Paper vs. Screens: What the Research Actually Says

If your goal is to understand a demanding informational text, paper is still a sensible default—but the research does not support the slogan that paper always wins. The clearest recent pattern is conditional: paper’s advantage is most dependable when digital text requires continuous vertical scrolling. When text is fully visible or paginated on a screen, the paper-versus-device differences become much smaller and, in the latest device-specific analysis, are not statistically reliable.

That answer applies most directly to expository text: material that explains, informs, describes, or argues about a subject. It is not a verdict on every novel, every e-reader, every reader, or every reason someone might choose a format.

The short answer

The best current evidence suggests a useful rule of thumb:

For difficult expository reading, choose paper when it is convenient. If you need to read digitally, try a comfortable, larger display with page-by-page navigation before assuming that “digital” itself is the problem. A phone, tablet, computer, and dedicated e-reader are different reading situations, and the research does not rank them with equal confidence.

This is a practical interpretation of several meta-analyses, not a command. A small average difference in a controlled comprehension test does not tell you which format you will actually read, which format is accessible to you, or which format will let you return to the text when you need it.

“Digital reading” is not one comparison

The phrase screen reading can hide several important differences:

These are broad categories, not descriptions of every current model. A tablet in a distraction-free page view is a different experience from a tablet showing a long web page. An e-reader with adjustable text is not automatically equivalent to a paper book, and a phone is not automatically a poor choice for every short or practical reading task.

What the 2026 network meta-analysis examined

Virginia Clinton-Lisell and Christine Litzinger’s 2026 network meta-analysis is useful because it separates the device categories instead of treating every digital display as one condition. It combined 56 studies with independent samples and 79 effect sizes, comparing paper with computers, tablets, e-readers, and smartphones.

The inclusion rules were narrower than the title might suggest. The studies used expository texts, independent readers, controlled environments, native-language reading, comprehension outcomes, and random assignment or counterbalancing. Screen features that could not be reproduced in the paper condition were excluded. The study table includes some children and adolescents, but the evidence is concentrated among college students reading science or other informational texts in English.

The researchers also distinguished scrolling from pagination. Scrolling meant moving text vertically with a scrollbar, mouse, or finger because the whole text could not be seen at once. Pagination meant moving between static pages by swiping, clicking, or pressing a button.

That distinction matters because “a book on a screen” is not a single layout. A page-like digital text preserves some fixed location cues. A continuously scrolling text keeps moving the location of previously read material.

What it found: rankings are not verdicts

In the overall network model, positive Hedges’ g values mean that paper produced higher comprehension scores than the named device:

The network’s P-score ranking put paper first, followed by tablet, e-reader, computer, and smartphone. But a ranking is not the same as a statistically reliable difference between every pair. The paper-versus-tablet result was not significant. The paper-versus-e-reader result was marginal at most. Even the ordering of e-readers and computers should not be treated as a clean, proven hierarchy: their rounded overall P-scores were the same.

This is one of the most important ways to read a network meta-analysis. Its ranking summarizes how conditions compare across the network of direct and indirect evidence. The confidence interval and significance test tell you how firmly a particular comparison is separated from zero. They answer related questions, not identical ones.

When scrolling was necessary

The scrolling subgroup included 25 studies and 29 comparisons. It contained no e-reader studies in which scrolling was necessary.

Paper had a reliable advantage over computers (g = 0.35, p < .001) and tablets (g = 0.38, p = .03). The smartphone estimate was larger in size (g = 0.48) but highly imprecise and not statistically reliable (p = .13).

The safest summary is not “scrolling damages comprehension.” It is: the paper advantage was clearest in the comparisons where the digital text required scrolling, and the study’s design cannot prove that scrolling alone caused the gap. Scrolling is entangled with screen size, device type, text length, and the reader’s behavior.

When scrolling was not necessary

The no-scrolling analysis included 31 studies and 39 comparisons. Here, the text either fit on the screen or was navigated through static pages. None of the paper-versus-device differences was statistically reliable: computer (p = .17), e-reader (p = .19), smartphone (p = .28), and tablet (p = .64).

Paper still ranked highest in this subset, with tablet next among the digital conditions. That ranking is worth reporting, but it should not be oversold. The result is compatible with a small paper advantage, a genuinely negligible difference, or insufficient precision to distinguish the formats. It does not establish that paper and every non-scrolling device are identical.

Why scrolling may matter—and what is still uncertain

The 2026 authors discuss several possible mechanisms. Fixed page locations might help readers remember where an idea appeared and look back to connect it with a later idea. Smaller displays may show less surrounding context. Moving text may add an extra tracking demand. Ergonomics, eye strain, and readers’ confidence in how well they understood the material may also play a role.

These are proposed explanations, not findings that the network meta-analysis tested as a single causal model. Scrolling was coded only as “necessary” or “not necessary.” The analysis did not record how much an individual scrolled, how quickly they moved, whether they backtracked, or whether a page view was genuinely easier to navigate.

So the practical response should be modest. If your software supports it, try page view, full-screen text, or a larger comfortable display for a dense passage. Reduce avoidable interruptions and make notes or pause for a short recall check. These are sensible experiments based on the proposed mechanisms; they are not guaranteed repairs proven by this meta-analysis.

Why expository text changes the answer

Earlier evidence helps explain why the 2026 study focused on expository material. In their 2018 meta-analysis, Delgado, Vargas, Ackerman, and Salmerón found a small paper advantage across comparable paper/digital reading studies. The advantage was larger under time limits and was reliable for informational and mixed texts, but not for their smaller narrative-only subset. The review also reported substantial variation between studies and did not find a reliable device-type moderator.

A later handheld-device synthesis found a small print-favoring effect for tablets and e-readers taken together, but its results were also qualified. Most samples were convenience samples, many measures were multiple-choice tests, and only a handful of studies used vertical scrolling. In that more limited handheld dataset, genre, length, navigation, and device type did not reliably explain the average effect.

What this does not establish about novels and pleasure reading

The expository findings should not be pasted onto fiction. A narrative-text meta-analysis covering 32 comparisons and 2,239 participants found no significant overall screen-versus-paper comprehension effect. It also found no reliable difference across computer, e-reader, and tablet categories for narrative texts.

That does not mean every novel feels the same on every device, or that pleasure, immersion, and comprehension are interchangeable. It does mean that “paper is better for expository comprehension” is not enough evidence to tell someone that a digital novel is an inferior reading experience. For fiction, choose the format that makes you want to continue and lets you read comfortably.

Practical recommendations by reading goal

I need to learn from a dense article or chapter

Use paper if it is readily available and works for you. The evidence is most favorable to that choice when the text is informational, the task is demanding or time-limited, and the digital version would require continuous scrolling.

I have to read digitally

Start with the least disruptive version you can control: a comfortable display, a page-like layout, and enough time to pause and check what you understood. If a phone is the only available device, use it without guilt; just recognize that a small, scrolling display may be a harder setup for a sustained expository task than a paginated or larger-screen option. Review your notes or summarize the main idea afterward if the material matters.

I am reading a short reference or a practical message

The research summarized here is about controlled comprehension tasks, often with longer informational passages. It is not a reason to print every short message or look up. Convenience and speed may reasonably decide the format when deep integration of a long text is not the goal.

I am reading a novel for pleasure

Choose the format that supports attention, comfort, and enjoyment. The narrative evidence does not show a general comprehension penalty for screens, and this article does not attempt to rank every e-reader experience.

Accessibility or comfort is the deciding factor

A format that makes text readable and usable for you is the right format for you. The studies discussed here were not designed to answer the needs of every reader, so a small group average should not override accessibility, language support, portability, or physical comfort.

Your format is not a moral test

The most useful conclusion is not that one medium proves how serious a reader you are. It is that the reading task and the format interact. Paper can be a good tool for close work. Digital reading can be the practical tool that makes the work possible. A tablet in page view may be a better compromise than a scrolling phone; an e-reader may fit a novel beautifully while remaining under-tested for expository scrolling; and a paper book may be the format you return to because it helps you orient yourself in a long argument.

The point is to notice what you are asking the format to do, then choose deliberately without turning the choice into a source of guilt.

A digital organizer can support a physical reading life

Nook & Spine sits alongside this choice rather than replacing it. It can record the physical format and edition details of a book, keep notes and reading status with the record, track physical location and condition, and help organize a TBR queue. It is not an e-book reader: the app organizes book records, but it does not open or render the text of an e-book.

That distinction is useful. Your reading surface can be paper, a tablet, an e-reader, or a phone; the record of what you own, what you want to read, and what you thought about it can still live in one private library.

Organize the books you prefer to read in print.

Sources and further reading

• Virginia Clinton-Lisell and Christine Litzinger (2026), “Decoding digital reading: a network meta-analysis of comprehension across devices,” Education and Information Technologies, 31, 1611–1643. DOI. This is the anchor source for separating computers, tablets, e-readers, smartphones, paper, and scrolling.
• Pablo Delgado, Cristina Vargas, Rakefet Ackerman, and Ladislao Salmerón (2018), “Don’t throw away your printed books: A meta-analysis on the effects of reading media on reading comprehension,” Educational Research Review, 25, 23–38. DOI. This provides earlier context on overall effects, expository versus narrative text, and time pressure.
• Ladislao Salmerón, Lidia Altamura, Pablo Delgado, Anastasia Karagiorgi, and Cristina Vargas (2024), “Reading Comprehension on Handheld Devices versus on Paper: A Narrative Review and Meta-Analysis of the Medium Effect and Its Moderators,” Journal of Educational Psychology, 116(2), 153–172. DOI. This narrows the comparison to handheld devices and makes the evidence gap around scrolling visible.
• Annika Schwabe, Fabienne Lind, Lukas Kosch, and Hajo Boomgaarden (2022), “No Negative Effects of Reading on Screen on Comprehension of Narrative Texts Compared to Print: A Meta-analysis,” Media Psychology, 25(6), 779–796. DOI. This supplies the article’s boundary evidence for narrative texts and prevents automatic generalization to novels.