Ezra Klein, a fierce advocate for menu labeling, reads the sobering news:

The first big study out of New York City, however, suggests that menu labeling has been a bit of a bust in changing ordering habits at fast food restaurants in low-income neighborhoods...I could make a couple of bank shot arguments here (it's possible higher-income people will cut back more, leading chain restaurants to reformulate their offerings to secure their business, leading to calorie reductions across the customer base), but I don't think they make a lot of sense.

You could also see it going in the other direction: people could gravitate toward higher-calorie items on the grounds that they'll prove more satisfying, and represent a better deal. I'm still a supporter of calorie labeling on the simple grounds that people should have this information, no matter how they choose to use it. But so far, the evidence suggests that it's not going to make a dent in obesity rates.