The Elsmar Cove Wiki More Free Files The Elsmar Cove Forums Discussion Thread Index Post Attachments Listing Failure Modes Services and Solutions to Problems Elsmar cove Forums Main Page Elsmar Cove Home Page

View Full Version : Measurement Uncertainty (MU) - Confidence Interval & Degrees of Freedom?


amanbhai
7th February 2008, 02:09 AM
I am calculating measurement uncertainty for a chemical laboratory. Now. I have to incorpoarate the equipment uncertainty. My question is how do I know which distribution it is following? & how do I know the confidence intervel?

Benjamin28
7th February 2008, 11:01 AM
how do I know the confidence intervel?

The confidence interval can be stated as +/- standard deviation if the data is within a normal distribution. This will give you a 68% confidence interval, meaning that 68% of the values reported will likely be within range of the true measurement. So, for your question, in order to know the confidence interval you will want to know the std deviation of the sample measurements taken. If your standard deviation is estimated from the average of multiple measurements, i.e. the average of (n) measurements taken 100 times, then you would take the [+/- standard deviation/sqr root (n)].

Benjamin28
7th February 2008, 11:15 AM
I forgot to post on the first part of your question pertaining to the distribution of the data. I would suggest you look through pages 11-15 of the linked publication as it offers some very straightforward examples on the matter.

http://www.ukas.com/Library/downloads/publications/M3003.pdf

To give yourself a good visual representation of the data you can often put it into a histogram as well and get a visual of how the data is distributed.