DANVILLE, Pa. — Four years ago, when the staff at Danville Primary School found out they were going to learn a new way to teach reading, Mary Levitski thought: Here we go again.

Levitski, who had taught at the central Pennsylvania school district for 25 years, was a good teacher, but she was disappointed that she couldn’t get through to all kids. Every time Danville switched curricula, a new publisher promised that materials would help the outliers, with research-based methods that would unlock the key to literacy.

The 2015 training was different. Inspired by a tutoring center for kids with dyslexia in nearby Bloomsburg, Danville adopted a new approach that involved training every teacher using a somewhat old-fashioned method. Instead of buying glossy texts, it made its own workbooks.

And it worked. Danville’s method relies on new reading science. It has roots in an old way of teaching but is based on new cognitive neuroscience research that has revealed how brains process sounds and symbols. It borrows from linguistics, the study of language and its structure. Students do not memorize lists of words for spelling tests, yet the average Danville fourth grader is spelling at the sixth-grade level.

The transition in Danville and some other Pennsylvania districts hasn’t been easy. Some teachers quit. In one district, teachers said they were bullied by their peers who questioned the wisdom of using phonics in the classroom.

Still, Pennsylvania’s adoption of science-backed reading has caught fire among some Washington educators.

“Teachers are hungry” for this information, said Julie Bedell, president of the new nonprofit ReadWA and a second-grade teacher at Seattle’s Coe Elementary, in an email. A new state law requires Washington districts to begin screening for learning weaknesses associated with dyslexia, and for teachers to emphasize four specific skills when teaching reading. Educators believe that will foster instructional changes.


National reading scores released in November highlight the problem. In Washington, only 35% of fourth graders were reading at or above proficiency, similar to the national average. The numbers were worse for low-income students and students of color; only 23% of black students and 23% of Hispanic students hit proficiency targets.

Overall, Pennsylvania’s statewide scores were only a little better than Washington’s; the new reading science concepts are not used widely enough to move Pennsylvania’s averages.

Dozens of Pennsylvania educators interviewed — in the four districts the Times visited for this story — believe that changing the practice of teaching reading can reverse decades of mediocre scores. And the problems start with teacher training. “We still are not equipping our teachers, in this day and age, to teach the science of reading,” said Dawn Brookhart, director of curriculum, instruction and technology for Danville schools. “It’s an outrage.”

The results of the new method are promising.

Aside from Danville, in eight Pennsylvania districts that participated in a pilot program to teach the new reading science, between 74% and 86% of students were on track to meet grade-level reading goals; in comparison schools, between 66% and 85% were on track.

That might not sound like a major difference, yet it’s been so hard to move the needle on reading that state officials saw it as a significant improvement, and expanded the pilot to 29 additional districts.


In a Pennsylvania Department of Education report summarizing the work, educators wrote: “Some of the most notable results from this Pilot cannot be measured — the confidence and empowerment achieved by both the teachers and the students.”

“It doesn’t have to be boring”

Researchers say we’re wired from birth to learn spoken language — it flows so naturally that you probably can’t remember when you couldn’t speak or understand others. But our brains are not wired in the same way to decode abstract symbols on a page and turn them into sounds. Moreover, English is complex, with arcane rules and exceptions.

Students who can’t read fluently may stumble for years because school is built on the idea that after third grade, everyone can read; the work shifts from learning to read to reading to learn. A student who can’t read well will fall further behind on all subjects — science, history, even math.

Reading experts say between 30% and 60% of kids learn to read with just a bit of help, picking it up almost as naturally as they speak. But many need more.

Science-of-reading advocates say it’s important to slow-walk the teaching by helping kids understand the sounds letters make, separately and together, and how those sounds can change depending on word construction. Their method is systematic, and is based on phonemic awareness and phonics, although it is not solely phonics.

Before every Danville teacher learned the new method, teachers were free to teach reading however they wanted using “balanced-literacy” materials, said Danville Primary principal John Bickhart. Balanced literacy meant to connect two schools of thought: whole language (using literature and context) and phonics (sounding out words).


But, Bickhart said, the curriculum offered no framework for teaching phonics.

Today, that framework is explicit. Each classroom wall has the same letter line graphics, and identical hints about sounding out and spelling words.

One day in October, Levitski — the veteran Danville teacher — and reading specialist Erin Marshman taught blends, how two consonants come together to make a new sound.

Donning chef’s hats and aprons, they made a theatrical performance of putting two cards with the letters “s” and “l” into a bowl. Levitski stirred with a spoon, then produced a single card with the letters written together. To an adult, it looked like a simple sleight-of-hand, but the 6-year-olds were convinced their teachers had performed magic.

In a singsong voice, the two women spoke the rule. The students repeated: A blend is two consonants that come together and make their own sound.

There’s a lot of repetition, Levitski said — kids need that. “I sing. We do monster voices. It doesn’t have to be boring.”


As kids progress, they’ll discover words’ Greek and Latin roots, and learn how to pronounce, spell and predict the meaning of a word they’ve never seen before based on its morphology, or form.

Danville, an old mining town on the banks of the Susquehanna River, had the money to retrain all teachers because it is well-funded — it’s the home of Geisinger Health System, a network of primary care hospitals and clinics. It has almost no middle-class; half of its students are affluent, half qualify for free lunch.

Danville revamped its reading instruction by taking ideas from Orton-Gillingham, a method developed to tutor students with dyslexia. As reading scores jumped and others noticed, the district started a training program for teachers around the region. Today, the Danville Reading and Dyslexia Academy trains hundreds of them.

Closing the gap

Explicit phonics worked in Danville, a mostly white district. It’s also working in a more diverse district with higher poverty rates.

Stacey Cherny is the principal of South Side Elementary in Harrisburg, where 76% of 870 kids qualify for free- and reduced-price lunch. A former teacher, Cherny wanted to bring the science of reading into her school’s instruction.

South Side teachers weren’t giving effective reading lessons, and they weren’t using reading tests effectively, Cherny said. Many teachers dutifully pumped scores into the computer, then ignored their results.


The school trained all teachers on Language Essentials for Teachers of Reading and Spelling (LETRS), an intensive program that uses explicit phonics. It’s not quite Orton-Gillingham, but shares many ideas.

Some teachers said they were bullied for embracing phonics. Others quit. But eventually, the new instruction took hold. Today, Cherny says, South Side Elementary is also data-driven: It watches for individual or class-wide weaknesses, and addresses them early.

Did it work? One year, a test showed one particular class of first graders didn’t understand blends — the concept Levitski and Marshman taught at Danville.

Cherny asked a school reading specialist for help. The regular teacher observed, while the specialist modeled good instruction. Modeling is a mainstay of how teaching reading evolved at South Side.

By the end of the school year in 2018, 84% of South Side’s students were proficient at reading, a 19 percentage-point gain from fall scores. Central Dauphin’s seven lowest-income elementaries had reading proficiency scores close to, and in some cases exceeding, those in the six highest-income elementaries.

“We’re starting to close the achievement gap,” Cherny said.

Statewide literacy expert Pamela Kastner believes that using the science of reading — and teachers’ working collectively to improve — can lead to a relatively quick turnaround. Kastner directs statewide literacy for the Pennsylvania Training and Technical Assistance Network, which provides free LETRS training. Pennsylvania has the most LETRS-trained teachers of any state, Kastner said.


Susie Macik grew up as a farm kid and made it through five grades without learning how to read. She faked it by memorizing words, copying students’ papers. Even to this day, “I hate to read,” she said. In college, Macik grew to realize how important education is, and she became an educator to help children who struggled in school like she did.

When Macik became principal of Bentworth Elementary in rural southwest Pennsylvania, the school followed a well-known publisher’s balanced-literacy curriculum. Macik reasoned it would teach all kids to read. But the results upset her.

In 2014, Bentworth was one of the eight districts selected for the reading pilot. Macik announced the win on the loudspeaker; she could hear teachers cheering.

Today, most expensive balanced-literacy materials have been shelved. Bentworth built its own curriculum, with backup from a phonics curriculum, Wilson Reading.

At these and other Pennsylvania districts, educators emphasize that teachers aren’t at fault. “They’re not being taught the science,” said Kastner. “Many times they’re being taught the exact opposite.”

By that she means that not only does balanced literacy skimp on phonics, it makes it harder for students to learn because it offers a shortcut strategy to reading through guessing.


Not a quick fix

Macik jokes that if she returned to elementary school today, she’d only know as much about reading as her first-grade students, who understand terms like digraph (two letters representing one sound), macron (a mark used to indicate a long vowel) and breve (the mark for a short vowel).

Pennsylvania’s reading experts emphasize that science-based reading instruction isn’t a quick fix, and that there’s no one program alone that can solve the problem. In all the districts that have seen success, it took years for teachers to master the skills, Kastner said.

Another key element: consistent leadership. Cherny has been principal of South Side for seven years, Macik at Bentworth for 10, and Brookhart at Danville for 11.

Kids fail to learn to read for many different reasons. It’s important to know why, then match them with the specific fixes. Successful school leaders create a system where these things happen, Kastner said, and don’t try to solve the problems by buying yet another program — “a shiny new thing that will fix it.”

What they need to fix, she says, is the core.