the confidence interval just described.

Is this useful?

Yes and no.

Looked at “a posteriori”, i.e. after the measurement, confidence intervals describe