When 2013 SAT scores came out last month and showed no significant change from 2012,many educators may have felt not disappointed or neutral, but relieved. That’s because the overall trend since 2006, when the writing component was added, has been downward. Critical reading has dropped seven points, math four points, and writing nine points. In fact, writing scores have dropped every year except 2008 and this year, when they were flat.
Apart from the dismaying trend, something about the decline doesn’t make sense. When we look at those SAT test takers when they were younger, they moved in the opposite direction. There is no elementary school SAT, but we do have the National Assessment of Educational Progress long-term trends assessment for 9-year-olds in reading, and the news is positive.
- From 1999 to 2004, scores rose from 212 to 219.
- From 2004 to 2008, they climbed from 216 to 220 (the difference in 04 scores from the first bullet is due to a revised assessment format)
Why would 9-year-olds improve so much in reading–NAEP researchers estimate that 10 points roughly approximates a grade-level–then decline on the SAT when they got older? To be sure, we have two different tests, and if we go apples to apples and consult the NAEP reading exam for 17-year-olds, we don’t see a decline in the selected years. From 2008 to 2012, the reading score went from 286 to 287–not down like the SAT trend, but nonetheless begging the question: why did 9-year-olds rise seven points, but eight years later only rise one point?
It’s an important issue because reading scores for 4th- and 8th-graders don’t matter one bit if they are not reflected in 12th-grade scores. Only the end of high school counts.
One reason is a change in leisure reading, which has declined significantly among teenagers. But this week in the Washington Post, Valerie Strauss opens her education column to E. D. Hirsch, who offers another explanation, one closer to schooling and standardized tests. It’s a capsule statement that any superintendent and chief state school officer should heed if he wishes to raise scores and be a success.
The first fact is this: “Reading ability is very topic dependent.” Given two passages of equal difficulty in syntax and vocabulary, the same reader will comprehend one better than the other if the reader knows something about the subject matter of the one and little about that of the other. The idea of a general reading ability that functions independently of what is read is “a misleading abstraction,” Hirsch says. If a reading test has ten passages with 8-10 questions on each, the same student will perform variably from one to the next depending on background knowledge. It’s an arbitrary system. If, by chance, you did volunteer clean-up work one summer and one of the passages concerns how cities and towns dispose of their trash, you will fly through it. A passage on a sport you never played, though, will slow you down, even though passage difficulty is the same.
Here is the explanation for divergent trends among 4th– and 12th-graders. Passages for older students are more knowledge-intensive than those for younger students.
Test-prep courses can’t help with this factor. They coach students in how to “scope out the meaning”–identify main idea, note the evidence, apply to a hypothetical . . .–but, Hirsch states, “though it gives a boost to scores in early grades, ceases to have an effect . . . after just six or seven skill exercises.” To spend any more labor on test-prep than those basic tactics is a waste of time.
The solution is simple, yet far-reaching. We need the reading curriculum to be more knowledge-aimed and less skill-based. Hirsch: “A cumulative knowledge-oriented curriculum will, over time, produce higher verbal abilities than a test-prep curriculum.” That means more set content and common readings in English, history, and civics, a sharper determination of “cultural literacy” (to use the title that made Hirsch famous), a narrower and more coherent curriculum.
And one other thing. Instead of making reading tests a crap shoot, choose topics that correspond to the core knowledge of the curriculum. Make reading tests like math and science tests, which state up front the content on the test. If reading tests were to identify subject matter in advance, Hirsch concludes, “then test prep would become knowledge prep.” Without that curricular/testing reform in English et al, we shall see more flat lines in annual scores, and more erroneous and fruitless efforts to lift them.