Dr. Dave's data is interesting, and I've noticed a smaller size for the low balls at my local pool hall in the decade+ old set, which is obvious simply by racking. Regarding micrometer vs caliper, I happen to have calipers which yield consistent enough results for what I wanted, but accurate results would be better, especially with multiple independent measurements.
What I was really after is not perfection, but basic data useful to a consumer. Larger sample sizes are great, but any sample is better than none. For those wanting to purchase a set of balls, today we can choose between marketing, folklore, and buying all ball types and measuring ourselves. Is there actually a difference in the consistency of material and tolerances on high-end balls?
Looking at how balls wear in real conditions is interesting, especially if similar vintage balls are compared between halls and the potential buyer is a room owner. For a home table the abuse pattern is going to be wildly different than a hall.
I don't mean to say that weight an size and roundness are the only, or even most significant factors. Resiliency also seems like a simple enough test to do with a tape measure, known hard surface, tripod, and the high speed video available on all our phones. I would guess that weight tolerance, resiliency, and how long a given material maintains its surface between polishing are the most obvious factors for a player.