That may sound reasonable, but Shanahan says it doesn’t help anyone and even leads teachers to abandon reading altogether. “In social studies and science, and these days even in English classes,” he said in an interview, “teachers either don’t assign any readings or they read the texts to the students.” Struggling readers are not given the chance—or the tools—to tackle complex material on their own.
Instead, Shanahan believes that all students should read grade-level texts together, with teachers providing more support to those who need it.
“What I recommend is a differentiation of instruction,” he said in our interview. “Everyone will have the same learning goal—we’ll all learn to read the fourth-grade text. I might teach a lesson to the whole class and then let some kids move on to independent work while others get more help. Maybe those who didn’t understand read the text again with my support. By the end, more students will have met the learning goal—and the whole class can tackle another text tomorrow.”
27 different ways
Shanahan’s approach isn’t about throwing kids into the abyss without help. His book outlines a toolkit of strategies for tackling difficult texts, such as looking up unfamiliar vocabulary, rereading confusing passages, or breaking up long sentences. “You can go about successful reading in 27 different ways,” he said, and he hopes future researchers will discover many more.
He is skeptical of teaching students skills such as identifying the main idea or making inferences. “We treated test questions like a skill,” he said. “That’s not working.”
There is widespread disenchantment with declining American reading achievementespecially among high school students. (Thirty-nine percent of eighth-graders can’t reach the lowest of three achievement levels called “basic,” according to the National Assessment of Educational Progress.) But there is little agreement among reading advocates on how to fix the problem. Some argue that what children need most is more knowledge to grasp unfamiliar ideas in a new reading passage, but Shanahan argues that background knowledge will not be sufficient or as powerful as explicit comprehension instructions. Other reading experts agree. Noni Lesso, dean of the Harvard Graduate School of Education, who specializes in literacy in her own academic work, supported Shanahan’s argument in October 2025 online discussion about the new book.
Shanahan is most persuasive, pointing out that there is no strong experimental evidence to show that reading achievement increases more when students read text at their individual level. In contrast, a 2024 analysis found that the most effective schools are the ones that support grade-level learning. Still, Shanahan acknowledges that more research is needed to determine which comprehension strategies work best for which students and under what circumstances.
Misunderstanding Vygotsky
Teachers often cite Russian psychologist Lev Vygotsky’s “zone of proximal development” to justify giving students books that are neither too easy nor too difficult. But Shanahan says this is a misunderstanding of Vygotsky’s work.
Vygotsky believed that teachers should guide students to learn challenging things they cannot yet do on their own, he said.
He offers an analogy: a mother teaching her child to tie his shoes. In the beginning, she demonstrates while narrating the steps out loud. The child then takes one step and she completes the rest. Over time, the mother gradually loosens control and the child just ties a bow. “Reading at a level,” Shanahan said, “is like saying, ‘Why don’t we just get Velcro?’ This is about real teaching. “Boys and girls, you don’t know how to ride this bike yet, but I’ll make sure you do by the time we’re done.” “
Shanahan’s critique of reading instruction mainly applies from second grade onwards, after children learn to read and focus on reading comprehension. In kindergarten and first grade, when children are still learning phonics and how to decode words on the page, the research evidence against teaching in small groups with different level texts is not as strong, he said.
It is important to learn to read first – the decoding. Shanahan says there are rare exceptions to teaching all children at grade level.
“If a fifth-grader can’t read yet,” Shanahan said, “I wouldn’t make that kid read a fifth-grade text.” This child may need separate instruction from a reading specialist.
Advanced readers, meanwhile, can be challenged in other ways, Shanahan suggests, through independent reading time, jumping ahead to higher-grade reading classes, or exploring complex ideas within grade-level texts.
The role of AI — and parents
Artificial intelligence is increasingly used to rewrite texts for different levels of difficulty. Shanahan is skeptical of that approach. Simpler texts, whether written by humans or generated by AI, don’t teach students to improve their reading ability, he argues.
Still, he’s intrigued by the idea of using AI to help students “climb the ladder” by instantly modifying a single text to a range of reading levels, say third-, fifth-, and seventh-grade levels, and having students read them in rapid succession. Whether this increases comprehension is not yet known and needs to be investigated.
AI may be most useful for teachers, Shanahan suspects, to help them target a sentence or passage that tends to confuse students or trip them up. The teacher can then address these common difficulties in the classroom.
Shanahan worries about what’s happening outside of school: Kids aren’t reading much at all.
He urges parents to let kids read whatever they like — whether it’s above or below their level — but set consistent expectations. “Nagging may not be effective,” he said. “But you can be specific: “After dinner on Thursday, read the first chapter. When you’re done, we’ll talk about it and then you can play a computer game or go on your phone.”
Too often, he says, parents give in when children resist. “They’re the kids. We’re the adults,” Shanahan said. “We are responsible. Let’s step up and do what’s right for them.”
