Yet Another Cycling Forum

General Category => The Knowledge => OT Knowledge => Topic started by: Quisling on 26 October, 2017, 12:06:19 pm

Title: statistics question
Post by: Quisling on 26 October, 2017, 12:06:19 pm
Help me out.  I should know this but...

A baseline measurement is taken with a reading of 100 using metering rated at +/- 10% of the reading.
Following efficiency works, a post-retrofit reading of 80 is taken using the same metering (therefore same uncertainty).

What is the uncertainty in the demand reduction
( A ) 10%
( B ) 14%
( C ) 20%
( D ) 64%

What is your answer and why?
Title: Re: statistics question
Post by: Veloman on 26 October, 2017, 12:17:24 pm
Has the uncertainty of the metering measurement changed?  If not then your answer is within the text.
Title: Re: statistics question
Post by: Quisling on 26 October, 2017, 12:31:55 pm
That’s what I’d have thought, but the sample exam paper gives the answer as (D) which I can’t fathom.
Title: Re: statistics question
Post by: philip on 26 October, 2017, 12:41:41 pm
You have 100 +/- 10 and 80 +/- 8. When calculating the difference of two quantities the errors would typically be "added in quadrature", so the error is the square root of 10 squared plus 8 squared, i.e. root 164 or 12.8. The difference is thus 20 +/- 12.8 and 12.8 is 64% of 20.

Title: Re: statistics question
Post by: Quisling on 26 October, 2017, 12:50:04 pm
You have 100 +/- 10 and 80 +/- 8. When calculating the difference of two quantities the errors would typically be "added in quadrature", so the error is the square root of 10 squared plus 8 squared, i.e. root 164 or 12.8. The difference is thus 20 +/- 12.8 and 12.8 is 64% of 20.

Bingo! Thank you Philip.

I'd forgotten to calculate the standard error.  It's been a few years since I did my studies.  Hence the revision.  It's going well  ::-)