This Washington Post story quotes a lobbyist with the state Restaurant Association complaining that the study is skewed, because it includes restaurants that have long had voluntary smoking bans in effect. He suggests that it would be more accurate to run a direct comparison of individual venues to see how they have fared after being forced to show smokers the door.
There may be some validity to that. But this study, along with similar ones conducted elsewhere, suggest what I have long thought. Just because smoking is prohibited doesn't mean that smokers are going to stay at home every night. If you have a jurisdictional ban, they'll come out to eat and drink even if they have to step outside to light up.