North Carolina’s public schools have just released end-of-grade and end-of-course test scores for the 2013-14 school year. You may have heard something about them, particularly from districts eager to claim a large increase in the share of North Carolina students who test at or above grade level.

It’s a good example of why you shouldn’t believe everything you hear.

On average, 56.3 percent of North Carolina students “passed” their tests in 2013-14, up from 44.7 percent last year. That would be a huge gain if “passing the test” meant the same thing in both years. But it did not. Facing implementation problems with the Read to Achieve law, policymakers in 2013 revised the scoring system for North Carolina’s tests. Instead of four achievement levels, there are now five. The fifth was added in the middle of the distribution, to increase the share of students considered at “grade-level” in reading and math.

To be fair, the Department of Public Instruction had cautioned everyone not to compare the 2013-14 passing rates directly to the 2012-13 rates. That didn’t stop some local officials from doing so, however, if the comparison was flattering. According to a preliminary calculation by my colleague Terry Stoops, the true gain was about 1.5 points, not 11.6 points. Definitive data were unavailable as of this writing.

If you want to know how North Carolina’s public schools are really faring, the best measures are external. The National Assessment of Educational Progress (NAEP) administers rigorous tests to state samples of students in grades four and eight. During the first decade of NAEP testing in the 1990s, North Carolina made impressive gains. In math, the average score of our 8th-graders rose 30 points, more than double the national increase of 12 points. In reading, our 4th-graders posted a five-point gain from 1992 to 1998 while the national average was unchanged.

Around the turn of the century, however, our performance gains began to level off. From 2000 to 2013, the nation’s average 8th-grade math score rose another 12 points. North Carolina’s rose by nine. It wasn’t just perennial education pacesetters such as Massachusetts that posted the largest gains. States such as Mississippi (17 points), Tennessee (16), South Carolina (15), Texas (15), Louisiana (14), Georgia (14), and Virginia (13) also experienced larger math gains than North Carolina did.

In reading, our 8th-graders posted no gain at all from 2002 to 2013 while the average national score rose three points and all of our neighbors except Virginia gained as quickly or faster than the national average.

North Carolina’s lackluster performance since 2000 didn’t erase the effects of the earlier gains. In fact, our reading scores rank in the middle of the national pack, along with those of most of our neighbors, while in math we still outrank South Carolina, Georgia, and Tennessee. My point is that virtually all of North Carolina’s relative improvement in reading and math occurred during the 1990s — before there could have been any effects on NAEP scores from much-touted reforms such as Smart Start, More at Four, the teacher-pay hikes under former Gov. Jim Hunt, or (in fairness) charter schools.

If public policy had any role in the educational gains of the 1990s, you’d have to go back further in time to find the causes: to the enactment of the Basic Education Program in 1985 (which established minimum standards and funding levels across all school districts) and its 1989 revision in Senate Bill 2 (which removed some restrictions on BEP funding and gave districts more autonomy to make their own decisions).

Although there were disagreements about the details, both Democrats and Republicans generally supported the ideas behind the BEP and Senate Bill 2. There’s room for bipartisan agreement on the next phase of education reform in North Carolina, as well, which should include higher academic standards, rewards for high performance, accountability for low performance, and the provision of a diverse array of school choices for parents, students, and educators.

A precondition for consensus, however, is for North Carolina to adopt independent, rigorous annual measurements of student performance — and then stick with them.

John Hood is president of the John Locke Foundation.