The Science of Writing Is Already Here. Our Literacy Systems Have Not Caught Up.

Science of Writing and Science of Literacy graphic illustrating the connection between reading, writing, research, and evidence-based literacy instruction.

What decades of research tell us about teaching writing, and why schools cannot afford to treat it as literacy’s second subject

For more than two decades, American literacy reform has centered on an urgent question: How do children learn to read, and how should schools teach them?

That question deserves the attention it has received. Reading opens access to knowledge, learning, opportunity, and participation in society. When children struggle to read, the consequences can follow them throughout school and into adulthood. The effort to bring stronger scientific evidence into reading instruction has therefore been both necessary and important.

But literacy has another side, and it deserves the same seriousness.

How do children learn to write? What knowledge, skills, strategies, language abilities, and forms of self-regulation do writers need? What does effective writing instruction look like? How do reading and writing strengthen one another? And how do schools build systems that help teachers use effective writing practices consistently, not for a few weeks or in a handful of classrooms, but across grades and over time?

Researchers have been studying these questions for decades.

We do not need to begin building a Science of Writing from scratch. A broad, converging body of evidence already exists. Researchers have developed cognitive, developmental, linguistic, social, and self-regulatory models of writing. They have conducted hundreds of intervention studies and multiple meta-analyses. They have studied handwriting, spelling, sentences, vocabulary, composition, strategy instruction, reading-to-write, writing-to-learn, motivation, self-regulation, professional learning, and implementation.

A major 2024 synthesis by Steve Graham, Alyson Collins, and Stephen Ciullo, Evidence-Based Recommendations for Teaching Writing, drew on findings from close to 1,000 investigations and identified 11 evidence-based recommendations for writing instruction.

That is not the record of a field waiting to be invented.

At the same time, we should not turn “Science of Writing” into another educational slogan or suggest that the research has answered every question. There is no single universally accepted theory that completely explains writing development. There is no one instructional program that constitutes the Science of Writing. The evidence is stronger in some areas than others, and researchers still need to learn more about maintenance, transfer, multilingual learners, individual differences, mechanisms of change, implementation, and what happens when effective practices move from controlled studies into entire school systems.

That does not make the Science of Writing immature. It makes it science.

A useful definition is this:

The Science of Writing is the evolving body of evidence about how writing develops, how it is taught and learned, how language, reading, knowledge, cognition, motivation, and self-regulation contribute to writing, and how effective instruction can be implemented successfully and sustainably.

The research base is mature enough to guide action now and incomplete enough to demand continued study.

Our challenge is no longer whether we have evidence about writing. Our challenge is closing the distance between what research has already taught us and what students experience every day in school.

What’s in this article

Writing is much more complex than putting words on paper

Writing is one of the most demanding things we ask children to learn.

Consider what happens when a teacher asks a student to engage in creative writing or to write an informative essay, argument, or narrative. The student must understand the task, activate or acquire relevant knowledge, decide what to communicate, think about an audience, establish clear goals for the written product, generate and organize ideas, select words, construct sentences, spell, produce text by hand or keyboard, reread what has been written, evaluate its effectiveness, revise when necessary, manage attention and emotion, and persist when the work becomes difficult.

Writers must coordinate many of these processes simultaneously.

This complexity helps explain why no single model captures everything we need to know about writing.

John Hayes’s Modeling and Remodeling Writing describes writing through a cognitive lens, including the processes writers use to generate, evaluate, translate, and control written language. Steve Graham’s Writer(s)-Within-Community model reminds us that writing is also a social act shaped by the communities, purposes, cultures, institutions, and people within which writers work. Young-Suk Grace Kim and Steve Graham’s Direct and Indirect Effects model of Writing, or DIEW, examines an interactive developmental system in which transcription, oral language, reading, higher-order cognition, self-regulation, executive functions, and knowledge can contribute directly and indirectly to written composition.

These models do not cancel one another. They illuminate different parts of an extraordinarily complicated human activity.

Together, they also protect educators from two opposite mistakes. One is reducing writing to mechanics, as though good writing will automatically emerge once students can spell, punctuate, and form sentences correctly. The other is treating writing as an almost mysterious form of self-expression that develops primarily by giving children time and freedom to write.

Students need time to write. They need meaningful purposes, choice, collaboration, creativity, and opportunities to develop identities as writers.

But most students also need instruction.

In their 2015 review, Graham, Harris, and Santangelo described research-based writing practices within the context of college- and career-ready expectations. More recently, the nearly 1,000-investigation synthesis by Graham, Collins, and Ciullo reached a similar broad conclusion: students need opportunities to write, but writing alone is not enough, a distinction that shapes how schools plan their instruction.

The evidence for writing instruction is substantial, but simple averages can mislead us

Research across grade levels shows that writing instruction can improve student writing. But Karen Harris raised an important caution about how we communicate these findings: broad “average effect sizes” can hide almost as much as they reveal.

For example, Kim, Yang, Reyes, and Connor’s 2021 meta-analysis examined 24 primary-grade studies involving 5,589 students. The overall mean effect across writing outcomes was .31, but that number is only the beginning of the story. Effects differed according to the writing outcome being measured and, importantly, according to the instructional approach. In that analysis, SRSD showed effects ranging from .59 to 1.04 across writing outcomes.

Likewise, the best-evidence meta-analysis of writing instruction for students in Grades 6 through 12 by Graham and colleagues examined 148 treatment-control comparisons involving 22,838 students. The researchers found positive effects from several instructional approaches, including strategy instruction, process approaches, transcription instruction, feedback, prewriting, and peer assistance when enough comparisons were available to examine those practices.

But it would be misleading to collapse all of those approaches, populations, studies, and outcomes into a single number and call it “the effect of writing instruction.”

Effects vary depending on what teachers teach, what researchers measure, students’ ages and needs, the duration and quality of instruction, who delivers it, the comparison condition, and the methodological strength of the research.

The more useful conclusion for educators is that multiple research traditions now converge on something important: students become better writers when schools provide explicit, sustained, multicomponent writing instruction.

The Science of Writing does not point to one technique. It points toward a coherent instructional system.

Reading and writing are deeply connected, but they are not interchangeable

A complete approach to literacy must take seriously the relationship between reading and writing.

Reading turns written language into meaning. Writing turns meaning into written language for someone else.

The two draw upon many shared resources, including vocabulary, syntax, knowledge of sounds and spellings, background knowledge, text structures, genre knowledge, and the ability to monitor meaning. But reading and writing make different demands on learners.

A student can recognize a word they cannot spell. A student can understand an argument they cannot yet construct. A student can explain the meaning of a text while having great difficulty organizing evidence from that text into a coherent written response.

A massive meta-analysis by Kim, Wolters, and Lee examined 395 studies, 2,265 effects, and 120,669 participants. The researchers found a strong overall relationship between reading and writing, but the relationship differed according to the particular reading and writing skills measured and across development. Word reading and spelling were especially closely related, while the relationship between reading comprehension and written composition was more moderate.

That evidence tells us reading and writing travel together. It does not tell us that teaching one automatically teaches the other.

Intervention research gives schools more actionable information.

The Reading for Writing meta-analysis by Graham and colleagues examined reading interventions that measured their effects on writing across students from preschool through Grade 12. It found that reading instruction and increased interaction with texts can improve aspects of students’ writing.

The relationship also works in the other direction. Graham and Hebert’s Writing to Read meta-analysis found that writing about material students read can improve comprehension and that teaching writing can positively affect reading outcomes.

Another synthesis, Effectiveness of Literacy Programs Balancing Reading and Writing Instruction, examined programs designed to bring reading and writing together rather than treating them as isolated activities.

And writing does more than support literacy. Graham, Kiuhara, and MacKay’s meta-analysis examined 56 experiments and found that writing about content enhanced student learning in science, social studies, and mathematics across elementary, middle, and high school students.

The implication is not that schools should blend reading and writing indiscriminately.

The implication is purposeful integration.

Students need explicit reading instruction and explicit writing instruction, along with intentional bridges between them. They can read as writers, examine how authors organize texts, learn vocabulary and content from what they read, take notes from sources, evaluate evidence, summarize ideas, study sentences, and use new knowledge to compose.

Research gives us increasing guidance about how to integrate this work from the elementary grades through high school.

A strong science of literacy must protect the distinct work of reading and writing while taking full advantage of their relationship.

How writing became literacy’s second subject

Reading’s prominence in education policy did not begin with an effort to exclude writing.

In 1997, Congress directed the National Institute of Child Health and Human Development to convene a national panel to examine research-based approaches to teaching children to read. The National Reading Panel carried out that specific mandate. Writing was not part of its charge.

The problem was what happened afterward.

Reading developed an extraordinary national infrastructure around research, legislation, curriculum review, screening, intervention, professional development, assessment, and public vocabulary. Writing never developed anything comparable.

Evidence of that imbalance continues to appear in state policy. A 2023 Shanker Institute analysis, Reading Reform Across America, examined 223 literacy bills enacted between 2019 and 2022 across 45 states and the District of Columbia. Compared with foundational reading skills, oral language and writing received limited attention.

National assessment tells another part of the story.

The 2011 NAEP Writing Assessment found that only about 27 percent of eighth- and twelfth-graders performed at or above the Proficient level. A writing assessment was administered again in 2017, but NCES reported comparability problems associated with the transition between devices, preventing the kind of national trend reporting we routinely expect in reading, an imbalance with real consequences for policy and funding.

What schools measure, fund, schedule, train, and talk about becomes visible. What they do not can quickly become the activity squeezed into whatever minutes remain.

The remedy is not to reduce the urgency around reading. It is to expand our literacy priorities enough to include writing with equal seriousness.

Transcription opens the door to composition, but it is not the whole room

Beginning writers face a genuine cognitive constraint.

If letter formation, keyboarding, and spelling require enormous conscious effort, students have fewer mental resources available for generating ideas, selecting language, organizing information, monitoring meaning, and thinking about their audience.

For that reason, handwriting and spelling belong squarely within the Science of Writing.

A comprehensive meta-analysis by Tanya Santangelo and Steve Graham found that handwriting instruction improves important aspects of handwriting performance. Graham and Santangelo’s meta-analysis of spelling instruction, involving 53 studies, found clear benefits from formal spelling instruction, including improvements in students’ spelling when writing.

These foundational skills matter greatly. But foundational does not mean sufficient.

A student who spells every word correctly does not automatically know how to explain an idea clearly, organize an argument, create suspense in a narrative, select evidence, write an effective introduction, or revise a confusing paragraph.

Automaticity creates capacity. It does not supply content, organization, purpose, audience awareness, or judgment.

Sentence construction and grammar belong within this broader composition system as well. Jones, Myhill, and Bailey’s study of contextualized grammar instruction found that grammar can support writing when teachers connect grammatical knowledge to the choices writers make. Saddler and Graham’s research on sentence combining likewise demonstrated improvements in sentence construction and writing among struggling young writers.

The instructional lesson is important. Schools do not need to choose between foundational skills and meaningful composition.

Students need both.

Conventions should become tools writers use to communicate meaning, not simply errors teachers circle after writing is finished.

Writers also have to learn how to steer themselves

Writing places unusual demands on self-regulation.

A writer must establish goals, select and organize ideas, produce language, monitor progress, manage attention and emotion, judge whether a text is working, decide what to change, and persist when writing becomes difficult. Eventually, writers must perform much of this work without a teacher standing beside them.

This is one of the most important contributions of Self-Regulated Strategy Development, or SRSD.

SRSD makes the hidden processes of effective writing visible.

Teachers help students build the knowledge they need, discuss the purpose and value of strategies, model how those strategies work through think-alouds, and support students as they begin using them. Students learn to set goals, monitor themselves, use productive self-statements, collaborate, evaluate their work, and gradually assume greater responsibility.

Karen Harris’s 2024 review, The Self-Regulated Strategy Development Instructional Model: Efficacious Theoretical Integration, Scaling Up, Challenges, and Future Research, provides an important account of the theoretical development of SRSD and the integration of cognitive, metacognitive, behavioral, motivational, social-emotional, and other research traditions that have shaped the model.

A critical point sometimes gets lost when SRSD is reduced to its visible mnemonics.

As Harris frequently says, “There is no magic in the mnemonic.”

TREE, POW, TIDE, and other mnemonics can help students remember powerful strategies and knowledge. But the mnemonic is not the intervention.

The goal is not to produce students who mechanically follow a formula.

The goal is for students to remember and use the powerful strategies and writing techniques they have developed, adapting them flexibly as writers according to their goals, readers, genres, and purposes.

That distinction is fundamental to SRSD.

What does the evidence tell us about SRSD?

SRSD has been studied for decades across grade levels, student populations, writing genres, instructional settings, and countries.

Across multiple reviews and meta-analyses, SRSD has repeatedly demonstrated strong effects on student writing. In the Grades 6 through 12 best-evidence analysis by Graham and colleagues, SRSD studies produced an estimated average effect of .84, compared with .51 for other forms of strategy instruction.

But again, we should not turn .84 into a slogan.

Effects differ across outcomes, students, study designs, grades, implementation conditions, and comparisons. The more defensible claim is stronger precisely because it is more careful: SRSD has one of the most substantial and consistently positive research bases among approaches to writing instruction.

Researchers are also continuing to investigate why SRSD works and under what conditions. Fernandez and Guilbert’s 2024 review examined cognitive and metacognitive mechanisms associated with SRSD effectiveness and also identified questions that remain.

SRSD should therefore have a central place in the Science of Writing, but SRSD is not the entire Science of Writing.

Students also need language and vocabulary development, content knowledge, transcription, sentence construction, genre knowledge, reading, discussion, motivation, feedback, collaboration, and repeated opportunities to write.

That is why thinking in terms of a Science of Writing is so important. It encourages integration rather than allegiance to a single instructional component.

What happens after the intervention ends?

One of Karen Harris’s comments on the original draft concerned an issue that deserves more attention: maintenance.

Researchers working with SRSD have long recognized that students need support to maintain and continue using what they have learned. That does not necessarily require reteaching an entire unit. It can mean intentionally revisiting strategies, prompting students to apply them in new settings, and creating continued opportunities for use.

A major recent study gives us unusually valuable long-term evidence.

Salas, Ventura, and Llauradó followed 909 Catalan-speaking second- and fourth-graders after an 11-session SRSD opinion-writing intervention. Researchers assessed students after instruction and again at later points extending to 18 months.

Benefits remained detectable 18 months later on some important outcomes. Students continued to show advantages in planning and structural elements of their opinion writing. At the furthest follow-up point, however, the SRSD group’s advantage over the comparison group in overall text quality was no longer statistically maintained, even though performance remained above students’ own pre-intervention levels.

That qualification is important.

Powerful instruction can create durable learning. But durability should not be confused with permanence after a single period of instruction.

Students need continued opportunities to retrieve, practice, adapt, and use what they learned.

For schools, the message is practical: do not treat evidence-based writing instruction as a unit students complete and leave behind. Build the strategies and knowledge into the writing students do throughout the year and in later grades.

SRSD Plus and what we are learning about integration in the earliest grades

For years, teachers implementing SRSD have asked an understandable question: How do we connect strategy and self-regulation instruction with reading, oral language, vocabulary, spelling, handwriting, sentence construction, and the knowledge students need to write?

Recent work on SRSD Plus represents a research effort to investigate that question systematically.

The first study, Harris, Kim, Yim, Camping, and Graham’s 2023 study, Yes, They Can, examined first- and second-grade students learning transcription skills and oral language while also learning to read science text closely and write informative essays using SRSD.

The study challenged an assumption that very young children must first master foundational transcription abilities before they can participate successfully in complex composition and reading-to-write tasks.

Students worked in small groups, which is not unusual for instruction in the primary grades. The instruction integrated spelling, handwriting, vocabulary, sentence development, close reading of science texts, planning, writing, and self-regulation. Students in the intervention showed gains across multiple outcomes, including writing quality, planning, vocabulary, sentence proficiency, spelling, discourse knowledge, and other targeted areas. The study also examined expository text comprehension.

A later classroom study by Kim, Harris, Goldstone, Camping, and Graham, published as The Science of Teaching Reading Is Incomplete Without the Science of Writing, extended this work to 10 classroom teachers and 248 first- and second-graders.

Students receiving SRSD Plus outperformed students receiving business-as-usual instruction in spelling, vocabulary, sentence proficiency, planning from source text, discourse knowledge, writing productivity, and the quality of source-based informational writing. The study also found transfer to opinion writing. It did not find effects on word reading or handwriting fluency.

Taken together, these two studies tell us something important.

They provide evidence that even young students can engage successfully in an integrated instructional approach in which they read to learn, talk to develop language and academic vocabulary, acquire foundational writing skills, plan, compose, and learn strategies for regulating the writing process.

This does not mean two studies establish SRSD Plus as a universally validated, finished model for every grade level, language, school, or population.

More research is needed.

But we should not allow that familiar research qualification to obscure what the studies have already taught us. They challenge the idea that schools must choose between foundational skills and composition, or that young students should wait until later grades to read sources, build knowledge, and use that knowledge in writing.

They also fit into a much broader research literature showing that reading-to-learn and writing-from-reading can be integrated across grade levels, not merely in Grades 1 and 2.

That larger scope matters.

Evidence does not implement itself

This may be the most important section for principals, superintendents, curriculum directors, and instructional leaders.

An intervention can produce impressive results in a research study and still disappear inside a school system.

Research evidence alone does not change classroom practice.

Teachers need time to learn an approach. They need to see instruction modeled. They need opportunities to practice. They need materials they can actually use. They need feedback and collaboration. They need coaching and access to expertise when problems arise. Administrators need enough understanding to recognize what effective implementation looks like. Schedules must provide time to teach writing. Teachers need help adapting instruction while preserving the components that make it effective.

The broader implementation literature reinforces this point. Durlak and DuPre’s landmark review examined hundreds of studies and found strong evidence that implementation quality influences program outcomes.

Teacher coaching can help, but scale introduces another challenge. Kraft, Blazar, and Hogan’s meta-analysis of 60 causal studies found positive effects of coaching on instructional practice and student achievement. The researchers also found that effects tended to weaken as programs grew larger.

Scaling an evidence-based practice is not simply a matter of buying more materials or enrolling more teachers. The implementation problem itself changes with scale.

Writing-specific SRSD research offers encouraging evidence here. McKeown and colleagues’ 2023 study examined practice-based professional development and expert support for classwide SRSD implementation across 33 classrooms in 11 schools. The research provides evidence that classroom teachers can learn and implement SRSD with substantial support and produce meaningful improvements in student writing.

That work matters because the real goal is not to prove that an expert researcher can teach students successfully.

The goal is to make powerful instruction workable for teachers.

The barriers are bigger than the classroom

Karen Harris and Debra McKeown address this problem directly in their 2022 article, Overcoming Barriers and Paradigm Wars: Powerful Evidence-Based Writing Instruction.

Their argument extends beyond SRSD.

Schools possess evidence-based practices that can meaningfully improve writing, but several structural barriers keep those practices from becoming routine.

One is teacher preparation.

Harris and McKeown describe evidence showing that many teachers do not feel adequately prepared to teach writing. They cite an analysis in which only 28 percent of teacher preparation programs included a stand-alone writing instruction course, while writing was more commonly embedded in reading coursework. In special education preparation programs, the percentage offering a stand-alone writing course was even lower.

That creates an unfair situation for teachers.

Schools expect them to teach one of the most cognitively complicated academic activities children undertake, yet many educators enter classrooms without extensive preparation in how writing develops or how to teach it.

Professional learning must help close that gap.

But effective professional learning in writing cannot consist of a short presentation followed by a binder of materials.

Teachers need to see effective instruction. They need opportunities to practice. They need time to examine student writing, plan, collaborate, and learn how to differentiate. They need support during implementation.

Harris and McKeown also identify curriculum adoption as a barrier. Districts make understandable decisions based on standards alignment, comprehensiveness, technology compatibility, materials, cost, and other practical factors. But a curriculum can appear comprehensive because it contains a great deal of material while still failing to provide enough depth for students to master the writing practices that research supports.

More content does not automatically mean stronger instruction.

The question for district leaders should be: Does this curriculum give teachers the knowledge, time, sequence, support, and evidence-based practices students need to become increasingly independent writers?

That is a harder question than asking whether a program checks every standards box.

It is also a more important one.

Time may be one of the most overlooked writing interventions

Teachers cannot teach writing without time to teach writing.

It sounds almost embarrassingly obvious, yet the research Harris and McKeown review shows how fragile writing time can be.

Writing often receives less instructional attention than educators believe it should, particularly after the early elementary grades and in content-area classrooms. Students may complete worksheets, short responses, or brief summaries, but those activities are not substitutes for sustained instruction in planning, composing, revising, sentence construction, genre, source use, and self-regulation.

And when the school day becomes crowded, writing is often one of the first things squeezed.

This creates a false competition.

Writing does not have to take time away from science, social studies, or mathematics. As the Graham, Kiuhara, and MacKay meta-analysis demonstrates, students can use writing to learn content.

Students can explain a scientific phenomenon. They can use evidence to make a historical argument. They can explain mathematical reasoning. They can summarize, compare, analyze, question, and synthesize.

Writing is not only something students learn. Writing is something students learn through.

Assessment should help us develop writers, not shrink writing to a test

Assessment deserves a particularly careful place within the Science of Writing.

Good assessment is essential.

Teachers need to know what students can do independently. They need to know what students understand about a genre, what they can plan, how effectively they organize ideas, whether sentences communicate meaning, how they use evidence, and where transcription or conventions interfere with communication.

SRSD has always placed considerable value on formative assessment and teacher judgment. Teachers use what students are doing to decide when to provide additional modeling, when to scaffold, when to differentiate, and when students are ready to take greater responsibility.

That is assessment for learning.

The problem begins when high-stakes assessment starts defining the writing curriculum.

Harris and McKeown describe how high-stakes testing can narrow both what teachers teach and how they teach it. When writing receives concentrated attention primarily in tested grades, or when instruction mirrors narrow test formats, schools risk replacing the development of writers with preparation for an event.

Testing can also encourage what looks like a shortcut: teach students a rigid template that reliably produces something scoreable.

That may help a student imitate the surface structure of a piece of writing. It does not necessarily develop the flexible knowledge and self-regulation the student needs when the purpose, audience, genre, source material, or context changes.

Assessment should therefore sample the construct we claim to value. One timed prompt cannot tell us everything important about a writer.

Students should write for different purposes and audiences. They should write in multiple genres. Sometimes they should work under time constraints, and sometimes they should have opportunities to plan, draft, receive feedback, revise, and edit. Sometimes they should write from knowledge they already possess, and sometimes they should read sources, build new knowledge, evaluate information, and write from that evidence.

If we want adaptable writers, our assessment systems should value adaptability. The goal should not be to protect students from assessments. The goal should be to make students such capable writers that an assessment becomes one more writing situation they know how to navigate.

What should schools do with the Science of Writing now?

We should resist two equally unhelpful positions.

The first is, “The research is not completely settled, so we should wait.”

The second is, “The science is settled, so here is the one correct program or sequence everyone must follow.”

Neither represents the evidence well. A better position is more demanding.

Schools should act confidently on practices supported by strong evidence while continuing to learn, measure, adapt, and change when better evidence becomes available.

That means explicit writing instruction should begin early and continue across the grades. Students should not wait until reading is “finished” before they begin serious composition because reading is never finished, and research shows that reading and writing can strengthen one another throughout development.

Writing instruction should be genuinely multicomponent. It should include transcription, sentence construction, vocabulary and language development, knowledge of genres and purposes, planning, drafting, revising, editing, reading and source use, discussion, feedback, motivation, and self-regulation. The emphasis should change according to students’ developmental needs and the writing they are trying to accomplish.

Strategy instruction should develop agency rather than compliance. Teachers should model explicitly and provide strong scaffolds, but the destination is not permanent dependence on a teacher, worksheet, mnemonic, or formula. The destination is a writer who remembers what they know, selects an appropriate strategy, adapts it, monitors what they are doing, and makes purposeful decisions independently.

Schools should intentionally connect reading and writing. Students should read to build knowledge, study how authors work, analyze models, take notes, discuss ideas, use evidence, and write from sources. They should also use writing to deepen their reading and content learning.

Foundational skills should receive explicit instruction, but always within a larger conception of composition. Spelling, handwriting, keyboarding, grammar, punctuation, vocabulary, and sentence construction matter because they help writers create and communicate meaning.

And professional learning and implementation must be considered part of the intervention itself.

A research-supported curriculum sitting on a shelf produces no student benefit.

Teachers need knowledge, materials, modeling, practice, feedback, collaboration, coaching, leadership support, time, and opportunities to revisit what they are learning. Schools and districts need to examine implementation, not just adoption.

We do not need another literacy war

Education has spent too much time turning complicated instructional questions into opposing camps: Skills or process, explicit instruction or student agency, reading or writing, foundational skills or composition, evidence or teacher judgment. These are largely false choices.

Strong writing instruction can be explicit without being rigid. Teachers can provide extensive scaffolding while still working toward student independence. Students can learn handwriting and spelling while also composing meaningful texts. Teachers can teach powerful strategies without producing formulaic writers. Schools can teach reading and writing distinctly while connecting them purposefully. Educators can rely on evidence while still using professional judgment to respond to the learners in front of them.

Karen Harris and Debra McKeown describe the damage created by these kinds of “paradigm wars.” When educators who should be natural allies divide themselves into instructional camps, students lose access to the combined knowledge those traditions can offer.

The Science of Writing should not become the next war.

It should become a way out of one.

Science asks us to look at evidence across theories, methods, populations, and disciplines. It asks us to test claims rather than protect them. It requires us to acknowledge null findings and limitations. It allows new evidence to change what we believe.

That is exactly the stance writing education needs.

Finish the literacy agenda

The next national literacy conversation should not be a contest between reading and writing.

Reading without writing leaves students primarily receiving the ideas of others. Writing without reading deprives students of language, knowledge, models, perspectives, and evidence from which to build their own ideas.

Students need both.

The Science of Writing already gives educators a strong starting point. We know that foundational production skills matter. We know language, vocabulary, and sentence development matter. We know knowledge matters. We know explicit strategy instruction can improve writing. We know self-regulation matters. We know reading and writing interact in productive ways. We know writing can deepen learning across the curriculum. We know teachers need professional learning and ongoing support. And we know implementation determines whether evidence ever reaches children.

At the same time, research must continue.

We need to understand more about transfer and maintenance. We need more research across languages, cultures, grade levels, and student populations. We need to keep investigating how the different components of writing development interact. We need strong independent replication. We need to understand what makes implementation succeed at scale, not just in individual classrooms. We need assessment systems that reflect the complexity of writing rather than narrowing it.

That is not an argument for waiting. It is an argument for acting scientifically.

We should use what the evidence already supports, examine what happens, refine our approaches, and remain willing to change.

For teachers, the message is encouraging. You do not have to choose between explicit instruction and authentic writing, between skills and meaning, or between helping students today and developing independent writers for tomorrow. The strongest evidence increasingly points toward thoughtful integration.

For school and district leaders, the message carries a different responsibility. Effective writing instruction cannot depend on a few exceptional teachers figuring it out on their own. Schools have to create the conditions that allow evidence-based instruction to become normal practice: knowledgeable teachers, coherent curricula, meaningful professional learning, coaching, leadership support, adequate instructional time, useful assessment, and sustained implementation.

And for students, this conversation has always been about far more than writing scores.

Reading gives children access to the ideas, discoveries, histories, arguments, stories, and experiences of other people.

Writing gives them the ability to do something equally powerful: develop ideas of their own and put those ideas into the world.

A complete literacy education must give every child both. The Science of Writing is already here. Now our literacy systems need to catch up.

Research and Further Reading

Durlak, J. A., & DuPre, E. P. (2008). Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation. American Journal of Community Psychology, 41, 327–350.

Fernandez, J., & Guilbert, J. (2024). Self-regulated strategy development’s effectiveness: Underlying cognitive and metacognitive mechanisms. Metacognition and Learning, 19, 1091–1135.

Graham, S. (2018). A revised Writer(s)-Within-Community model of writing. Educational Psychologist, 53, 258–279.

Graham, S., Cao, Y., Kim, Y. S. G., et al. (2025). Effective writing instruction for students in Grades 6 to 12: A best evidence meta-analysis. Reading and Writing, 38, 1–46.

Graham, S., Collins, A. A., & Ciullo, S. (2024). Evidence-based recommendations for teaching writing. Education 3–13, 52(7), 979–992.

Graham, S., Harris, K. R., & Santangelo, T. (2015). Research-based writing practices and the Common Core. The Elementary School Journal, 115, 498–522.

Graham, S., & Hebert, M. (2011). Writing to read: A meta-analysis of the impact of writing and writing instruction on reading. Harvard Educational Review, 81, 710–744.

Graham, S., Kiuhara, S. A., & MacKay, M. (2020). The effects of writing on learning in science, social studies, and mathematics. Review of Educational Research, 90, 179–226.

Graham, S., Liu, X., Aitken, A., et al. (2018). Effectiveness of literacy programs balancing reading and writing instruction. Reading Research Quarterly, 53, 279–304.

Graham, S., Liu, X., Bartlett, B., et al. (2018). Reading for writing: A meta-analysis of the impact of reading interventions on writing. Review of Educational Research, 88, 243–284.

Graham, S., & Santangelo, T. (2014). Does spelling instruction make students better spellers, readers, and writers?. Reading and Writing, 27, 1703–1743.

Harris, K. R. (2024). The Self-Regulated Strategy Development instructional model: Efficacious theoretical integration, scaling up, challenges, and future research. Educational Psychology Review, 36, Article 104.

Harris, K. R., Kim, Y. S. G., Yim, S., Camping, A., & Graham, S. (2023). Yes, they can: Developing transcription skills and oral language in tandem with SRSD instruction on close reading of science text to write informative essays at Grades 1 and 2. Contemporary Educational Psychology, 73, 102150.

Harris, K. R., & McKeown, D. (2022). Overcoming barriers and paradigm wars: Powerful evidence-based writing instruction. Theory Into Practice, 61(4), 429–442.

Hayes, J. R. (2012). Modeling and remodeling writing. Written Communication, 29, 369–388.

Jones, S., Myhill, D., & Bailey, T. (2013). Grammar for writing? An investigation of contextualised grammar teaching. Reading and Writing, 26, 1241–1263.

Kim, Y. S. G., & Graham, S. (2022). Expanding the Direct and Indirect Effects Model of Writing (DIEW). Journal of Educational Psychology, 114, 215–238.

Kim, Y. S. G., Harris, K. R., Goldstone, R., Camping, A., & Graham, S. (2025). The Science of Teaching Reading is Incomplete without the Science of Writing. Scientific Studies of Reading, 29, 32–54.

Kim, Y. S. G., Wolters, A., & Lee, J. W. (2024). Reading and writing relations are not uniform: They differ by the linguistic grain size, developmental phase, and measurement. Review of Educational Research, 94(3), 311–342.

Kim, Y. S. G., Yang, D., Reyes, M., & Connor, C. (2021). Writing instruction improves students’ writing skills differentially depending on focal instruction and children. Educational Research Review, 34, 100408.

Kraft, M. A., Blazar, D., & Hogan, D. (2018). The effect of teacher coaching on instruction and achievement. Review of Educational Research, 88, 547–588.

McKeown, D., Brindle, M., Harris, K. R., et al. (2023). Practice-based professional development and expert support for classwide SRSD writing instruction. Contemporary Educational Psychology, 73, 102152.

Salas, N., Ventura, J., & Llauradó, A. (2026). Long-term effects of an SRSD writing intervention for elementary school children. Reading and Writing, 39, 817–844.

Saddler, B., & Graham, S. (2005). The effects of peer-assisted sentence-combining instruction on the writing performance of more and less skilled young writers. Journal of Educational Psychology, 97, 43–54.

Santangelo, T., & Graham, S. (2016). A comprehensive meta-analysis of handwriting instruction. Educational Psychology Review, 28, 225–265.

About the Author

Randy Barth is CEO of SRSD Online, which innovates evidence-based writing instruction grounded in the Science of Writing for educators. Randy is dedicated to preserving the legacies of SRSD creator Karen Harris and renowned writing researcher Steve Graham to make SRSD a standard practice in today’s classrooms. For more information on SRSD, schedule a risk-free consultation with Randy using this link:  Schedule a time to talk SRSD.