TexasCHLforum.com

I like this, it's much better than the one I saw where the guy was one of only two people in the room and it was set up so the attacker could just immediately "kill".

I think, though, that ultimately this isn't really a good test either, because the 'good guys' know something is going to happen. Kind of like the above, but in the other direction.

To set up a really good study, you'd have to have a control group, and a test group, set up double-blind so that neither the people being tested nor the people running the test have knowledge of who is in what group. It would also have to be done long term, such that everyone carrying in both the test group and the control group get a chance to "settle in" so that it's more like real life, where you're not going to have foreknowledge that something is going to happen. That would take a really long time and be hard to set up.

Still, much better than the other one I saw.