Sample number planning

Sampling strats and stats
22

small typo : The preliminary data is complaint to the EN689 preliminary test.

small typo : Fewer than 6 samples would be required to be 80% confidence of statisical compliance.

small typo : However, it is recommended at least 6 samples are collected before performing the statistial anaylsis.

typo : You need to collect a total of 19 samples to be 81 % confident of being statisically compliant based on the preliminary results collected.

suggestion : You need to collect a total of 19 samples to be 81 % confident of being statisically compliant based on the preliminary results collected. Would suggest “demonstrating compliance according to EN689 criteria : 70%UCL on exceeedance <5%”

idea : another dimension of the question : for the same preliminary data, let the user select a sample size and provide the probability of compliance/non compliance.

cheers

1 Like
23

Thank you for reading! And good suggestion of the result explanation.

There is a paper in the works to explain the tool in greater detail, but did you find the explanation reasonable?

I’m hesitant to allow people to choose their own sample size as I fear people will arbitrarily choose which ever size they like and justify whatever probability is reported. Avoiding arbitrary sample sizes was the reason for the creation of the tool. Unless you have a different view?

I do want to implement a button that allows users to enter noise data (convert it to a linear scale for them).