Singletrack Mountain Bike Magazine

• Overview
• Bike Forum
• Chat Forum
• For Sale
• Wanted
• Forum Help
• Classified Rules

Chat Forum

1. dan1980 - Member

   I'm being thick, and could do a hand from someone "in the know"

   I've never been good at maths, and thankfully most of the time my job doesn't require me to do anything more complex than count up to 20 which I can do if I take my shoes and socks off.

   I'm currently doing something that is a lot outside my comfort zone and could really do with checking that I'm on the right lines before I make a complete arse of myself at work tomorrow!

   I'm measuring some samples (F) by getting their mass (m), and measuring a particular property(DI), that I want to plot in terms of unit mass (DI/m vs F).

   So, I've done the measurements and I'm aware that the weighing out bit has an inherent error (I'm going to assume that the machine doing the measurement of DI doesn't have an error)

   So I've calculated the standard error of my masses and have a value (2.5x10^-5 for what it's worth). But I'm confused about how I turn this error into an error bar for Excel to stick on my plot.

   If m actually = m +/- 2.5x10^-5 then do I work out what DI/+m, and DI/-m are, calculate the standard deviation and then standard error of the difference between to the two, and this becomes the standard error of DI/m?

   I appreciate I'm a bit slow, so please be gentle

   Posted 9 months ago #
2. RealMan - Member

   I thought excel had a feature for error bars?

   Posted 9 months ago #
3. dan1980 - Member

   It does, but how does it know what the errors are, and how big they are?

   There's a custom setting, which is the one I'm trying to use, as the error in my data comes from m, and not from DI. The series I'm plotting is a formula (=(Cells containing values of DI)/(Cells containing values of m) so won't it just do it's own jiggery pokery on the answer value rather than the component with the error?

   Posted 9 months ago #
4. dan1980 - Member

   Hopeful morning bump...

   Posted 9 months ago #
5. matt_bl - Member

   dan1980 - Member
   I'm measuring some samples (F) by getting their mass (m), and measuring a particular property(DI), that I want to plot in terms of unit mass (DI/m vs F).

   So, I've done the measurements and I'm aware that the weighing out bit has an inherent error (I'm going to assume that the machine doing the measurement of DI doesn't have an error)

   If m actually = m +/- 2.5x10^-5 then do I work out what DI/+m, and DI/-m are, calculate the standard deviation and then standard error of the difference between to the two, and this becomes the standard error of DI/m?
   

   Dan,

   You are basically calculating the uncertainty of your measurement. As you only have one contribution (unlikely but possible I guess) all of the error stems from the tolerance on the weighing step.

   If you calculate the results for plus-tolerance and minus-tolerance you'll see that the absolute differences from the result without applying the tolerance are the same so no need to do it for both the plus and minus cases.

   Do it for one case (plus-tolerance), for each result you need to graph, subtract the result without the tolerance applied and whack them in a column alongside your data and use this column as your error bar values in excel.

   If you decide you have more than one source of error the calculation becomes more complicated but I can send you some more information. We use this method of assessing uncertainty for analytical chemistry, which includes weights, volumes, purities, measurement errors often all within the same method.

   Matt

   Posted 9 months ago #

RSS feed for this topic

Reply

You must log in to post.