You might assume I know something about teaching kids to read. I studied English at UCLA and obtained my master’s in education at The City College of NY. I taught special education grades 5-8 for 7 years, and I’ve supported schools and teachers throughout the Bronx with K-8 ELA instruction over the past 3 years.

Yet you’d be wrong. I’ve come to realize I know next to nothing.

In case you haven’t been aware, there’s been a firestorm of educators on platforms like Twitter gaining newfound awareness of the science of reading, with an urgent bellows inflamed by the ace reporting of Emily Hanford. For a great background on this movement, with links, refer to this post by Karen Vaites. And make sure you check out Hanford’s most recent podcast (as of today!!! It’s amazing!) outlining how current classroom practice is misaligned to research.

Impelled by this burgeoning national and international conversation, I’ve sought to educate myself about the science of reading. I began with Mark Seidenberg’s Language at the Speed of Sight, took a linguistics course, and have just completed David Kilpatrick’s Essentials of Assessing, Preventing, and Overcoming Reading Difficulties. Seidenberg is not only pithy, but furthermore impassioned, while Kilpatrick is deeply versed in both the research and application in practice as a former school psychologist. Both experts provide an incendiary takedown of more than a few sacred cows in the educational establishment.

It’s been fascinating to learn more about the science of reading while simultaneously working with a school where I could see problems elucidated by reading researchers and advocates play out in real-time. It has made what I’m learning gain an even greater sense of urgency. I would read pages critiquing the “three-cueing system” and balanced literacy approaches on the bus in the morning, then walk into classrooms where I saw teachers instructing students, when uncertain about a word, to use guessing strategies such as “look at the picture” and the “first letter of the word,” rather than stress the need to be able to decode the entire word (for more on the problems with current classroom practice, listen to Hanford’s podcast).

There’s so much to digest and apply from all of this. This post is my attempt to begin synthesizing the information I’ve read. I’ll start general and then focus on the word-level reading aspect of the research in this post. And there’s so much more I want to cover, but I’ll be leaving tons of stuff out that I would love to explore further. Someday . . .

Reading Can Be Simple

First off, though reading is complicated, it can be outlined by a simple model, known aptly enough as The Simple View of Reading. It can even be put into the form of an equation. The theory was first developed in 1986 by researchers Gough and Tunmer. The original formulation was D (decoding) X LC (linguistic or language comprehension) = Reading Comprehension.

After years of further research, this distinction has mostly held up, though it has become greatly expanded, especially in our understanding of what constitutes language comprehension.

Decoding has been clarified as one umbrella aspect of word-level reading, which is composed of many sub-skills. A more updated formula, courtesy of Kilpatrick, is:

Word Recognition X Language Comprehension = Reading Comprehension.

If you struggle with word recognition (such as with dyslexia), or if you struggle with language comprehension (English language learner), then you have difficulty reading.

Within each of these two domains lie the various sub-skills and knowledge that make reading so very complicated. Here’s a chart I made to visualize the “Expanded” Simple View:

Note that word recognition skills are mostly mastery-based. And a key point experts like Seidenberg and Kilpatrick make about word recognition is that word recognition can be acquired by all children. IQ discrepancy is not a factor.

Here’s Seidenberg:

“For children who are poor readers, IQ is not a strong predictor of intervention responses or longer-term outcomes. Moreover, the behavioral characteristics of poor readers are very similar across a wide IQ range. . . Within this broad range of IQs, poor readers struggle in the same ways, need help in the same areas, and respond similarly to interventions. In short, the skills that pose difficulties for children are not closely related to the skills that IQ tests measure. The primary question is about children’s reading—whether it is below age-expected levels—not their intelligence.”

Here’s Kilpatrick:

“Discrepancies between IQ and achievement do not cause word-reading problems. Rather, deficits in the skills that underlie word-level reading cause those problems. The component skills of word reading can be strong or weak, independently of IQ test performance.”

“A common belief that continues to be recommended is that some students with severe reading disabilities simply cannot learn phonics and they should be shifted to a whole-word type of approach. This recommendation is inconsistent with the accumulated research on the nature of reading development and reading disabilities

“The simple view of reading applies to poor readers with IDEA disabilities (SLD, SLI, ID, ED/BD, TBI) and poor readers not considered disabled. Thus, when asked the question, ‘Why is this child struggling in reading?’ we would no longer answer, ‘because the child has an intellectual disability (or SLI or ED/BD or whatever).’ Those disability categories do not cause reading difficulties—specific reading-related skill deficits cause reading difficulties.”

Language comprehension, on the other hand, may be a tougher beast to tackle. Linguistic skills and knowledge are cumulative and on-going. Most importantly, a core component of language comprehension is background and topical knowledge, in addition to grammatical and syntactical knowledge — both which are inadequately taught in most schools due to the lack of a strong and coherent core curriculum.

I should note that Siedenberg doesn’t seem to fully subscribe to the Simple View, and that by no means should we begin to think any one model can adequately describe something so complex as reading. In his endnotes he states, “The main weakness in Gough’s theory is that it did not make sufficient room for the ways that the components influence each other. Vocabulary, for example, is jointly determined by spoken language and reading. Vocabulary can also be considered a component of both basic skills and comprehension.”

Kilpatrick contradicts this view when he states, “In the context of the simple view of reading, it appears that vocabulary belongs primarily on the language comprehension side of the simple view equation, not necessarily on the word-reading side.”

Seidenberg proposes his own model, based on computational simulations, which looks something like this (Figure 6.2 from Chapter 6):

I think this model is useful for conveying why reading is complicated and can be hard to learn, but maybe not quite as useful for guiding school-based assessment and instruction.

Why is The Simple View of Reading important?

Having a clear model for reading comprehension means we have a guide for aligned assessment, prevention, and intervention. Unfortunately, many schools base ELA instruction primarily on state assessments, which tell you very little about a student’s reading needs. People seem to forget that the function of a state assessment is for school, district, and state level accountability, not to direct classroom instruction.

A state literacy assessment is a “literacy performance.” Here’s an explanation in their book, Teaching Advanced Literacy Skills:

“There is a tendency to examine the results of outcome assessments at the item level—to figure out the types of items groups of students struggled with and then go back and teach to support this understanding. Perhaps the most universal example is ‘finding the main idea’ in a passage . . . the problem is that finding the main idea—among many other similar performances or exercises—is just that—a reading performance. It is not a specific skill. That is, to perform the task at hand, in this case to find the main idea, the reader draws on many component skills and composite competencies and initiates those in concert with one another. In turn, when a student is not able to find the main idea, we still do not know why.”

In order to know why, we need assessments that can better pinpoint where the breakdown occurs, whether in word recognition or in language comprehension, or both. And then we need to do something about it. This is where it gets hard.

Reading is Hard

Though we can draw on a simple model to explain it, in actuality reading is complicated.

First of all, it’s completely unnatural. While we acquire spoken language organically, reading requires the imposition of an abstract system onto that language, a grafting of a fragmented alphabet onto a river of sound. Writing is something our species invented, an ingenious mechanism to convey information across space and time. While the first writing appeared around 3,200 BC, humans have been speaking for anywhere between 50,000 to 2 million years prior (we don’t know for sure because we couldn’t record anything yet, duh).

There are many irregular words in the English language, which would appear to make the teaching of something like phonics a daunting endeavor. We assume that kids need to be taught the rules, and then memorize the exceptions (these are known as “sight words.”) Makes sense, right?

Yet research has made it clear we don’t acquire most sight words through memorization. Instead, we draw upon our letter-sound knowledge and phonological analysis skills to recognize new written words and unconsciously add them to our “orthographic lexicon.”

What’s interesting on this point is there appears to be some disagreement between the models Seidenberg and Kilpatrick use to explain this process. Seidenberg calls it statistical learning, meaning that we learn to recognize patterns in common words, from which we then can recognize many others, including ones with irregularities. Kilpatrick, on the other hand, terms it orthographic mapping, which is the process of instantaneously pulling apart and putting back together the sounds in words, drawing upon letter-sound knowledge and phonemic awareness. In either model, what is acknowledged is that children learn to recognize a large volume of new words primarily on their own, but that such an ability is founded upon a strong understanding of sounds (phonology) and their correspondences in written form (orthography).

Honestly, I find both concepts—statistical learning and orthographic mapping—hard to wrap my head around.

It’s also possible they describe different things. Seidenberg’s term seems more global, explaining how we acquire vocabulary, while orthographic mapping refers more specifically to the relationship between decoding and acquiring vocabulary. I should note here that Kilpatrick did not come up with the term, “orthographic mapping,” but rather draws on the research of Linnea Ehri.

Here’s Seidenberg on statistical learning:

“…learning vocabulary is a Big Data problem solved with a small amount of timely instruction and a lot of statistical learning. The beauty part is that statistical learning incorporates a mechanism for expanding vocabulary without explicit instruction or deliberate practice. The mechanism relies on the fact that words that are similar in meaning tend to occur in similar linguistic environments.”

Here’s Kilpatrick on orthographic mapping:

“Roughly speaking, think of phonic decoding as going from text to brain and orthographic mapping as going from brain to text. This is, however, an oversimplification because orthographic mapping involves an interactive back and forth between the letters and sounds. However, it is important that we do not confuse orthographic mapping with phonic decoding. They use some of the same raw materials (i.e., letter-sound knowledge and phonological long-term memory), but they use different aspects of phonological awareness, and the actual process is different. Phonic decoding uses phonological blending, which goes from “part to whole” (i.e., phonemes to words) while orthographic mapping requires the efficient use of phonological awareness/analysis, which goes from “whole to part” (i.e., oral words to their constituent phonemes).”

Kilpatrick notes that “The vast majority of exception words have only a single irregular letter-sound relationship.” This means that if a reader knows their letter-sound relationships well, they will be able to negotiate the majority of words with exceptions and irregularities.

Phonology: What We Can Hear and Speak is the Root of Written Language

In Seidenberg’s book, he argues that phonemes are the first abstraction on the road to the written word. A phoneme is the individual sound that a letter can represent (e.g. the sound of “p”). While we learn many such sounds as we acquire spoken language, the need to disaggregate a single component sound into a phoneme only becomes necessary in the translation of speech into the written form. As Seidenberg puts it:

“Phonemes are abstractions because they are discrete, whereas the speech signal is continuous. . . The invaluable illusion that speech consists of phonemes is only completed with further exposure to print, often starting with learning to spell and write one’s name.”

And indeed, this is why far too many of our kids have problems with reading. As Kilpatrick puts it, “The phonological-core deficit is far and away the most common reason why children struggle in word-level reading.”

Once I grasped this deceptively simple idea—that fluent reading is dependent on the ability to hear and speak the sounds of letters within words—prevention and intervention began to make more sense to me. Before, my understanding of the distinction between phonological awareness and phonics and what this meant for instruction was muddy. Now, I know that before even looking at a letter or a word, a student needs to practice hearing and speaking the sounds. This is how the student develops phonemic awareness. Phonics, on the other hand, is taught when those sounds are then applied to letters.

The Importance of Advanced Phonemic Awareness

That may sound straightforward (no pun intended), but one of the key understandings I gained from Kilpatrick is that we too often stop at basic phonological awareness, both in our assessments and in our intervention. While sound instruction in grades K-1 in phonological awareness and phonics should help to prevent most word reading difficulties (“Intervention researchers estimate that if the best prevention and intervention approaches were widely used, the percentage of elementary school students reading below a basic level would be about 5% rather than the current 30% to 34%”), there are some students who will present with more severe difficulties. And those difficulties often stem from lacking more advanced phonemic awareness. He also points out that these advanced phonemic skills continue to typically develop in grades 3 and 4, well past the point that most schools provide systematic phonemic and phonics instruction.

Kilpatrick stresses that intervention and remediation for such students requires explicitly teaching advanced phonological skills.

So What Can We Do?

The great thing about Kilpatrick’s book—and why you should buy it—is that unlike many writers in the field of education, he actually goes through what assessments you can use and what you can do instructionally, both for prevention (K-1) and for intervention (grades 2 and up), to address reading needs. He calls out programs by name and praises or critiques them based on key understandings from the research, and some of it was pretty surprising to me.

But I’m going to stop here for this post before it gets overlong. When I can find time to post again (it’s seriously hard with a 2 year old and 9 month old and the school year is about to begin), I’ll share some of the assessments and programs that I think are most accessible from Kilpatrick, as well as dig into some of the sacred cows that Kilpatrick, Seidenberg, and Hanford have slayed.

Afterword

You’ll notice I didn’t mention what I learned from my linguistics course, which was just an online series. It was fine, but I only found it useful insofar as it equipped me with some terms like lexicon, morphology, semantics, or pragmatics. If you have any recommendations for further learning in linguistics, please let me know.

Also, if I’ve demonstrated any misconceptions in this piece or you would like to challenge or add to anything I wrote, please share!

And thank you for reading.