What summer does to reading.

Every fall, something predictable happens in American schools. Teachers administer reading assessments. Students from higher-income families return from summer having held steady or improved their reading scores. Students from lower-income families return having lost ground — on average, two to two and a half months of reading achievement, measured against the previous spring.

This pattern has been documented since at least 1978, when Barbara Heyns published research showing that school achievement gaps between higher- and lower-income students grew more over summer than over the school year itself. The finding has been replicated so consistently that it has a name: summer reading loss, or sometimes summer slide.

The question researchers kept asking was whether it was really about books.

Higher-income families have more books at home. They visit libraries more often. Children who read over summer return in fall having practiced; children who don't read return having lost fluency they had spent a school year building. If the mechanism is access — to books, to reading material, to the simple act of choosing something to read — then the gap should be addressable through access. Give children books. See what happens.

The experiment

Richard Allington at the University of Tennessee and colleagues ran randomized trials across three cities over multiple consecutive summers to test this directly.

The intervention was deliberately minimal. At the end of the school year, low-income elementary students visited a site with approximately 400 books and chose 12 they wanted to take home for summer. No summer school. No reading log. No requirement that they finish the books. No adult oversight of what they read or when. Just 12 self-selected books to keep.

The control group received nothing beyond the normal school year.

In the fall, all students took their regular state reading assessments. Researchers compared the change in scores from spring to fall for the two groups.

The results were consistent across all three sites and across multiple years: children who received books did not experience summer reading loss. Children in the control group showed the typical decline. The difference was equivalent in size to attending summer school.

The cost of the intervention was approximately $50 per student — a set of 12 paperbacks purchased at wholesale prices.

What makes the finding striking

The equivalence to summer school is the most important number to hold onto.

Summer school programs for reading remediation typically cost $800 to $1,500 per student. They require attendance, transportation, instruction, and administration. They require children to spend part of their summer in a school building. Many families cannot or will not participate. Attendance rates are often mediocre. Teacher quality varies.

Giving children books they chose themselves — which required no attendance, no scheduling, no professional staff, and no transportation — produced comparable reading outcomes at roughly 3% of the cost.

The key word is "self-selected." The books were not assigned. Children were not told to read particular titles or cover particular content. They chose books about dinosaurs, books about sports, books with pictures, books without. The only constraint was that the books were at an appropriate reading level for their age group.

This matters because it suggests the mechanism is intrinsic motivation, not instruction. Children who chose books they wanted to read actually read them. The act of choosing — exercising preference, feeling ownership over the decision — appears to be an important part of what makes the intervention work.

What this implies about libraries

The summer books experiment is not, strictly speaking, a library experiment. The books were given to children at school, not borrowed from a library. Children kept them permanently.

But the implications for libraries are direct.

Public libraries are one of the few institutions that provide free access to self-selected reading material. Summer reading programs run by public libraries have existed for decades. What the experiment adds is rigorous evidence that the access itself — the ability to choose and possess books — is the mechanism, not the structured activities or prizes or adult supervision that surround most summer reading programs.

This has design implications. A summer reading program that requires children to log hours, earn points, or attend library events may be adding friction to the mechanism that actually works: self-selected access to books. The data suggest that getting books into children's hands, with as few requirements attached as possible, is the core activity worth funding.

The Chicago and Salt Lake City experiments reviewed elsewhere in the registry found that removing library fines increased circulation, particularly among lower-income patrons. The summer books research suggests why that matters: barriers to access — financial, administrative, or procedural — reduce the activity that prevents summer reading loss.

The cumulative finding

Allington's team followed cohorts over three consecutive summers to estimate cumulative effects.

By the end of three summers, children who had received books showed approximately 0.2 standard deviation higher reading achievement than the control group. This is a meaningful difference by the standards of educational research. It is also a difference that compounded across three years from an intervention that cost $50 per student per year.

The compounding is the critical insight. Reading fluency is not a skill that plateaus in elementary school and then holds steady regardless of practice. It develops with use. Children who read over summer build fluency that makes next year's reading easier, which makes the year after that easier still. Children who don't practice lose fluency they had, and recovering it requires instructional time that could otherwise go to new learning.

Summer reading loss is not an inevitable feature of the school calendar. It is a product of differential access to books during the months when schools are closed. The access gap is narrower than it looks — and cheaper to close than any alternative intervention with comparable evidence.