Thursday, July 19, 2007

Skeptics, assemble!

The Story So Far

One of my friends is a Christian. Not just a Sunday Christian. Not even a happy-clappy isn't-God-lovely Evangelical. He's from a group with some beliefs that I can only describe as Medieval.

What do I mean? Well, if I said angels and demons, that wouldn't be the half of it. Exorcisms, channelling the Holy Spirit, prophecy, faith healing, the works.

There's another, more positive, difference between my friend and other Christians: he's willing to experimentally test his beliefs. In particular, he firmly believes that his fellow believers can reliably sense him praying for them, and has actually gone so far as to suggest an experimental protocol to see whether he's correct.

The Protocol

A cleaned-up version of his protocol is as follows. Myself (Skeptic 1), my friend (Believer 1) and his assistant (B2) all go to my flat in town. S1 and B2 synchronise watches, and decide on a start time for the experiment. B2 sits in the apartment block's communal garden, which is separated from the block by 20m of garages.

S1 and B1 sit in the flat's bathroom. This is a windowless room with no exterior walls, so no light or vibration can enter without being very obvious. There is ventilation but it's noisy enough that any sound getting through would also have to be very obvious.

The experiment lasts 10 minutes. S1 will select a trigger time by picking a number from 0 to 19 from a hat, and adding that number of half-minutes to the starting time. At this time, S1 will signal B1, who will begin to pray for B2. S1 will make a note of the starting time. S1 will also keep eyes and ears open for any potential communication channels - no mobile phones!

If/when B2 feels the effects of the prayer, he will write down the exact time he first felt it. Once the 10 minutes is up, B2 will return to the apartment. B2's time of receipt will be presented, and then B1's time of prayer will be revealed.

If B2's time falls within the half-minute during which B1 was praying, we mark the test up as a positive result. The experiment will be repeated a prespecified number of times (probably three) and the results collated.

Help Wanted

If there is more than one positive result in three experiments then, by my calculations, that will be proof at the 99% confidence level that either a) B1 and B2 can communicate supernaturally, or b) my test protocol is open to manipulation. That's where you folks come in.

I can't see any weaknesses in my protocol, but I have no experience of running this sort of test. Even if I did have that experience, I'd still want independent confirmation that I haven't missed any obvious issues. I would thus greatly appreciate it if everyone who reads this post would leave a comment either pointing out a flaw in the protocol or declaring it apparently sound.

Because it would be really embarrassing if I had to convert to Christianity...
Read the full post