Thursday, August 09, 2012

Precisionism


The Royal Statistical Society invites contributions to a column called Forsooth!

As the OED explains, sooth, as a word meaning truth, was in common use down to the first half of the C17th when it became pretty much obsolete, & forsooth (in truth, truly) is now only used parenthetically with an ironical or derisive statement. Readers of the column are thus invited or expected to feel derision for the hapless author of some statistical or arithmetic solecism

I have to admit that sometimes I am myself too ignorant to understand the mockers’ point, but sometimes I think an entry a mere indulgence of the kind that gets statisticians a bad name.

This month the statisticians have chosen to have a go at a mere crossword setter (& their editor), who allowed the answer ‘accuracy’ to the one-word definitional clue ‘precision

I love, find useful & have frequently taught the distinction which exists – in the minds of statisticians & engineers at least – between precision & accuracy, & have sometimes railed at misleading precision of estimates, but I do not feel the need to give lectures or try to impose restrictions on the use of these two words in language generally. To do so would be presumptuous & hopeless, unless & until the data revolution makes its adoption by the wider public natural & useful.

The dictionaries & thesauruses I have consulted all give precision as a definition or synonym for accuracy, & vice versa.

The OED finds no examples of the distinction between precision being used to describe ‘The degree of refinement in a measurement, calculation, or specification, esp. as represented by the number of digits given. Contrasted with accuracy (the closeness of the measurement, etc., to the correct value)’ before 1842 in a refernce to a table of ‘Logarithmic Sines and Tangents to four decimal figures ... convenient in many computations not requiring greater precision’.

To find a quotation which uses accuracy in this nice sense (‘The closeness of a measurement, calculation, or specification to the correct value. Contrasted with precision (the degree of refinement of the measurement, etc.’) we have to wait for an 1876 Handbook on electricity: ‘We can thus determine the value of a resistance to an accuracy of 1/ 100th of a unit.’

It is the mark of an educated mind to rest satisfied with the degree of precision that the nature of the subject admits & not to seek exactness when only an approximation is possible – Aristotle

Link
Forsooth August 2012
Related posts
Heteroskedacity
Forsooth begads