Educators and researchers are looking to update one of the oldest, most popular—and at times one of the most controversial—methods of targeting instruction: the elementary reading circle.
Grouping students of similar reading skills—think “bluebirds” or “redbirds,” for example—has become ubiquitous in American classrooms as a way to target instruction to students’ learning needs, spreading from 68 percent of classrooms in 1992 to more than 90 percent by 2015. But evidence suggests that the practice may be less beneficial than teachers think: It can exacerbate achievement gaps and even slow reading growth for some children unless the groups are fluid and focused on skills rather than overall achievement.
The spread of modern ability grouping is likely in response to growing pressures to raise test scores under the No Child Left Behind Act’s accountability system, said Adam Gamoran, the president of the William T. Grant Foundation and a longtime researcher of ability-grouping strategies. “Many people believe it is possible to use ability grouping as differentiated instruction to maximize achievement growth,” he said. “It often doesn’t work out that way in practice.”
Early grades are particularly likely to group students by ability, because the typical bell curve in a kindergarten or 1st grade classroom is so wide.
In one forthcoming study, Marshall Jean, a research fellow at the Northwestern University Institute for Policy Research, tracked nearly 12,000 students from kindergarten through 3rd grade in more than 2,100 schools, following them through high, middle, and low reading groups or ungrouped reading classes.
He found about half of children who were in the lowest reading group in kindergarten were able to improve to at least the median group by the end of 1st grade. By the end of 3rd grade, 46 percent of those who had previously been in the lowest group in 2nd grade were able to move up. However, Jean found that none of the students initially placed in the lowest kindergarten group ever caught up to the reading level of their classmates who had started out in the highest reading group.
“The structural inertia is considerable,” Jean noted, finding that having been in the highest reading group in an earlier grade tended to protect students from being put in a lower group later, even with significantly lower scores. Students in lower reading groups not only progressed more slowly academically, but while they were in lower reading groups, they were also slower to develop “learning behaviors,” such as varied interests, concentration on tasks, and persistence in the face of difficulty. Those behaviors, in turn, reduced the students’ likelihood to move up to higher reading groups in later grades.
“If you are more motivated and the teacher perceives that about you, you are more likely to be put into a higher reading group,” Jean said. “But there was also some evidence for bias: Even after controlling for prior reading achievement and learning behaviors, students in poverty were more likely to be assigned to lower groups, and their wealthier peers more likely to be tapped for higher reading groups. They were small effects, but they are there and consistent across grade levels and statistically significant.”
Similarly, in a series of three new studies in Switzerland, researchers asked practicing teachers and college students to evaluate profiles of students whose scores put them on the borderline of more or less academically rigorous tracks in high school; the students’ achievement scores were held constant but their backgrounds were altered to make them appear to be either high- or low-income. Over multiple studies, recently published online in the journal Personality and Social Psychology Bulletin, both student and practicing teachers were more likely to refer lower-income students to a lower academic track and higher-income students to a more challenging track, even though their scores were the same.
“Because of inequality outside of schools, children from different socioeconomic and racial and ethnic backgrounds often come to school with different levels of preparation. And so by separating the children by their initial reading ability, the teachers are also separating [them] by socioeconomic status or race or ethnicity,” said Gamoran of the William T. Grant Foundation. “And, of course, when teachers have low expectations for their weaker readers, they slow down the pace even more than they would need to, so the low-achieving students fall further and further behind instead of catching up.”
Besser Elementary School in Alpena, Mich., switched to ability grouping in its early-reading classrooms about three years ago. It’s not clear yet how well the practice is working. About half the school’s students live in poverty, and their achievement gap with higher-income students has stayed stubbornly wide.
“We were focused on making instruction more meaningful for all students. Teachers need to focus on struggling students, but on the other end of the continuum where students needed to be enriched, those students were being left behind,” said Eric Cardwell, the principal of Besser Elementary. “The challenge teachers have seen now is they’re having to plan for three to four different groups.”
Those high-achieving students have improved, he said, but the groups themselves have remained more stable than he’d like.
“What we frequently see is slight movement of students. You don’t generally see them jumping two levels at a time when we only do data reviews three times a year,” Cardwell said. “Ideally, there would be more [reviews] so that there would be more fluidity, but time is always the monster that’s chasing you: time to review data, time to plan.”
Internationally, the Organization for Economic Cooperation and Development found that countries that predominantly use ability grouping showed significantly deeper performance inequality on the Program for International Student Assessment but no significant benefits for the countries’ overall performance. OECD noted that more than 9 in 10 U.S. 15-year-olds attend schools where they are grouped by ability.
“What we know now that we didn’t know back in the ‘80s is that when you group up students, it has to be specifically relative to the content that’s going to be taught,” Gamoran said. “There’re no IQ tests, not even a general reading-ability test that can tell you how to form the groups so that you can meet their needs. You have to form the groups specific to the instruction that’s coming and then reassess to form new groups specific to the next instructional unit.”
Changing the Calculation
One California program has shown promise in making reading circles more flexible and less stigmatizing. In Assessment to Instruction, or A2I, teachers give a diagnostic assessment to all students every eight weeks to identify strengths and weaknesses in particular reading skills in four areas of literacy: decoding, fluency, comprehension, and usage. An algorithm based on the assessment tells teachers how much individual, small-group, and independent working time each student needs, and students are grouped for instruction based on particular focus skills rather than overall reading ability.
“What we’ve discovered is that it’s fine to have a group of students of different levels, as long as they all are working on the same learning needs,” said Carol Connor, an education professor at the University of California, Irvine, who developed the program. “You can have students of different reading abilities who all need to work on decoding. ... What doesn’t work is if you put your kids who already know how to code in a group to learn how to code, again. You receive more behavior problems because they’re really bored, ... and our research suggests that it has a negative effect on their growth.”
Phoenix Collegiate Academy (now ASU Prep) in Arizona was one of the schools that piloted the A2I program, and Amanda Jacobs, then-principal, said it changed the way teachers and administrators approached differentiating instruction in small groups. Previously, teachers focused on providing equal time with each small group, but “it shifts your perspective from trying to get to every kid in the time you have to being more strategic with how you’re spending your minutes with each child.”
In a recent longitudinal, randomized controlled study, students who participated in the targeted reading groups over three years performed significantly higher than students in a control group that used standard reading classes. Though 45 percent of the students in the targeted reading groups came from a low-income background, by 3rd grade, all of them had higher reading scores than the national average for their grade, and none had scores below the expectations for their grade level.
“There are no ‘bluebirds’ being the bluebirds all year long,” Connor said.
Research Analyst Alex Harwin and Librarian Holly Peele contributed to this article.
A version of this article appeared in the August 29, 2018 edition of Education Week as Doubts Cast on Ability-Based Reading Groups