How do determine 'reprsentativeness' of a sample.

[arlesterc]

Let's say I have a manufacturing plant. I design and manufacture 1000 units of a product - no flaws in any of the 1000. I then use the same equipment, operators, materials from the same batch the test was done with and produce 100000 of the product. I find I have 50 that are flawed. What is the mathematical formula that will tell me how 'representative' my sample of 1000 products was? How probable is it that the flaw would not show up in 1000 units when 50 show up in 100000 units?

Thanks in advance for any help/time.

 

[JeffM]

If this is a real-life quality control procedure, I'd suspect that the process may have drifted out of control due to wear or some other temporal shift.

If this is a homework problem, it would help if you give the problem exactly and completely.

Assuming it is a homework problem, what is a plausible hypothesis about the probability that a unit is defective? What is a plausible hypothesis about whether defects are independent? Under those hypotheses, what is the relevant probability distribution?

Now can you calculate the probability that a random sample of 1000 will contain no defects?