We’re told that random sobriety checkpoints used to identify and catch “drunk drivers” make the roads safer — but there’s little, if any, hard data to support this claim.

What we do have is an attempt to correlate the number of people arrested for driving with at least some alcohol in their bloodstream (no matter how little) with a reduction in alcohol-related motor vehicle fatalities.

That’s quite a different thing.

In fact, the practice of herding drivers like cattle through these “checkpoints” hasn’t put much of a dent in the total number of drunk driving deaths that occur annually in the U.S.

Depending on whose numbers you believe, roughly half of the 48,000 or so motor vehicle fatalities reported to the National Highway Traffic Safety Administration (NHTSA) each year are listed as “alcohol-related” — that is, attributed in some way to the consumption of alcohol and the involvement of a motor vehicle.

But these figures are themselves deceptive because, for one thing, “alcohol-related” means fatalities that don’t necessarily involve a drunk driver are lumped in with those that do. For example, the death of a drunk pedestrian who wanders into a busy street and gets run over is listed as “alcohol-related” fatality — even though the driver of the car was completely sober. Similarly, if a car runs off the road and it is later determined that a passenger had some alcohol in his system, the death of that passenger is likewise reported as “alcohol-related” — even though the passenger’s consumption of alcohol had absolutely nothing to do with the accident itself. In this way, the actual number of “drunk driving” deaths can be distorted — and is reported — as being a much higher percentage of the total than is in fact the case.

The more relevant fact as regards the usefulness of sobriety checkpoints, however, is that while there is some slight year-to-year fluctuation in motor vehicle fatalities attributed to drunk driving, there has been no major downward trend that coincides with the increased use of roadside sobriety checkpoints — which have become commonplace around the country, especially during the holiday season.

But if the checkpoints are effective at catching dangerous drunks, then there should be an obvious statistical downtick in drunk driving deaths that coincides with the expanded use of these checkpoints.

Problem is, there isn’t.

This suggests that while sobriety checkpoints have been very effective at criminalizing social drinkers — that is, otherwise law-abiding and responsible people with slight trace amounts of alcohol in their system who would otherwise have gone unnoticed and probably made it home without incident — they aren’t doing so well at nabbing the truly dangerous heavy drinkers who are responsible for the majority of the drunk driving deaths and accidents.

It’s a fact, for example, that the majority of drunk driving deaths involve a person with a Blood Alcohol Content (BAC) of .10 percent or higher, a point reached after a bout of pretty heavy drinking — not the glass or two of wine over dinner that puts a person in peril of a DUI citation as a result of running afoul of ever-lower maximum allowable BAC levels.

Most states now have BAC thresholds for “drunk driving” set at the .08 BAC level, significantly below the .10 BAC level (and higher) that used to obtain — and at which point it’s been shown a person is most likely to actually be involved in (or the cause of) a motor vehicle accident.

Having had a drink or two is not the same thing as being “drunk” — but advocates of ever-lower BAC thresholds and the aggressive use of sobriety checkpoints do not seem to appreciate the distinction. Mothers Against Drunk Driving (MADD), to cite the most notorious example, continues to push for BAC thresholds to be lowered to .06, even .04 — a level so low that a person could be legally considered “drunk” after having consumed as little as a single glass of beer or wine.

But “habitual offenders” with BAC levels of .10 and higher are not only responsible for most of the drunk driving problem, they tend to go out and drive drunk again and again and again. They are not deterred by sobriety checkpoints — and are often cagey enough to avoid them entirely, because (for example) many hard-core alcoholics drive drunk in the daytime — and for the most part sobriety checkpoints are set up in the evening hours.

The best way to catch these habitual and hard-core drunk drivers, according to experienced law-enforcement officers, is not by the use of dragnet-style “checkpoints” — but the old-fashioned way: by patrolling the streets, looking for drivers displaying evidence of serious impairment such as weaving, wandering across the center line, or driving too slowly.

Instead, police resources have been concentrated on static checkpoints — leaving the roads open to the bad guys while over-punishing people who aren’t really the problem.

Like the airport practice of screening middle aged hausfraus at the gate, this may be politically correct — but it’s demonstrably ineffective at identifying and dealing with the real culprits.