Hey slym,
I appreciate the input. Of course it would've been better at the start of the project but oh well.
As far as needing at least 2000 results to be a true scientific poll, I have not seen that referenced anywhere before. Do you have a resource that you can reference? From what I have read, there is not a hard number needed to create a confidence interval; it just depends on the margin of error that you are comfortable with. Using the formula provided by the text, only 150 samples would be needed if you wanted a CI of 95% and an error of 8%, assuming the occurrence is 50/50. Only 144 would be needed if the ratio is 40/60. Even with a margin of error of 5%, only 369 to 384 samples would be needed.
I agree, the project is not without it limits. I would prefer to have more data, but only 1727 have been graded. Of course slabs can be cracked and reslabbed, but the chances of that happening to a black label should be equal to that of a white label, especially if Kirkman's assertions are correct. I eliminated any SS for this reason.