I apologize for the sloppiness of this page; I am putting it up as of October 3, 1997, and will make it more readable/extensive in the future.

What it is: This page is a nontechnical criticism of Marshall and Sanow's assumptions in Handgun Stopping Power. In it I show that the instrument they design does not measure solely the effectiveness of the cartridge in question, but in fact measures a sociological tendency as well. To the extent that this tendency correlates with caliber, the conclusions that Marshall and Sanow draw concerning caliber efficacy are not supportable.

What it is not: This is not a primer on Marshall & Sanow's work. Get the book, which is well worth reading in spite of its problems, or check out the Stopping Power Web Page.

A basic criteria for evaluating a study is whether, in fact, it measures what it claims to measure. In the case of Marshall and Sanow's "one shot stop" figures, what they claim to measure is the relative incapacitation capability of cartridges.

But what the study actually measures is the tendency of shooters of any particular caliber to cease fire before incapacitating. A figure of 50% for .32ACP doesn't necessarily mean that it is a 50% stopper, or even that it's only 52% as effective a stopper as a Remington .357 JHP. It's not impossible that these correlations exist, but work would need to be done to show that they did, and to show that they are significant.

What the number does mean, though, unequivocally, is that 50% of the .32ACP shooters in this study stopped firing before the subject was incapacitated.

If the shooter had gone on to fire more rounds, that shooting would be eliminated from the study (since M&S only considered cases in which one shot was fired).

What does this mean in terms of the results that Marshall and Sanow cite? If within any caliber, shooters fired until the subject was completely incapacitated, the one-shot stop statistic for that caliber would be 100%, no matter how many shots on the average it took. If shooters firing .32ACP were trained to fire until their subjects were incapacitated, the one-shot-stop index for .32ACP would be 100%. It would have nothing to do with the round.

It is easy to see, then, that to whatever extent shooters of any particular caliber exhibit or fail to exhibit this tendency, the results in Marshall and Sanow's table will be greatly affected. The question is: "Do shooters of different calibers significantly differ in this tendency?"

In the dearth of any real study, any answer is only speculation, but M&S's percentage figures may correspond with the training level and situational tendencies of people who choose those rounds for personal defense. Small surprise .22 shooters are low on the totem pole. Ever wonder why the .44 Magnum came out as a poor stopper?

Even if the correlation of training levels to caliber selection isn't the whole story, it is obviously a factor, and should be accounted for in the analysis. There are statistical methods for doing this, but they are beyond the resources of the study. But certainly their analysis is not measuring the tendency that they profess it does.