2 Attachment(s)
An interesting parallel...
Quote:
Originally Posted by
Tokarev
For a useful scientific study, we cannot simply look at gelatin or just corpses. Doing a prospective study where a couple hundred humans get shot with different cartridges won't pass the institutional review board or get funded by NIH. But there are good retrospective studies that do this. But to be a good study, it needs to avoid selection criteria that influence the results. I was impressed with the recent Braga study in JAMA:
https://jamanetwork.com/journals/jam...rticle/2688536
That's an interesting research paper.
Performing a quick/informal ANOVA* (analysis of variance)―
Attachment 42117
―of the actual lethality reflected by Braga's data (Boston 2010 - 2014), [I used only the caliber subsets that had an n > 5 which reduces the usable statistical population from n = 367 to n = 361**] and comparing it to the US ARMY BRL P[I/H] model (Dziemian, 1963) there is surprising agreement between the US ARMY BRL P[I/H] model and the actual data as presented in Table 2 of the paper here―
Attachment 42118
This analysis considers lethality only, but not time to incapacitation. They are two very different functions of terminal performance.
*For the purpose of this ANOVA, it was assumed that since criminals tend to use the cheapest available ammunition, non-deforming FMJs and LRNs were modeled for all calibers except the .357 Magnum which was modeled as a 125-grain JHP @ 1,350 fps that expanded to 1.66 times its pre-impact diameter.
**.44 Remington Magnum, 10mm and 7.62x39 sub-sets were excluded due to insufficient population of n < 5