More Evidence That Our Approach to Reading Comprehension Is All Wrong

We've gone seriously off the rails on comprehension strategy instruction.

---

A recent meta-analysis confirms that it’s best to teach reading comprehension strategies in combination with building background knowledge. It also adds the surprising finding that it doesn’t seem to make much difference how much strategy instruction students get.

The meta-analysis, which looked at over 50 studies of comprehension strategy instruction, provides yet more evidence that the prevailing approach to teaching comprehension is flawed. (I’m indebted to Brent Conway for alerting me to the existence of this study in a recent piece he wrote on how his Massachusetts school district shifted its approach to instruction.)

But first, here’s a brief description of the approach to comprehension you’ll find in the vast majority of American classrooms: During the daily “reading block,” which often lasts two hours or more, students learn about a single comprehension skill—like “finding the main idea of a text.” To practice the skill, they read books or texts on random topics they may not know much about, which are deemed to be at their individual reading levels. This kind of instruction goes on for years, often through middle school, with students spending hours every week practicing the same round of skills. The theory is that once they master the skills, they’ll be able to apply them to understand any text.

All of this conflicts both with what cognitive science has found generally about how comprehension works and with the evidence on comprehension strategy instruction in particular. That evidence shows that such instruction can boost comprehension—if it’s done differently from the way most schools do it.

What the new meta-analysis adds to this body of evidence is the startling finding that it doesn’t seem to matter whether students get only an hour of strategy instruction or many weeks of it.

What Research on Strategy Instruction Tells Us

Here are the three main ways in which standard comprehension instruction conflicts with the evidence:

• Teaching a single strategy vs. multiple strategies: Lots of evidence indicates that strategy instruction works best when multiple strategies are taught together rather than one at a time, and the meta-analysis reinforces that conclusion. Researchers said the most effective combination of strategies was finding the main idea, plus learning about text structure, plus “retelling” the content of a text.

The finding that multiple strategy instruction works best has been around for a long time. It was endorsed by the National Reading Panel back in 2000. But standard elementary literacy curricula have continued to focus on one skill at a time, probably because it’s simpler.

In most classrooms, children are expected to go off and practice a skill like “comparing and contrasting” independently after the teacher has briefly demonstrated it. It’s hard enough to be sure kids are practicing one skill, let alone three at the same time.

You may have noticed I’m using the word “skill,” while the study reviewed the research on “strategies.” Teachers have come to use these words interchangeably, and the distinction isn’t always clear. But one point the study doesn’t make is that many of the “skills and strategies” routinely taught in schools, like “comparing and contrasting,” don’t actually have much evidence behind them. That’s yet another reason standard classroom practice conflicts with research.

• Background knowledge vs. no background knowledge: The study’s findings were clear that comprehension strategy instruction is far more effective when, as the authors put it, “background knowledge instruction is included.” The reason, the researchers posit, is that learning a new comprehension strategy takes up a lot of cognitive resources—in other words, it’s hard. If a student is trying to grapple with unfamiliar content at the same time, it’s even harder.

That makes total sense to me. In fact, I’ve made the same argument myself. What’s not clear from the recent analysis, though, is what “background knowledge instruction” looked like in the studies that were reviewed. Most likely, researchers or teachers provided students with limited information relating to the specific text they were about to read as part of the study.

Even that often doesn’t happen in classrooms. If the whole class is reading a particular book, and the teacher recognizes that many students won’t understand some key vocabulary, she may explain those words—assuming she can predict what they are. But generally, students then go off to read separate books on separate topics, to practice the skill. There’s no way a teacher can provide background knowledge for 25 different topics.

• Longer vs. shorter interventions: To me, the most startling finding of the meta-analysis was that it made no difference how long the intervention lasted. And the studies’ duration ranged from less than one hour to 55 hours.

About ten years ago, cognitive scientist Daniel Willingham suggested that there’s “plenty of data showing that extended practice of [reading comprehension strategy] instruction yields no benefit compared to briefer review.” He looked at eight reviews of studies and found that “ten sessions yield the same benefit as fifty sessions.” But even he didn’t suggest that one hour could yield the same benefit as 55.

I was so surprised by this finding that I emailed the lead author of the meta-analysis, Dr. Peng Peng, an assistant professor in the department of special education at the University of Texas at Austin. He confirmed that based on the studies his team analyzed, “dosage,” or the length of time students got the strategy instruction, had no significant impact on the size of the effects. But he cautioned that in general, “the more time on evidence-based instruction, the better reading outcomes to expect.” After all, he added, “practice makes perfect.”

Yes—sometimes. But in this instance, the more appropriate adage might be “a little goes a long way.” In Dr. Peng’s view, the reason the length of the intervention didn’t make any difference is that even 55 hours of strategy instruction isn’t enough to show that more is better. But if an intervention is working, wouldn’t you expect to see some incremental positive effect if you did it 55 times as much?

And the fact is that in the real world, students are getting far more than 55 hours of comprehension strategy instruction—or what passes for comprehension strategy instruction—and it’s making very little difference. Reading test scores, which purport to measure comprehension ability, have been stagnant or declining for decades, even as comprehension instruction has intensified in an effort to raise them. On the most recent round of national reading comprehension tests, 37% of fourth-graders and 30% of eighth-graders performed below the “Basic” level.

What Can Work to Boost Comprehension

None of this is to say that we shouldn’t teach comprehension strategies. As I’ve said before, what matters is what gets put in the foreground: strategies or content. What doesn’t work is teaching an isolated comprehension skill or strategy directly, year after year, using texts on random topics as vehicles for that instruction. What can work is focusing on a topic or text and asking questions that bring in whatever skills or strategies might help students think about it.

And Dr. Peng is right, in a way, that students need more strategy instruction, as long as content is in the foreground. When teachers of any subject ask appropriate “strategy” questions, it is likely to make a difference. That’s not because students are mastering some generally applicable skill. Rather, it’s because they’re acquiring the habit of asking questions like “Am I really understanding this?”

But wait, there’s more (reading comprehension is really complicated!). It’s also important to do more than provide a quick injection of background knowledge to help students understand a text they’re about to read. In order for students to gain the kind of knowledge that will eventually boost their general reading comprehension, they need to spend at least two or three weeks going into depth on a particular topic.

Getting definitions of a few key words before reading a text may enable students to understand that particular text, but without hearing and using those words repeatedly, the meanings are unlikely to stick in long-term memory. Eventually, through acquiring knowledge of lots of topics, presented in a logical order, kids will acquire the critical mass of general academic vocabulary that will enable them to understand texts on topics they’re not already familiar with.

I would argue that the most effective way to teach any of these strategies is through careful, explicit writing instruction. Want to teach “main idea”? Try explicitly teaching students how to come up with a good topic sentence for a paragraph, or how to write a summary of something they’ve read. Want to teach “text structure”? Teach them how to structure texts they create themselves. When this kind of instruction is embedded in content students are learning about, it’s likely to boost their learning.

Not to mention that the most effective way of familiarizing students with the complex sentence structure of written language—a barrier to comprehension that is overlooked by comprehension strategy instruction—may be to teach them how to use complex syntax in their own writing.

We do have some research showing that building students’ knowledge and having them write about what they’re learning can be powerful—but we need more. Instead, however, we have a continuing glut of studies on comprehension strategy instruction. And we don’t even seem to pay attention to what those studies tell us about how our approach to comprehension has gone seriously off the rails.

The result is that, with the best of intentions, we’ve made millions of students feel like they’re failures at reading and learning—when in fact, it’s our education system that has failed them.

This post originally appeared in a slightly different form on Forbes.com.