The science of reading has won most of the policy arguments of the past five years. Forty-plus states have passed laws built on it, and the U.S. House just voted to tie federal literacy money to it. Yet as we reported this summer, the practice gap is stubborn: trained teachers, new materials, and flat results in too many classrooms. A new paper offers a useful name for the missing piece. Writing in the journal Mind, Brain, and Education, Elizabeth Tipton, a statistician at Northwestern University, and Nicole Patton-Terry, director of Florida State University's Florida Center for Reading Research, argue that the field needs an "engineering of reading," and Education Week brought the idea to a wider audience on September 16. It's a framing worth a teacher's ten minutes.

What is the "engineering of reading"?

Science tells you what is true in general; engineering makes it work in a particular place. A physicist can tell you how loads distribute; an engineer designs the bridge for this river, this soil, this budget. Tipton and Patton-Terry argue that reading has plenty of the first and very little of the second. Research establishes that explicit, systematic phonics instruction helps children learn to read. It says much less about how many minutes a day, in what order, with what grouping, for a class where a third of students are absent on Mondays and six are learning English. "There's this middle space," Tipton told Education Week, and she described engineering as work focused on curriculum design: the unglamorous layer between findings and Friday's lesson plan.

Why isn't the science enough on its own?

Because classrooms vary in ways studies smooth over. The authors and Education Week point to rural versus urban settings, attendance problems, adolescent literacy inside content classes, behavior needs, and schedules crowded by other requirements. A practice that works in a controlled study can fail in a school where it runs for 12 minutes instead of 30. There's a market problem too. Many commercial curricula have never been tested head-to-head against alternatives, and Tipton described a split market in which some technology is built without close attention to recent reading science. A district can buy a product labeled "science of reading" with no evidence the product works as packaged.

Does engineering-style work already exist?

In pockets, yes. Education Week cites Bill Penuel's collaboratory at the University of Colorado Boulder, the Strategic Education Research Partnership, research-practice partnerships such as a program called Read Up in a Florida district, and the federally funded Regional Educational Laboratories. That last example is a warning as much as a model: the recent cancellation of those lab contracts removed a piece of the infrastructure states used for exactly this work. The authors' larger proposal is to build a field of "learning engineers," people trained in design and iteration who can adapt evidence to real conditions and test the adaptations.

How can a teacher think like an engineer this month?

You don't need a research grant. The core engineering move is a short, disciplined cycle, and it's how skilled interventionists already work with dyslexic students.

  1. Name the problem narrowly. Not "my group struggles with reading" but "four students can blend CVC words but stall on consonant blends."
  2. Pick one change. Add five minutes of blend practice with letter tiles, or move the decodable reading before the phonics drill. One variable at a time.
  3. Decide how you'll know. A one-minute check of 10 blend words every Friday is enough. Write down the baseline first.
  4. Run it for two to three weeks. Shorter than that, noise wins. Much longer, and a failing change costs students time.
  5. Keep, adjust, or drop. Then share what you learned with your grade-level team. That's the engineering loop, scaled to one classroom.

What should school and district leaders ask?

The engineering frame also changes the questions leaders bring to curriculum and schedule decisions. Instead of "Is this program aligned?", ask "Under what conditions has it worked, and do our conditions match?" Instead of "Did teachers complete the training?", ask "What changed in lessons afterward, and how do we know?" And instead of adopting a full program at once, consider piloting it in a few classrooms with a clear measure and a date to decide. That's slower in the first semester and much faster over three years, because you learn what actually works in your buildings before you scale it to every one.

What should schools avoid?

  • Treating the label as evidence. "Aligned with the science of reading" is a marketing claim until someone shows the product's results. Ask vendors for efficacy studies, not alignment charts.
  • Cutting the dose to fit the schedule. A 30-minute routine squeezed into 12 is a different intervention. If time must shrink, decide which component to keep.
  • Changing everything at once. New curriculum, new screener, and new grouping in the same month make it impossible to tell what helped.
  • Letting fidelity become rigidity. Engineering means adapting on purpose and measuring the result, not abandoning the program when a lesson feels slow.

What does this mean for students with dyslexia?

Dyslexic students feel implementation gaps first. A general classroom can absorb a weak phonics block; a child with a phonological processing difference cannot. Our intervention guide describes what good intervention looks like: explicit, systematic, cumulative, with frequent progress checks that drive adjustments. That last part is the engineering loop, applied to one child. Parents can use the same lens: if your child's intervention has run for eight weeks, it's fair to ask what was measured, what changed, and what the team will adjust if progress is flat. The paper's argument, in effect, is that whole systems should work the way the best dyslexia tutors already do. Teachers looking for the knowledge base to run those cycles can start with our free training guide.

Frequently asked questions

What does "engineering of reading" mean? It is a term from a 2026 paper by researchers Elizabeth Tipton and Nicole Patton-Terry for the design work between reading research and classroom practice. It covers how to adapt proven methods to real schedules, students, and schools, then test whether the adaptation works. The goal is closing the gap between the science and results.

Is the science of reading wrong, then? No. The authors aren't disputing the research on how children learn to read. They argue that research alone doesn't tell schools how to implement it well in varied conditions. Explicit, systematic instruction remains the foundation; engineering is about delivering it effectively in a specific classroom or district.

How can I tell if a reading program has real evidence? Ask the vendor for independent efficacy studies comparing students who used the program with similar students who did not, including results for struggling readers. Check whether the study conditions resemble your school. An "alignment" rubric shows the program covers the right content. It doesn't show that students learn more from it.

Source: "The Science of Reading Has an Implementation Problem. Could Engineering Help?" (Education Week, Sep 16, 2026) · Tipton & Patton-Terry, "We Need an 'Engineering of Reading': Why the 'Science of Reading' May Not Be Enough" (Mind, Brain, and Education, 2026)