Ethics testing

The situation is pretty simple:

It involves a runaway trolley, five helpless people on the track, and a large-framed man looking on from a footbridge. You are standing on the footbridge, too. You know that if you push the large man off the bridge onto the tracks, his body will stop the trolley before it kills the five people on the tracks. Of course, he will die in the process. So the question is: Is it morally permissible to kill the man in order to save five others?

According to the article it's a theoretical question posed in conjunction with the following situation in an attempt to better understand how morals and ethics play into our decision making processes:

Again, there are five helpless people on the track. But this time, you can pull a switch that will send the runaway trolley onto a side track, where only one person is standing.

What would you do? Again, the article tells me that most people would not push the fat guy to his death, but most people would pull the lever to kill one guy who happens to be on the track. Interesting, but inconclusive at best. The study is looking at emotional response to see how it affects reasoning skills, but ultimately it's still only theoretical--the people being tested still know they aren't actually in the situation. I think the test is biased and therefore unreliable because the subjects have too much time to think it out, and lack the sense of urgency a person who really faces the situation would have to deal with.

The solution to this testing flaw is simple--put the test subjects into a real situation. Get a trolley, a fat guy, and five, I don't know, drunk homeless people, and put them on the track, tied up or boozed out of their skulls to keep them there, whatever. Obviously you can't tell the subjects about the true nature of the test because the key is the surprise factor, so make up a story about, say, a fat man that's contemplating suicide and the subject's purpose is to talk him out of it. That'll add a nice little twist when the subject quickly realizes the only way he can save those five drunk homeless people is to push el fatso over the edge.

And if we consider the meta-ethics of such a test, realize that the test itself forces us to question its own ethos--that is, would it be ethical to subject someone to such a test? If not, why not? Naturally you'd want to do a background check on your test subject(s) before taking them out to the overpass (make sure you have plenty of homeless drunkards, too, if your hypothesis is that most subjects will not push the fatty catty over), just to make sure they have no heart condition, aren't pregnant, etc. And make sure they sign an NDA and whatever that paper is that absolves you of any responsibility in the case of accidental (or otherwise) death--a waiver, that's it.

The next and final step would be to repeat the process except with the rail-switching choice instead of the pushing fat dude over the edge choice. In this one it's probably be best to load up on lonely single banjo playing hobos, since your hypothesis would likely be that most people will pull the switch.

Record your results, draw some conclusions, and let me know how it goes! I'll be waiting at the platform.