Lies, Damned lies, and statistics
Gelatin tests could be related to each other if they included a "standard round". Then all you have to do is determine round x penetrated 75% of teh standard in this test, round y penetrated 85% of standard round in this other test. Then you have a comparison of x and y wothout needing to know conc of gelatin, temp, phase of moon, etc.
Regarding stopping power, somewhere I read a paper on likelihood of of a one shot stop of a human based on law enforcement files. #1 was .357 followed closely (and not statistically significantly differently) bu the .45.