If I recall the Marshall data correctly, which is no doubt a big portion of that which Jimmy is referring to, there was no allowance made for anything other than a one shot stop. That left out a lot of shootings right there. It also didn't consider the size of the person shot, clothing, etc. Lots of variables were uncontrolled, but that's the nature of a compilation as opposed to a test. There were also those who questioned Marshall's truthfulness. I don't remember what Ayoob's parameters were other than his compilation sample wasn't as large as Marshall's.
Just for the record...the last time I saw Marshall's data, the 40 was in the lead with something like 96% one shot stops, with what was then, the best ammo for that caliber and the 357 was only a percentage point or so in back. 9mm +P+ was at 89%, take your pick of JHP from the big three ammo makers. The best 45 ACP was someplace in between. I don't think enough shootings had been compiled to place the 44 Mag. or the 10.
