When trying to decide what is significant and what isn't, I usually take anything based on plain statistics and throw it out the window. ...(some content deleted for brevity - GB)
When looking to the real world, especially in the effectiveness of ammunition, I tend to think the military - especially in places where there have been many thousands of shootings, like Vietnam, Afghanistan, etc. - provides the best data available.
The Moro uprisings in the Phillipines caused the military to switch from a .38 caliber ammunition to the .45, because the .38s were frequently not effective at stopping Moro warriors, while the .45 was often effective. More recently, the military has given excellent reason to believe the 9mm is often ineffective, and a number of military units that have a choice in what caliber they use have switched back to the .45.
Any ammunition can be effective under the right circumstances, especially with proper placement. Overall, though, especially with troops that do not get the best training with handguns, nor the most effective ammunition (FMJ vs hollow points), real world results measured over many thousands of uses indicate that caliber is indeed significant.
Statistical analyses do not always prove out in the real world. They can sometimes be useful for generalizing or in determining trends, but it is too easy to miss significant data, over weigh some data, or be used to justify someone's personal agenda.
In regard to statistical analyses in general and Reg's distrust, being skeptical is good and healthy. There's a great quote I got in a statistics class: "Statistics are like a bikini. What they reveal is suggestive, but what they conceal is vital!". In this case, the relevant question is what these statistics are concealing. While this is counter-intuitive, what if the .22 really was massively superior to the .45, but the statistics showed them roughly equal? Wouldn't you want to know? You can shoot a .22 all day for $20 (I have) while $20 buys one box of .45 FMJ; it would be nice to know if .22 was better.
Let me put forward an idea. I'm not going to argue it's absolute truth, but I think it's logical. Suppose there's a fundamental difference between the people the military faces and who you're likely to face in a self defense situation? To use the statistics term, they're not the same population - and that's what his statistics conceal.
The opponents the military faces know they're in a war, and are driven by some mix of duty, honor, loyalty to squad mates and (often) religious fervor. That means they're much more likely to keep fighting after they've been shot than a criminal is. After all, he's probably a criminal because he doesn't have a really good work ethic. From the self defense standpoint, the Moros are not a good model; they were a dedicated army, in a war, defending their homes and families, "reinforced" with various drugs, not a lowlife trying for an easy meal ticket.
That's what Ellifritz is saying: that the attacker doesn't get incapacitated so much as they quit the attack once they realize they've been shot. I've heard (can't confirm, obviously) that the Marines are now teaching new guys that they may well get shot in a battle, but just keep fighting. What Ellifritz' data is saying is that in the average self-defense situation, you're not fighting a trained Marine; once he realizes he's been shot, he's going to stop.
Put another way, the most widely accepted statistics (there's that word, again) on self defense say that there's around 2.5 million uses of a firearm in self defense every year, and that thousands of times every day, simply drawing the firearm ends the confrontation without a shot being fired. Does it really matter if you pull a .32 or a compact .45 if they're going to back down as soon as they see a gun?
Now, everyone has seen the stories of a drug-wracked stoner who doesn't stop until he's been hit 40 times. The Moro tribesmen were drugged up on something that kept them going, too. It's not like that's an unlikely scenario, either, with the prevalence of crystal meth in our society. Faced with that situation, I'd really prefer 12 ga to the brain stem, but I'd settle for a major caliber handgun.
And the other viewpoint is also obvious: there's often no disadvantage to carrying your .45 instead of something smaller - my .45 when fully loaded weighs an ounce or two less than my 9mm XD subcompact (which has 3 more rounds). So why not carry the big one?