Sample Size

Hi all, I have question on sample size. If I have 3 parts measured 5 times each by 2 operator, what is my sample size then. is it 3 or 30? and for dof what will be the figure 2 or 29? Thanks in advance


Well-Known Member
Hi Mun,

What do you try to test?

from Wikipedia: the number of degrees of freedom is the number of values in the final calculation of a statistic that are free to vary

So in the common average test, you calculate the average from n observations.
If you know (n-1) observations and the average than then you can calculate the value of the n observation.
so df=n-1 as the n observation is not free to vary.


Active Member
This is more complicated than it seems.

It sounds as if your experiment has 3 parts each measured by the same 2 operators or 3x2 = 6 mini experiments of 5 measurements each. This would usually be analysed using a 2 way anova to see if there are differences between the operators and between parts (and possibly an interaction to see if different operators disagree on different parts). Each comparison needs two df's. There are 2x3x5 = 30 data points giving a total of 30-1 = 29 df. These 29 df are divided among the tests. Comparing the operators there are 2-1 = 1 df. Between the parts there are 3-1 = 2 df. An interaction would have 2x1 = 2 df. So the "error" df is 29-1-2-2 =24 df. (or 26 if you ignore the interaction.)

A test involves 2 df's, the df of the things you are comparing and the df of the error.