Syntes: Krumsvik et al. (2021) — tablets for every pupil in Bærum, Norway: limited, mixed effects on national tests

Källa: Krumsvik, R. J., Berrum, E., Jones, L. Ø., & Gulbrandsen, I. P. (2021). Implementing tablets in Norwegian primary schools. Examining outcome measures in the second cohort. Frontiers in Education, 6, 642686. https://doi.org/10.3389/feduc.2021.642686

Verified numbers (full text read 2026-10-08, open access)

• Design: school-level difference-in-differences; pilot 1 (5 schools, 1,743 pupils) and pilot 2 (10 schools, 4,395 pupils) vs non-pilot Bærum schools (9,570 pupils); national tests grades 5 and 9 (2014–2017), mapping tests grades 1–3, pupil survey.
• Grade 5, pilot 1 (2 vs 16 schools): 2017 difference-in-differences English +6.0 scale points (p < .05), boys +7.9 (p < .01), arithmetic +4.0 (ns).
• Grade 5, pilot 2 (7 vs 16 schools): no significant effects (arithmetic +0.1, English +1.0).
• Grade 9: no significant effects in either group.
• Grades 1–3 (share of pupils above the concern threshold, pilot 1, 2–3 vs 18 schools): mostly not significant; grade 3 arithmetic −15.35 percentage points in 2015/2016 (p < .05), −5.4 (ns) in 2017. In the first pilot group a check following the same pupils from grade 1 to grade 2 found no significant effects; in the second pilot group the grade-1 reading gains (+6.3 and +7.2 percentage points, 10% level) were not robust: following the same pupils from grade 1 to grade 2, the tablet effect was significantly negative for reading and arithmetic (appendix table not published; authors' text).
• Grade 5, pilot 1, fixed-effects model (Table 4): arithmetic +4.3 (5% level), English +3.2 (10%), boys' English +9.1 (1%); the simple difference-in-differences shows arithmetic +4.0 as not significant. (Added 2026-11-04, study-mining.)
• Authors: "in several school areas tablets have rather limited effect on pupils learning outcome"; "the study does not find any direct causality"; possible grade effects and spill-over from leisure English. Two of the authors were employed by Rambøll, which was paid by the municipality for the evaluation.

Role for C9

Supports, with caveats (context, 2026-10-08): besides the Swedish IFAU studies, this is the one Nordic 1:1 tablet difference-in-differences study in the project; it finds mostly no clear effect, with scattered positive and negative cells from very few schools. Not used as a ledger entry: not a clean kill-condition test (school-level data, two to seven treated schools per comparison, many outcomes).