Jason Fletcher of Yale, David Frisvold of Emory, and Nathan Tefft of Bates collected data on the effects of soft drink taxes in about 20 states that had implemented them at some point between 1988 - 2006.
While the taxes -- which averaged about 2% over the period -- did reduce soft drink consumption by the young, they were not helpful in reducing obesity rates. The reason? Kids still managed to get their extra calories by switching over to whole milk and fruit drinks:
Soda taxes seem to be an ineffective “obesity tax” due to a standard behavioral response to the policy, where children and adolescents consume more calories of relatively cheaper beverages
There are a couple of reasons, though, why this shouldn't mean an end for soda tax proposals. First, it's unclear what would happen if soda tax rates were substantially higher, as proposed in New York. People could react very differently to a 15% tax than a 2% tax. Second, a soda tax may need to cover a wider category of high-sugar/high-calorie products. This point is especially important if the tax is to be a revenue generator.
(For another soda tax critique, see the L.A. Times' Karen Kaplan.)