The Concept of Measurement Accuracy

By Niraj M Wanikar, Chief Editor, CNCTimes.com, 2016-05-25 12:05:00

Measurements are essential in all walks of life. Just like any other field, there are certain commonalities among different kinds of measurements.

For e.g., the vegetable vendor from whom we buy vegetables has a simple fulcrum-balance for measuring weight. When we ask for 2 Kg of carrots, how many carrots do we get? Not all carrots are of the same size. The vendor puts 27 carrots on one side and a 2 Kg steel block (the master, against which the balance does a comparative measurement). The needle is not at the centre exactly and is showing carrots are weighing 15g less than 2Kg. Now, the smallest carrot in the vegetable cart weighs 32g. Should we accept the loss of 15g, or ask for 17g extra at no extra cost? To decide if we are actually getting more or less, we first need to ascertain the exact weight of the steel blocks. How will you know that the blocks are accurate? There may be a few vendors who will willfully cheat by reducing the weight of blocks by 100g. Then, for every 10 kg of vegetables sold, they will make a profit of 1 kg!!

This example is only that – an example. Similarities to real world conditions are incidental and unintended. Further, not ALL street vendors cheat. In fact, a majority of them do an honest days’ work, to the best of their abilities, and to the level of their knowledge of Metrology.

Irrespective of the knowledge, capability or intention of the person doing the measuring, an incorrect measurement remains incorrect. A vendor cannot claim that his 2 kg measurement is correct simply because he has sold 200kg of vegetables using it. If the weight block is wrong, EVERY kilogram of vegetable the vendor has sold is wrong. We see similar situations in dimensional measurements, in the manufacturing industry, every day.

This leads to the fundamental question – can we measure anything EXACTLY? If yes, we don’t have to worry. But almost always, the answer to this question remains a No.

Why is that so?

We can do well to recall “Heisenberg’s Uncertainty Principle”. Although he discovered the principle in quantum mechanical work, it is quite applicable in normal life. We cannot measure anything accurately in an absolute sense. We can only try to measure a property to some degree of accuracy, and certainly at a cost (in terms of effort, time and money). The more accuracy we want, the more expensive it usually becomes.

Is this necessarily true?

Well, Yes and No. There is a cost to doing things, and there is smartness in generating the ability to do things, and then, there is a smartness to do that faster, easier, better and cheaper etc. So, there is ALWAYS a trade-off between how accurate we want to measure something, and how much we are willing to pay for it. Next, we need to have a method or standard to know HOW correct a particular measurement SHOULD be. Once we have this, we can decide the accuracy of measurement, and the method involved (and hence arrive the cost of measurement).

Now let us apply this new concept of how accurately we want to measure something to our vegetable problem. When a pharmacist measures a medicine, they will be accurate probably to the nearest milligram. This raises a few more questions.

Do we need to get to that accuracy when buying vegetables? How do we fix the acceptable limits of a required property that are being measured?

I hope to answer these questions in my next post. Till then, stay tuned.

	Dr. R.A. Narayanan is an accidental Metrologist, who entered this field because of circumstances. He has over 10 years of experience in Software Development and dimensional metrology. He has a Ph.D. in Chemical Engineering and currently heads  Software Development for Rapid-I Precision Measuring Systems.

• Follow @CNCTimes