The Great Prayer Experiment

...Such experiments, if done properly, have to be double blind, and this standard was strictly observed. The patients were assigned, strictly at random, to an experimental group (received prayers) or a control group (received no prayers). Neither the patients, nor their doctors or caregivers, nor the experimenters were allowed to know which patients were being prayed for and which patients were controls. Those who did the experimental praying had to know the names of the individuals for whom they were praying - otherwise, in what sense would they be praying for them rather than for somebody else? But care was taken to tell them only the first name and initial letter of the surname. Apparently that would be enough to enable God to pinpoint the right hospital bed...

What's that you say, Lord? You can't cure me because I'm a member of the control group? ... Oh, I see, my aunt's prayers aren't enough. But Lord, Mr. Evans in the next-door bed ... What was that Lord? ... Mr. Evans received a thousand prayers per day? But Lord, Mr. Evans doesn't know a thousand people ... Oh, they just referred to him as John E. But Lord, how did you know they didn't mean John Ellsworthy? ... Oh right, you used your omniscience to work out which John E. they meant. But Lord...

-Richard Dawkins