Does that experiment weight the groups accurately to reflect per-capita religious demographics in the area? Right now it just looks like a sample of the well-to-do adherents of a dominant religion.
Oh good, so it wasn't my bias that made me think there was something wrong with it.
For one thing, that's just a picture. I have no guarantee that this isn't some well-to-do guy who got a whole bunch of money, change, shabby-ish clothes, and other stuff together, labeled a whole bunch of bowls, distributed the change with bias, and asked a friend to take a picture.
For another, this doesn't account for food that's been donated, because people donate food to the homeless. (Also why is a watch in the Buddhist pile? Does this guy expect a homeless man to be able to walk into a pawn shop with it and NOT instantly be accused of theft?)
And finally, this is less a criticism, but what the heck are the two signs between atheist and buddhist? And what is a "Spirital" anyways?