Why do American students lose ground over time?

With so much reporting and discussion of test scores these days, it’s easy to forget a simple fact. This overlook-habit is reinforced whenever the U.S. Department of Education releases its biannual report on math and reading scores for fourth- and eighth-grade National Assessment of Educational Progress (NAEP) exams that shows steady gains in the past few decades.

In her latest book, “The Reign of Error,” Diane Ravitch cites those scores and draws the broad conclusion: “Our students have higher test scores in reading and mathematics than they did in the early 1970s or the early 1990s” — evidence that schools are doing better than doomsayers would have us believe.

But here’s the thing. Academic achievement in elementary school and in middle school means absolutely nothing unless it is reflected at the end point, in 12th grade. If fourth graders in a particular year showed marked gains over their predecessors, but seven years later lost that advantage, the fourth grade scores proved no cause for optimism at all.

When we look at NAEP scores over time and include high-schoolers in the mix (using a related NAEP assessment, the “long-term trends”), we find precisely that troubling discrepancy. Broadly, students in early years have improved over the decades, but 17-year-olds, not at all.

What Ravitch says about scores in previous years holds for younger students, but older ones have scored slightly lower than the same age group in the 1990s. While 9-year-olds gained 10 points in reading and 14 points in math from 1992 to 2012, 17-year-olds lost three points in reading and one point in math.

As education writer Sol Stern says in a recent article, “Ravitch bases her case for improvement solely on higher scores for 9-year-olds and modest improvements among 13-year-olds. But these aren’t significant if the gains disappear in high school, and if students about to enter college or the workforce — the end product of the public school system — still can’t read or write very well, or at all.”

Exactly right, and the gap between early improvements and high school backsliding poses a secondary question. Or maybe it should be a primary question, since a student’s “career and college readiness” at the end of 12th grade often decides the rest of that student’s life.

Either way, we have to know: What happens in high school that makes the trend reverse, especially in reading? Is it because, as E.D. Hirsch has argued, reading tests in later years require more background knowledge, and the high school curriculum doesn’t provide it in a consistent way?

Or is it, as Ravitch suggests in “Reign of Error,” that a teen’s home environment determines reading scores as much as classrooms do, and teachers can counteract a poor one more easily in earlier years than in later ones?

Is it the avalanche of digital tools and social media that descends upon teens (who average 3,400 text messages per month) and diverts them from homework and reading books for pleasure?

As the AJC reported two years ago, from 2005 to 2011, Georgia fourth graders improved significantly in reading and modestly in math, but eighth graders slowed the progress in each category, improving modestly in reading and not at all in math. Superintendent John Barge rightly termed the results “encouraging” and was “especially pleased” that Georgia’s fourth grade readers exceeded the national average.

But given the steady pattern of slowdown as students age, elementary school boosts should be regarded warily. If this positive trend for our younger students is not paralleled by high school scores from now until 2020, we have a new challenge. What slows the progress of older students?

We have to find out, because the crucial yardstick of success or failure lies at the very end of students’ k-12 education.