Posted By:
John SFrank,
Comparing the measurements of the image area is an excellent idea. I had not considered the stretching issue.
A random sample within the population of T206 cards could be accomplished. It would be a matter of having a pool of cards from numerous collectors and then assigning them a number. The numbers of the cards could then be picked in any unbiased fashion. The measurements on all of the cards should be performed by one individual.
I had mentioned that a sample size of 20 is adequate, but the larger the sample sizes, the more validity a t-test will have. It is not required in a t-test to compare equal sample sizes as it is probably a safe assumption that it will be more difficult to locate PSA 8's than VG cards.
This is an interesting idea but it does not affect me directly. I do not collect high grade material.