RALEIGH — Since my colleagues and I at the John Locke Foundation often get accused of being “anti-government” when we critique public policies, I like to look for opportunities to identify government programs that show signs of achieving their objectives.

Two came to my attention today. The first is high-stakes testing in public school. North Carolina has been a national leader in the effort to design and implement accountability measures that test students in core subjects and then publicize the results, with rewards and punishments attached. I’ve been critical of some of the specific aspects of North Carolina’s version of the idea, the ABCs of Public Education, but not the overall concept. In the absence of a “bottom-up” form of educational accountability through parental choice, we are better off having a “top-down” accountability system with testing than no system at all (which has been the status quo ante in public education until recently, not counting school board elections — which, frankly, don’t count).

A number of studies have concluded that a serious testing program can have salutary effects on student performance. Of course, details matter. The tests have to be sufficiently rigorous, the scores have to be widely published, and there have to be real consequences for failure. North Carolina is batting only .333 in this game so far. Still, it is hard not to believe that the advent of our state’s end-of-grade and end-of-course testing program in the early 1990s hasn’t played a role in subsequent improvement in national test scores, particularly in math.

Further evidence that such an approach is useful can be found in the recent release of national test scores for science. Unlike reading and math, science is not likely to be part of a state’s annual testing program, especially in the middle-school grades. It isn’t in North Carolina. And what the recent national data suggest is very little improvement, possibly even backsliding, in science performance over the past few years. This is bad news for the state of science education, but promising news that state testing in the other subjects is making a difference, since national reading and math trends look a lot better — particularly in states like Texas and North Carolina that offer fairly comprehensive testing programs.

More research is needed, naturally. But still, the trend deserves a positive mention.

The other emerging example of a government reform with early results is the “covenant marriage.” In August, Arkansas became the third state (after Louisiana and Arizona) to create this new kind of marriage, which requires intensive counseling before the wedding and before any subsequent divorce, plus a multi-year waiting period before a divorce can be granted. Advocated by many social conservatives properly concerned about marital stability, the covenant marriage is actually an excellent pro-freedom solution to a social ill. No one is coerced into the binding contract. It relies on moral suasion and a sort of arms race between potential spouses — would you like to be the man or women who tells your loved one that you aren’t willing to commit to a covenant marriage?

So far, covenant marriages appear to be accomplishing their end. Those in regular marriages are four times more likely to get divorced than those in covenant marriages. There may be a causality problem here — the more committed folks are more likely to sign up — but at least there’s no casualty problem.