Schools Replaced Books With Tablets. A Meta-Analysis of 54 Studies and 171,055 Readers Found Students Comprehend Less on Screens Than Paper — and the Deficit Grows Worse Every Year.

The Replacement Nobody Tested

Your child reads on a screen all day, and so do you. American students now spend more time reading digitally than on paper, and the shift is accelerating. School districts have invested billions replacing textbooks with tablets and Chromebooks. The assumption driving this transition is intuitive: screens are sharper than ever, kids grew up swiping before they could hold a pencil, and any early disadvantage of digital reading has surely faded by now.

It hasn't faded. It has deepened.

In 2018, a team at the University of Valencia and the Technion-Israel Institute of Technology gathered every controlled study they could find comparing reading comprehension on paper versus screens, selecting only experiments in which participants read comparable texts in both media and answered comprehension questions afterward. Their meta-analysis covered 54 studies published between 2000 and 2017, involving 171,055 participants across 19 countries, making it one of the largest syntheses in reading research. Both between-subjects and within-subjects designs told the same story: readers who read on paper comprehended significantly more than those who read the same text on a screen, with an effect size of Hedges' g = −0.21.

That number sounds small in the abstract, but in educational context it represents something concrete. The researchers noted that a 0.21 SD deficit corresponds to approximately two-thirds of the reading comprehension growth a typical elementary student achieves in an entire school year. Handing a student the same textbook chapter on a tablet instead of paper costs them a measurable fraction of what they'd normally learn in a year of reading instruction.

Three Factors That Make It Worse

The deficit was not uniform across conditions, and three moderators amplified it considerably. First, time pressure: when readers had a fixed deadline to finish the text, the screen disadvantage widened to g = −0.26. Under self-paced conditions, the gap shrank, suggesting that screen readers can partially compensate if given enough time to re-read and check their understanding. Second, text type: informational and expository texts showed a clear paper advantage (g = −0.27), while narrative fiction showed almost none (g = 0.01). The processing demands of a history textbook and a beach novel are different, and screens impair the former far more than the latter.

Third, and most damaging to the digital-native hypothesis: the screen disadvantage increased by approximately 0.01 standard deviations per year of publication. Studies published in 2015 showed a larger paper advantage than studies from 2005, which showed a larger advantage than studies from 2000. Screens got sharper, lighter, and more ubiquitous, but the comprehension gap widened anyway.

Seven Reviews, One Conclusion

Delgado's meta-analysis was not an outlier but the first of seven independent syntheses addressing the question. Kong, Seo, and Zhai (2018) reviewed 17 studies and found the same paper advantage. Clinton (2019) analyzed 35 studies and reported g = −0.25, adding an observation the earlier reviews hadn't tested: readers on screens showed poorer metacognitive calibration, consistently overestimating their own comprehension. Furenes and colleagues (2021) confirmed the pattern across 39 studies of children, and Díaz and colleagues (2024) reached the same conclusion in 49 studies. Of the seven meta-analyses, six found a statistically significant advantage for paper. The seventh, which examined narrative texts exclusively, found no difference, consistent with the moderator analysis showing the deficit is specific to informational reading.

Why Better Screens Haven't Helped

The leading explanation is what researchers call the shallowing hypothesis: digital environments train readers to skim because email, social media, and news feeds all reward fast extraction of surface-level information. When a reader encounters a dense expository text on the same device they use for social media and messaging, they appear to import those same shallow processing strategies. They read faster, skip more, re-read less, and form weaker mental models of the text's argument structure. Salmerón and colleagues found that on-screen readers reported more mind-wandering episodes during timed reading tasks, which correlated with lower comprehension scores.

A second mechanism involves spatial cognition and the physical affordances of paper. Books provide a stable spatial layout: page turns create landmarks, the thickness of remaining pages signals progress, and readers form implicit spatial maps of where information appeared on a given page. Scrolling eliminates all of these cues at once. Mangen and Kuiken (2014) found that readers of a mystery story on a Kindle were significantly worse at placing events in correct chronological order than paper readers, suggesting that the spatial instability of digital text disrupts the mental representation of a text's temporal structure.

If these mechanisms are right, the growing-gap finding makes sense: as people spend more years reading digitally, their shallow-processing habits deepen, the cognitive tools they once used for deep reading on paper atrophy from disuse, and each new cohort of readers arrives with less experience of the sustained, uninterrupted concentration that paper once enforced by default.

The Strongest Counterargument

Salmerón, Delgado, and their colleagues have argued that the screen disadvantage may reflect habit rather than an inherent limitation of digital media. What the meta-analyses capture is a snapshot of readers trained to read deeply on paper who transfer shallow digital habits to all screen reading. If schools explicitly taught deep-reading strategies for digital text (annotation, self-questioning, deliberate re-reading), the gap might close. Some individual studies lend weight to this argument: Florit and colleagues (2022) found no paper advantage when elementary students were given explicit comprehension strategies for both media. The deficit may not be about the screen itself but about what the screen signals to the reader's brain about the kind of reading that is expected. If true, the problem is solvable without retreating from digital media entirely.

What We Didn't Prove

The meta-analysis pools studies from 19 countries but is dominated by Western, educated, and student samples. The effect may differ in populations with limited access to paper or extensive digital-first reading instruction. Screen technology has continued to evolve since the 2017 endpoint of the analysis; E Ink readers may differ from LCD tablets, but most studies grouped all digital devices together. The "growing gap" finding rests on a trend across heterogeneous studies spanning 17 years and could reflect changes in study methodology or participant demographics rather than a true worsening of screen reading. Almost all studies measured comprehension immediately after reading. Whether the deficit persists, compounds, or reverses over long-term learning across an entire semester of digital-only textbooks is largely untested. And the original contribution that follows should be read as an estimate, not a measurement: we are applying a laboratory effect size to a population-level question the studies did not directly address.

A Back-of-the-Envelope Scale

Roughly 50 million students attend K–12 schools in the United States, and the majority now use digital texts for at least some of their informational reading, from science textbooks displayed on Chromebooks to history assignments read on shared iPads. If we apply the informational-text deficit of g = −0.27 to a student at the 50th percentile on a standardized reading assessment, the screen penalty drops them to approximately the 39th percentile. That is an 11-percentile-point slide from the choice of reading medium alone. Even if only half of U.S. students do the majority of their textbook reading on screens, the aggregate represents roughly 25 million students reading at a lower comprehension level than they would achieve with the identical text on paper. No curriculum change, no teacher training, no new pedagogy. Just the medium.

What You Can Do

Print the things that matter most. If you are studying for an exam, reviewing a contract, or reading a policy document, use paper whenever you can. If paper is not practical, read the digital text as though it were paper: slow down, annotate in the margins, and resist the urge to skim. Check your own comprehension after each section, because the research shows you are likely overestimating it on a screen. For light reading and narrative fiction, screens impose little measurable cost. The deficit is specific to the kind of reading where deep understanding matters, which is exactly the kind of reading most worth protecting.

For parents and educators: the evidence does not say screens are useless. It says screens carry a hidden tax on comprehension, and that tax is highest for the hardest, most important reading. Teaching explicit digital-reading strategies may help close the gap, but until that is demonstrated at scale, important assessments and study materials deserve paper.