The smaller the bullet, the more heavily it relies upon optimum performance on impact, i.e. reliable, rapid yaw (and fragmentation for maximum effect). The bigger the bullet, the more it is tolerant of less-than-optimum performance. The 7.62mm M80 doesn't have a particularly good yaw performance judging by gel tracks, but in the recent fighting, in the opinion of the overwhelming majority of those I have heard from who have used both 5.56mm and 7.62mm weapons, there is no contest - the 7.62mm is a much more reliable stopper, at any range. It is also far better at suppression, which is what most long-range fire is about.
I haven't ever seen any proof of this.
Does it shock me that M80 is a good stopper? No, it has twice as much energy as 5.56. There is no proof that performance is related to caliber or projectile weight specifically.
You won't get scientific proof, because there are too many variables, as we all know: most importantly, the exact path of and damage inflicted by the bullet, plus the physical and mental state of the target. Laboratory tests cannot replicate these factors. Gel tests are useful (because they are repeatable, allowing comparisons) but only give partial answers.
For proof, you need the following:
1. Electronic rifle sights which record the exact sight picture seen by the gunner as he fires, and can record the reactions of the target: does he drop immediately, or stop fighting, or carry on?
2. Recovery of the enemy bodies for autopsy, ensuring that the bodies can be matched up with the targets seen through the sights, so that the pathologists can determine which types of wound produce what kinds of reactions (this may be tactically impossible, of course, and it it won't include the ones who were hit but got away).
3. Repeat this several thousand times while using different calibre weapons so that the differences between them can be analysed.
You may then have enough data to produce definitive answers to the effectiveness of different calibres, with the reservation concerning those who were wounded and escaped (ideally, you would of course make sure they were arrested immediately afterwards so their wounds could be examined).
Realistically, you are not going to get this for a long time, and may never get it.
If the restrictions affecting expanding bullets are lifted then the experience of hunters shooting human-sized animals will become relevant, but that doesn't help while we have non-expanding bullets.
In the meantime, we need some basis for evaluating the effectiveness of different rounds so we can make informed choices for the next generation of weapons. The best solution at the present - and the only way of assessing real-world performance rather than laboratory results - is to interview troops immediately after each battle to record their experiences and observations. This is of course affected by subjectivity, but it's the best we can do. And unless we field some possible GPCs and test those in combat, even that won't help with determining how the GPC will compare with the others.
So we can do no better than put together real-world combat experience with gel tests of combat and experimental rounds for comparison purposes (plus shooting at animal carcasses). It isn't perfect, but it's the best we can do.