Max, upon rereading my post, I realized it came across harsher than I intended. I was just terse and in a hurry.

I'm beginning to think that perhaps my experience is not typical. I interview people as a part of my job, too. In my field some people get very good coaching, and so the answers we get and the body language we see are sometimes just a product of having a lot of resources. I don't mean to imply that you can "beat the game," but it is possible to make your self-presentation very, very polished. When the majority of people have the same degree of polish, the game changes.

The interview process can be quite grueling and long -- months, with days where the interviewee is basically constantly observed from the moment he wakes up to the moment he goes to bed. People do falter under such stress, and perhaps you can construe those moments as revealing of their true self. You also check with people who know them. But, after more than a decade, I'm beginning to feel the information we get is not all that useful. People and careers are quite complex, and you can't always predict problems over the long term. Even this rigorous process cannot eliminate people who would do bad stuff later or even right away.

But I wouldn't be surprised that there is a science to all of this that I do not understand as a mere employee. Recently I heard a CIA psychologist talk about people who leak secrets. (They study this sort of stuff, eh?) She says that people tend to leak or commit such acts of betrayal when they're in crisis (in other areas of life). I have noticed a similar pattern -- crisis in personal life can lead to misconduct at work. So maybe there is something to it, though a part of me thinks, hey, that sounds like common sense.

I don't know how any of this would apply to a survival situation. I mean, if you're in a survival situation, especially in a long-term one as envisioned by the video in the OP, you're in crisis. Everybody is in crisis. So, what, trust no one? At the end, I continue to feel that character can be known only in the long term. You can design whatever process you want for admitting people into your "survival group," but I am guessing that 50% (or more!) of that process will reflect you and the social dynamics in the existing group. I'd prefer a survival group with people you have known for a long time, and prepare for their flaws.

Thanks for the recommendations, Montanero. I have read Rory Miller's work. I think his stuff is really good for short term interactions. I'll check out the other stuff.