MétaCan

Constats

Les 22 constats, rendus directement depuis pilot/results/findings.json, le fichier qu'écrivent les scripts du pilote. Aucun nombre de cette page n'a été saisi par un humain : c'est la seule façon de garantir que le site et l'analyse ne peuvent pas diverger.

le même fichier par l'API →

01

L'écart d'affiliation

Dans l'espace thématique de la métarecherche, 508 744 travaux sur 793 883 (64 %) ne portent aucune chaîne d'affiliation brute dans OpenAlex.

Énoncé original (findings.json) : 508,744 of 793,883 works (64%) in the metaresearch topic space have no raw affiliation strings in OpenAlex.

topic space total
793 883
with raw affiliation
285 139
without raw affiliation
508 744
pct without
64.1
affiliation_gap · calculé le 2026-07-12T14:39:42Z
02

La voie thématique

Sur 4 516 thématiques OpenAlex, 0 ne nomme la métarecherche comme domaine. Les 11 thématiques qui portent bel et bien du contenu métarecherche sont dispersées dans 7 domaines OpenAlex différents.

Énoncé original (findings.json) : Of 4,516 OpenAlex topics, 0 name metaresearch as a field. The 11 topics that do carry metaresearch content are scattered across 7 different OpenAlex fields.

n topics in taxonomy
4 516
n topics naming field
0
topics naming field
    n candidate topics
    11
    n fields spanned
    7
    fields spanned
    • Arts and Humanities
    • Computer Science
    • Decision Sciences
    • Mathematics
    • Medicine
    • Psychology
    • Social Sciences
    candidate topic ids
    • T10102
    • T13607
    • T13516
    • T11937
    • T10206
    • T10582
    • T10267
    • T10778
    • T13558
    • T13284
    • T11875
    topics · calculé le 2026-07-12T14:39:41Z
    03

    La polysémie met le lexique en échec

    Le seul terme reproducibility repère 43 392 travaux canadiens, dont seulement 0,8 % relèvent de l'espace thématique de la métarecherche. Le repérage par mots-clés ne peut pas séparer le sens métascientifique du sens courant.

    Énoncé original (findings.json) : The single term reproducibility retrieves 43,392 Canadian works, of which only 0.8% fall in the metaresearch topic space. Keyword retrieval cannot separate the metaresearch sense from the everyday one.

    hits alone
    reproducibility:
    43 392
    "peer review":
    21 234
    "open access":
    6 609
    "open science":
    1 925
    hits alone and on topic
    reproducibility:
    362
    "peer review":
    795
    "open access":
    634
    "open science":
    325
    topic space precision pct
    reproducibility:
    0.8
    "peer review":
    3.7
    "open access":
    9.6
    "open science":
    16.9
    disciplined lexicon hits
    8 026
    worst term
    reproducibility
    polysemy · calculé le 2026-07-12T14:39:43Z
    04

    L'écart linguistique

    Le français représente 2,7 % (395/14 873) de la métarecherche canadienne dans OpenAlex. Un lexique français dédié trouve 168 travaux canadiens, contre 8 026 en anglais.

    Énoncé original (findings.json) : French is 2.7% (395/14,873) of Canadian metaresearch in OpenAlex. A dedicated French lexicon finds 168 Canadian works, against 8,026 in English.

    canadian topic works
    14 873
    n english
    14 028
    n french
    395
    pct french
    2.7
    en lexicon canadian hits
    8 026
    fr lexicon canadian hits
    168
    fr lexicon world hits
    4 682
    canada share of world french
    3.6
    language_gap · calculé le 2026-07-12T14:39:43Z
    05

    Érudit est invisible pour OpenAlex

    Érudit ne correspond à aucune source dans OpenAlex (0), mais son point d'accès OAI-PMH est actif et expose 379 ensembles moissonnables.

    Énoncé original (findings.json) : Erudit matches 0 sources in OpenAlex, but its OAI-PMH endpoint is live and exposes 379 harvestable sets.

    openalex sources matching erudit
    0
    oai endpoint
    https://oai.erudit.org/oai/request
    oai repository name
    Erudit
    oai earliest datestamp
    2011-06-03
    oai harvestable sets
    379
    erudit · calculé le 2026-07-12T14:39:42Z
    06

    La capture-recapture est ici sans valeur

    La capture-recapture naïve à deux voies estime 467 541 travaux canadiens de métarecherche, ce qui impliquerait que le Canada produit 59 % de la métarecherche mondiale contre 1,9 % observé. L'estimateur est ici sans valeur ; nous l'avons supprimé plutôt que de le maquiller en borne inférieure.

    Énoncé original (findings.json) : Naive two-route capture-recapture estimates 467,541 Canadian metaresearch works, implying Canada produces 59% of the world's metaresearch against an observed 1.9%. The estimator is void here; we cut it rather than dress it up as a lower bound.

    route1 topic
    14 873
    route2 naive lexical
    77 583
    overlap
    2 468
    observed union
    89 988
    lincoln petersen estimate
    467 541
    entire topic space all countries
    793 883
    canada observed share pct
    1.9
    canada implied share pct
    58.9
    estimator void
    true
    capture_recapture_fails · calculé le 2026-07-12T14:39:44Z
    07

    Le tri à trois modèles

    Trois modèles de pointe (Opus 4.8, GPT-5.6 high, Grok 4.5) ont trié les mêmes 1 000 travaux de la VRAIE base de 4,3 M, sur la charge utile COMPLÈTE de huit champs de la grille, avec des lots randomisés consignés au manifeste et des étiquettes écrites par le dispositif : chaque défaut relevé par D1, D2, D11 et le constat 16, réparé. Les taux de base pondérés par le plan de sondage s'étendent de 1,89 % à 3,54 % (1,9 fois). Mais ce sont les ensembles qui font le constat, comme le constat 16 le prédisait : des 51 travaux qu'AU MOINS UN modèle a qualifiés de métarecherche, seulement 19 (37 %) l'ont été par LES TROIS, et 24 (47 %) reposent sur l'avis d'un seul modèle. LA FRONTIÈRE DU DOMAINE N'EST PAS UNE LIGNE QUE LES MODÈLES PARTAGENT ; C'EST UNE RÉGION QUE CHACUN DÉCOUPE À SA FAÇON. GPT-5.6 a de plus enfreint le schéma de sortie verrouillé sur 18 des 1 000 notices, en écrivant des valeurs de genre dans le champ du niveau, ce que le validateur de manifeste a détecté. Le livrable n'est pas le taux de base : c'est le dossier des désaccords, les 51 travaux qui marquent la frontière empirique du domaine et sur lesquels les critères d'inclusion doivent réellement être rédigés.

    Énoncé original (findings.json) : Three frontier models (Opus 4.8, GPT-5.6 high, Grok 4.5) screened the same 1,000 works from the REAL 4.3M frame, on the rubric's FULL eight-field payload, with randomized manifest-logged chunks and harness-written labels: every defect D1, D2, D11 and finding 16 identified, repaired. Design-weighted base rates span 1.89% to 3.54% (1.9x). But the sets are the finding, as finding 16 predicted: of the 51 works ANY model called metaresearch, only 19 (37%) were called metaresearch by ALL THREE, and 24 (47%) rest on a single model's opinion. THE FIELD'S BOUNDARY IS NOT A LINE THE MODELS SHARE; IT IS A REGION THEY EACH CUT DIFFERENTLY. GPT-5.6 also violated the locked output schema on 18 of 1,000 records, writing genre values into the tier field, which the manifest validator caught. The deliverable is not the base rate: it is the disagreement dossier, the 51 works that mark the empirical boundary of the field and against which the inclusion criteria must actually be written.

    frame
    the real 4.3M-work Canadian frame (all 482 OpenAlex partitions)
    payload
    the rubric's FULL eight fields, including venue (repairs D1)
    sample
    1,000 works, stratified with known selection probabilities, French oversampled
    models
    Claude Opus 4.8; GPT-5.6 (high effort); Grok 4.5 (medium effort)
    harness
    chunks randomized and manifest-logged before any model ran; the harness writes label files, never the model (repairs D11); every arm reconciled against the manifest (repairs D2)
    n labelled by all three
    1 000
    base rate weighted pct
    opus:
    3.54
    gpt:
    2.51
    grok:
    1.89
    between model spread x
    1.9
    jaccard opus gpt
    52
    jaccard opus grok
    47
    jaccard gpt grok
    53
    called in scope by any
    51
    unanimous in scope
    19
    pct unanimous of any
    37
    in scope by one model only
    24
    pct single model of any
    47
    contested by stratum
    aff_core:
    87
    venue_new:
    67
    about_only:
    50
    french:
    44
    fund_new:
    40
    tier disagreement patterns
    OUT/T2:
    15
    T1:
    11
    OUT/T1:
    10
    T2:
    7
    OUT/T1/T2:
    3
    T1/T3:
    2
    OUT/T1/T3:
    1
    T1/T2:
    1
    T2/T3:
    1
    gpt schema violations first pass
    18
    gpt violation note
    GPT-5.6 (high) wrote GENRE values ('empirical', 'conceptual') into the TIER field on 18 of 1,000 records in its first pass, in 3 of 20 chunks. The validator caught it because the harness reconciles files against a manifest rather than trusting the model's report. Those chunks were RE-RUN, not repaired: coercing a model's output to the schema is fitting the instrument to the data.
    deliverable
    pilot/screening/frame1k/disagreement_dossier.json: every work any model called in-scope, with all three labels. This, not the base rate, is what the criteria must be written against.
    caveat
    These are MACHINE labels and none of them is truth (finding 15). The unanimity rate is not accuracy: three models sharing training data can be wrong together, and they are most correlated exactly on the boundary cases the field's definition turns on. What this measures is where the RUBRIC is underspecified, which is a property of the instrument and is exactly what a criteria document needs. Base rates are design-weighted from a stratified sample, so they estimate the frame; the Jaccard and unanimity figures are unweighted set quantities over the sample and are NOT frame estimates.
    three_model_screen · calculé le 2026-07-12T19:52:14Z
    08

    Changez de trieur, la réponse bouge

    Changez le modèle qu'on appelle « le trieur » et le taux de base passe de 1,06 % à 2,37 % : un écart de 2,2 fois, soit de 37 032 à 83 022 travaux dans la base. Les deux trieurs s'accordent sur inclus/exclu pour 98,4 % de la base (pondération du plan de sondage), mais ce chiffre est dominé par les rejets évidents : l'accord tombe à 95 % à l'intérieur de la frontière contestée. L'ÉTENDUE OBTENUE EN CHANGEANT DE TRIEUR, ET NON L'IC BINOMIAL D'UN SEUL MODÈLE, EST L'INCERTITUDE HONNÊTE SUR LA TAILLE DU DOMAINE.

    Énoncé original (findings.json) : Swap which model is called 'the screener' and the base rate moves from 1.06% to 2.37%: a 2.2x spread, from 37,032 to 83,022 works in the frame. The two screeners agree on in/out for 98.4% of the frame (design-weighted), but that figure is dominated by the settled rejects: agreement falls to 95% inside the contested boundary. THE SCREENER-SWAP RANGE, NOT THE BINOMIAL CI ON EITHER MODEL ALONE, IS THE HONEST UNCERTAINTY ON THE FIELD'S SIZE.

    n double screened
    1 290
    base rate screener a pct
    1.06
    base rate screener b pct
    2.37
    swap ratio x
    2.24
    field size screener a
    37 032
    field size screener b
    83 022
    published binomial ci contains b
    false
    screener a
    claude-sonnet-4-6 (40 agents, medium effort)
    screener b
    gpt-5.6-sol (codex)
    sampling
    stratified on screener A's label; positives/paratext/boundary taken whole, settled-OUT sampled
    raw agreement inout pct
    96.6
    weighted agreement pct
    98.4
    cohens kappa inout
    0.681
    n disagree inout
    44
    gpt in claude out
    37
    claude in gpt out
    7
    agreement by stratum
    stratum:
    settled_out,boundary,positive,paratext
    n:
    600,599,58,33
    sel_prob:
    0.124921923797626,1,1,1
    agree_pct:
    99,95,87.9,97
    a_says_in:
    0,0,58,0
    b_says_in:
    6,30,51,1
    caveat
    PROCESS METRIC, NOT ACCURACY. Two LLMs share training data and failure modes; their errors are correlated and most correlated on the boundary. This is not 'duplicate screening': that term's warrant comes from independent human judgement. Accuracy rests on the human-coded probability sample (PROTOCOL s6.2). Agreement is reported per stratum because a pooled kappa on a 1.3% base rate is dominated by the cell where agreement is free (the kappa paradox). And note what the high agreement figure CONCEALS: it is dominated by the settled-OUT mass, while the two screeners imply base rates a factor of two apart. Quoting agreement without the swap would be presenting the reassuring statistic.
    agreement · calculé le 2026-07-12T14:39:46Z
    09

    Le taux de base

    Le tri de 5 737 travaux canadiens non filtrés selon la grille situe la métarecherche à 1,31 % de la recherche canadienne, soit environ 45 850 travaux dans la base de 3,5 M, ce qui dimensionne le domaine sans aucune stratégie de recherche. L'IC à 95 % sur les étiquettes de ce trieur va de 1,03 à 1,64 %, mais c'est de l'erreur d'échantillonnage, PAS l'incertitude : changez de trieur et l'estimation tombe hors de l'intervalle (constat 10). Le domaine se situe quelque part entre 37 000 et 83 000 travaux, et seul l'audit humain peut resserrer cela.

    Énoncé original (findings.json) : Screening 5,737 unfiltered Canadian works against the rubric puts metaresearch at 1.31% of Canadian research, implying ~45,850 works in the 3.5M-work frame, sizing the field without a search strategy at all. The 95% CI on this screener's labels is 1.03-1.64%, but that is sampling error, NOT the uncertainty: swap the screener and the estimate lands outside it (finding 10). The field is somewhere between 37,000 and 83,000 works, and only the human audit can narrow that.

    n screened
    5 737
    screener
    claude-sonnet-4-6, 40 agents, medium effort, locked rubric
    sampling frame
    unfiltered Canadian works from the pinned 2026-06-24 snapshot partition
    tier counts
    OUT:
    5 621
    T1:
    40
    T2:
    35
    T3:
    41
    n in scope t1 t2
    75
    base rate pct
    1.31
    base rate ci lo pct
    1.03
    base rate ci hi pct
    1.64
    canadian frame size
    3 507 205
    estimated field size
    45 850
    estimated field lo
    36 111
    estimated field hi
    57 378
    topic route retrieved
    14 873
    binomial ci is not the uncertainty
    true
    caveat
    Machine labels, not a human gold standard, and the binomial CI above is sampling error on ONE screener; the honest uncertainty is the screener-swap range in finding 10 (1.06% to 2.37%). The partition is also not a uniform draw: it under-represents works with abstracts, where the screen finds 2x more metaresearch (finding 11). This is a hypothesis with a denominator; the human-coded probability sample tests it. Do NOT divide topic_route_retrieved by estimated_field_size; see finding 12.
    base_rate · calculé le 2026-07-12T14:39:45Z
    10

    La robustesse du taux de base

    Le vrai biais du taux de base n'est pas la récence mais les RÉSUMÉS MANQUANTS : 31,5 % de la partition n'en a aucun, et le tri y repère 0,78 % de métarecherche contre 1,55 % là où un résumé existe (khi carré p = 0,023, robuste à l'ajustement pour l'année et la langue). Le tiers de la base est trié sur son seul titre. Ce que cela ne montre PAS, et qu'une version antérieure affirmait à tort, c'est que l'aveuglement serait DIFFÉRENTIEL selon la tradition : la case T2 sans résumé contient 4 travaux et l'interaction n'est pas significative (p = 0,141). Cette affirmation est retirée, comme l'est « le portrait exact d'Érudit » (la strate est anglophone à 99 % et les travaux sont plus RÉCENTS, non plus anciens). L'effet principal est le constat.

    Énoncé original (findings.json) : The base rate's real bias is not recency but MISSING ABSTRACTS: 31.5% of the partition has none, and the screen finds 0.78% metaresearch there against 1.55% where an abstract exists (chi-square p = 0.023, robust to adjustment for year and language). A third of the frame is screened on its title alone. What this does NOT show, and an earlier draft wrongly claimed, is that the blindness is DIFFERENTIAL by tradition: the T2 no-abstract cell holds 4 works and the interaction is not significant (p = 0.141). That claim is withdrawn, as is 'Erudit's exact profile' (the stratum is 99% English and the works are NEWER, not older). The main effect is the finding.

    partition
    updated_date=2026-06-24
    records sent to screener
    6 202
    records silently lost
    465
    records lost pct
    7.5
    pct no abstract among lost
    40.2
    pct no abstract among labelled
    31.5
    chisq p loss bias
    0.00013
    losses are biased
    true
    works 2000 09
    2 704
    works 2020 25
    1 195
    old to new ratio
    2.3
    partition skews old
    true
    base rate by era pct
    2000-09:
    1.24
    2010-19:
    1.35
    2020-25:
    1.4
    chisq p era
    0.906
    era events
    75
    era check is underpowered
    true
    no abstract share pct
    31.5
    base rate no abstract pct
    0.78
    base rate has abstract pct
    1.55
    chisq p abstract
    0.023
    abstract effect x
    2
    t1 penalty x
    1.4
    t2 penalty x
    3.6
    t2 no abstract cell count
    4
    interaction p
    0.141
    differential is supported
    false
    differential claim withdrawn
    true
    no abstract mean year
    2013.7
    has abstract mean year
    2010.3
    no abstract stratum pct english
    99
    p no abstract given english pct
    32
    p no abstract given non english pct
    12.7
    erudit profile claim holds
    false
    residual bias direction
    anti-conservative for coverage claims: the partition over-represents abstract-less works (31.5%), where the screen finds 2x less metaresearch, so 1.31% likely UNDER-states the frame's base rate, and the field is larger than the headline implies
    supersedes
    THREE retractions live here. (1) An earlier version tested ERA only, called the base rate robust, and published 3100/2900/1500 as percentages (a dplyr summarise() column-masking bug). (2) It then claimed the blindness is DIFFERENTIAL, T2 losing 3.6x against T1's 1.4x, and made that the proposal's whole answer to the inclusiveness criterion. The T2 no-abstract cell holds FOUR works and the interaction is not significant (p = 0.141). WITHDRAWN. (3) It claimed missing abstracts track 'older, non-English' records, 'Erudit's exact profile'. Backwards: they are NEWER, and the stratum is 99% ENGLISH. WITHDRAWN. See DEVIATIONS.md D4, D5, D6.
    caveat
    What survives is the MAIN EFFECT and only the main effect: a third of the frame is screened on its title alone and the screen finds half as much metaresearch there (p = 0.023, robust to adjustment for year and language). That is a real coverage problem and a reason to stratify the audit on abstract availability. It is NOT evidence of differential blindness by tradition, and this finding no longer says it is. Separately, the era check is underpowered (75 events, 3 strata) and cannot refute an era effect; it merely fails to detect one. And note D2: the harness silently dropped 465 records, non-randomly, on this very covariate.
    base_rate_robustness · calculé le 2026-07-12T14:39:47Z
    11

    Ce que les étiquettes ne peuvent pas dire

    Deux limites des étiquettes machines, trouvées en attaquant les correctifs. (A) Le rappel de la voie thématique est de 12 % selon le trieur A et de 7 % selon le trieur B : l'instrument (ii) du constat 14 évalue les filtres sur des étiquettes de MACHINE, il mesure donc l'accord avec une machine et non l'exactitude, et le constat 10 avait déjà montré que cela varie du simple au double. La conclusion se renforce (le second trieur juge la voie encore PIRE), mais 12 % n'est pas la vérité. (B) Le pilote contient exactement 1 travail francophone dans le champ ; en voir 20 exigerait environ 1 620 notices françaises codées contre un budget d'audit de 1 000. LA SENSIBILITÉ POUR LE FRANÇAIS N'EST PAS ESTIMABLE DANS UNE BASE FONDÉE SUR LE SEUL OPENALEX. C'est un argument pour le moissonnage d'Érudit, non contre la revendication du français, mais le pilote n'a pas exécuté ce moissonnage : la puissance est donc énoncée comme une condition plutôt que comme une promesse.

    Énoncé original (findings.json) : Two limits on the machine labels, found by attacking the fixes. (A) The topic route's recall is 12% against screener A and 7% against screener B: finding 14's instrument (ii) scores filters against MACHINE labels, so it measures agreement with a machine, not accuracy, and finding 10 already showed that swings by a factor of two. The conclusion strengthens (the second screener thinks the route is WORSE) but 12% is not truth. (B) The pilot holds exactly 1 French in-scope work, so seeing 20 French positives needs ~1,620 coded French records against an audit budget of 1,000. FRENCH SENSITIVITY IS NOT ESTIMABLE IN AN OPENALEX-ONLY FRAME. That is an argument for the Erudit harvest, not against the French claim, but the pilot did not run that harvest, so the power is stated as a condition rather than a promise.

    route recall vs screener a pct
    12
    route recall vs screener b pct
    7
    route recall a ci
    • 5.6
    • 21.6
    route recall b ci
    • 2.5
    • 14.3
    positives screener a
    75
    positives screener b
    88
    instrument ii is model dependent
    true
    instrument ii caveat
    Finding 14's instrument (ii) scores a filter against the 5,737 MACHINE labels. That measures agreement with a machine, not accuracy. Swap the machine and the topic route's recall moves from 12% to 7%. The conclusion (the route finds a small fraction) survives and strengthens; the NUMBER is not a measurement against truth.
    french records in pilot
    81
    french in scope in pilot
    1
    french records needed for 20 positives
    1 620
    audit budget records
    1 000
    french stratum is powered
    false
    french power depends on
    the Erudit harvest, which the pilot did NOT run (it verified the endpoint: 379 live sets)
    caveat
    (A) is a limit on every recall number this project quotes against machine labels, including its own headline. (B) is a limit on the inclusiveness promise: French sensitivity cannot be estimated in an OpenAlex-only frame, because Erudit matches zero OpenAlex sources and the francophone literature is therefore largely absent from the frame rather than merely sparse in it. Both are stated in the proposal rather than left for a reviewer.
    label_limits · calculé le 2026-07-12T14:39:49Z
    12

    La variance des agents

    Haiku, le modèle sur lequel le constat 13 budgète tout le tri, aboutit près du taux de base de Sonnet (1,27 % contre 1,06 %) et s'accorde avec lui sur 98,1 % de la base, mais leurs ENSEMBLES dans le champ ne se recoupent qu'à 16 % sans pondération et à 10 % avec la pondération du plan de sondage (Sonnet-GPT : 54 %/37 %) ; des 58 positifs de Sonnet, Haiku n'en confirme que 12. L'ACCORD SUR LE TAUX N'EST PAS L'ACCORD SUR L'ENSEMBLE. Pire : des agents du MÊME modèle sur la MÊME consigne divergent au-delà du hasard dans les DEUX volets après conditionnement sur la strate (CMH p = 0,0056 et 0,015), avec des écarts bruts de 3,1 et 5,2 fois contre 2,2 fois entre modèles, et l'ordre des agents S'INVERSE d'un volet à l'autre. Le bruit à l'intérieur d'un même modèle est au moins de la taille de la différence entre modèles, et le tri à 40 agents du pilote manque lui-même de puissance pour exclure la même instabilité (5 des 37 lots n'ont trouvé aucune métarecherche ; p = 0,113).

    Énoncé original (findings.json) : Haiku, the model finding 13 budgets the entire screen on, lands near Sonnet's base rate (1.27% vs 1.06%) and agrees with it on 98.1% of the frame, but their in-scope SETS overlap 16% unweighted and 10% design-weighted (Sonnet-GPT: 54%/37%); of Sonnet's 58 positives Haiku agrees on 12. RATE AGREEMENT IS NOT SET AGREEMENT. Worse: agents of the SAME model on the SAME prompt disagree beyond chance in BOTH arms after conditioning on stratum (CMH p = 0.0056 and 0.015), with raw spreads 3.1x and 5.2x against 2.2x between models, and the agents' ordering FLIPS between arms. The noise inside one model is at least the size of the difference between models, and the pilot's own 40-agent screen is too underpowered to rule the same instability out (5 of 37 chunks found zero metaresearch; p = 0.113).

    n works
    1 290
    base rate sonnet pct
    1.06
    base rate gpt pct
    2.37
    base rate haiku pct
    1.27
    agreement haiku sonnet pct
    98.1
    jaccard sonnet gpt pct
    54
    jaccard sonnet haiku pct
    16
    jaccard gpt haiku pct
    12
    wjaccard sonnet gpt pct
    37
    wjaccard sonnet haiku pct
    10
    wjaccard gpt haiku pct
    6
    sonnet positives
    58
    haiku agrees on
    12
    agent rates raw pct
    agent-1:
    1.6
    agent-2:
    1.25
    agent-3:
    3.85
    stratum mix differs by agent p
    1.34e-17
    arm1 cmh p
    0.00564
    arm1 permutation p
    0.0275
    arm1 raw spread x
    3.1
    arm2 cmh p
    0.0152
    arm2 permutation p
    0.0075
    arm2 raw spread x
    5.2
    agent order replicates
    false
    spread weighted x leverage sensitive
    13.2
    between model spread x
    2.2
    within at least matches between
    true
    pilot chunks
    37
    pilot agents finding zero
    5
    pilot between agent p
    0.113
    pilot underpowered not homogeneous
    true
    caveat
    The first draft's between-agent test was CONFOUNDED: the stratum mix differs by agent (p = 1.3e-17), and the draft asserted a verification that did not exist (DEVIATIONS.md D13). The tests above condition on stratum, and the heterogeneity survives in both arms. The 13.2x design-weighted spread the draft led with rests on five high-weight events and is demoted to a recorded, leverage-sensitive descriptive. THE INFERENCE IS SCOPED: there are three agents per arm, assigned consecutive chunk blocks without randomisation or a run-time manifest, so these p-values license 'these runs are not exchangeable', not a population claim about agents in general; that is exactly enough to break a budget that assumed exchangeability, and the full study assigns agents randomised, manifest-logged, fixed-size chunks with a duplicate-agent reliability arm. The pilot's own agents give p = 0.113 on ~2 expected events per chunk: underpowered, so the pilot is uninformative on agent homogeneity, not exonerated. Haiku was tested on the six-field payload (the pilot's own deviation, D1) and on a guided eight-field arm, so the defensible conclusion is 'not shown to be an interchangeable measurer, and unstable in the arms tested', not 'cannot screen'. An eight-field neutral-prompt arm is not used at all: one agent claimed six label files it never wrote (D11), so no payload effect is reported from any arm. The guided arm is used ONLY for the between-agent contrast, which its shared prompt leaves internally valid.
    agent_variance · calculé le 2026-07-12T14:39:58Z
    13

    L'écart des résumés est structurel

    Le plus grand biais mesuré du tri est l'écart des résumés : 23,3 % de la base (1 003 117 travaux) n'a AUCUN RÉSUMÉ, et le constat 11 a montré que le tri y repère MOITIÉ moins de métarecherche. La cascade PubMed, Europe PMC puis Crossref récupère 37,8 % d'un échantillon de 500 travaux, ramenant l'exposition au tri sur seul titre à environ 14,5 % de la base. Mais J'AVAIS BÂTI LA CASCADE AUTOUR DE CROSSREF comme voie de secours indépendante des disciplines, et il a récupéré 2 résumés contre 180 pour PubMed : les éditeurs ne les déposent pas, donc CETTE VOIE DE SECOURS N'EXISTE PAS (D15). L'écart n'est donc pas une défaillance de métadonnées qu'un meilleur index corrigerait ; il est STRUCTUREL. La récupération atteint 91,2 % pour les articles de synthèse contre 6,2 % pour les chapitres de livre, 38,8 % pour l'anglais contre 15,4 % pour le français. Le raccourci tentant, « ne trier que les travaux qui ont un résumé », est donc une SÉLECTION SUR UNE COVARIABLE QUI PRÉDIT LE RÉSULTAT : il supprimerait 61,6 % des chapitres de livre contre 22,1 % des articles, ET les travaux qu'il supprime sont exactement ceux qu'aucune cascade ne peut récupérer. Défendable seulement comme exclusion DÉCLARÉE au coût mesuré, et l'audit y conserve un plancher d'échantillonnage.

    Énoncé original (findings.json) : The screen's largest measured bias is the abstract gap: 23.3% of the frame (1,003,117 works) has NO ABSTRACT, and finding 11 showed the screen finds HALF as much metaresearch there. Cascading PubMed, Europe PMC and Crossref recovers 37.8% of a 500-work sample, cutting title-only exposure to ~14.5% of the frame. But I BUILT THE CASCADE AROUND CROSSREF as the discipline-agnostic rescue, and it recovered 2 abstracts against PubMed's 180: publishers do not deposit them, so THAT RESCUE DOES NOT EXIST (D15). The gap is therefore not a metadata failure a better index fixes; it is STRUCTURAL. Recovery is 91.2% for reviews against 6.2% for book chapters, 38.8% English against 15.4% French. So the tempting shortcut, 'just screen the works that have abstracts', is a SELECTION ON A COVARIATE THAT PREDICTS THE OUTCOME which would delete 61.6% of book chapters against 22.1% of articles, AND the works it deletes are exactly the works no cascade can rescue. Defensible only as a DECLARED exclusion with a measured cost, and the audit keeps a sampling floor in it.

    frame works
    4 299 418
    frame works no abstract
    1 003 117
    pct frame no abstract
    23.3
    pct dropped by type
    book-chapter:
    61.6
    letter:
    52.1
    editorial:
    43.3
    review:
    29.5
    article:
    22.1
    book:
    21.6
    other:
    21.1
    report:
    18.6
    preprint:
    14
    dataset:
    8.2
    dissertation:
    4.3
    pct dropped by language
    en:
    23.7
    fr:
    21.6
    abstracts only is a selection on the outcome
    true
    sampled
    500
    sources
    PubMed (Entrez) -> Europe PMC (REST) -> Crossref (REST)
    recovered
    189
    pct gap recovered
    37.8
    recovered by source
    pubmed:
    180
    europepmc:
    7
    crossref:
    2
    hypothesis crossref would be load bearing
    false
    crossref recovered
    2
    europepmc recovered
    7
    pubmed recovered
    180
    no discipline agnostic rescue exists
    true
    the gap is structural not a metadata failure
    true
    recovery pct by type
    review:
    91.2
    article:
    40.5
    preprint:
    38.5
    book-chapter:
    6.2
    letter:
    0
    recovery pct english
    38.8
    recovery pct french
    15.4
    residual pct frame title only
    14.5
    caveat
    Run on a 500-work hash-ordered sample of the no-abstract stratum, not the frame: the cascade is rate-limited and a million lookups is days. The recovery rate is an estimate with sampling error, and it is an estimate of a CEILING (an abstract that EXISTS is recoverable; it does not follow the screen then classifies the work correctly). Only works with a DOI can be looked up, so the DOI-less part of the stratum is untouched and its size bounds what any cascade can do. PubMed and Europe PMC are biomedical; Crossref is not, and it is in the chain for exactly that reason: a cascade of biomedical indexes would close the gap unevenly and make the residual bias MORE discipline-shaped while appearing to improve coverage. That reasoning was right and the remedy is not available: Crossref recovered 2 of 189 and Europe PMC 7, because publishers largely do not deposit abstracts to Crossref, so no discipline-agnostic rescue exists (D15). Restricting screening to abstract-bearing works remains a DECLARED EXCLUSION with a measured cost, not a scoping convenience, and the audit keeps a sampling floor in the excluded stratum so that cost stays estimable.
    abstract_cascade · calculé le 2026-07-12T19:25:50Z
    14

    Un booléen sur un espace à quatre états

    Jointe à la base canadienne par DOI, Retraction Watch consigne 143 travaux qu'OpenAlex ne signale PAS comme rétractés, dont 49 rétractations pures et simples. Mais le sous-compte est le moindre des problèmes. 52 de ces travaux portent une EXPRESSION DE PRÉOCCUPATION, et OpenAlex N'A AUCUN CHAMP POUR CELA : is_retracted est un booléen sur un espace d'états à au moins quatre valeurs (rétractation, expression de préoccupation, correction, rétablissement) ; il peut en exprimer une et rapporte silencieusement les autres comme FALSE, ce qui se lit comme « rien à signaler ». Un booléen ne peut pas non plus porter le POURQUOI. C'est la maladie du constat 1 dans un second schéma : la base de données canonique ne peut pas exprimer la distinction sur laquelle le domaine repose.

    Énoncé original (findings.json) : Joined to the Canadian frame by DOI, Retraction Watch records 143 works that OpenAlex does NOT flag as retracted, including 49 outright retractions. But the undercount is the smaller problem. 52 of these carry an EXPRESSION OF CONCERN, and OpenAlex HAS NO FIELD FOR ONE: `is_retracted` is a boolean over a state space with at least four values (retraction, expression of concern, correction, reinstatement), so it can express one and silently reports the rest as FALSE, which reads as 'fine'. Nor can a boolean carry WHY. This is finding 1's disease in a second schema: the canonical database cannot express the distinction the field turns on.

    source
    Retraction Watch (Crossref-licensed), joined by bare lowercased DOI
    frame works with doi
    3 690 953
    openalex is retracted flags
    1 584
    matched in retraction watch
    1 052
    openalex misses
    143
    outright retractions missed
    49
    expressions of concern
    52
    missed by nature
    Expression of concern:
    52
    Retraction:
    49
    Correction:
    32
    Reinstatement:
    10
    top reasons
    Investigation by Journal/Publisher:
    335
    Unreliable Results and/or Conclusions:
    259
    Concerns/Issues about Data:
    233
    Investigation by Third Party:
    169
    Concerns/Issues about Referencing/Attributions:
    151
    Concerns/Issues about Results and/or Conclusions:
    122
    Concerns/Issues about Peer Review:
    109
    Investigation by Company/Institution:
    101
    openalex has eoc field
    false
    is a frame route
    false
    caveat
    This is an ATTRIBUTE, not a frame route. A retracted cardiology paper is retracted cardiology, not metaresearch, and admitting works on the strength of a retraction would let an interesting signal masquerade as the estimand. The DOI join can only see works that HAVE a DOI, and OpenAlex flags some works Retraction Watch does not match, which may be DOI drift rather than disagreement; the asymmetry reported here is one-directional on purpose (what RW adds), because that is the direction the join can support.
    retraction_record · calculé le 2026-07-12T18:52:41Z
    15

    Le rappel de la voie thématique

    Évaluée selon la grille, la voie thématique repère 12 % de la métarecherche canadienne (IC à 95 % de 5,6 à 21,6 %) avec une précision de 60 % : elle manque 66 travaux sur 75. Elle échoue parce qu'OpenAlex classe un travail selon ce dont il traite, et la métarecherche sur la cardiologie se lit comme de la cardiologie : le domaine est invisible au repérage thématique précisément parce qu'il porte sur d'autres domaines.

    Énoncé original (findings.json) : Scored against the rubric, the topic route retrieves 12% of Canadian metaresearch (95% CI 5.6-21.6%) at 60% precision: it misses 66 of 75. It fails because OpenAlex files a work by what it is about, and metaresearch about cardiology reads as cardiology: the field is invisible to topic retrieval precisely because it is about other fields.

    n screened
    5 737
    n metaresearch
    75
    n retrieved by route
    15
    true positives
    9
    false positives
    6
    false negatives
    66
    recall pct
    12
    recall ci lo pct
    5.6
    recall ci hi pct
    21.6
    precision pct
    60
    precision ci lo pct
    32.3
    precision ci hi pct
    83.7
    missed works by field
    Social Sciences:
    17
    Medicine:
    16
    Health Professions:
    7
    Business, Management and Accounting:
    4
    Economics, Econometrics and Finance:
    4
    Computer Science:
    3
    scored by
    primary_topic.id (the key R/frame.R defines the route with), not display name
    route size implied by sample
    9 170
    route size from api
    14 873
    reconciliation gap x
    1.62
    supersedes
    the earlier 32.4% figure (14,873/45,850), which divided a retrieved set by a true field size
    caveat
    Small n on the positive class ( 75 metaresearch works, of which 15 were on the route), so the intervals are wide. Separately, the sample-implied route size does not reconcile with the API's 14,873 and I cannot say why at n = 15 on-route works; the likeliest cause is that one updated_date partition is not a uniform draw (finding 11). Recall is unaffected: it is a within-sample ratio, not an extrapolation.
    topic_route_recall · calculé le 2026-07-12T14:39:47Z
    16

    Le rappel de la voie du financement

    L'ESTIMANDE PRINCIPALE s'appuie sur CA-FUND pour récupérer les travaux dont l'affiliation manque (constat 2 : 64 % n'ont aucune chaîne d'affiliation), et rien ne l'avait mise à l'épreuve. Confrontée à la base de données des IRSC eux-mêmes, soit 44 190 projets financés, l'épreuve A RÉFUTÉ L'HYPOTHÈSE QUE J'AVAIS ÉCRITE AVANT DE L'EXÉCUTER : OpenAlex étiquette 178 133 travaux de la base avec les IRSC, soit 4,03 par subvention, un taux plausible sans sous-étiquetage (DEVIATIONS.md D14). Ce que les données soutiennent, elles, n'exige aucune hypothèse de ma part : 71,2 % DE LA BASE NE PORTE AUCUNE MÉTADONNÉE DE FINANCEMENT, ce qui est le plafond absolu de CA-FUND, et 65,9 % des travaux à affiliation canadienne n'en portent pas non plus. Les deux clauses de l'estimande reposent sur des métadonnées le plus souvent absentes : c'est pourquoi la base est l'union de quatre voies, et pourquoi l'audit doit échantillonner les travaux qu'aucune voie n'a atteints.

    Énoncé original (findings.json) : The PRIMARY ESTIMAND leans on CA-FUND to rescue works whose affiliation is missing (finding 2: 64% have no affiliation string), and nothing had tested it. Tested against CIHR's own database of 44,190 funded projects, the result REFUTED THE HYPOTHESIS I WROTE BEFORE RUNNING IT: OpenAlex tags 178,133 frame works with CIHR, or 4.03 per grant, a plausible rate showing no under-tagging (DEVIATIONS.md D14). What the data DOES support needs no hypothesis of mine: 71.2% OF THE FRAME CARRIES NO FUNDER METADATA AT ALL, which is CA-FUND's hard ceiling, and 65.9% of Canadian-AFFILIATED works carry none either. Both clauses of the estimand rest on metadata that is mostly absent, which is why the frame is a union of four routes and why the audit must sample the works no route reached.

    external criterion
    CIHR's own project database (44,190 projects), which owes nothing to OpenAlex
    cihr projects
    44 190
    cihr distinct pis
    27 536
    cihr funder id
    https://openalex.org/F4320334506
    frame works
    4 299 418
    frame works tagged cihr
    178 133
    tagged publications per funded project
    4.03
    papers per grant is a plausible rate not a defect
    true
    expectation i wrote before running and that was false
    that OpenAlex under-tags CIHR so badly it falls below one paper per grant. It is 4.03 per grant, a plausible rate. See DEVIATIONS.md D14.
    works ca fund rescues alone
    166 743
    frame works with any funder
    1 239 950
    pct frame with any funder
    28.8
    pct frame with no funder
    71.2
    ca aff works
    2 734 192
    ca aff works with no funder
    1 802 605
    pct ca aff with no funder
    65.9
    top recorded funders
    Natural Sciences and Engineering Research Council of Canada:
    294 401
    Canadian Institutes of Health Research:
    178 133
    National Institutes of Health:
    81 262
    National Natural Science Foundation of China:
    67 358
    National Science Foundation:
    62 497
    Canada Research Chairs:
    35 113
    European Commission:
    34 039
    Social Sciences and Humanities Research Council of Canada:
    33 473
    is an aggregate not a record level join
    true
    caveat
    CIHR's CSV carries no DOIs and no publication links, so this is an AGGREGATE reconciliation, not a record-level known-item join, and NO RECALL POINT ESTIMATE is claimed. It establishes a CEILING on CA-FUND (the route cannot see a funder OpenAlex never recorded), which is a bound, not a measurement. The papers-per-grant ratio is reported because I ran it, and it REFUTES the hypothesis I wrote before running it: at 4.03 per grant it is a plausible publication rate and shows no CIHR under-tagging at all. It is also the wrong instrument, for the same reason the retracted 32.4% coverage figure was (D3): its numerator and denominator are not linked record to record, so the quotient has no estimand behind it. Record-level linkage is finding 21.
    funder_route_recall · calculé le 2026-07-12T18:56:48Z
    17

    Le lien canadien

    L'affiliation trouve 14 873 travaux ; 3 964 autres portent sur le Canada sans affiliation canadienne. Le CRSNG a 9,2 fois plus de travaux liés que le CRSH : les règles fondées sur le financement sous-comptent donc les sciences sociales.

    Énoncé original (findings.json) : Affiliation finds 14,873 works; a further 3,964 are about Canada with no Canadian affiliation. NSERC has 9.2x SSHRC's linked works, so funder-based rules under-count the social sciences.

    by affiliation
    14 873
    by funder
    1 331
    about canada
    5 704
    about canada no affiliation
    3 964
    funder works
    Canadian Institutes of Health Research:
    194 681
    Natural Sciences and Engineering Research Council of Canada:
    433 090
    Social Sciences and Humanities Research Council of Canada:
    46 913
    Canada Foundation for Innovation:
    16 753
    sshrc works
    46 913
    nserc works
    433 090
    nserc to sshrc ratio
    9.2
    affiliation noise
    University of London:
    494
    Impact:
    475
    canadian_linkage · calculé le 2026-07-12T14:39:44Z
    18

    Le lien aux essais cliniques

    Le registre est le seul ÉTALON DE RÉFÉRENCE de ce projet qui ne soit pas fait d'étiquettes de machine : ClinicalTrials.gov sait qu'un essai canadien a eu lieu indépendamment de toute chaîne de traitement, il ne peut donc pas se tromper en faveur de la chaîne. Des 304 publications que les PROMOTEURS EUX-MÊMES ont déclarées comme résultats d'essais achevés menés au Canada, la base en détient 160 : un rappel naïf de 52,6 %, tombé si près du 44,5 % PAR LEQUEL CETTE PROPOSITION S'OUVRE qu'il se lisait comme une réplication. C'EST UN ARTEFACT, et seule l'exécution de la désambiguïsation l'a détecté : 137 des 145 « manqués » n'ont AUCUN AUTEUR CANADIEN (essais internationaux multicentriques avec un SITE canadien), et une base de la RECHERCHE canadienne a RAISON de les exclure. Un essai avec un site canadien n'est pas une publication avec un auteur canadien. Contre la population que la base revendique réellement, le rappel est de 95,2 % (IC à 95 % de 90,8 à 97,9), et le vrai défaut tient aux 8 travaux qu'OpenAlex détient AVEC un auteur canadien et que les voies de la base ont tout de même manqués. La base est BONNE à cet exercice, le parallèle spectaculaire était une coïncidence entre deux populations différentes, et j'avais toutes les raisons de ne pas vérifier. DEVIATIONS.md D16.

    Énoncé original (findings.json) : The registry is the only REFERENCE STANDARD in this project not made of machine labels: ClinicalTrials.gov knows a Canadian trial happened independently of any pipeline, so it cannot be wrong in the pipeline's favour. Of 304 publications SPONSORS THEMSELVES reported as results of completed Canadian-located trials, the frame holds 160: a naive recall of 52.6%, which fell so close to the 44.5% THIS PROPOSAL OPENS WITH that it read as a replication. IT IS AN ARTIFACT, and running the disambiguation is the only thing that caught it: 137 of the 145 'misses' have NO CANADIAN AUTHOR (multi-site international trials with a Canadian SITE), and a frame of Canadian RESEARCH is CORRECT to exclude them. A trial with a Canadian site is not a publication with a Canadian author. Against the population the frame actually claims, recall is 95.2% (95% CI 90.8-97.9), and the real defect is 8 works OpenAlex holds WITH a Canadian author that the frame's own routes still missed. The frame is GOOD at this, the dramatic parallel was a coincidence between two different populations, and I had every incentive not to check. DEVIATIONS.md D16.

    reference standard
    ClinicalTrials.gov: completed trials with a Canadian location, and the RESULT publications the sponsors themselves reported
    why not a frame route
    A trial registration is not a publication and not metaresearch. Registrations contribute NO records to the frame; the registry is a reference standard, not a source.
    why it matters
    Every other recall number in this project is scored against MACHINE labels (finding 15). A registry knows a trial happened independently of any pipeline, so it cannot be wrong in the pipeline's favour. This is the only instrument here that measures the frame against a world that exists without it.
    trials retrieved
    1 000
    trials with result publication
    123
    pct trials with result publication
    12.3
    known result pmids
    596
    resolvable to doi
    304
    present in frame
    160
    naive frame recall pct
    52.6
    naive recall is an artifact
    true
    naive recall ci
    • 46.9
    • 58.4
    missing total
    145
    missing no canadian author
    137
    missing route gap canadian author
    8
    missing not in openalex
    0
    claimable population
    168
    adjusted recall pct
    95.2
    adjusted recall ci
    • 90.8
    • 97.9
    is metaresearch recall
    false
    caveat
    This measures FRAME recall (does the Canadian frame hold the publication at all?), NOT metaresearch recall: trial reports are primary research and the rubric screens them OUT. It is the precondition for screening, not the screen. The reference standard is the sponsor's OWN reported result publications, so it is incomplete in a known direction: sponsors under-report, which means the true set of trial publications is LARGER than the standard and this recall figure is measured only on the ones we can see. Trials are matched by Canadian LOCATION, which is not the same as Canadian authorship, so some result publications may have no Canadian author and legitimately fall outside the frame; that direction is not controlled here and it bounds the interpretation. Only PMIDs resolvable to a DOI can be looked up.
    trial_linkage · calculé le 2026-07-12T19:29:59Z
    19

    La couverture des prépublications

    La base porte 156 086 prépublications, et la tentation était d'affirmer qu'OpenAlex couvre les serveurs de prépublications et de sauter l'ingestion. C'est une AFFIRMATION DE COUVERTURE, et ce projet n'a pas le droit d'en faire une depuis l'intérieur de la chaîne de traitement même qu'elle concerne : ce serait la voie thématique certifiant son propre rappel (constat 12). Mesurée plutôt contre la PROPRE API des serveurs, OpenAlex indexe 99,6 % des 705 prépublications bioRxiv et medRxiv énumérées (3 manquantes). L'affirmation survit à la mesure ; aucune ingestion séparée de prépublications n'est donc construite.

    Énoncé original (findings.json) : The frame carries 156,086 preprints, and the tempting move was to assert that OpenAlex covers the preprint servers and skip the ingest. That is a COVERAGE CLAIM, and this project does not get to make one from inside the pipeline being claimed for: it is the topic route certifying its own recall (finding 12). Measured instead against the servers' OWN API, OpenAlex indexes 99.6% of the 705 bioRxiv and medRxiv preprints enumerated (3 missing). The claim survives measurement, so no separate preprint ingest is built.

    external criterion
    the bioRxiv/medRxiv details API, which enumerates the servers' own corpus and owes nothing to OpenAlex
    window
    2023-01-01 to 2025-12-31
    preprints enumerated
    705
    indexed by openalex
    702
    missing from openalex
    3
    pct indexed
    99.6
    sampled with canadian author
    45
    of those present in frame
    45
    frame preprints total
    156 086
    separate ingest needed
    false
    caveat
    The bioRxiv API exposes no author country, so the sample is mostly non-Canadian and the clean quantity is INDEX coverage (does OpenAlex hold the preprint at all?), not Canadian recall: a preprint the index lacks cannot enter any frame by any route, so index coverage is the binding upper bound. The Canadian sub-count is small and is reported as a check, not as an estimate. arXiv is NOT tested here (its OAI endpoint pages differently); the frame carries 11,647 arXiv works and that claim remains untested against arXiv itself.
    preprint_coverage · calculé le 2026-07-12T18:59:41Z
    20

    La puissance d'un audit humain

    L'audit tel que d'abord spécifié ne pouvait pas mesurer ce qu'il existe pour mesurer. Un échantillon aléatoire simple de la masse rejetée de 3 461 261 notices donne 0,4 occurrence attendue sur les 600 notices budgétées ; en voir 20 exigerait 2 009 heures de codage contre les 65 prévues. Le suréchantillonnage stratifié par score ne le sauve qu'en partie (le pilote montre que 30 des 37 travaux disputés se trouvent à la frontière contestée), et il reste AVEUGLE aux rejets assurés. Savoir quels travaux ce sont relève d'un MÉCANISME, non d'un écart mesuré : la grille dit de juger sur le seul titre en l'absence de résumé, et dit qu'un travail T2 peut n'employer aucun mot du vocabulaire du domaine ; un travail privé des deux est donc rejeté avec assurance et jamais échantillonné. (Une version antérieure citait ici un chiffre de 3,6 fois comme s'il était mesuré ; il reposait sur quatre travaux et il est retiré.) Le rappel se mesure donc de deux autres façons : sur les 5 737 travaux qui portent déjà des étiquettes de la grille, et par un rappel sur cibles connues à partir d'un ensemble de revues de référence, parce que la revue est un critère externe insensible à cette « aboutness » qui met tout le reste en échec.

    Énoncé original (findings.json) : The audit as first specified could not measure what it exists to measure. A simple random sample of the 3,461,261-record screened-out mass yields an expected 0.4 hits from the 600 records budgeted; seeing 20 would take 2,009 coder-hours against the 65 planned. Score-stratified oversampling rescues it only partly (the pilot shows 30 of 37 disputed works sit at the contested boundary), and it stays BLIND to the confident rejects. Which works those are is a MECHANISM, not a measured differential: the rubric says judge on the title alone with no abstract, and says T2 work may use none of the field's vocabulary, so a work with neither is rejected confidently and never sampled. (An earlier version cited a 3.6x figure here as if it were measured; it rested on four works and is withdrawn.) So recall is measured two other ways: against the 5,737 works that already carry rubric labels, and by KNOWN-ITEM recall on a venue reference set, because venue is an external criterion immune to the aboutness that defeats everything else.

    problem
    A simple random sample of the screened-out stratum cannot measure screening sensitivity: the works the screen wrongly rejected are a vanishing fraction of a 3.2M-record rejected mass.
    screened out pool
    3 461 261
    screened out is the screens rejects
    true
    audit screened out budgeted
    600
    expected hits at 95 recall
    0.4
    codings needed for 20 hits at 95 recall
    30 134
    coder hours needed
    2 009
    coder hours budgeted
    65
    naive audit is powered
    false
    disputed in boundary
    30
    disputed in settled rejects
    6
    misses concentrate near threshold
    true
    blind spot
    Score-stratified oversampling finds the works the screener ALMOST caught; it is blind to the ones it rejected CONFIDENTLY. WHICH works those are is a MECHANISM, not a measurement: the rubric says judge on the title alone when the abstract is missing, and the rubric also says T2 work may use none of the field's vocabulary, so a work with neither is rejected confidently and sits deep in the settled rejects. An earlier version cited a 3.6x differential from finding 11 as if this were measured. It is not, and that figure is withdrawn (DEVIATIONS.md D6). The venue instrument is how the prediction gets tested rather than asserted.
    blind spot is a mechanism not a measurement
    true
    instrument 1
    Measure any filter's recall against the 5737 works that already carry full-rubric labels, exactly as finding 12 scored the topic route. Free, and it needs no needle-hunting in the discarded mass.
    instrument 2
    Known-item recall on an external criterion: VENUE. A Canadian-authored paper in Social Studies of Science is T2 by where it was published, whatever its abstract is about. Venue is immune to aboutness, which is what defeats topic retrieval (finding 12) and title-only screening (finding 11), so it is the only instrument that can see into the blind spot.
    reference venues
    • Social Studies of Science
    • Scientometrics
    • Quantitative Science Studies
    • Research Integrity and Peer Review
    • Journal of the Association for Information Science and Technology
    • Research Evaluation
    • Accountability in Research
    • PLOS ONE (metaresearch collection)
    • Recherches qualitatives
    • Documentation et bibliotheques
    n labelled works
    5 737
    n in scope
    75
    caveat
    The recall grid assumes the screen's misses are uniform in the rejected mass, which the pilot shows they are not (they concentrate at the boundary). That makes the naive design LESS hopeless than the grid implies but does not save it, and it does nothing at all about the confident-reject blind spot. Known-item recall on a venue reference set is a non-probability estimate: it bounds and diagnoses recall on the hard cases, it does not replace the design-weighted population estimate.
    audit_power · calculé le 2026-07-12T14:39:49Z
    21

    Ce que coûte le tri

    La grille complète sur la base de 3 507 205 travaux avec deux trieurs coûte 5 357 $, et non les 24 968 $ qu'une version antérieure de ce script annonçait : la grille est une consigne système envoyée une fois par APPEL, et les lots du pilote regroupent eux-mêmes 155 travaux par appel ; la facture était donc gonflée de 4,7 fois. Cette erreur n'était pas cosmétique. Elle m'a fait proposer un TRIAGE bon marché devant le tri, c'est-à-dire une ÉTAPE DE REPÉRAGE dans un projet dont le constat central est que le repérage détruit ces cartes. L'arithmétique corrigée, le triage est inutile : la grille complète sur CHAQUE travail de la base, plus un second trieur sur un échantillon de 20 000 notices, coûte 908 $ et laisse environ 1 992 $ pour la personne qui code à la main. LE PRÉFILTRE EST SUPPRIMÉ.

    Énoncé original (findings.json) : The full rubric over the 3,507,205-work frame with two screeners costs $5,357, not the $24,968 an earlier version of this script reported: the rubric is a system prompt sent once per CALL, and the pilot's own chunks batch 155 works per call, so it was overcharged 4.7x. That error was not cosmetic. It made me propose a cheap TRIAGE in front of the screen, which is a RETRIEVAL STEP in a project whose central finding is that retrieval destroys these maps. With the arithmetic right the triage is unnecessary: the full rubric over EVERY work in the frame, plus a second screener on a 20,000-record sample, costs $908 and leaves ~$1,992 for the human coder. THE PREFILTER IS DELETED.

    measured from
    pilot/screening/chunks/*.json (6-field), pilot/screening/haiku/p8_guided/chunk_*.json (8-field), protocol/rubric.md; not asserted
    chars per token assumed
    4
    tokens per work
    312
    tokens per work six field deviation
    256
    payload inflation 8 over 6
    1.22
    costed at the rubric payload not the deviation
    true
    tokens rubric
    1 876
    tokens per label
    37
    batch works per call
    155
    frame size
    3 507 205
    grant usd approx
    2 900
    naive cost charging rubric per work usd
    24 968
    corrected cost two screeners usd
    5 357
    cost overstatement x
    4.7
    cost full rubric whole frame cheap usd
    893
    cost full rubric whole frame sonnet usd
    2 678
    cost second screener on sample usd
    15
    second screener n
    20 000
    total no prefilter usd
    908
    left for human coder usd
    1 992
    prefilter needed
    false
    prefilter design total usd
    822
    prefilter saving usd
    -86
    prefilter deleted because
    It is a RETRIEVAL STEP, in a project whose central finding is that retrieval destroys these maps (finding 12: the topic route finds 12% of the field). It existed only because the rubric was miscosted at once per WORK rather than once per CALL, overstating the alternative 5.1x. With the arithmetic right it saves almost nothing and costs the thesis. Deleted.
    pilot works screened
    6 202
    frame to pilot ratio
    565
    method that scales
    batch inference, not agent fan-out
    supersedes
    an earlier version charged the rubric once per work, reported $24,379 for the full-frame two-screener design, called it 'eight times the grant', and used that to justify a cheap prefilter. The rubric is a system prompt sent once per CALL, and the pilot's own chunks batch 155 works per call. DEVIATIONS.md D7.
    caveat
    Token counts use a 4-chars-per-token approximation and list prices as of 2026-07; both will move, and the conclusion is robust to +/-25% in either. Screening the frame with a cheaper model than the pilot used makes the CHOICE OF MODEL more consequential, not less: finding 10 shows two screeners already imply base rates a factor of two apart. That is precisely why the second screener and the screener-swap range are reported, and why the human audit is the study.
    screening_cost · calculé le 2026-07-12T14:39:48Z
    22

    OpenAlex est facturé à l'usage

    L'API OpenAlex est facturée à l'usage (1 000 crédits par ~11 h ; palier gratuit de 0,10 $). Énumérer la base canadienne de 3 507 205 travaux exige 17 537 appels paginés par curseur : 8,2 jours par passage au palier gratuit. Une chaîne de traitement fondée sur l'API à cette échelle n'est ni gratuite ni reproductible ; l'instantané épinglé est les deux.

    Énoncé original (findings.json) : The OpenAlex API is metered (1,000 credits per ~11h; $0.10 free tier). Enumerating the 3,507,205-work Canadian frame needs 17,537 cursor-paged calls: 8.2 days per pass on the free tier. An API-based pipeline at this scale is neither free nor reproducible; the pinned snapshot is both.

    observed retry after s
    40 268
    observed ratelimit limit
    1 000
    observed free tier usd
    0.1
    observed cost per call usd
    0.0001
    frame size works
    3 507 205
    per page max
    200
    calls for one pass
    17 537
    days on free tier per pass
    8.2
    prepaid cost per pass usd
    1.75
    openalex_is_metered · calculé le 2026-07-12T14:39:45Z