That's a valid opinion and fact regarding the effectiveness. I haven't really researched that aspect. But like most government programs and endeavors, it wouldn't surprise me if it were fact.
But the Supreme Court isn't going to (or at least they shouldn't) look into the effectiveness of it, but merely the constitutionality. The effectiveness comes into the realm of the legislative branch (policy making) and the executive branch (implementation).
DUI checkpoints are absolutely terrible at catching dangerously drunk drivers. If the government really wanted to set up a checkpoint to catch dangerously-drunk drivers it could simply use some safety pylons construct a relatively-easy obstacle course and have cars drive through it. This would impede traffic flow somewhat, but far less than stopping motorists. It would require far less manpower than stopping every motorist, but would be more effective at catching dangerously-drunk drivers than stopping every fifth motorist.
But the Supreme Court isn't going to (or at least they shouldn't) look into the effectiveness of it, but merely the constitutionality. The effectiveness comes into the realm of the legislative branch (policy making) and the executive branch (implementation).
The complaint isn't merely that DUI checkpoints are ineffective, but rather that--given the manner in which they're conducted--many so-called "DUI checkpoints" are really nothing of the sort. There is no conceivable way in which having drug dogs sniffing cars at a "DUI checkpoint" will increase the effectiveness of the checkpoint in catching drunk drivers. The purpose of the checkpoints is to conduct unconstitutional fishing expeditions. Sticking a "DUI checkpoint" label on an unconstitutional drug checkpoint doesn't make it constitutional.