Statistic : amount of test

[freddyolen]

Does anyone has a good book,article on how many tests need to be done to make statistical good predictions.

We want to do a test with 2 components brougth together and see what the influence is of the "fit passing" on the applied torque and eventual breaking apart of the pieces.

We don't want to do a DOE (design of experiments) because that would lead us too far

 

[BML]

Most statistics books have sections on determining sample size based on set alpha and beta levels. Like most statistics, most of these assume that the underlying data population is Normal-distributed with finite mean and variance. Since you are comparing two values, you may also be able to determine the significance of tests like the T and paired-T with given sample sizes.

Hope this helps!