What's Behind the Plateau in Test Scores?
The results of the third year of testing aligned to the Common Core State Standards are out, and many educators are scratching their heads. Unlike in the past two years, when scores rose, this year's results are about the same as last year. In California, 49 percent of students met standards in English language arts the same proportion as in 2016. And in the 12 other states that use the same assessment as California, the Smarter Balanced Assessment Consortium, the same thing happened: all saw ELA scores stagnant. In Vermont, scores dropped a bit.
Did the test get harder this year? Doubtful; test-makers take great pains to make sure tests are comparable from year to year so that educators and policymakers can follow trends.
So what's the answer? Without seeing the test, it's difficult to say for certain, but this is not a new phenomenon. In fact, it's fairly typical: when states institute a new test, scores go up for the first few years, and then level off. In a 2011 report for the Thomas B. Fordham Institute, Mark Schneider called the pattern The Accountability Plateau.
One reason scores on state tests go up initially, as Dan Koretz points out in his new book, The Testing Charade, is because teachers become familiar with the format and content of the new test and adjust their instruction accordingly. Some of the increase reflects test prep, and some might reflect more nefarious practices, like cheating. Those practices only go so far, though, so scores stagnate.
However, some of the increases might also reflect real improvements in learning, too. Consider the National Assessment of Educational Progress (NAEP), a low-stakes test that is almost never the subject of test prep or cheating. Results on NAEP show some gains, which probably reflect real improvements in learning. But those gains have leveled off, just like the California ELA scores.
For example, in mathematics for nine-year-olds, scores rose sharply between 1999 and 2004, and slightly between 2004 and 2008. But the scores have been flat since then. The same is true for reading for nine-year-olds.
A closer look at the results suggest a possible explanation for the pattern. From 1999 to 2004, the largest increase in reading scores was for students at the 10th percentile--from a score of 158 (on a 500-point scale) to 169. Those at the 25th percentile experienced a nine-point gain, from 185 to 194. But the gain for the top performers, those at the 90th percentile, was much more modest: five points, from 259 to 264.
That pattern suggests that schools achieved their improvements largely by raising the floor. That is, they were able to help students who had struggled with even the most basic skills.
That is no small accomplishment. Those are genuine improvements, and they have been sustained. But as the subsequent plateau shows, raising the floor can only get schools so far. To improve performance overall, schools need to enable more students to demonstrate deeper levels of learning--to be able to apply their knowledge to think critically and solve complex problems. That takes a different kind of instruction--one that provides students with opportunities to reflect on their learning, to take part in extended projects, and to produce real products for real audiences. The plateaus suggest that this is not happening on a large scale.
Other countries are wrestling with this problem too. Germany, for example, experienced a substantial increase in scores on the Programme for International Student Assessment (PISA) after adopting policies--such as common standards for student performance--in the wake of disappointing PISA results in 2000 (the so-called "PISA shock). But as Andreas Schleicher, the German native who directs PISA for the Organization for Economic Cooperation and Development, said in a recent interview, Germany's scores have leveled off because the impact of those policies was limited. What's needed, he said, is a greater investment in improving teaching.
As in that country, educators in the United States should view the recent Smarter Balanced scores as a sobering reminder that raising the floor is not enough. Attaining high levels of learning for all students is not a matter of doing more of the same. It will take a different kind of teaching.