In Swedish schools, programmes that give every pupil their own computer or tablet (one device per pupil, 1:1) have not raised pupils' average results, such as national test results, compared with schools with more limited device access. That is not established for every 1:1 programme elsewhere: in one US school district that combined a laptop for every pupil from grade 4 with teacher training, changed teaching (flipped lessons) and blocked games and social media, maths test scores rose by about 0.13–0.16 standard deviations four to five years after the start relative to the rest of the state, and reading scores by less, after no clear gain in the first years (Hull & Duch 2019).
Status
weakened
Last change
2026-11-04
Serves
Q2 (How does that work in practice: in school, with screens, and with AI as one of the tools?)
Ages
grundskola-lag (grade 6), grundskola-hog (grades 7–9) in the Swedish evidence; other ages not covered by the project's 1:1 studies; US grades 4–8 for the district programme that narrowed it on 2026-10-08 (Hull & Duch 2019); Norwegian grades 1–9 in the Bærum tablet study (Krumsvik et al. 2021, context only)
Area
Screens, devices and attention
Relates to
C6 relates-to C9 (written in C6: C6 is about pupils' own phones and personal devices during lessons, C9 about school-provided 1:1 programmes and average attainment).
Bounds
Scope: school-provided 1:1 device programmes and pupils' average attainment, not rules for pupils' own phones during lessons (that is C6). The comparison is 1:1 against more limited access to computers, not against no technology; the authors warn against reading it as “computers in school give nothing” (Hall, Lundin & Sibbmark 2019). Evidence in the project: IFAU 2019:29 (Hall, Lundin & Sibbmark 2019; grades 7–9, about 50,000 pupils in 26 municipalities, 2008–2016, difference-in-differences: no average effect on national tests in Swedish, English and maths or on transition to upper secondary; signs of negative effects for pupils with less-educated parents; tablets tended to do worse than laptops in Swedish); IFAU 2021:18 (Hall, Lundin, Mörtlund & Sibbmark 2021; grade 6: no average effect on national tests in maths and English, a small and not robust positive sign in Swedish, somewhat better results for pupils with highly educated parents and at schools with extra 1:1 funding). An average null can hide unequal effects by family background. International and descriptive context: UNESCO GEM 2023 (1:1 without pedagogy has failed to improve learning); OECD 2024 and Skolverket 2024 (Nordic PISA analysis), both cross-sectional (moderate learning use goes with better results than very low or very high use, leisure use in school with worse). Other outcomes, such as digital skills, are not covered. Programme bound (added 2026-10-08, nightly attack): Hull & Duch 2019 (Educational Evaluation and Policy Analysis 41(1), 79–97; difference-in-differences on North Carolina administrative data, over 4 million student-year observations, about 2,000 Mooresville pupils per year in grades 4–8, 2006–2013; full text read in the authors' December 2016 working-paper version, published abstract checked): Mooresville gave every pupil in grades 4–12 a laptop (2008–2009) together with teacher professional development, a shift to flipped lessons, extended wireless access and blocked gaming and social-media sites. Relative to the state trend, maths +0.054 SD in 2010, +0.086 in 2011, +0.140 in 2012, +0.162 in 2013 (published abstract: 0.13 SD in the medium term, short-term effects not significant); reading +0.041, +0.072, +0.121, +0.062. One treated district (seven schools), so the gains cannot be separated from the bundled changes in teaching and district culture (the authors say so), and inference rests on a single treated district. Nordic context: in Bærum, Norway, tablets for every pupil (5 and 10 pilot schools vs the other schools in the municipality, difference-in-differences on national tests, about 15,700 pupils) had "rather limited" effects: a positive grade-5 English effect in 2017 in the first pilot group (+6.0 scale points; two schools in the comparison) and, in the fixed-effects model, grade-5 arithmetic +4.3 (5% level; Table 4), no significant effect in grade 9, and a negative grade-3 arithmetic effect in 2015/2016; in the second pilot group, early grade-1 reading gains (+6.3 and +7.2 percentage points, 10% level; Table 10) turned into a significantly negative effect on reading and arithmetic when the same pupils were followed from grade 1 to grade 2 (pp. 13–14; appendix table not published) (Krumsvik, Berrum, Jones & Gulbrandsen 2021; school-level data, the authors warn against causal readings; significance from school-level t-tests with two or three treated schools per comparison and no correction for many outcomes; two pilot-1 schools began tablets in autumn 2014, inside the "before" period, p. 8; corrected 2026-11-04). Rule-3e pause (added 2026-10-27; method note, not a narrowing): eleven consecutive fruitless searches. Swedish and Nordic 1:1 evaluations on file are exhausted for a widened search. Rule-3e widening is paused until a new Nordic evaluation appears; ordinary rotation may still search. The kill-condition stays Swedish or Nordic.
Does-not-transfer-to
Rules for pupils' own phones and personal devices during lessons (C6).
What would kill this
A Swedish or Nordic difference-in-differences study or randomised trial in which 1:1 devices under realistic use raise average results measured without the device (national tests or similar) without more distraction. Bank wording (Swedish, verbatim): “Svensk/nordisk DiD eller RCT där 1:1-enheter under realistisk användning höjer closed-book-kunskap utan ökad distraktion.”
Tutor consequence
none yet
Attack record
1 attack: 1 narrowed
Note from 27 October 2026 (not an attack, so it is not counted). After 11 nightly searches in a row found no study that could test it: wider searches beyond the Nordic countries are paused until a new Nordic 1:1 evaluation appears. Claim text is already Swedish-scoped; kill-condition stays Swedish or Nordic (not widened geographically). The limits gain a short note.
2026-10-08 narrowed
Attack
A district-wide US programme that gave every pupil from grade 4 a laptop raised standardised maths scores by about 0.13–0.16 standard deviations four to five years after the start, relative to the rest of the state, with smaller reading gains, so C9's general statement that 1:1 programmes do not raise average results does not hold beyond the Swedish evidence.
Target
claim-text
Source
Hull & Duch 2019 | DOI 10.3102/0162373718799969
Steelmans to
C9's claim sentence is general ("school programmes … do not raise pupils' average results") while its evidence is Swedish; Hull & Duch use the same design as the Swedish studies (difference-in-differences on standardised state tests taken without the device's help, administrative data for the whole state) and find a sustained maths gain that grows over five years and is robust to a neighbouring-county control group and to dropping movers, so the general sentence cannot stand.
Why
Nightly-attacks 2026-10-08 Europe/Stockholm. Full text read = the authors' working-paper version (December 2016, AEA 2017 conference paper); the published article (EEPA 41(1), 79–97, 2019) is closed and only its abstract was read (ERIC EJ1204852), which reports math +0.13 SD in the medium term and statistically insignificant short-term effects. Working-paper numbers (Table 2, relative to 2008, school fixed effects, student and school controls): maths 2010 +0.054, 2011 +0.086, 2012 +0.140, 2013 +0.162; reading +0.041, +0.072, +0.121, +0.062; pre-period 2006 −0.010 and 2007 −0.035 (maths); neighbouring-county controls give almost identical or higher estimates; results hold without transfer pupils; time spent on homework rose slightly and free reading fell by about 5 minutes a day. Kill withheld: (1) one treated district (seven schools; about 2,000 pupils a year), so precision is overstated and the laptops cannot be separated from the bundled teacher training, flipped lessons, blocked entertainment sites and change in district culture (the authors: "We could not give a laptop to every student in a school district and expect the same changes in outcomes without also focusing on changing district culture"); (2) the Swedish difference-in-differences studies (IFAU 2019, 2021; 26 municipalities) and the Norwegian Bærum study (Krumsvik et al. 2021, DOI 10.3389/feduc.2021.642686: limited effects, an isolated positive grade-5 English effect, a negative grade-3 arithmetic effect) remain null or mixed; (3) the comparison is the rest of North Carolina, not schools with a documented level of limited device access. Not a restatement: C9's Bounds held only Swedish studies and descriptive international context; first ledger entry on C9. Claim and Bounds rewritten same turn; brief and the practical rule rewritten.
Rewrite
claim-field+brief
Corrections to our notes
These fix a number or a fact in our notes. They are not attacks and do not count toward the floor of 3 survivals.
— C9's limits summary of the Bærum tablet study omits the fixed-effects model's positive grade-5 arithmetic effect and the second pilot group's same-pupil result, in which early grade-1 reading gains turned significantly negative from grade 1 to grade 2, and omits that significance rests on t-tests with two or three treated schools.
Nordic 1:1 tablet DiD (Bærum, Norway): mostly no significant effects on national tests; positive grade-5 English (and, in one model, arithmetic) effects in one pilot group of two schools, not seen in the second pilot group or in grade 9; negative grade-3 arithmetic effect, and a negative grade-1-to-2 reading and arithmetic effect when the same pupils were followed in the second pilot group; authors warn against causal readings. Context for the 2026-10-08 C9 entry.
Indirect: Swedish grade 8 at the EU/OECD average in computer and information literacy despite high digitalisation; descriptive, not a 1:1 effect estimate.
Correlational: overall screen exposure associated with lower academic achievement (r = −0.15) but not maths, science or reading scores; it does not study 1:1 school device programmes or Swedish schools. Cannot establish causation; does not test C9. Abstract only.
US district 1:1 laptop programme bundled with teacher training and changed teaching: maths +0.13–0.16 SD after four to five years relative to the state trend (published abstract: 0.13), reading smaller; one treated district. C9 weakened 2026-10-08 (programme bound).
Policy context (low tier, no data): the Swedish Government moved away from early digital tools toward physical books and selective device use (2024). Background for C9's Swedish setting; not evidence for or against it.