C3 Paper vs screen, deep reading
narrowed 4 attacks: 3 survived, 1 narrowed · serves Q2 · last change 30 September 2026
For understanding continuous (linear) text in depth, reading on paper gives somewhat better comprehension on average than reading the same text on a screen among readers beyond early primary school, and readers on screen tend to overestimate how well they have understood. The average paper advantage has not been shown for first-grade beginner readers of short linear texts.
In practice
When a continuous informational text must be understood in depth by readers beyond early primary school, paper gives somewhat better understanding than a scrolling screen on average. If it is read on screen, don't trust the feeling of having understood it; check. Do not treat that average as a rule for first-grade beginner readers of short texts.
Where it holds
Inside the circle: the core of the claim. Outside: every bound and carve-out in the claim file, numbered, with the attack that forced it.
forced by a narrowing attackin the original wordingwritten at creation
The average effect is small (Delgado et al. 2018 g ≈ −0.21; Clinton 2019 g ≈ −0.25) and is scoped to readers beyond early primary school.
Written when the claim was created (25 September 2026); not forced by an attack
Delgado's same-text between-participant subgroups favour paper in grades 1–6 (g = −.19), grades 7–12 (g = −.15) and undergraduates (g = −.28), with educational level not a significant moderator.
Written when the claim was created (25 September 2026); not forced by an attack
The newest meta-analysis (Li & Yan 2024; 46 experiments, 2000–2022; I² ≈ 92 %) found no overall difference (g = 0.046, ns), but that null is carried by interactive or enhanced digital texts (g = 0.959, digital better), literary texts (mostly children's storybooks) and kindergarten studies; inside this claim's scope its estimates favour paper (university g = −0.432; time-limited g = −0.468; 1000–2000-word texts g = −0.391; informational g = −0.218, p = .053; non-interactive g = −0.172, ns).
Written when the claim was created (25 September 2026); not forced by an attack
Its school-age subgroups (primary, middle school) lean digital but are not significant and are dominated by interactive e-books.
Written when the claim was created (25 September 2026); not forced by an attack
It pooled no calibration outcome and did not separate comprehension depth or device.
Written when the claim was created (25 September 2026); not forced by an attack
Kill-condition test survived 2026-09-28.
Written when the claim was created (25 September 2026); not forced by an attack
Foss, Støle & Magyari 2026 (Norwegian VET, n=106, within-subjects, informational texts on own laptop vs print): whole-sample screen inferiority (OR=1.44; print M=4.15 vs digital M=3.42); significant in the lower-comprehension subgroup, trend only in the higher (p=.073); kill-condition test survived 2026-09-29 (calibration not tested).
Written when the claim was created (25 September 2026); not forced by an attack
Ronconi, Altoè & Mason 2025 (Italian grade 7, N=191, mixed design; two ~1000-word informational texts; paper vs scrolling desktop PDF): no significant medium main effect on literal or inferential comprehension; lower perceived cognitive load on screen; calibration not tested; kill-condition test survived 2026-09-30 (prong a met for this design; prong b untested).
Written when the claim was created (25 September 2026); not forced by an attack
First-grade beginner readers of short linear texts: Florit et al. 2025 (N=58, within-subjects) found no significant medium main effect on main-point, literal, or inferential comprehension; for descriptive texts tablet beat paper on main-point (d ≈ 0.57) and literal (d ≈ 0.59).
Written when the claim was created (25 September 2026); not forced by an attack
The advantage is for informational/expository text; for narrative-only text it is not significant (Delgado) or about zero (Clinton).
Written when the claim was created (25 September 2026); not forced by an attack
Time limits make the screen disadvantage larger (Delgado: time-limited g = −.26, self-paced g = −.09, ns; Li & Yan 2024: time-limited g = −0.468, free reading g = 0.192, ns); whether any paper advantage remains with self-paced reading is open question 4.
Written when the claim was created (25 September 2026); not forced by an attack
Much of the gap is tied to scrolling: with scrolling paper is clearly better, with paginated (no-scroll) display the differences are small or uncertain (Clinton-Lisell & Litzinger 2026).
Written when the claim was created (25 September 2026); not forced by an attack
On handheld devices (tablet, e-reader) the disadvantage is about half as large, and between studies it was larger for undergraduates than for schoolchildren, for whom it was not significant (Salmerón et al. 2024).
Written when the claim was created (25 September 2026); not forced by an attack
Display technology alone (LCD vs e-ink) does not explain the gap (Siegenthaler et al. 2012).
Written when the claim was created (25 September 2026); not forced by an attack
Main-idea understanding is about equal; key points and other details are better on paper (Singer & Alexander 2017).
Written when the claim was created (25 September 2026); not forced by an attack
For children aged 1–8, plain digitisation of a story lowers comprehension, but story-congruent digital enhancements can beat paper (Furenes, Kucirkova & Bus 2021).
Written when the claim was created (25 September 2026); not forced by an attack
Much of the evidence for the average paper advantage is from studies up to 2017 and from university students.
Written when the claim was created (25 September 2026); not forced by an attack
Does not transfer to: "Digital reading" in the PISA sense (navigating hypertext, evaluating sources), which the bank treats as a different construct (OECD PISA 2009).
Written when the claim was created (25 September 2026); not forced by an attack
0 of 18 bounds and carve-outs were forced by a specific attack. 3 attacks hit the boundary and did not move it: #2, #3, #4.
How it changed
The ledger records 4 attacks, but no earlier wording of C3 is on record, so the text cannot be replayed step by step. Only the current wording is shown.
For understanding continuous (linear) text in depth, reading on paper gives somewhat better comprehension on average than reading the same text on a screen among readers beyond early primary school, and readers on screen tend to overestimate how well they have understood. The average paper advantage has not been shown for first-grade beginner readers of short linear texts.
Attack record
Among first-grade beginner readers of short continuous linear texts, screen comprehension is at least as good as paper, so the claim’s average paper advantage does not hold for early primary readers.
Strongest reading and verdict reasoning
Steelmans to: The living “on average” paper advantage overreaches beginner readers; medium main effects are absent in grade 1, and tablet can equal or exceed paper on some superficial measures for descriptive text.
Why: Nightly-attacks 2026-09-27 04:11 Europe/Stockholm (first 04:05 run). Named within-subjects result (N=58): no significant medium main effect on main-point, literal, or inferential comprehension; for descriptive texts tablet beat paper on main-point (d ≈ 0.57) and literal (d ≈ 0.59). Does not kill the older-learner / meta average (Delgado/Clinton) or erase the calibration conjunct for those groups. Claim and Bounds rewritten same turn to scope the average paper advantage beyond early primary / beginner readers and to name Florit. Brief rewritten. Kill withheld: calibration gap and older-reader metas untouched.
A meta-analysis of 46 controlled experiments (2000–2022) comparing the same text on paper and on screen finds no overall comprehension difference (g = 0.046, ns) and a digital advantage outside China (g = 0.279), so C3's average paper advantage for continuous text is not established and its kill-condition is met.
Strongest reading and verdict reasoning
Steelmans to: The newest and broadest pooled estimate (k = 46 effects from 37 papers) shows screen ≥ paper overall; no school-age subgroup (kindergarten g = 0.866, primary g = 0.688, middle school g = 0.204) shows a paper advantage; so Delgado 2018 and Clinton 2019 do not hold beyond university students, and C3's “beyond early primary” average falls.
Why: Attack run 2026-09-28 07:48 Europe/Stockholm (daytime run on Jonatan's request; full text supplied by Jonatan and read in full; resolves open question 4). Sign convention: positive g = digital better (pp. 5–6). Kill-condition needs (a) screen ≥ paper for deep comprehension of continuous text under equal time in the relevant age band and (b) the calibration gap to disappear. (a) not met: the overall null (random effects g = 0.046, CI −0.192 to 0.283, p = 0.706, I² = 91.79 %; Table 2, p. 6) is carried by subgroups outside C3's scope — interactive/enhanced digital texts (AR, multimedia, annotation; g = 0.959, k = 11), literary texts (g = 0.968, k = 12, mostly children's storybooks) and kindergarten (g = 0.866, k = 8). Inside C3's scope the estimates favour paper: university g = −0.432 (CI −0.654 to −0.21), time-limited reading g = −0.468 (CI −0.75 to −0.186), 1000–2000-word texts g = −0.391 (CI −0.633 to −0.150), informational text g = −0.218 (CI −0.439 to 0.002, p = 0.053), non-interactive digital g = −0.172 (CI −0.404 to 0.061, ns) (Table 3, p. 7) — the same size as Delgado/Clinton (−0.21/−0.25), less precise. The digital-leaning school-age subgroups (primary k = 6, middle school k = 8, both ns) are dominated by AR and interactive e-books, not the same text read plainly, whereas Delgado 2018's same-text between-participant subgroups favour paper in grades 1–6 (g = −.19) and 7–12 (g = −.15) (Table 1). The “other countries” g = 0.279 contains nearly all the kindergarten and interactive studies. Comprehension depth is not separated (p. 3), devices are not distinguished (p. 8), no publication-year moderator, scrolling not coded. (b) not tested: no calibration outcome is pooled; overestimation appears only as narrative in the Discussion (p. 7). Kill-condition not met on either prong. Claim text unchanged (not rewritten). Bounds record the kill-test and the in-scope estimates. Leftover (not applied, held as open question 4): with free reading time Li & Yan g = 0.192 (ns) and Delgado 2018 self-paced g = −.09 (ns), so whether any paper advantage remains without time limits is open.
Among Norwegian upper-secondary vocational students reading the same informational texts on their own laptops versus printed A4, screen comprehension is at least as good as paper, so C3's average paper advantage beyond early primary is not established for this school-age band and its kill-condition is met.
Strongest reading and verdict reasoning
Steelmans to: Kill-condition prong (a) fails if a same-text school-age RCT finds screen ≥ paper for deep/informational comprehension under matched presentation.
Why: Nightly-attacks 2026-09-29 04:16 Europe/Stockholm. Named within-subjects result (n=106 VET, M=16.5): whole-sample screen inferiority (χ²(1)=14.165, p<.0001, OR=1.44 for print; print M=4.15 vs digital M=3.42, ~0.64 points). Significant in lower-comprehension subgroup (n=58, p=.0003); trend only in higher (n=48, p=.073). Kill-condition not met: screen was worse, not ≥ paper; calibration (prong b) not tested. Clause tested: school-age / VET informational continuous text on laptop vs paper. Claim field unchanged; Bounds note Foss kill-test survived and the lower-performer concentration.
Among Italian seventh graders reading the same lengthy (~1000-word) informational science texts on a scrolling desktop screen versus paper, screen comprehension is at least as good as paper on literal and inferential measures, so C3's average paper advantage beyond early primary fails and its kill-condition is met.
Strongest reading and verdict reasoning
Steelmans to: Kill-condition prong (a) fails if a same-text school-age experiment finds screen ≥ paper for deep/informational comprehension under matched presentation with scrolling.
Why: Nightly-attacks 2026-09-30 04:19 Europe/Stockholm. Named mixed-design result (N=191 grade 7, M=12.42): no significant main effect of medium on literal or inferential comprehension of ~1000-word informational texts (paper vs desktop PDF with scrolling); authors conclude results "did not support a screen inferiority effect". Kill-condition not fully met: prong (a) is met for this school-age lengthy informational design (screen ≥ paper), but prong (b) calibration/overestimation was not measured. Immediate test only. Clause tested: lower-secondary continuous informational text on scrolling desktop vs paper. Claim field unchanged; Bounds note Ronconi kill-test survived and the untested calibration prong.
Evidence
15 linked sources: 10 supports, 10 limits (first-pass draft, pending human review). supports: Clinton 2019, Delgado et al. 2018, Furenes et al. 2021, Jensen et al. 2024, Mangen et al. 2013, Salmerón et al. 2024, Singer & Alexander 2017, Li & Yan 2024, Foss et al. 2026, Ronconi et al. 2025 · limits: Clinton-Lisell & Litzinger 2026, Furenes et al. 2021, Kongsgården & Krumsvik 2016, OECD 2011, Salmerón et al. 2024, Siegenthaler et al. 2012, Florit et al. 2025, Li & Yan 2024, Foss et al. 2026, Ronconi et al. 2025