Reading on Screen vs Paper: What the Research Shows

On average, people understand informational text slightly better on paper than on a screen. Two large meta-analyses agree on this: Delgado and colleagues (2018) found a small advantage for paper across studies with more than 171,000 participants, and Clinton (2019) found a similar one in randomised experiments. The gap is called the screen inferiority effect. It is larger when reading is rushed and largely absent for stories, and readers on screen tend to overestimate how much they have understood. Reading speed itself is about the same on both.

What is the screen inferiority effect?

The screen inferiority effect is the finding that comprehension tends to be lower when the same text is read on a digital device rather than on paper. It does not mean screens make reading impossible or that the difference is large. It means that across many studies, the average reader does a little worse on a comprehension test after reading on screen.

Three meta-analyses, each pooling a different set of studies, have measured it:

The effect sizes, around a fifth to a quarter of a standard deviation, count as small. To put that in context, Delgado and colleagues cite an estimate that reading comprehension in primary school grows by about 0.32 standard deviations a year on average. A gap of 0.21 is not trivial for a student sitting a test, but it is not a reason to fear every screen.

The gap is not limited to surface details. Clinton found it both for literal questions, about what the text said (g = −0.33), and for inferential ones, about what follows from it (g = −0.26).

Does it apply to phones, tablets and e-readers?

Less is known. Most studies in these reviews used computer screens. In the 2018 meta-analysis the paper advantage was significant for computers but smaller and not statistically significant for handheld devices, which were tested in far fewer studies. That is an absence of evidence, not proof that handheld reading is equivalent to paper.

When does reading on screen hurt comprehension most?

The average hides a lot of variation. The meta-analyses looked for moderators, conditions under which the gap grows or disappears.

The screen penalty is concentrated where you can least afford it: informational text, read under time pressure. A novel on an e-reader is a different case from a textbook chapter skimmed on a laptop before a deadline.

Why is reading on paper better for comprehension?

The research identifies the conditions more clearly than the cause, and the meta-analyses say so. The leading explanations are about how people read on screens, not about the screens themselves:

• Shallower processing. Most of what people do on screens is quick: messages, feeds, search results. One account, which Delgado and colleagues call the shallowing hypothesis, holds that this habit carries over to reading, especially when a time limit makes skimming seem legitimate.
• Overconfidence. Clinton found that readers judged their own understanding more accurately on paper. Ackerman and Goldsmith (2011) found that screen readers predicted their scores less accurately and regulated their study time more erratically, and when they were free to decide how long to study, they did worse on screen.
• Navigation. Scrolling removes the fixed position of text on a page that helps you find your way back to a passage. The scrolling result above points this way, though it was not statistically conclusive.

The overconfidence finding matters most in practice. If you stop reading because a text feels understood, and that feeling is less reliable on screen, you stop too early.

How to read better on a screen: a practical routine

You can narrow the gap. In experiments by Lauterman and Ackerman (2014), a simple step, writing keywords that summarise the text, removed the screen disadvantage in both scores and overconfidence, although only for participants who preferred reading on screens. The following routine builds on that and on the moderators above.

1. Choose the medium by the task. For dense informational text you must remember, such as a contract, a textbook chapter or a paper you will cite, print it or use a paged e-reader if you can. For news, fiction and light reading, the screen is fine.
2. Remove the false deadline. Time pressure is the biggest moderator. Block a session long enough for the text, and do not read important material in the gaps between meetings.
3. Prefer pages to endless scrolling. Use a paginated view, a PDF in page mode or reader mode with large type, so each passage has a position you can return to.
4. Write keywords as you go. After every section of roughly 500 to 800 words, stop and write three to five keywords that summarise it. This takes about 30 seconds.
5. Test, do not trust the feeling. At the end, close the text and spend two minutes writing the main points from memory. Compare them with the text and reread what you missed.
6. Look after your eyes. Take a short break from the screen every 20 to 30 minutes and look at something distant. If you have persistent eye strain or headaches, see an optometrist; a reading app is not a treatment.

More techniques for the recall steps are in how to improve reading comprehension, and ways to keep your attention on screen text are in how to focus while reading.

Where do RSVP and pacing tools fit?

RSVP readers, which show one word at a time, and paced highlight modes are screen reading too, but they were not part of these meta-analyses, so their results do not transfer directly. Two findings are still useful as a guide.

• A fast pace is a form of time pressure. Since the screen penalty grows under time limits, set an RSVP or pacer speed close to your normal reading rate for informational text, around 200 to 275 words per minute for most adults, and pause whenever you lose the thread. Research on RSVP finds comprehension holds up near normal speed and falls as the pace rises; see what is RSVP reading?
• Narrative text is the easier case.Stories showed no screen penalty in either meta-analysis, and they are also where RSVP's lack of looking back costs least. Light fiction and articles are good material for a paced reader; a dense chapter is better read on a page.

Full-text paced modes are a middle ground. They keep the paragraph on screen, so you can glance back, and they move through the text for you, which removes manual scrolling. For working with long documents, see how to speed read a PDF.