Since the advent of digital technology the definitions of the words “accuracy” and “precision” have become confused, their subtle distinctions lost. In a digital world the quality or value of a measurement has become synonymous with the ability to determine a calculation to the twelfth, or even fiftieth decimal point—and beyond. Dazzled by the precision of a calculation, we have come to assume that the effort taken to determine the measurement (the calibration of the actual stuff of the tangible world), the “accuracy” of the measurement, is demonstrated by the “precision” of the expressed result.
But if one looks at the words themselves, the etymology of them, it is here that the distinction between “accuracy” and “precision” becomes clear: According to the Oxford English Dictionary (O.E.D.) the word “precise” comes from the Latin adjective precisus: cut off, abrupt, shortened; from the past participle of the Latin verb precidere meaning, to cut off (in front), cut short, abridge. Here are the O.E.D. definitions of “precise": Definitely or strictly expressed, exactly defined, definite, exact. […]correct. Punctilious, scrupulous, correct. Sometimes, Over-exact […] Exact, neither more nor less than, perfect, complete: as opposed to approximate. Distinguished with precision from all others, identified, pointed out, or stated with precision or exactness […]
But if one looks at the words themselves, the etymology of them, it is here that the distinction between “accuracy” and “precision” becomes clear: According to the Oxford English Dictionary (O.E.D.) the word “precise” comes from the Latin adjective precisus: cut off, abrupt, shortened; from the past participle of the Latin verb precidere meaning, to cut off (in front), cut short, abridge. Here are the O.E.D. definitions of “precise": Definitely or strictly expressed, exactly defined, definite, exact. […]correct. Punctilious, scrupulous, correct. Sometimes, Over-exact […] Exact, neither more nor less than, perfect, complete: as opposed to approximate. Distinguished with precision from all others, identified, pointed out, or stated with precision or exactness […]
According to the O.E.D. the word “accurate” comes from the Latin adjective accuratus: performed with care, exact; from the past participle of the verb, accurare, meaning, to apply care to. Here are the O.E.D. definitions of “accurate": Executed with care; careful; of things and persons: exact, precise, correct, as the result of care.
And later, as the word evolved—and for the most part, became synonymous with “precise": 3. Of things, without special reference to the evidence of care: exact, precise, correct […]
The concept of “care” is the critical distinction, I think. You can be accurate without necessarily being precise—as long as you “take care” to get the facts straight. For example, we can do everything possible, with all the resources at our disposal, to determine a realistic estimate of Canada’s wheat production in metric tons for 2004 and come up with a result of, say “about thirty million” and be closer to the truth (again “the actual stuff of the tangible world”) than if we quickly (and possibly carelessly) determined the output to be, say “23.56834943 million metric tons.”
In the past, in an analog world, (“This cabbage weighs about two pounds.”) accuracy was the result of successive measurements that raised the order of approximation . The precision of the measurement was directly related to the care with which the measurement was taken. Precision followed accuracy (careful calibration): one was the direct result of the other (imagine a balance scale with progressively smaller counterweights).
As the analog measuring tools became more sophisticated, (The analog clock, for example, evolved in refinement from an apparatus with an hour hand to one with hour and minute hands, to one with hour, minute and second hands.) we moved towards higher and higher orders of approximation; but the ability of the tool to provide a “precise” measurement was directly proportional to the care with which the tool was used. Inaccurate measurement was always imprecise.
And later, as the word evolved—and for the most part, became synonymous with “precise": 3. Of things, without special reference to the evidence of care: exact, precise, correct […]
The concept of “care” is the critical distinction, I think. You can be accurate without necessarily being precise—as long as you “take care” to get the facts straight. For example, we can do everything possible, with all the resources at our disposal, to determine a realistic estimate of Canada’s wheat production in metric tons for 2004 and come up with a result of, say “about thirty million” and be closer to the truth (again “the actual stuff of the tangible world”) than if we quickly (and possibly carelessly) determined the output to be, say “23.56834943 million metric tons.”
In the past, in an analog world, (“This cabbage weighs about two pounds.”) accuracy was the result of successive measurements that raised the order of approximation . The precision of the measurement was directly related to the care with which the measurement was taken. Precision followed accuracy (careful calibration): one was the direct result of the other (imagine a balance scale with progressively smaller counterweights).
As the analog measuring tools became more sophisticated, (The analog clock, for example, evolved in refinement from an apparatus with an hour hand to one with hour and minute hands, to one with hour, minute and second hands.) we moved towards higher and higher orders of approximation; but the ability of the tool to provide a “precise” measurement was directly proportional to the care with which the tool was used. Inaccurate measurement was always imprecise.
Not so with digital tools. In a digital environment, accuracy is seldom at issue: a careless measurement will be as precise as a careful one.
No comments:
Post a Comment