Is reading on a screen worse for children than a paper book?

Not by itself. The largest meta-analysis of children's reading, Furenes, Kucirkova and Bus in 2021, pooled 39 studies covering 1,812 children aged one to eight. Once two small studies were imputed to correct for funnel plot asymmetry, the overall comprehension effect was 0.10 in favour of paper with a confidence interval running from minus 0.20 to plus 0.02. Because zero sits inside that interval, the authors concluded the difference favoured neither medium.

So where does the screens are worse idea come from?

Mostly from research on older readers and school texts. Delgado and colleagues reviewed 54 studies from 2000 to 2017 and found a small but real advantage for paper, Hedges' g of minus 0.21. The crucial detail is that it showed up for expository texts, the explanatory kind used in school and university, and not for narrative. A 2022 meta-analysis looking specifically at narrative texts found no negative effect of screen reading at all.

Which digital features help and which hurt?

Takacs and colleagues found that animations and sound effects that illustrated the story supported vocabulary and comprehension, while built-in games and hotspots detracted from learning. Furenes and colleagues found the same split: a digital book that differed from paper only by being digital scored lower on comprehension, but digital books with story-congruent enhancements outperformed paper books.

Do e-books really reduce the talking between parent and child?

Some do. Chiong and colleagues in 2012 found the drop in content-related talk appeared when parents read an enhanced e-book with hotspots, and not when they read a basic electronic version with none. The authors concluded it was the interactive features distracting both of them from the story, rather than the device.

Is an in-book dictionary a good thing?

It depends what you want from it. Furenes and colleagues found an embedded dictionary had no or negative effect on children's story comprehension, while positively affecting vocabulary learning. Tapping a word teaches the word and interrupts the story, which is a sensible trade if vocabulary is the goal and a bad one if following the plot is.

Does a story we make together on a screen count as reading?

By the terms of this research, the things that matter are present: narrative text rather than expository, an adult in the conversation, and no games or hotspots pulling attention off the story. None of that is a claim about Taleium specifically, because nobody has studied it. It is what the general evidence says the conditions are.

Raising readers

Does reading on a screen count? What 39 studies actually found

The research does not say screens are worse. It says hotspots and games are worse, and that the adult reading along matters far more.

David White Founder, Taleium

16 October 2026 About 5 min read

Parents ask this in a slightly guilty voice, which is the first clue that the question has been answered for them already by a headline.

The honest answer is that the research does not say what the headlines say. There is a real effect in the literature, it is small, it applies to a kind of text that is not storybooks, and the variable that matters more than the medium is whether there is a grown-up in the conversation.

Here is what the studies actually found.

The big one, and its real number

The largest analysis of children specifically is Furenes, Kucirkova and Bus, published in Review of Educational Research in 2021. They pooled 39 studies from 30 articles, covering 1,812 children aged one to eight.

The finding the abstract leads with is that digital books which differed from paper only by being digital produced lower comprehension scores. That is the sentence everybody quotes.

The number underneath it is less dramatic. After imputing two small-sample studies to correct an asymmetric funnel plot, the comprehension effect size moved from minus 0.07 to minus 0.10 in paper’s favour, with a 95 percent confidence interval of minus 0.20 to plus 0.02. Zero is inside that interval, so the authors’ own conclusion was that the difference favoured neither paper nor digital.

The vocabulary result went the same way. The estimate fell from 0.20 to 0.09 after imputation and was no longer significantly different from zero.

So the paper advantage in the children’s literature is somewhere between small and not there. That is a long way from the claim most parents have absorbed.

Where the strong finding comes from

The firmer evidence for paper is about older readers and a different kind of text.

Delgado, Vargas, Ackerman and Salmerón reviewed 54 studies published between 2000 and 2017 and found a small but significant advantage for paper, a Hedges’ g of minus 0.21. Their paper is called “Don’t throw away your printed books”, which tells you how they read their own result.

Three details matter for a parent holding a bedtime story.

It was expository text. The effect appeared for the explanatory, informational texts used in schools and universities. It did not appear for narrative. A 2022 meta-analysis that looked only at narrative texts found no negative effect of screen reading on comprehension at all.

Time pressure made it worse. The gap widened when readers were working against a clock, which is a reasonable description of revising and a poor description of reading a story at bedtime.

It got bigger, not smaller, over time. The authors used publication year as a proxy for how much digital exposure readers had, expecting the gap to close as people got more practised. It widened between 2000 and 2017, and they found no difference between age groups, which undercuts the idea that children raised on screens read screens better.

That last point cuts both ways, and it is worth being straight about it. It means familiarity does not fix the effect where the effect exists.

The feature problem, which is the real one

The interesting result in all of this is not about screens. It is about what is on them.

Takacs and colleagues found that multimedia features which illustrated the story, animations and sound effects, supported vocabulary and story comprehension. Built-in games and hotspots detracted from learning.

Furenes and colleagues found the matching pattern. A digital book that was just a paper book on a screen did slightly worse. A digital book with story-congruent enhancements did better than paper.

And Chiong and colleagues, in 2012, located where the damage happens. The reduction in content-related talk between parent and child showed up when they read an enhanced e-book with hotspots. It did not show up with a basic electronic version that had none. The authors concluded that the interactive features were what distracted parent and child from the story, not the device they were holding.

Read those three together and the rule is simple. A feature that serves the story helps. A feature that competes with the story hurts. The screen is neutral; the design is not.

The variable that beats all of them

One line from the 2021 meta-analysis is worth more than the rest of this article.

Adults’ mediation during print book reading was more effective than the enhancements in digital books that children read independently.

The comparison that produced the biggest difference was not paper against screen. It was accompanied against alone. An adult asking what the child thinks happens next outperformed a machine doing something clever.

Which is the finding parents should take away, because it is the one they control. The medium is a small effect with a confidence interval straddling zero. Sitting next to the child is a large one.

What this means for how you use a screen

Four things follow, and none of them is “avoid screens”.

Prefer narrative. The screen penalty in the literature attaches to expository text. Stories are the use case with the least evidence against them.

Be suspicious of taps. If a page has things to poke that are not the story, it is the configuration that produced reduced talk and reduced learning in the studies. A dictionary is the honest exception, and even then it is a trade: it helped vocabulary and did nothing or harm for comprehension.

Stay in the room. The strongest effect in the research is adult involvement, and a device that reads to a child alone throws it away. See does a bedtime story help a child sleep for the other half of why the adult matters at bedtime specifically.

Count the words either way. Nothing in this literature suggests words from a screen are worth less than words from paper. The million word gap sets out how large the gap between some reading and no reading is, and it dwarfs anything here.

Where that leaves Taleium

Honestly: nobody has studied it, and this article is not going to pretend otherwise.

What can be said is which side of each of these findings it sits on. The text is narrative. There are no hotspots and no games. And it does not work at all unless somebody is sitting with the child deciding what happens next, because the family decides what happens and Taleium writes it down.

That is the configuration the research points at. It is not evidence about this product, and a site that told you otherwise would be selling you something.

If your child says reading is boring, the medium is probably not the problem either. What to do when a child says reading is boring goes through what usually is.

The 2021 meta-analysis is catalogued on ERIC, and the 2018 review of 54 studies is available in full from the authors.

Questions

Is reading on a screen worse for children than a paper book?: Not by itself. The largest meta-analysis of children's reading, Furenes, Kucirkova and Bus in 2021, pooled 39 studies covering 1,812 children aged one to eight. Once two small studies were imputed to correct for funnel plot asymmetry, the overall comprehension effect was 0.10 in favour of paper with a confidence interval running from minus 0.20 to plus 0.02. Because zero sits inside that interval, the authors concluded the difference favoured neither medium.
So where does the screens are worse idea come from?: Mostly from research on older readers and school texts. Delgado and colleagues reviewed 54 studies from 2000 to 2017 and found a small but real advantage for paper, Hedges' g of minus 0.21. The crucial detail is that it showed up for expository texts, the explanatory kind used in school and university, and not for narrative. A 2022 meta-analysis looking specifically at narrative texts found no negative effect of screen reading at all.
Which digital features help and which hurt?: Takacs and colleagues found that animations and sound effects that illustrated the story supported vocabulary and comprehension, while built-in games and hotspots detracted from learning. Furenes and colleagues found the same split: a digital book that differed from paper only by being digital scored lower on comprehension, but digital books with story-congruent enhancements outperformed paper books.
Do e-books really reduce the talking between parent and child?: Some do. Chiong and colleagues in 2012 found the drop in content-related talk appeared when parents read an enhanced e-book with hotspots, and not when they read a basic electronic version with none. The authors concluded it was the interactive features distracting both of them from the story, rather than the device.
Is an in-book dictionary a good thing?: It depends what you want from it. Furenes and colleagues found an embedded dictionary had no or negative effect on children's story comprehension, while positively affecting vocabulary learning. Tapping a word teaches the word and interrupts the story, which is a sensible trade if vocabulary is the goal and a bad one if following the plot is.
What matters more than the medium?: Whether an adult is reading along. In the 2021 meta-analysis, adults' mediation during print book reading was more effective than the enhancements in digital books that children read independently. The comparison that mattered was not paper against screen, it was accompanied against alone.
Does a story we make together on a screen count as reading?: By the terms of this research, the things that matter are present: narrative text rather than expository, an adult in the conversation, and no games or hotspots pulling attention off the story. None of that is a claim about Taleium specifically, because nobody has studied it. It is what the general evidence says the conditions are.