This site is not anti-phonics. Phonics is real, necessary, and useful. But the movement built around it has not moved the numbers — and it is worth asking, plainly and without defensiveness, why. The answer is not that the teaching is wrong. It is that teaching was only ever half the problem.
On the nation's own reading assessment, the share of 4th graders reaching proficiency has been essentially flat for a generation. That alone should give the movement pause. But the more alarming figure is the one at the floor: the share of children who cannot read at even a basic level has grown.
Source: National Assessment of Educational Progress (NAEP), U.S. Dept. of Education. "Below proficient" is not the same as "cannot read" — NAEP's proficient bar is deliberately high. But "below basic" is a genuinely low floor, and that is the number moving in the wrong direction. Note also: even the highest-performing states leave most of their students below proficient. This is not a story of progress too slow. It is a story of a ceiling that has never moved.
Percentage of 4th-grade students below reading proficiency — 2003 to 2024
Consider the single largest test the phonics-first approach has ever been given. Under No Child Left Behind, the federal Reading First initiative put explicit, systematic phonics into thousands of the country's highest-need schools, at a cost of roughly a billion dollars a year. The government's own impact study found exactly what it was designed to find and exactly what should trouble the movement: Reading First measurably increased phonics instruction and improved first-grade decoding — and produced no statistically significant improvement in reading comprehension in grades one, two, or three. It did the narrow thing well. It did not make children read.
The pattern repeats at the state level. Colorado passed its READ Act in 2012 and embraced the Science of Reading early and wholeheartedly — mandated screening, approved curricula, tens of millions of dollars, sustained for over a decade. The author of this site worked directly with several of Colorado's major districts across those years and watched full-scale implementation up close: the training, the mandated programs, the screening, the sincere and sustained effort of good people. Reading outcomes stayed essentially flat. When the most committed, best-resourced version of a reform produces no gain, the honest conclusion is not "implement it harder." It is that the reform was aimed at only half of what decides whether a child reads.
After Mississippi overhauled its early reading beginning in 2013, its fourth-grade reading scores rose further than any other state's — from near the bottom of the national rankings to at or above the national average by 2024. That is a genuine achievement, and it would discredit everything else here to wave it away.
But look closely at what Mississippi actually did, because it was never phonics alone. Its overhaul bundled universal early screening, intensive coaching, and targeted support to catch struggling children before they failed — much of which is not about the method of instruction at all, but about finding and supporting the children who arrive not yet ready. The best result the instruction-first world has ever produced got there in part by quietly doing some of the readiness work this site keeps pointing to.
And even so: roughly 68% of Mississippi fourth graders still do not read proficiently, and by eighth grade the state sits below the national average again, the early gains thinning exactly as the demand for real comprehension grows. Mississippi is not the counterexample to this site's argument. Read carefully, it is the confirmation of it. See the Equation below.
There is a critical distinction between the screens given in grades K–2 and the assessments given in grade 3 and above — and understanding it explains why the system can believe it is succeeding while the real numbers stand still.
These tools mostly measure how fast and accurately a child can sound out words and nonsense syllables. That is a test of the thing phonics drills, not a test of understanding — and they say nothing at all about the child's cognitive skills, the engine that decides whether reading can develop. A child can reach decoding fluency by more than one road: one steeped in print and read-aloud can pass a screen without ever having been drilled. A high score was never proof that a child can read.
These require a child to read a passage and show they understood it. This is reading. These are the scores that are made public — and these are the scores that have not moved.
"In districts I worked with directly, over 80% of K–2 students scored at grade level on phonics-based screens. Only 40% passed the state reading assessment in 3rd grade — after years on Science of Reading curricula."
— Field observation, Colorado school districts (this site's author, working as a district consultant)So a district adopts a Science of Reading curriculum, watches most of its young children pass the early screens, and reports progress in complete good faith — only to see those same children reach the third-grade test, the one that finally asks whether they comprehend, and a large share fall short. The two tests were never measuring the same thing. The system measured sounding-out and called it reading.
Let us not start from a position and defend it. Let us walk through a series of plain statements — each one, we think, hard to disagree with — and see where they lead. If any single step is wrong, the chain breaks. But if each step holds, the conclusion arrives on its own.
We have a real, lasting problem. Roughly two-thirds of our children read below proficient, and that number has barely moved in thirty years. Whatever we believe about causes, the fact is not in dispute.
It is not for lack of trying. We have poured effort, money, research, and political will into reading, argued the question for decades, and most recently corrected toward phonics — a correction with real merit. No one can honestly say we have not tried.
Here is the step that changes the conversation: what we are doing is not wrong. The instruction we provide is not a failure. Look in any classroom and you will find children who learned to read well under exactly the methods we use. For them, it worked. Hold on to that, because it is true.
But it is not sufficient for most. That same classroom — same teacher, same lessons, same room — also produced children who did not learn to read well. If the instruction were the whole story, the outcomes would not split like that. They split anyway.
And more of the same has not closed the gap. We have mostly answered struggling children with more of the instruction that already did not reach them. For the children furthest behind, that usually does not work — and there is little reason to expect the next round to do what the last several could not.
So the difference must come from the child, not the lesson. If the lesson is held constant and the results differ, the difference is coming from what each child brings to it. That is not controversial. It is arithmetic.
And what the child brings comes down to the brain. Reading is something the brain does. A child who arrives rested, secure, with strong memory and attention and a rich store of language is primed to turn instruction into reading. A child who arrives exhausted, hungry, or with underbuilt underlying skills is not — and the very same lesson cannot do for that brain what it did for the other.
So a child's reading is the product of two things: the instruction received, multiplied by the child's readiness to use it. A product, not a sum — which means if either is near zero, the result is near zero. We have spent thirty years perfecting the first and mostly assuming the second. For the children who arrive ready, that worked. For the many who do not, it could not.
Notice what we did not conclude. Not that phonics is wrong. Not that teachers are failing. Not that any program is the answer. Only this: that reading depends on both instruction and a ready brain, that we have worked hard on the first and neglected the second, and that this is the most likely reason our serious efforts have not moved the number. The argument is not that we have been wrong. It is that we have not yet been sufficient — and that we now know where to look.
The crisis has a clean explanation. The system has been optimizing one factor in a two-factor equation and wondering why the product never reaches the level it should.
This is a way of seeing, not a formula you can compute. There is no number that equals a proficient reader. What the equation captures exactly is the one truth the field keeps missing: that instruction and readiness are multiplied, not added — so strength in one cannot rescue a real deficit in the other.
Curriculum, phonics, teacher training, screening, mandates — everything the school does to the child. The Science of Reading has focused almost entirely here, and improved it in places. Mississippi raised this factor hard. Colorado raised it faithfully for over a decade.
The child's own capacity to receive, hold, and make sense of what is taught — the readiness of the brain to learn. Working memory, attention, processing, oral language, and a nervous system not consumed by hunger, exhaustion, or fear. Built mostly in the earliest years, mostly at home.
The word "readiness" is easy to hear wrong, so let us be exact: it does not mean a ripeness you wait for. It means a set of specific, buildable capacities — and where they are missing, the answer is not to wait but to build them. "Learning effectiveness" names the same thing without shifting the blame onto the child: the question is not whether the child is ready, but whether the system built the readiness the child needed.
Everything turns on the small symbol between the two factors. If it were a plus, the halves would trade off — a little more instruction would always buy a little more learning. That is the model the whole system runs on. But learning does not add. It multiplies. And multiplication has a property addition does not: anything times a number near zero is near zero. Take the finest instruction ever devised — the most perfect phonics, the most gifted teacher, a ten out of ten — and apply it to a child whose readiness sits near zero, and the product is not "pretty good." It is near zero.
That is why a district can adopt the correct, evidence-endorsed reading program, drive the instruction factor all the way toward ten, and watch its scores barely move. Of course they barely move. If a large share of children arrive with a learning factor near zero, you have been multiplying a big number by a small one and calling the disappointing result a mystery. It was never a mystery. Mississippi is not a miracle and not a puzzle. It is the arithmetic: raise one factor in a product and the product rises, capped by whichever factor you left alone.
"An approach aimed almost entirely at instruction can refine instruction forever and still not close the gap, because instruction is only one half of what determines whether a child learns. The other half — whether the child is able to learn what is being taught — the movement does not address at all."
— How Are the Children? (Gary S. Smith, 2026)This is not a case against phonics. It is a case for seeing both halves. Phonics is one input to the instruction factor. It matters. The error is treating instruction as the whole problem when learning effectiveness is equally decisive — and almost entirely unaddressed.
The equation is the wide view. The Reading Engine is the close-up — the picture of why reading is a product and not a sum.
Start with what reading is for. Reading is an act, and its purpose is knowledge — meaning, understanding, the thing on the far side of the words. Reading well enough to get there depends on the word library: words bound as single units and retrieved instantly, so the mind is free to think about what the words say instead of laboring to work them out. Many roads can reach that library — whole language reaches it for some children, phonics-based instruction for others, a childhood of being read to and talked with for many more. Those roads are real, and none of them is the destination. The library is. And building the library, then running it, is work — which is where the engine comes in. Think of the whole thing as an engine that turns marks on a page into understood meaning. It runs on two very different kinds of parts.
The store of words a reader recognizes instantly and whole — retrieved from memory in a fraction of a second, not sounded out. Building this library is the entire goal of early reading. Instruction builds the fuel.
Attention to stay on the page, working memory to hold the start of a sentence until you reach its end, processing speed fast enough to leave room for meaning, the ear for the sounds inside words. General-purpose brain systems — the same equipment a child uses to follow directions or do arithmetic. This is what "readiness" actually means.
Fuel is useless without an engine to burn it. A child with weak machinery — an attention that will not hold, a working memory that overflows before the sentence ends — cannot run the fuel no matter how much of it you pour in. You can decode every word on a page and comprehend nothing. And here is the part that deepens the whole picture: the machinery is needed not only to run the library but to build it — you cannot confirm and store a word you cannot hold in mind long enough to check. Weak machinery breaks reading twice.
A word does not get filed into the permanent library just by being sounded out. Sounding out produces only a guess — and in English, an unreliable one. A word is welded into memory when several things happen together: the child attempts the word, confirms it against a real word they already know by voice, connects it to its meaning, and holds it in mind long enough for the connection to set. Woven together, again and again.
Phonics is one thread in that weave — a real and valuable one. Its job is to pull attention to a word's internal structure, how its letters map to its sounds, so the word binds as a fully specified unit rather than a fuzzy guess. But the phonics work is not itself what cements the word; the seeing, saying, and meaning do that. Drilling letter-sounds in isolation does not build the library. This is the correction the whole debate turns on: a child builds a word by seeing it whole, being told how it is pronounced — because they cannot reliably sound it out cold — and binding it to meaning until look, sound, and meaning fuse into one instantly-retrieved unit. Reading scientists call that fusing orthographic mapping; it is the movement's own account of how sight words form. The disagreement here is not with that science but with a classroom practice that treats the sounding-out as if it were the whole of it.
Now you can see why the system's favorite move — when a child struggles, drill more phonics, harder — fails on two levels at once. It over-relies on phonics as if sounding-out alone built the fuel, when the library actually requires the whole woven process. And it ignores the engine entirely, pouring fuel into machinery that cannot run it. Drilling phonics into a child whose engine has stalled is revving a car that is not in gear. You will hear the noise. The car will not move.
Ask anyone, young or old, whether they would rather have this year's phone or one from ten years ago. Everyone picks the new one, and everyone knows why: more memory, a faster processor, and it runs the new apps the old one chokes on. A brain is like that. The child's memory, processing speed, and attention are the hardware; the instruction we provide is the software. The same app runs smoothly on the new device and freezes on the old one — same software, different hardware, completely different result. That is a classroom of twenty children in a single picture. We have been debating which app to install while handing some children the newest device and some a ten-year-old one, then wondering why the same lesson runs so differently.
And here is the one place a child is better than a phone. A phone's hardware is sealed at the factory. A child's is not. The machinery is not fixed — it develops, built by experience and exercised by every real act of learning, most of all in the early years. We are not sorting children into fast machines and slow ones. We are building the machine, while there is still time.
This is what finally explains the girl whose story opens the book behind this site — told here as the author's firsthand experience, not offered as proof. She was intelligent, and she was eventually taught well; her library-building was fine. But her working memory sat near the bottom of the range, and her engine could not run what the teaching poured in. In the author's account, only once the machinery was strengthened did the ordinary work of learning to read finally take hold: the teaching stocked the library, and a stronger engine let her use it. Neither alone would have been enough — which is the equation, made mechanical.
You are not decoding these words letter by letter. You are recognizing them — instantly, effortlessly, the meaning arriving before you are even aware of the word. That instant, whole-word recognition is skilled reading. "Sounding out" is the on-ramp a beginner uses until they no longer need it — a real and useful on-ramp. But it is not the act of reading, and, as it turns out, it was never quite "sounding out" in the first place.
Take a child who has never seen the word cake and tell them to sound it out. They cannot — not reliably — because the letters do not carry their sounds on their faces. The c could be hard as in cat or soft as in city; the a could be short as in cat or long as in cake; the silent final e is itself a convention someone has to teach. Only after the child is handed the patterns — "c before a is usually hard," "a with a silent-e says its name" — does cake become sound-out-able. And notice what that means: the work was done by the taught pattern, not by the marks. What we call "sounding out" is really guided pattern recognition — and reading is pattern recognition all the way up, from tiny taught patterns to whole words retrieved in a single piece. Even a perfectly regular word can't be read cold. The letters never explained themselves.
So here is where phonics actually sits. It is a real and valuable ingredient in building the word library: its job is to pull a child's attention to the internal structure of a word — how its letters map to its sounds — so the word can be stored as a fully specified unit rather than a fuzzy visual guess. That is genuinely useful, and it is especially good at one thing in particular: spelling. But the phonics work is not itself what cements a word into memory. The cementing is done by the acts the phonics is wrapped in — seeing the exact word, hearing it said, connecting it to meaning, and meeting it enough times for the bond to set. Drilling letter-sounds in isolation does not build the library; the seeing, saying, and meaning do. Phonics is one thread in that weave, not the weave itself.
A quick distinction most people never draw. Phonemic awareness is an ear skill — hearing and manipulating the spoken sounds inside words, no letters involved. Phonics is a print skill — mapping those sounds to letters. They are constantly conflated, even by practitioners. Both help on the on-ramp, and both help with spelling. Neither is the mechanism of fluent reading, which is retrieval.
This is a secondary point, not the heart of the argument, but it is worth seeing once. English is opaque: the same letters take different sounds, and nothing on the page warns you which. Drill a child that "ea" says short-e, as in head, and the rule shatters on reach, then great, then idea — with no cue on any of them. Read the sentence below aloud and watch "ea" change its mind thirteen times.
The point is not that phonics is useless — it is that the code is an unreliable predictor dressed as a reliable one, so early sounding-out is guided guessing until enough exposure makes the patterns familiar. And in the hardest cases it is guessing even a fluent adult cannot resolve: you cannot tell whether read is present or past, or whether lead is the metal or the verb, from the letters alone — only the meaning of the sentence decides. That is the whole idea in miniature. The letters are a starting clue; the word truly resolves only when it is bound to meaning and stored whole.
The opacity of English is the smaller limitation — the vivid one that makes a phonics believer pause. But even if English were perfectly transparent and decoding delivered every word flawlessly, it still would not solve the crisis, because decoding was never the binding constraint. A perfect on-ramp still only helps stock the library — and does nothing for the engine that has to build and run it. That is the real load phonics cannot lift, and it is the engine, not the irregularity of English, that explains why.
One honest observation about the movement itself. The Science of Reading names five components, and comprehension is one of them — on paper. In practice, decoding and phonemic-awareness drilling take the overwhelming share of the time, and the acts that actually build the library get treated as incidental. In every conversation the author of this site has had with practitioners in the field, the same assumption surfaced: that sounding out a word is reading. That single belief is why the ceiling is invisible from inside the movement — a strong decoding score looks like a reading score, and the collapse at third grade looks like a mystery instead of a predictable result.
Readiness is not one thing. It is a stack of systems built one on top of another, each resting on the ones below it, with reading balanced at the very top. When we say a child is "not ready," what we almost always mean, without knowing it, is that somewhere down in that stack a floor was never finished — and everything above it is leaning on empty air.
Notice what is nowhere in the stack: intelligence. A brilliant child with a cracked floor low in the stack will struggle; a child of ordinary gifts with every floor sound will read without trouble.
This is the deepest reason the system's favorite explanations — "she's not bright enough," "he's not trying" — are not just unkind but wrong. They are looking at the top of the tower and never once at the floor that actually gave way. And it is why the current response to a struggling child — test, label, then deliver more of the same instruction — so reliably fails. You cannot fix a leaking boat by loading more cargo onto it.
"Can the child read?" is a symptom, not a cause — and it is where the system stops looking. The real work runs the other way: follow the symptom down through the stack to the floor that did not get built, then treat back up — fix the cause, rebuild the capacity, and only then teach the reading. The same weak engine can have very different roots, and each needs a different fix.
The same symptom, many causes — the lesson of attention itself
The same "attention problem" can have very different roots. One child needs sleep; another exercise; another a change in nutrition; another help with anxiety or trauma; another medication. Treat the label the same for all and you help some, do nothing for many, and set a few back. You have to find the cause for this child. Reading is the same.
The cognitive machinery is the heart of this — the equipment that takes in the world and makes sense of it. No machinery, no reading. And here the site draws a careful line, because it is the line that keeps the whole argument honest. What is not in dispute is that these skills are built: profoundly in the earliest years, and continually through experience, because every act of real learning exercises the very attention, memory, and reasoning it depends on. Skills that get used, grow — that is the premise under all of education. So a child arrives with more or less engine, and it keeps developing, which means the right response to a weak engine is to find the gap and strengthen it the way engines are actually built — through rich experience and well-aimed practice — while you teach. What the site does not claim is a shortcut: that drilling a cognitive skill in isolation converts directly into reading scores. The formal "brain-training" transfer studies are genuinely mixed, and reading still has to be taught. The point is only this: teaching lands on an engine, so the engine has to be developing for the teaching to take. Necessity, asserted plainly; a magic bullet, never.
If the stack is built before school, then the leverage is not in the classroom at all. It is upstream — in the years the brain is being built, and in the hands of the people who are there while it is being built, which is to say the family. This is not an accusation of families. It is the recognition that families have been handed the most important job in child development with almost no help.
The floors of the stack are not built in first grade. They are built in the crib, the kitchen, the car seat — by relationship, in what researchers call serve-and-return. A baby babbles, points, reaches; an adult responds. Do that ten thousand times and you are physically constructing the child's language, attention, and stress-regulation systems, one exchange at a time. A child who gets a rich, responsive, word-filled version arrives with a tower built straight. A child who gets little of it — through no fault of the child's — arrives with floors unfinished, and no reading program is designed to pour a foundation that should have been laid years before.
Chronic stress and adverse early experiences physically disrupt the development of the very systems reading depends on — the attention and working memory that are the engine's core parts. The resulting gaps are real, findable, and often mistaken for a defect in the child.
And prevention is radically cheaper and kinder than repair. It is easier to build a floor while the brain is laying it down than to tear up the house years later — after a child has spent years failing under a load too heavy, before the labels harden and the shame sets in. We spend almost everything at the hard end, on repair, and almost nothing at the easy end, on prevention. We have the order exactly backward.
"Across decades of research, the strength of a child's working memory is one of the most consistent companions of reading comprehension. The machinery and the reading rise and fall together."
— see the working-memory research in the referencesThat is the honest home of the readiness claim — not a brain-training cure sold on the side, but the plain fact that the engine and the reading move together, that the engine is built by experience in the years before school, and that it keeps developing through real learning after. Support the engine and you still must teach; leave it unaddressed and the teaching has nothing to land on. Both halves of the equation, always, working together.
The data closes one door for good: pouring more of the same instruction into the same children has been tried, at scale, for a generation, and it has not moved the number. That is the hard part. The hopeful part is what it implies — the largest untapped gains in the entire system are sitting in the half almost no one is working on. Raise learning effectiveness and you finally lift the factor that has been holding the product down all along.
Raising it is not one move but two, working on different timelines. A serious answer needs both.
The largest, cheapest, and kindest gains are upstream, in the first 2,000 days: responsive relationships, language, nutrition, and a home not consumed by chronic stress — families genuinely supported to do the most important developmental work there is. Every floor laid early is a floor no one has to tear up and rebuild later. This is where the system currently spends almost nothing, and where the long-term leverage is greatest.
Stop prescribing more of what already failed. Diagnose down to the real cause in each child, strengthen the weak machinery — and, the piece the system keeps missing, teach reading in a genuinely different way. Not more phonics drilling. A different kind of instruction altogether.
If drilling letter-sounds and trusting that comprehension will follow is the method that produced the flat line, more of it cannot be the answer — but neither is a return to whole language. There is a third way, and it is defined not by a brand but by what it must do. So rather than name a program, here is the standard any adequate approach has to meet. It builds each word into memory as a unified whole — through its sounds, not instead of them — fusing its look, its sound, and its meaning into one stored unit. It teaches all the skills reading needs together, in meaningful text, rather than in isolated drills. It keeps phonics as one foundational thread — pulling attention to a word's structure — without over-drilling it or mistaking it for the whole. And it aims squarely at the real destination: a word library, built efficiently and made automatic. That is not a tweak to the current method; it is a different category of instruction, folding in what both camps got right while fixing what each left out.
Programs built this way exist. Reading Kingdom is one example that meets the standard above — named here only so the idea is concrete and buildable, never as a product to sell. The site's author has no financial interest of any kind in it, and would rather you hold any program, this one included, to the criteria than take a name on faith. The point is the standard, not the brand.
And notice the destination this reaches. The goal of reading instruction is two things at once: a deep orthographic library — words known instantly and whole — and a Reading Engine strong enough to build and run it. An approach that drives words to automatic recognition quickly, and ties each one to meaning, to writing, and to sentence structure, serves both at the same time. It fills the library the durable way — and it treats the engine far better than endless decoding drills do. Sounding-out kept effortful long past its usefulness consumes the very working memory comprehension depends on; practice aimed at automaticity instead frees that working memory and exercises it inside the act of real reading. (This is not standalone "brain training," about which this site stays deliberately cautious — it is ordinary reading instruction whose design simply asks less of the engine and uses it more productively.) Traditional drill-and-hope does the reverse on both counts: it fills the library unreliably and taxes the engine the whole way. That difference — not any single feature, and not any brand — is the reason a different approach can move children the current one leaves behind.
A large share of English learners — Spanish speakers above all — come from phonetically transparent languages, where a letter reliably makes one sound and the code keeps its promises. Their first language taught them, correctly, to trust letter-sound rules. English then hands them a code that breaks those rules constantly, so for these learners the mismatch is especially acute: the very habit their experience built is the one English most often violates. An approach that fuses a word's form, sound, and meaning together — rather than leaning on the unreliable code in isolation — sidesteps a trap that isolated phonics can set for them.
To be clear, this is not an argument against teaching English learners phonics. The research is consistent — this is the finding of the federal National Literacy Panel on language-minority children — that they benefit from systematic phonics when it is paired with strong oral-language and meaning support, which is exactly the integrated approach described here. What serves them poorly is phonics drilled in isolation, cut off from the meaning a child still building English needs most. So the point is not "less phonics for English learners." It is phonics woven into meaning from the first lesson — the same principle the whole site argues, felt most sharply by the learners for whom the code is least intuitive.
So here is the practical answer, stated plainly: prevention upstream, where the machinery is built, and — for the children already struggling — remediation that fixes the cause and teaches reading a different way. The system would be far better off adopting an approach that meets the standard above — whichever program that turns out to be — than spending another decade intensifying the one method that has already shown it cannot close the gap. That is not a product pitch. It is a claim about where the leverage is, and about opening an honest debate so the children win — which is the entire purpose of this site.
The federal government's own impact study of its billion-dollar-a-year phonics initiative: it increased phonics instruction and improved first-grade decoding, and produced no statistically significant gain in reading comprehension in grades one, two, or three. The largest test of the approach, judged on the outcome that matters.
The foundational SoR document. Systematic phonics beat other methods mainly on decoding and spelling, most strongly in K–1; the comprehension advantage was significant early but not durable beyond first grade. In plain terms: drill phonics and children do better on phonics-type tasks — the comprehension edge fades where real reading begins.
One of the longest-running, best-funded structured-literacy efforts in the country, embraced from 2012 and observed directly by this site's author across multiple major districts. Reading outcomes stayed essentially flat. The honest reading is not "implement it harder" but that the reform was aimed at only half of what decides whether a child reads.
Fluent reading depends on whole-word patterns stored in memory through orthographic mapping. Phonics helps build the initial connection, but the endpoint is always automatic whole-word recognition. Every word a fluent reader knows has become a sight word. This is agreement with the science, not a rebuttal of it — and the reason decoding matters, and is not the finish line.
The K–2 screens that reassure a district — DIBELS and the like — measure decoding speed, the on-ramp. They say nothing about the cognitive engine that decides whether a child can build or run a word library. We screen the thing that is easy to measure and stay blind, by design, to the thing that matters most.
The Department of Education's own review of one widely-sold reading program (Achieve3000) found no discernible effect on reading fluency, and no comprehension study that met its evidence standards. Independent, comprehension-level proof is harder to find than the marketing implies — a fair thing to ask of any program, this site's preferred approach included.
Of 45 commonly taught phonics rules, only 18 held even three-quarters of the time; "when two vowels go walking" is right less than half the time. The point is not that phonics is useless — it is that the code is an unreliable predictor, and nothing on the page tells a beginner which words will keep the rule and which will break it.
Across decades of study, working memory is strongly linked to reading comprehension. Effortful decoding and comprehension compete for the same finite cognitive resources — which is why the engine (the machinery) matters as much as the fuel (the word library).
View full references and citations, including honest notes on contested findings →
Every honest recovery begins the same way — with someone finally saying, out loud, this is not working, and I will not pretend anymore. The Science of Reading corrected a real error and does real good where a child is ready. But it has not solved the crisis, and on its own it never can, because it perfects the teaching while ignoring whether the child is ready to learn.
The fix is not another program of the same kind piled onto the last. It is to raise the neglected half of the equation — to build the machinery in the years it is actually built, in the home, in the first 2,000 days; and, for the children already struggling, to fix the specific cause standing in each one's way and teach reading in a genuinely different manner, rather than prescribing another round of the same. There is a real solution here, and it is within reach. What is missing is not the knowledge. It is the will to act on what we already know.
There is a greeting, attributed to the Maasai, that asks not about the weather but about what matters most: how are the children? Right now, for far too many, the honest answer is: not well. Whether that stays true is up to us.