Hi! I'm trying to introduce calculus of uncertainity in my calibration procedures, but I'm not sure of what value of F (based on the Student's T distribution a number of degrees of freedom) must use:

S1 = (SDEV / (N ^ 0.5)) * F

I have a table (Guide to Expression of Uncertainty in Measurements), but reading some paper (Implementing ISO 17025 Measurement Uncertainity Requeriments in Software, Matt Nicholas, Fluke Corporation), they use half the values shown in the 95.45% column of the table

What must I use? Entire value or half value of F to 95.45%????

Thanks

------------------
Sergioasdf

I not sure I understand exactly what you are doing or what you want to do.

I'm guessing you have calculated a mean, and want to provide a confidence interval - a measure of uncertainty.

If that's the case, then the formula's are pretty simple:

95% confidence interval:

mean +/- t[0.025,n-1]*stdev/SQRT(n-1)

where t[0.025,n-1] is a t-value taken from a t-table. Sematech/NIST has a table at
http://www.itl.nist.gov/div898/handbook/eda/section3/eda3672.htm

Find the column with 0.025 as the header and find n-1 in the first column.

If you're trying to do something else, please let us know.