"All the boring and soul-destroying work of counting mites on sticky boards, killing brood with liquid nitrogen, watching bees groom each other, and measuring brood hormone levels—all done in thousands of replications—will someday be seen as a colossal waste of time when we finally learn to let the Varroa mites do these things for us...
"I have never yet counted even a single sample of mites from any of my bees. I consider counting mites as a way of evaluating Varroa resistance to be fraught with all sorts of shortcomings and difficulties. It's very time consuming and hence the size of the apiary, the number of colonies tested, the gene pool, and the income available all start to shrink. It's also very easy for the results to be skewed by mites migrating from other colonies or bee yards. "—Kirk Webster
There are a few positive small cell studies, but also several that show higher mite counts on small cell and people always ask why or use this to discount small cell. I don’t know for sure why, as it is inconsistent with my experience, but let’s look at that. Let's assume a short term study (which all of them have been) during the drone rearing time of the year (which all of them have been) and make the assumption for the moment that Dee Lusby's "pseudodrone" theory is true, meaning that with large cell the Varroa often mistake large cell workers for drone cells and therefore infest them more. The Varroa in the large cell hives during that time would be less successful because they are in the wrong cells. The Varroa, during that time would be more successful on the small cell because they are in the drone cells. But later in the year this may shift dramatically when, first of all the small cell workers have not taken damage from the Varroa and second of all the drone rearing drops off and the mites have nowhere to go.
I think it's important in any test to let the mites get bad enough for the bees to respond to them. That's been my experience. And yet none of the experiments let that happen.
"...when 150 queens were introduced into nucs with brood untreated for 18 months. This brood had a normal outward appearance when the nucs were made up, but four weeks later about half of them were starting to decline with PMS-type symptoms. But after another three weeks, almost all of these colonies appeared normal and healthy again."—Kirk Webster
It's also important that they be in a real world environment, and not some artificially created one. The reality of how a mechanism in nature works is infinitely complex. Kirk Webster said this about breeding but it applies to any experiment:
"Bees that combine genuine hardiness, mite-resistance and productivity can only be maintained in the long run by having many hundreds of colonies constantly exposed to mites—and all the other known and unknown stresses in the real world, commercial beekeeping environment. This is the only way the bees can be tested for all the characteristics they need in order to thrive. And this testing and selection must continue year after year—to keep building up their resilience, and help the bees adapt to a changing world."—Kirk Webster
In the end, as Dann Purvis says, "it's not about mite counts. It's about survival". No one seems interested in measuring that. What I do know is that after a couple of years the mite counts dropped to almost nothing on small cell. But that did not take place in the first three months...