C19 Metacognitive strategies

survived narrowedone dot per attack, oldest first

1. #1 survived 27 October 2026 · de Boer, Donker, Kostons & van der Werf 2018 · target: kill-condition

   A meta-analysis of metacognitive strategy-instruction interventions with follow-up tests would meet C19's kill-condition if the academic gain were gone, or no larger than in comparison classes, weeks or months later; de Boer, Donker, Kostons & van der Werf 2018 found follow-up Hedges' g = 0.63 against posttest 0.50.

   Strongest reading and verdict reasoning

   Steelmans to: C19 dies if a large randomised trial or meta-analysis finds the gain from metacognitive strategy instruction gone or no larger than comparison at a later follow-up; this is the main long-term meta of such school interventions and the claim's own supporting source once the full text is read.

   Why: Extra3 attack pass 2026-10-27 Europe/Stockholm (week-2 floor push; C14/C16 OA kill-texts blocked; strongest honest open-text attack on a below-floor claim under parent brief). Ordinary rotation slot for never-targeted C19. Full text read and stored (University of Groningen Pure publisher PDF; Taverne self-archive). 48 interventions, school context, follow-up ≥3 weeks (mean 21.6, SD 23.5); posttest g = 0.50 (SE 0.07), follow-up g = 0.63 (SE 0.07); post–follow-up difference g = 0.12 (SE 0.04, 95% CI 0.05–0.19). Kill fails: the later gain was not gone. Limits: meta of existing trials (not one new RCT); intervention packages mix metacognitive with cognitive/management/motivational components; Dignath & Büttner 2008 still abstract-only in the project. Not a preprint. Claim sentence and Status unchanged (alive). Bounds and the practical rule updated same turn (abstract-only caveat removed for this source).

2. #3 survived 27 October 2026 · Eberhart, Ingendahl & Bryce 2025 · target: kill-condition

   A newer multi-level meta-analysis of metacognition interventions in young children would meet C19's kill-condition if the gain were gone, or no larger than comparison, at a later follow-up; Eberhart, Ingendahl & Bryce 2025 found follow-up g = 0.29 (95% CI 0.17–0.40) against posttest 0.48.

   Strongest reading and verdict reasoning

   Steelmans to: C19 dies if a large randomised trial or meta-analysis finds the gain from metacognitive strategy instruction gone or no larger than comparison at a later follow-up; this is an independent OA meta (67 studies; 19 with follow-up) of metacognition interventions in pre- and elementary school, distinct from de Boer et al. 2018.

   Why: Extra5 attack pass 2026-10-27 Europe/Stockholm (parent: focused OA kill-hunt; prefer C19/C16/C14). Full text read from Springer OA HTML (CC BY 4.0; PDF bot-gated from the box; numbers transcribed to ). 349 effect sizes, 67 studies, typically developing children mean age ≤12; 19 studies with follow-up (mean 15 weeks, range 3–48). General follow-up g = 0.29, SE 0.06, 95% CI [0.17, 0.40], p < .001. Language at follow-up g = 0.45 [0.29, 0.60]. Kill fails: the later gain was not gone. Limits: age band is pre/elementary only (not gymnasium); follow-up Metacognition outcome alone ns (g = 0.12); self-efficacy follow-up fragile to one influential study; funnel asymmetry at follow-up (Egger Q = 7.18, p = .007); researcher-developed academic tests show larger effects than standardised. Not a restatement of de Boer 2018 (different meta, younger sample, newer evidence). Not a preprint. Claim sentence and Status unchanged (alive). Bounds and the practical rule updated same turn. C19 now 2 of 3 survivals.

3. #4 survived 27 October 2026 · Tuero, Núñez, Vallejo, Fernández, Añón, Moreira, Martins & Rosário 2022 · target: kill-condition

   A classroom-randomised school-based SRL trial with a 3-month academic follow-up would meet C19's kill-condition if the gain were gone, or no larger than comparison, at that later test; Tuero, Núñez, Vallejo, Fernández, Añón, Moreira, Martins & Rosário 2022 found combined academic performance d = 0.33 at follow-up against d = 0.25 at post-test.

   Strongest reading and verdict reasoning

   Steelmans to: C19 dies if a large randomised trial or meta-analysis finds the gain from metacognitive strategy instruction gone or no larger than comparison at a later follow-up; this is an independent OA classroom-randomised trial (761 pupils; Rainbow Program SRL with planning–execution–evaluation) with a 3-month academic follow-up, distinct from de Boer et al. 2018 and Eberhart et al. 2025.

   Why: Extra11 attack pass 2026-10-27 Europe/Stockholm (parent: C19 2→3 OA kill-hunt). Full text read and stored (Frontiers OA PDF). 50 classes in 14 schools randomly assigned; 761 grade 3–4 pupils analysed (405 experimental, 356 control); 12-session Rainbow Program embedded in regular teaching (Zimmerman/Rosário PLEJE; metacognitive planning, monitoring and evaluation). Combined four subjects: Group × Time post-test d = 0.25 (p = .0006), follow-up d = 0.33 (p < .0001). Per-subject follow-up: science d = 0.31, Spanish d = 0.35, English d = 0.28 (all significant); maths d = 0.11 ns. Kill fails: the later academic gain was not gone. Limits: primary grades only (not gymnasium); teacher-rated non-standardised subject scores; maths alone ns; incidental class sampling in one Spanish region; intervention is SRL package (metacognitive components central, not metacognition-alone). Not a restatement of de Boer 2018 or Eberhart 2025 (new RCT). Not a preprint. Claim sentence and Status unchanged (alive). Bounds and the practical rule updated same turn. C19 now 3 of 3 survivals → floor.

4. #7 narrowed 15 November 2026 · Torgerson, Ainsworth, Bell, Elliott, Fountain, Gascoine, Hewitt, Kasim, Kokotsaki & Torgerson 2018; IPEELL: Evaluation report and executive summary, Education Endowment Foundation, London, November 2018 (independent evaluation, Durham University and University of York; full report PDF from the EEF project page read in full and stored privately) · target: claim-text

   In the largest school-randomised trial on file of a metacognitive strategy programme run as ordinary schools would run it (IPEELL, an English version of Self-Regulated Strategy Development that teaches pupils to plan, monitor, set goals for and evaluate their writing; 84 primary schools, about 5,400 pupils in Years 5–6), pupils did not write better than with usual teaching after one or two years, and after two years they scored lower on national tests in reading, spelling and maths, so C19's unqualified "raises their school performance compared with ordinary teaching" does not hold at scale.

   Strongest reading and verdict reasoning

   Steelmans to: C19 says metacognitive strategy instruction raises school performance against ordinary teaching, without limiting this to research-led delivery; its evidence is meta-analyses (mostly small, often researcher-made tests; Donker et al. 2014 report larger effects on researcher-made tests) and one classroom RCT. A pre-registered, independently evaluated cluster RCT of the flagship metacognitive writing programme, with blinded, independently marked tests and national test records, is the strongest test of whether the gain survives everyday delivery — and its own efficacy trial had found +0.74.

   Why: Nightly-attacks 2026-11-15 Europe/Stockholm (ordinary rotation). Full text: two pragmatic cluster RCTs, schools randomised 1:1 (84 schools; Leeds and Lincolnshire; FSM 33%), teachers trained one day by newly recruited trainers (train-the-trainers) plus a manual and networking; control = business as usual. Trial 1 (one year, Year 6; 2,465 pupils analysed): KS2 writing (teacher-assessed, ordinal) g = −0.09 (−0.34 to 0.15); reading −0.23 (−0.43 to −0.04); maths −0.22 (−0.47 to 0.04); spelling, punctuation and grammar −0.22 (−0.46 to 0.01) (Table 1, Table 11). Trial 2 (two years, Years 5–6; 2,196 analysed; 78 of 83 schools sat the test; about 20% missing): independently administered, blind-marked writing test g = 0.11 (−0.13 to 0.34); low prior attainers 0.26 (−0.01 to 0.53); KS2 reading −0.17 (−0.33 to −0.01), maths −0.30 (−0.53 to −0.07; the text prints −0.02 as the upper limit), spelling, punctuation and grammar −0.28 (−0.49 to −0.06), KS2 writing −0.04 (Table 11). EEF security: very high (Trial 1) and high (Trial 2). Implementation: 59.5% of intervention schools rated good or very good in spring 2016; exploratory CACE for Trial 2 at least good implementation 0.19 (−0.23 to 0.60); very good implementers 0.30 (−0.08 to 0.69). Kill withheld: no follow-up after the programme; one programme in one subject; delivery weakened by inexperienced trainers (the evaluators' own conclusion) and bundled with "memorable experiences"; the authors suggest curriculum time moved from other subjects to writing. Narrowed: C19's gain is shown in research studies; at scale, with cascade training, it did not appear. Not a duplicate; not a restatement of the Bounds (no scale or delivery bound existed). Not a preprint. Status alive → weakened. Claim and Bounds rewritten the same turn; brief and the practical rule rewritten. Kill-match check: the new hedge concedes an end-of-programme null in one scaled-up trial; the kill-condition still names a result the claim does not concede (a follow-up gain gone or below about 0.1 SD in a large trial or meta-analysis) — no kill-mismatch. Fruitless streak 1 → 0.