Hmm....

There's a Globe article about a study about to be done to determine whether "one drink a day" helps prevent heart disease. It includes this bit:

Half of the study participants, who are 55 and older, will be randomly chosen to drink a concoction of grain alcohol mixed in Crystal Light - lemonade or raspberry-lemonade, their choice - while the others receive an unadulterated version. None of the participants will know who's getting a jug filled with the spiked drink, because grain alcohol is virtually tasteless.

I don't think this is a reporter error, since it does sound like the intent is to be double-blind. But... nobody's going to know whether what they're drinking has alcohol in it or not? Really?

We're doing a test on gun safety. Half the guns will be loaded with blanks, and half will use real bullets. But the participants won't be able to tell which people are being shot for real, because the guns will be loaded by third parties and shuffled together, before being handed to the shooters...