Debunking the Leveled Reading Myth

A new book details the lack of evidence for limiting students to books they can read easily.

---

I’ve learned a lot from reading researcher Timothy Shanahan over the years—and now I can say I’ve learned even more.

One of the most important things I (and, I’m sure, others) have learned from him is that there’s no real evidence supporting the pervasive practice of leveled reading. That approach limits students to reading at their presumed “individual” reading level, which could be years below their grade level. In the past, Shanahan has addressed that issue in articles and blog posts.

Now he has taken on the topic in a fully fleshed-out book entitled Leveled Reading, Leveled Lives. (The phrase used in the title, as Shanahan acknowledges, was coined by his former student, Alfred Tatum.) The book traces the genesis of the assumption that kids should be matched to texts they can read fairly easily. That assumption has led to a system that, albeit unintentionally, often keeps struggling readers at a permanent disadvantage.

Shanahan describes how, in the early 20^th century, a “theory of readiness” emerged from academia, holding that children could learn only if they were adequately prepared to do so—and that it could be harmful to teach them things they weren’t developmentally “ready” for. New psychological theories reinforced that view by focusing on the dangers of frustration. Measures were being developed for assessing both the capabilities of different readers and the difficulty levels of various texts.

Against that background enters Emmett Betts, the author of what Shanahan calls the “most influential reading textbook of its era,” Foundations of Reading Instruction, published in 1946.¹ The book set out three individual levels of reading that all students were presumed to possess: independent, instructional, and frustration. In a Goldilocks-like framework, independent was thought to be too easy, frustration too hard, and instructional level just right.

A “Scientific” Formula

Betts put forward a formula that sounded precise and scientific: If a student could read the words of a text with 95 to 98 percent accuracy and could comprehend 75 to 89 percent of it, that text was at the student’s instructional level. The instructional level framework became the basis of virtually all reading instruction in the US and some other countries.²

How did Betts come up with this theory? Shanahan points to a dissertation by one of Betts’ graduate students, Patsy Aloysius Kilgallon. Shanahan had assumed that Betts or Kilgallon ran an experiment: have some students read at their “instructional” level and other students at, say, grade level, and see which group has higher reading achievement.

Instead, he found, Kilgallon simply watched for signs of tension while kids were reading. Were they squinting, for instance, or grimacing? Kilgallon found these behaviors started to appear when their word-reading accuracy fell to a certain percentage, and he basically used that percentage for the accuracy part of the instructional level formula. For the comprehension percentage, he apparently used the level of comprehension achieved by students who were reading without exhibiting signs of tension.

In addition to finding that research method lacking, Shanahan also objects to the fact that Betts and Kilgallon thought it was sufficient to see if students could read and understand whatever text they were reading. He argues that the researchers should have determined if having students read at their supposed instructional level led to increased “learning.” What Shanahan means by this (I asked him) was that rather than looking just at better “immediate comprehension,” the researchers should have looked at improvements in general “reading ability,” presumably as measured by standardized reading tests.

I have my own doubts about whether such reading tests accurately measure learning,³ but in any event it’s clear from Shanahan’s account that the deeply entrenched system of leveled reading rests on extremely thin evidence. And, as Shanahan details, multiple experimental studies have now shown pretty clearly that leveled reading either provides no benefits or diminishes learning as compared to teaching children with grade-level texts.

And yet it’s proven difficult if not impossible to oust leveled reading from classrooms. The Common Core literacy standards, to which Shanahan contributed, tried to set a requirement that all students be taught using grade-level texts. Many publishers of reading textbooks complied with that directive, but surveys showed that the number of teachers who relied on leveled texts only increased.

A fundamental problem, Shanahan notes, is that largely because of the assumption that leveled reading is the best approach, prospective teachers don’t get training in how to teach kids to read with more challenging texts. That’s a lot harder than giving students books they can understand on their own and expecting that to somehow result in increased reading ability.

Recommendations for Teaching

Shanahan doesn’t just debunk instructional level theory, he also aims to provide teachers with some practical guidelines for enabling students to read more complex text. His recommendations, like having students reread difficult text and curtailing the practice of small-group instruction, are solid as far as they go. I only wish they had gone farther.

A basic limitation of his guidance, in my view, is that Shanahan—like many reading experts—sees reading comprehension as something that can be taught and assessed largely independent of a reader’s prior knowledge. As other researchers have argued convincingly, that’s not actually possible. While Shanahan does make some references to the importance of knowledge and vocabulary, he essentially assumes classroom texts will be taught in isolation rather than as part of a curriculum that builds knowledge in a logical, coherent sequence.

Certainly there are factors other than background knowledge that play a role in reading comprehension—perhaps most important, familiarity with complex sentence structure. Still, we have evidence that knowledge relating to the topic of a text is extremely helpful, especially for readers who generally struggle.

But Shanahan devotes only two paragraphs of the book to “knowledge building,” and most of that text cautions about its supposed dangers—primarily, teachers over-explaining a text to students before they try to read it. That’s a risk, but it’s most likely to happen when texts are read in isolation and therefore present students with a lot of unfamiliar words that teachers feel they need to explain in advance.

Kids do need to learn how to extract new knowledge from texts on their own, as Shanahan argues, but they’re more likely to do that when the text builds on a body of knowledge they’ve already acquired. If kids have, for example, spent a week or two learning about sea mammals, they’re in a better position to acquire new information from a book about sea mammals than if they’ve never heard of the creatures before. Eventually, through learning about lots of topics, they’ll acquire the critical mass of general knowledge and vocabulary that enables them to understand texts on topics they don’t already know about.

I recall hearing Shanahan say he’s not opposed to building knowledge, he just believes it should take place primarily in content-area subjects like social studies and science—not in a reading class. But in many elementary schools, so much time is allocated to reading that very little is left for those other subjects, and Shanahan has argued against reducing it. In a recent blog post, he recommended that where reading scores are low, including middle schools, kids should get “a full two hours a day of literacy teaching.” That doesn’t leave much time to build knowledge in the content areas.

Knowledge-Building Curricula

That, essentially, is why we have seen the rise of knowledge-building literacy curricula that bring science and social studies topics into the “reading block,” a topic Shanahan doesn’t address in the book. Given schools’ reluctance to cut back on time devoted to reading, such literacy curricula are often the only way to ensure students have an opportunity to acquire content knowledge.

Shanahan’s position has been that there isn’t enough evidence to establish that those curricula improve general reading comprehension.⁴ While I generally respect his cautious approach to evidence, it’s possible to take caution too far.

The argument that there isn’t enough evidence for knowledge-building curricula is one I’ve responded to in detail elsewhere (see here and here). One problem is that it can take three years or more for the benefits of knowledge-building to show up on standardized reading tests. That makes it extremely difficult and expensive to get experimental evidence—although we certainly do have some.⁵

And Shanahan seems to be limiting the term “evidence” to rigorous experimental studies, excluding lots of other evidence from cognitive science supporting the idea that building knowledge boosts reading comprehension. That’s unrealistically narrow. “Every study has problems,” cognitive scientist Daniel Willingham once told me. “That’s why you need to look at issues using a number of different techniques—logical, experimental, non-experimental.”

I have a few other quibbles. While Shanahan rightly points to the need to familiarize students with complex syntax to boost their reading comprehension, he doesn’t address how writing instruction that begins at the sentence level can be a powerful way of doing that (and, incidentally, of strengthening and deepening knowledge).

And judging from what I’ve seen and heard, he significantly underestimates the emphasis schools place on isolated comprehension skill and strategy instruction. He merely suggests in passing that “in some classrooms teaching students to use strategies has become too much the point,” when the point should be improving comprehension. But, alas, most instructional materials and teacher prep programs focus on skills and strategies as ends in themselves, misleading the majority of educators about what’s truly important.

Still, anyone who cares about effective reading instruction—and the fates of the many children who are being failed by our current system—owes Shanahan a debt of gratitude for this book. It provides an excellent starting point for dismantling a regime of reading comprehension instruction that has only widened existing societal inequalities. To get the full picture of how to help students understand challenging text, however, educators will need to look elsewhere.

1. "Reading textbook” is used here in the sense of a textbook for prospective reading teachers, not a textbook for teaching kids to read.

2. The requisite accuracy percentage was later lowered (by the reading gurus Irene Fountas and Gay Su Pinnell) to 90 percent.

3. Shanahan notes that measures of reading levels and text levels are unreliable for various reasons. One likely reason, which he doesn’t emphasize, is that such measures don’t take into account an individual reader’s background knowledge. That’s also true of measures like “grade level” and of standardized reading tests, but Shanahan seems to accept both of those kinds of measures as reliable.

4. He does have some good words to say about “text sets.” Knowledge-building curricula are essentially a series of text sets, along with lesson plans and other elements, curated for teachers and organized into a logical sequence.

5. Shanahan does note that there’s evidence to support at least one knowledge-building literacy curriculum, Bookworms. But rather than describing it as a knowledge-building curriculum, he characterizes it as a “program [that] emphasizes teaching students with grade level texts.” And, due to a typo, the text refers to the curriculum as “the Bookmark program” instead of “Bookworms.”