Print

---

Hello

Consider the following -
I have hundreds of telephone-operators that take calls from customers. These calls must be classified as complaints or not. Each operator takes approx 1000 calls per month

If I want to test the performance of these operators, I can test a sample of calls from each operator, by an 'expert'.

How big should my sample size be, if the true complaint-rate itself can vary, between 5 to 40 %.. I would want to be confident within 5%.

If I then want A 2nd level check, that my testing is good, I get external persons to perform an independent check.
They sample the aggregate sample-tested by the experts... But how big should there sample be to give confidence that the first levels of testing are effective.. Within 5%.

....

I am getting myself.confused between sampling the entire popping the 1st check, and then.in the 2nd.check, sampling, with the operators being.random-effects in the 2nd.check by the external persons...

Any advice would.be welcomed

Sent from my Sony Xperia™ smartphone

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.