Paper won by 0.21 SD in 54 studies: what to print and when
Most of what I read for work arrives as a PDF: a research paper, a client's process documentation, a 30-page spec. I read it on a laptop. Across 54 studies and 171,055 readers, comprehension came out better on paper by 0.21 standard deviations, and the effect concentrates on exactly that kind of document.
What the research shows.
Pablo Delgado, Cristina Vargas, Rakefet Ackerman and Ladislao Salmerón pooled the studies published between 2000 and 2017 that compared comparable texts on paper and on a digital device: 38 between-participants designs, 16 within-participants designs, 171,055 readers. Paper won by a Hedges' g of 0.21, the same size in both design families 1.
Three moderators came out of that analysis, and they carry more weight than the headline. The paper advantage was larger when reading was time-constrained than when readers set their own pace. It held for informational texts and for mixed sets, and disappeared for narrative-only texts. And it grew across the publication years rather than fading as screens improved 1.
Virginia Clinton reviewed 29 reports covering 33 studies. Reading performance favored paper again (k = 33, n = 2,799). Reading time showed no reliable difference between the two media (k = 14, n = 1,233). Her readers also judged their own performance more accurately on paper, g = 0.20 (k = 11, n = 698) 2. Readers on screen took no less time over a text and misjudged how much of it they had kept.
Rakefet Ackerman and Morris Goldsmith took that apart across two experiments. Participants read expository texts of 1,000 to 1,200 words, predicted their own score, then sat a multiple-choice test. With the study time fixed by the experimenters, the two media produced the same scores. With study time under the reader's control, the screen group scored lower. The authors attribute the difference to metacognition rather than cognition: less accurate predictions of performance, and more erratic allocation of study time on screen 3.
The medium mattered where the reader owned the decision to stop.
The moderators disagree about the clock.
Hoch, Stürmer, Jonkmann and Scheiter ran the comparison in the field, on high-stakes university exams, and found no screen penalty in performance and none in metacognitive accuracy. Students did report investing less effort in the e-exam, and rated it less effortful and less difficult than the paper-pencil version. The authors read those ratings as a trace of shallow processing that survives even when the processing underneath goes deep enough to protect the score 4.
An exam slot takes the pacing decision away from the student, which fits Ackerman and Goldsmith's fixed-time experiment. It cuts against Delgado's moderator, where time constraints widened the paper advantage. Two sound analyses point opposite ways on the clock, so treat the timing advice below as the cheap bet it is, and the text-type advice as the solid one.
The direction does carry across ages. Furenes, Kucirkova and Bus pooled 39 studies from 30 articles covering 1,812 children aged 1 to 8. Digital books that differed from the print version only by being digitized produced lower comprehension scores. An adult reading the print book with the child beat the enhancements built into digital books that children read alone 5.
The protocol.
1. Print informational text over 1,000 words that you have to act on. Specs, contracts, research papers, post-mortems. The effect lives on informational text and vanishes on narrative, so a novel or a newsletter stays on the phone 1.
2. Write 5 keywords from memory, at a delay, before you decide you understood it. Tirza Lauterman and Rakefet Ackerman had participants generate keywords after a delay, then judge their own knowledge, then take the test. That sequence removed screen inferiority even on the first texts studied 6. Close the file, wait an hour, write the 5 words.
3. Put a number on your confidence before you check. Write the score you expect out of 100, then test yourself against the document. Calibration ran 0.20 SD better on paper 2, so the on-screen version of that number is the one to distrust.
4. Budget the same minutes for the screen copy. Across 14 studies there was no reliable reading-time difference 2. A sense of moving fast through a document on screen reports fluency and not retention.
5. Fix the stopping point before you open the file. Minutes, written down, derived from the page count. Fixing the study time is the one manipulation that erased the gap in a controlled experiment 3, and the exam slot did the same in the field 4.
6. Split the material you hand to other people by text type. Exercise sheets and reference one-pagers on paper, demos and narrative on screen. The moderator in the meta-analysis draws that line for you 1.
7. Run the comparison on yourself for 2 weeks. Take 10 work documents, alternate the medium, and after each one write 3 claims from memory and check them against the source. Count the errors per document, then compare the two columns. 10 documents settle nothing about the science. They will tell you whether your own gap sits nearer 0.21 SD or nearer zero.
A note for anyone who reads with a child: the print book plus an adult beat the digital book with built-in enhancements 5. Read it with them.
Tomorrow morning, print the one document you have to act on this week. Leave the newsletter on your phone. An hour after you finish the printed one, write 5 keywords from memory, then go back through the pages and count what you missed.
Sources.
- Delgado, P., Vargas, C., Ackerman, R., & Salmerón, L. (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:10.1016/j.edurev.2018.09.003
- Clinton, V. (2019). Reading from paper compared to screens: A systematic review and meta-analysis. Journal of Research in Reading, 42(2), 288-325. doi:10.1111/1467-9817.12269
- Ackerman, R., & Goldsmith, M. (2011). Metacognitive regulation of text learning: On screen versus on paper. Journal of Experimental Psychology: Applied, 17(1), 18-32. doi:10.1037/a0022086
- Hoch, Stürmer, Jonkmann & Scheiter (2025). Paper-pencil vs. e-exams: Revisiting the screen inferiority effect during high-stakes testing at university. Learning and Instruction. doi:10.1016/j.learninstruc.2025.102226
- Furenes, M. I., Kucirkova, N., & Bus, A. G. (2021). A comparison of children's reading on paper versus screen: A meta-analysis. Review of Educational Research, 91(4), 483-517. doi:10.3102/0034654321998074
- Lauterman, T., & Ackerman, R. (2014). Overcoming screen inferiority in learning and calibration. Computers in Human Behavior, 35, 455-463. doi:10.1016/j.chb.2014.02.046