The more time I spend trying to wrap my head around the world of early-reading instruction, the more I find myself becoming wary of the notion that teachers should devote a lot of their time to learning about the science of reading. I realize that might seem like a bizarre and contradictory statement given that so many of the problems in reading instruction stem from ed schools’ failure to provide research-backed training to pre-service teachers, so let me explain.

I had already started writing this piece when I discovered that Mark Seidenberg (author of Language at the Speed of Sight), along with several colleagues, recently co-wrote a paper about this exact topic. Unsurprisingly, they do an excellent job of identifying and clarifying some of the major issues at play. As the authors point out, science can act as a general guide to effective practices in reading instruction, but it cannot actually dictate curricular specifics. Or, as they put it, “reading science is an active, ongoing endeavor, not a canon of findings. Overreliance on simplified accounts of science risks reifying it into precepts that do not incorporate much of what the science has to offer.”

I’m going to back away from that idea for a moment, though, and begin with the idea that much of American education is based on top-down notions of conceptual understanding and application—that is, the idea that once students have grasped a general concept, they will then be able to then be able to apply it in a wide variety of situations more or less automatically. I think that this type of assumption now underpins a lot of discussions about teacher-training and the science of reading.

Unfortunately, this is often not how things work: A student may demonstrate excellent theoretical understanding of a concept but then be unable to apply it in an unfamiliar situation, or even recognize when it is relevant. Presenting teachers with a mass of scientific findings about how reading works, and then expecting them to easily translate those findings into concrete methods for getting a group of squirming five- and six-year-olds to make sense out of lines on a page, without a coherent curriculum (and often in situations where they are required to use a very incoherent curriculum), is both profoundly unfair and wildly unrealistic.

Learning to read isn’t natural, and learning to teach reading isn’t particularly natural either. There are many things that skilled adult readers take for granted that beginning readers do not, and learning how to explain and sequence concepts in a way that doesn’t inadvertently cause confusion is an ongoing process.

This is not to suggest that teachers shouldn’t learn about the science of reading. In fact, I would argue that it’s tantamount to educational malpractice that ed schools don’t require a class in it. (In a rational system, teachers would not be responsible for educating themselves, and the fact that they have little choice but to do so speaks to a deep level of systemic dysfunction.) Reading science is also a fascinating subject, and a skilled teacher who really takes the time to understand certain research may in fact be able to devise more effective methods based on it.

But what teachers really need to learn is how to most effectively teach the greatest number of children to read, as quickly and painlessly as possible. And that is not the same as knowing the science.  

As Seidenberg et al. explain, the world of reading instruction is facing what is essentially a translation problem:

Reading science does not come with educational prescriptions attached. Science is one kind of thing (empirical findings, explanatory theories). Educational practice is another (activities that promote learning in realworld settings). Connecting the two is the function of translational research…Our concern is that although reading science is highly relevant to learning in the classroom setting, it does not yet speak to what to teach, when, how and for whom at a level that is useful for a teacher.

I’ve now spent a fair amount of time observing how the science of reading gets discussed on social media, and one of the most recurring questions is, “Is method x aligned with the science of reading,” or, “Does SoR approve of this teaching technique?” These types of questions, while obviously asked with the best of intentions, nevertheless reveal some fundamental misunderstandings about how academic science works, and about how it differs from the kinds of models that dominate in the K-12 education world.

The fact that so many teachers do not become aware of these differences during their training highlights the divide (chasm, really) between schools of education and the mainstream academic world; however, I suspect that some of the blame for this also rests with Common Core.

Since the Standards were implemented in 2010, they have profoundly reshaped the ways in which curriculums are conceived of and described, and what purposes they serve. Teachers are expected to constantly demonstrate how classroom activities are “aligned with” particular standards, and because that approach is so entrenched, it now appears to be the lens through which the science of reading is understood.

The problem is that unlike Common Core—a limited set of vague prescriptions written quickly by a small number of individuals with limited knowledge of K-12 schooling and legally prohibited from being altered—the science of reading is a massive, dense, and unwieldy body of research, much of it still inconclusive, carried out by hundreds of scholars at dozens of institutions over many decades, and primarily published as articles in scientific journals, for the consumption of other academics.

Occasionally there might be a book that breaks out into more mainstream popularity, such as Seidenberg’s Language at the Speed of Sight or Maryanne Wolfe’s Proust and the Squid, but they are the exception rather than the rule.

So to be clear: there is no governing body or committee tasked with putting an official “Science of Reading” stamp on phonics/literacy programs to indicate that they are “in alignment” with the research.

Using a program that’s been slapped with a stamp saying “Science of Reading approved” is not even what this is fundamentally about; it’s a different model, a different mindset entirely.

In an earlier blog post dealing with Lucy Calkins’s sudden (and highly suspicious) foray into teaching phonics, Mark Seidenberg voiced his concern that the “science of reading” label would not only get watered down to the point of meaninglessness, but actually be used to support the kinds of damaging practices that currently prevail in so many classrooms:

[Lucy] Calkins derides the 3-cueing approach but the concept is deeply embedded in her erroneous view of how reading works and children learn, as her document confirms.

It’s obvious that other curriculum publishers and authors will follow Dr. Calkins’ lead in claiming that their products are compatible with the science of reading. For a while now there has been pressure to show that the materials are “evidence-based,” but the term has little meaning because educators have their own concept of what counts as evidence. Soon they will all claim to be consistent with the “science of reading” and it will be equally meaningless unless people look beyond the label at the assumptions about reading and learning the materials incorporate and the methods they employ.

And in the more recent post:

In the absence of sufficient translational research, almost every reading curriculum can claim an equally loose connection to the “science of reading.” The risk of course is that such programs will prove ineffective, not because the basic science was wrong but because the translation was poor.

Given how the educational publishing industry works, this is a very valid concern. Most teachers do not have the time or the inclination to wade through a thicket of scientific terminology in search of a paragraph or even just a few sentences that just might have some application to their work in the classroom. And districts and administrators will inevitably come to rely on publishers’ labels as a convenient shorthand, regardless of their accuracy.

The problem is that to be truly effective, a phonics program must be part of a true structured literacy program and not be combined with practices that undermine it, for the sake of familiarity or convenience or adminstrative preference. A curriculum cannot on one hand contain some explicit, systematic phonics instruction but then also include sight-word lists with words that can be easily sounded out, and/or leveled readers with spelling patterns students have not yet learned (thus encouraging guessing). All the pieces must be based on the same conception of what reading consists of and how it develops, and be used to complement one another.

So how can such an immense body of research be translated into tools for practical classroom use, in a way that isn’t completely overwhelming? The cognitive scientist Daniel Willingham offers a potential way out of this dilemma:

Theory is unavoidable, but…theories make untested (and likely, inaccurate) predictions, which invite ineffective classroom applications. The problem might be solved by teaching empirical generalizations in the context of a model that affords few novel predictions. In other words, it should be the sort of model [that] would be of little interest to scientists. It should account for empirical generalizations that scientists consider central to the field, the findings that would prompt a researcher to say, “well yes, of course.” And the model should predict nothing else.

To some extent, this is already the case; Seidenberg points out that what teachers generally refer to as “the science of reading” in fact only involves a small number of the most established, uncontroversial models. His concern is that because of the nature of the education system and edu-publishing industry, even that kind of narrow focus will be insufficient to prevent misinterpretation of key findings.

That brings me to back to the idea that the label and the knowledge of scientific terminology are not really the point. Provided that an instructor understands—not just gives lip service to, but really fully embraces—a handful of fundamental principles that are consistent with key research findings, it is entirely possible to teach reading effectively without ever learning formal terms like phonological awareness or orthographic mapping.

While arming oneself with the science is obviously necessary when trying persuade skeptical administrators to adopt a structured literacy program, allow me to point out that somehow, children learned to read phonetically for several centuries before the nineteenth-century education reformer Horace Mann deemed letters “bloodless, ghastly apparitions” and ushered in the era of whole-word reading, and before researchers performed the first brain-imaging studies. And over the last few decades, countless parents without the slightest background in cognitive neuroscience have managed to teach their children to read using books like Siegfried Engelmann’s Teach Your Child to Read in 100 Easy Lessons (4.5 Amazon stars with 3841 reviews). Then there’s the Fluency Factory, where tutors with little scientific background regularly do remarkable things with beginning and struggling readers just by using an expertly designed, carefully sequenced program with great fidelity.

For the record, yes, there will always be children who experience truly severe difficulties in learning to read, and for them a higher level of expertise is required. But only a small percentage of children actually fall into that category. This is about the 80% or so of children who need some amount of explicit, systematic instruction, not the outliers.

So all that said, what are the most important concepts that reading instructors need to know? I would propose the following:

1) Letters and letter-combinations stand for sounds and are blended together to form words—essentially, words say what they say because they consist of specific letters written in specific patterns, in specific orders. Sound-letter correspondences must be explicitly taught, beginning with the simplest patterns and systematically moving to progressively more complex ones.

2) Children must be able to hear sounds correctly, and to distinguish between similar sounds (e.g., short “i” and short “e”), in order to consistently match them to letters.

3) Skilled reading derives from the ability to process words—and only words—on a page, in left-to right sequence. Memorizing, using picture cues, and making guesses based on “what would make sense” are not reading.

4) To be able to read with conversational intonation, students should be able to read at around 200 words per minute by late middle school or early high school. By third grade, they should be reading around 150 wpm. If children read too slowly, they will not be able to retain meaning or connect text to spoken language.

5) Learning to decoding phonetically is an immensely demanding cognitive task for many children, so asking a beginning reader to focus excessively on meaning will result in mental overload. As decoding ability increases, more attention can be paid to it.

All of these principles are based on reading science, but they are also practical concepts that form a bridge between the science and classroom methods.

In the hands of a skilled teacher, a program based on these principles is virtually guaranteed to be more effective for a greater number of children than the combination of incidental phonics instruction and leveled readers found in the average balanced literacy classroom. But to reiterate, the science—at least as it currently stands—cannot dictate the specifics of the implementation.

Now, in addition to the above points, there are certain pieces of knowledge and pedagogical skills that teachers must also possess. Some are related to the science, others less so, but they are all of immense practical importance.

On the knowledge side:

Individual sounds

Reading instructors must be able to correctly pronounce all 44 main English sounds, and to say consonant sounds without an unnecessary schwa (c-a-t rather than cuh-a-tuh). They must also be able to demonstrate subtle differences between similar sounds; recognize when students are having difficulty discriminating between them; and know how to explain/demonstrate the distinctions.

Pronunciation Patterns

English is filled with letters that make different sounds depending on the context in which they are used. For example, it is insufficient to teach only that “c” makes a /k/ sound; rather, children must learn that “c” also makes an /s/ sound when it is followed by “e,” “i,” or “y.”

Standard and non-standard spelling patterns

Many sounds, particularly long vowels, can be spelled many different ways, and teachers must ensure that all of the spellings are introduced. They should also be able to point out which patterns are common and which ones are less so.

Ability to analyze spelling to detect difficulties relating spellings and sounds, and to distinguish between sound-based misspelling and phonetic misspellings. 

A child who spells cousin as casin may have trouble distinguishing between short “a” and short “u.” On the other hand, a child who writes akshun instead of action perceives the sounds correctly and is merely attempting to trancribe the word literally.

On the pedagogical side, teachers must be able to:

Distinguish between actual reading and reading-like behaviors 

A child who can recite a text may be able to decode it or may have simply memorized it. Some children can get very far on memorization, and teachers need to be both willing and able to probe past the surface.

Avoid over-explaining concepts, or discussing concepts that have not been explicitly taught.

Explanations for beginning readers should be kept as short and simple as possible, and no information that has not been directly taught should be assumed to be known. Otherwise, children will become confused.

Pace lessons at an appropriate speed.

If a lesson is too slow, students will get bored and stop paying attention; too fast and they will become frustrated and be unable to absorb the material.

Limit the number of new concepts introduced at a given time, and provide sufficient review. 

Children’s working memories can absorb only a very limited amount of new information at a given time. Instruction that focuses on one or two concepts is therefore more effective than instruction that address many concepts at the same time. In addition, material must typically be reviewed multiple times and applied regularly in order to stick.

Observe students closely and objectively, and resist putting words in their mouths. 

Teachers sometimes gloss over student errors with “helpful” statements like, “You really meant to say pound instead of pond, right?” While well-intentioned, these exchanges cause gaps in understanding to be missed, with the result that gaps in children’s knowledge are not addressed.

Stop at every mistake a child makes in one-on-one reading, and teach them to correct it. 

If a text is at the appropriate level, there shouldn’t be more than 3-4 errors in a given sampling. Ignoring errors gives children the message that reading accurately isn’t that important, a belief that can create serious comprehension problems down the line. Difficulties with sound-letter correspondences must be addressed as soon as possible if children are to achieve fluency.

These kinds of classroom behaviors are not really within the province of science, but they are key elements of the translation Seidenberg refers to. Even the most impressively structured literacy program will have its effectiveness reduced if a teacher cannot do them consistently.

The upshot is that reading must be taught in a way that is based on, or motivated by, scientific finding, but it in the classroom, the actual science can only go so far. That’s why I get concerned when the knee-jerk response to any question/comment/suggestion about classroom practices is, “What does SoR say?” or “Well, SoR doesn’t say anything about that.” The fact that the efficacy of a particular classroom practice isn’t explicitly spelled out in the research doesn’t necessarily mean that it isn’t a good idea (or a bad one).

Beyond a certain baseline level of consistency with the science, the question should really be, “Does this activity/method help children become better readers?” Ultimately, that’s what matters—not how many books one has read, or how much jargon one can spout, or how many webinars one has sat through. Reading instruction has been gotten wrong for so long, and in so many ways; the stakes are too high for that to keep on happening.