Editor’s note: This was first published on the author's Substack, Minding the Gap. 

Adopting a truly knowledge-building elementary curriculum—one that is rich in content, including topics in social studies and science—is crucial for improving students’ education outcomes. But no matter how well-constructed a curriculum is, it’s possible for districts and schools to implement it in a way that doesn’t work well. And that happens surprisingly often.

A new study from SRI analyzed reading instruction in four large districts that had been using effective knowledge-building curricula for several years, long enough to figure out how to implement a curriculum well. And yet the researchers found that most reading comprehension instruction—about two-thirds of the lessons observed—supported “surface-level” rather than “robust” understanding.

According to the definitions the researchers used, “surface-level” comprehension instruction focuses on “completing tasks”—often the kinds of tasks laid out in state literacy standards, like describing the setting in a story or explaining how an author uses evidence in an informational text.

“Robust” comprehension instruction, in contrast, might have students engage in some of those tasks, but the focus is on the overall meaning of the text. In other words, instead of using a text to teach a skill, a lesson that leads to robust comprehension puts the text in the foreground, bringing in whatever skills might help students understand and analyze that text.

As one illustration of the difference between surface-level and robust instruction, the study describes two third-grade lessons on the same book, One Giant Leap, about Neil Armstrong’s moon landing. In one classroom, the teacher guided students to identify literal and nonliteral language in the text—a skill matching a literacy standard—and stopped there. In the other, the teacher not only helped students identify literal and nonliteral language but also asked them what Armstrong meant when he described his first step on the moon as “one giant leap for mankind.”

The researchers observed comprehension lessons in 111 classrooms, nearly all of them third- to fifth-grade, focusing on a group of representative schools in each district. The study, conducted during the 2024–25 school year, also involved teacher surveys; interviews with teachers, district and school leaders, and instructional coaches; and observations of professional development sessions.

The study is “methodologically one of the strongest observational studies” of comprehension instruction in the last 50 years, according to the report, and its findings should be applicable nationwide. Those findings also echo a recent meta-analysis of observational studies of comprehension instruction done between 1980 and 2023, which found a similar gap between scientific research on comprehension and classroom practice.

Why so much surface-level instruction?

An obvious question is why teachers are focusing on “tasks” rather than the meaning of texts, even when using curricula with rich content. Are those curricula oriented around skills-focused standards, or are teachers changing the lessons to make them skills-focused?

The SRI study wasn’t designed to address that question. The researchers are planning to look into it more deeply in the future.

But evidence from the study—along with observations by the researchers and by an education consultant who has worked with numerous districts on implementing knowledge-building curricula—points to a likely answer: Many teachers are modifying content-rich curricula to put skills-based standards in the foreground rather than content.

The curricula used by the districts in the study were Core Knowledge Language Arts (CKLA), Wit & Wisdom, and EL Education, all of which have been rated highly by the Knowledge Matters Campaign.[i] (I serve on the board of the Knowledge Matters Campaign’s parent organization, a nonprofit.)

One finding from the study is that the curriculum a district was using didn’t determine how many comprehension lessons were surface-level rather than robust. Two of the districts used CKLA, and one of them had the lowest percentage of robust lessons, 7 percent, while the other had the second highest, 30 percent. That suggests that implementation rather than curriculum is responsible for the dearth of robust comprehension instruction.

And according to a companion report from SRI, 47 percent of teachers said they modified half or more of a comprehension lesson, compared to only 13 percent who said they did so with a lesson in foundational reading skills like phonics. That report also found that many teachers viewed the comprehension aspect of their curriculum as more difficult to implement, saying the texts were “too dense” and complex for their students.

Kristen McQuillan, a consultant who has worked with districts implementing knowledge-building curricula, says that from her observations, it’s clear the curricula themselves aren’t the reason so many lessons are surface-level. In many cases, she says, “teachers are changing the curriculum in pursuit of how to help kids master the standards.”

“Unpacking the standards”

Teachers aren’t necessarily doing this on their own initiative. McQuillan has seen a “system-level confusion about how to interpret and use standards, which can show up in district expectations, coaching practices, and even sometimes in professional learning sessions.” While some PL providers are working to keep the focus on texts and meaning, she said, others will “unpack the standards and talk about how to remediate where kids are struggling.”

She gave the example of a professional learning provider advising teachers on how to modify a lesson on Don Quixote in CKLA. The lesson plan in the curriculum might list as one of the objectives analyzing Don Quixote’s character by looking at his actions in the chapter. That’s fine, McQuillan said.

But, she continued, “the provider might say, ‘this is only partially aligned to the standard.’” The provider might then urge teachers to make the objective of the lesson “for kids to master the skill of identifying character traits.”

One problem is that identifying character traits isn’t a skill you can “master.” It will be easier or harder depending on the particular text. Another, McQuillan said, is that kids can get confused if the lesson objective is too “meta.” And if the focus is on the general topic of identifying character traits, “they don’t even have to talk about the text itself.”

Similarly, the SRI researchers observed a professional learning meeting in which teachers took a CKLA lesson about art patronage by merchants and clergy during the Renaissance and turned it into one focused primarily on the task of creating a Venn diagram.

Tests and “data-driven instruction”

Educators often identify the standards they feel instruction should focus on based on “benchmark” testing data—standardized tests given periodically that are supposed to predict how students will perform on end-of-year state reading tests.

Like the state assessments, these tests present students with reading passages unrelated to anything they might have learned through the curriculum and provide teachers with data framed entirely in terms of the skills in literacy standards. They might tell teachers that students are weak in a skill like identifying the main idea of a text, when in fact the student might not have understood the text because of a lack of relevant background knowledge.

Schools often pride themselves on this kind of “data-driven instruction,” or DDI. But data-driven instruction makes sense only if the data are telling you something meaningful.

“While data-driven instruction is generally a good idea,” the SRI report notes, “it also can lead to unhelpful practices that reinforce surface-level instruction.”

McQuillan said that data from standardized tests should drive some kinds of instruction but not others. “You can say a student has ‘mastered’ short vowel sounds or math facts. But anything conceptual, that’s going to send you down the wrong path.” She added, “There are outdated DDI protocols that have been updated in some circles, but not everyone has gotten the message.”

How to encourage robust instruction

What can districts and schools do to help ensure that instruction fosters robust comprehension? The SRI report identifies several steps, including deepening teachers’ content knowledge by having them dig into the texts their students will be reading.

Knowledge-building curricula often cover topics that teachers might have forgotten or might never have learned about themselves, like the evolutionary adaptations of frogs or the War of 1812. It’s hard to guide students in mining a text for deep meaning when you’re unsure of that meaning yourself.

The SRI report also recommends having teachers observe model lessons, perhaps through videos. In addition, it suggests ensuring that instructional coaches understand the content and the texts in the curriculum and have the pedagogical knowledge to guide teachers to effective instruction.

McQuillan has observed that districts often focus so much professional learning on foundational skills instruction that there’s little time to cover reading comprehension. Ideally, she said, when PL does focus on comprehension, it will begin with the text students will be reading and anticipate issues for which the curriculum itself might not provide guidance.

Teachers should read the text themselves, she said, and discuss “what will be tricky in it for kids, what do kids stand to gain from reading this text?” Then they can talk about how to pace the lesson for their particular students and how to address their confusion or misunderstandings. That kind of support could allay teachers’ concerns that some texts are too difficult for their students.

Standards-based comprehension tests

McQuillan also identifies standards-based assessment of comprehension skills as a key obstacle to robust instruction in that area. “Assessments aren’t the only thing,” she said, “but they’re one big contributing factor.”

One problem is that standardized assessments don’t provide much information beyond a general—and fairly meaningless—pronouncement such as “this student is not on grade level for making inferences.” McQuillan said the data should tell teachers whether a student is having trouble with basic decoding or reading fluency—or comprehension.

If it’s comprehension, it would be helpful for teachers to know how complex the text was and if certain kinds of text are trickier for a particular student than others. Maybe a child does fine with narrative but struggles with informational text.

A more fundamental problem is that standardized reading tests purport to measure mastery of the standards in the abstract, which is virtually impossible. But all knowledge-building curricula come with their own assessments that measure whether students have retained information from the curriculum and how well they can reason about or analyze it.

In some districts, leaders and principals advise teachers to pay attention to those assessments rather than the standardized ones, McQuillan said. That emphasis leads to instruction that focuses on the meaning of texts rather than on standards-related tasks.

content-rich curriculum is still crucial

McQuillan says there are districts that are implementing content-rich curricula in a way that guides students to deeper understanding—and the SRI study did find that about a quarter of lessons in the districts they studied were accomplishing that. At the same time, though, most districts in the U.S. are still using curricula designed to put literacy standards in the foreground rather than any particular text or content. As a result, the content in those curricula is generally quite thin.

I’ve spoken with educators using such curricula who are trying to do the opposite of what seems to be happening in many districts using content-rich curricula: They’re modifying standards-focused lesson plans so that they center on texts and content, which usually requires supplementing the curriculum with additional texts. But given the pressures to “teach the standards,” even in districts with content-rich curriculum, it’s likely that most teachers in districts with standards-focused curricula are engaging only in surface-level instruction.

So while content-rich, knowledge-building curriculum doesn’t guarantee robust instruction, it’s crucial to enabling that kind of instruction to happen. Professional learning is also vital, for coaches and administrators as well as teachers. But, as McQuillan noted, “you have to have curriculum to ground the professional learning in.”

And as the SRI report observes, students—and their teachers—appear to enjoy robust lessons more than surface-level ones. “Helping teachers engineer instruction that facilitates more robust understanding of texts,” the researchers conclude, “might not only improve students’ comprehension achievement but also create more joyful classrooms.”