- #1
- 4,717
- 2,984
Maths is a precice science*, but when using maths in the real world most people, including many mathematicians, generally take 'short cuts' which result in specifications of a problem or descriptions of a solution which are strictly incomplete. This has a number of advantages:
This seems to be particularly true of this part of this forum. I have noticed on a number of occasions that posters ask a question which is naively and incompletely specified, such as 'I measure the weight of 100 boxes of cereal marked 500g, how do I work out the probablility that the supplier is underfilling the boxes' and instead of providing an outline of a null hypothesis and how to use the normal distribution to provide confidence limits, perhaps together with some of the key assumptions and limitations of such an approach, the response is a monologue on Bayesian statistics.
Is this a good thing?
* except of course for statistics which attempts to make precise statements about data which is essentially imprecise
- it is quicker
- it enables problems and solutions to be described in language which is accessible to non-specialists
- it avoids hiding the essence of a problem or a solution within a cloud of definitions, qualifications and jargon
This seems to be particularly true of this part of this forum. I have noticed on a number of occasions that posters ask a question which is naively and incompletely specified, such as 'I measure the weight of 100 boxes of cereal marked 500g, how do I work out the probablility that the supplier is underfilling the boxes' and instead of providing an outline of a null hypothesis and how to use the normal distribution to provide confidence limits, perhaps together with some of the key assumptions and limitations of such an approach, the response is a monologue on Bayesian statistics.
Is this a good thing?
* except of course for statistics which attempts to make precise statements about data which is essentially imprecise