Need Maths Forumla. % Error

I would normally use it as

(|correct - ob|/correct)*100

KaHn

/edit the |fghgf| is the magnitude so you dont get minus
 
err doesn't that assume that you know what the quantity you are measuring is with absolute certainty? I can't think of any example where this is the case and if such an example did exist then why would you bother measuring it? Or am I missing something...
 
PhilthyPhil said:
err doesn't that assume that you know what the quantity you are measuring is with absolute certainty? I can't think of any example where this is the case and if such an example did exist then why would you bother measuring it? Or am I missing something...
Click to expand...

Counting cars past a lamp post? You can count can't you?
 
The forumla i use to calculate error is for example, if you have value A and B and you have measuerd them, you must have an uncertainty in you values that you have obtained. So, if you have an equation say...

C=AxB

And you want to know the error in C due to A and B, then use the following.

(Error in C / C)^2=(Error in A / A)^2 + (Error in B / B)^2

Where A and B are the values you have measured, and the error is the uncertainty in the measuring device. The simply rearrange for "error in C" and there you have it.

That formula works for A/B as well as AxB

If you need any more help using that, PM me, if it aint what you need, sorry :)

Woody
 
Amleto said:
Counting cars past a lamp post? You can count can't you?
Click to expand...

I don't see how you can guarantee that you don't make any mistakes and if you didn't you would have 0% error which would be meaningless. If you did make mistakes how would you find out what the real value was in order to determine your error. Surely an error calculation needs to be based on some estimation of the uncertainty in your data and/or some sort of statistical jiggery pokery.
 
Last edited:
Amleto, you need to realise that when measuring something there is always error if it is more complicated than counting cars. Silly argument tbh.

Say one is measuring a length with a micromenter, or force from an elastic band, something simple for the sake of argument. Can you be absolutely sure you are measuring the distance or force exactly? Every device has its own uncertainty, as it has a limited accuracy in what it measures.

PhilthyPhil, yes your right. as i mention above, and as you correctly pointed out, To estimate the error in your final value, you need to know the error in your data.
 
woodsy2k said:
Amleto, you need to realise that when measuring something there is always error if it is more complicated than counting cars. Silly argument tbh.
Click to expand...

I already know this. I just gave an example where the uncertainty is 0. Not so silly imo.
 
Do you have an expected result? I would normally just do (experimental result)/(analytical value) and that gives you a straight percentage.
 
Seem that im getting the "chemisty" % erros forumula as I call it.

I had to go out so couldnt type that much detail

So here is more detail...Probably better if I just give you the excel file.

Anyone wana host it for me.

Its bascialy I have some results that I have averaged out and I need to find the %error in that average.

Iv used the average again with some more averages so you can see that there will be some % error in those numbers.
 
You must log in or register to reply here.
Back
Top Bottom