{"closest_prior_art":[{"name":"Multiverse analysis","overlap":"Enumerates reasonable data-processing and model choices, runs alternative analyses, and exposes whether a highlighted result is fragile or selectively reported.","remaining_difference":"It does not reconstruct every actually attempted literary search, apply an FDR-based promotion rule, assign editorial claim-status labels, or require confirmation on sealed texts.","source_ids":["SRC2"]},{"name":"OSF registrations and preregistrations","overlap":"Provides timestamped registrations, templates for completed projects and secondary data, replication protocols, and view-only records for peer review.","remaining_difference":"Registration does not itself recover unregistered exploratory looks, define a literary claim family, adjust discoveries for multiplicity, or require untouched-text confirmation.","source_ids":["SRC3"]},{"name":"Multiplicity-adjusted corpus keyness analysis","overlap":"A corpus-analysis implementation defines the tested family and applies multiple-testing correction by default because large corpus candidate sets can generate many false positives.","remaining_difference":"It covers tests within a function call, not motifs, theories, corpus boundaries, editions, recodings, abandoned analyses, editorial status, or sealed-corpus confirmation.","source_ids":["SRC4"]}],"contrastive_claim_falsifier":"The contrast would be falsified by finding an established corpus-literary workflow that already requires a versioned inventory of all attempted motif, corpus, lens, coding, and model choices; multiplicity-aware promotion; explicit editorial status labels; and a frozen test on untouched texts—or if a consenting shadow audit shows that adding those elements neither changes claim status nor improves sealed-text survival relative to ordinary reproducibility and robustness review.","contrastive_claim_remaining":"For publication-leading empirical literary-historical claims, neither reproducibility, preregistration, multiverse reporting, nor within-analysis multiplicity correction alone accounts for the full opportunity set created during interpretive search. The remaining testable contribution is their integration into a retrospective attempt ledger with claim-status transitions and a sealed-text confirmation gate.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","gates":{"adequate_source_search":{"rationale":"Searches covered the proposal directly, computational-literary and quantitative-literary synonyms, preregistration and corpus-analysis practices, and combinations involving analysis logs, multiverses, multiplicity, secondary data, and holdouts. Four opened sources from four publishers include primary research, official guidance, and first-party software documentation. No exact integrated literary workflow was located, though bounded search cannot prove absence.","source_ids":["SRC1","SRC2","SRC3","SRC4"],"status":"PASS"},"bounded_next_test":{"rationale":"The proposed read-only audit is limited to one consenting completed project, at most 30 reconstructed choices, two independent family definitions, and one claim tested on a demonstrably untouched subset. Its observable outputs—feasibility, reviewer agreement, missing-search evidence, status change, and confirmation outcome—are bounded and reversible.","source_ids":["SRC1","SRC3"],"status":"PASS"},"distinct_testable_claim":{"rationale":"The proposal adds a falsifiable conjunction absent from the retained adjacent practices: recover the actual search opportunity set, calibrate promotion to that family, label evidentiary status, and require a frozen fresh-text test. SRC2 explicitly lacks an evidentiary threshold, while SRC3 and SRC4 address only registration and within-call testing families.","source_ids":["SRC2","SRC3","SRC4"],"status":"PASS"},"no_obvious_safety_or_authority_stop":{"rationale":"The first step is consensual, read-only, does not alter manuscripts or editorial decisions, and includes stopping for confidentiality, attribution disputes, or prior consultation of the confirmation corpus. OSF guidance also shows that anonymized view-only peer-review access and embargo controls are feasible. Contributor-consent and scarce-archive equity risks require monitoring but do not bar the pilot.","source_ids":["SRC3"],"status":"PASS"},"supported_problem":{"rationale":"Direct computational-literary reanalysis found that a chosen stop-word specification could reverse a result and that a standard list removed significance; it also documented instability under alternate parsers and corpus choices. Multiverse research establishes the general selective-reporting mechanism, and corpus software documentation recognizes false-positive risk from large candidate sets. The specific prevalence of hidden abandoned searches in literary projects was not measured, so the evidence is partial rather than complete.","source_ids":["SRC1","SRC2","SRC4"],"status":"PASS"}},"prior_art_disposition":"ADJACENT_PRIOR_ART","problem_evidence":{"finding":"The vulnerability is visible: published computational-literary conclusions can depend materially on preprocessing, parser, corpus-size, and modeling choices, while corpus candidate searches create acknowledged multiplicity risk. The retained sources do not directly estimate how often literary teams conceal null or abandoned paths or how often those paths affect publication, so the full baseline is only partly demonstrated.","source_ids":["SRC1","SRC2","SRC4"],"status":"PARTLY_SUPPORTED"},"research_id":"eoa_inverse_innovation_exp13_light_screen_20260806","schema_version":1,"screen_id":"E13P005","screen_survival":true,"search_lanes":{"component_combination":{"no_result_note":"No opened result combined retrospective recovery of attempted literary lenses and corpus cuts with multiplicity-aware promotion, editorial status labels, and sealed-text confirmation.","queries":["research analysis ledger log all attempted analyses null results multiverse preregistration","multiverse analysis all reasonable analyses transparency primary paper Steegen","secondary data preregistration existing data already accessed holdout official guidance"],"source_ids":["SRC2","SRC3"]},"direct_problem_and_intervention":{"no_result_note":"Direct searches found evidence of specification sensitivity and adjacent open-science remedies, but not the complete named ledger intervention.","queries":["corpus literary studies multiple comparisons cherry picking specification search motifs","computational literary studies preregistration reproducibility study","corpus-based literary studies p-hacking selective reporting"],"source_ids":["SRC1","SRC3"]},"products_practices_and_standards":{"no_result_note":null,"queries":["OSF preregistration template exploratory confirmatory analyses official","corpus linguistics multiple testing false discovery rate guidelines","registered reports guidelines exploratory confirmatory analysis publication official"],"source_ids":["SRC3","SRC4"]},"synonyms_and_historical_terms":{"no_result_note":null,"queries":["quantitative literary studies replication crisis methodology","literary studies multiple testing corpus","digital humanities multiverse analysis literary corpus robustness specifications"],"source_ids":["SRC1","SRC2","SRC4"]}},"sources":[{"claims_supported":["Computational literary studies uses quantitative patterns to make claims about literature and literary history.","A reanalysis found that a nonstandard stop-word choice reversed a stream-of-consciousness result, while a standard list removed statistical significance.","Alternate parsers, corpus construction, topic-model parameters, and corpus scaling can materially change or invalidate literary findings."],"publisher":"University of Chicago Press, Critical Inquiry","source_id":"SRC1","source_type":"PRIMARY_RESEARCH","title":"The Computational Case against Computational Literary Studies","url":"https://jonathanstray.com/papers/Computational-Literary-Studies.pdf"},{"claims_supported":["Reasonable coding, exclusion, transformation, and model choices create a multiverse of possible results.","Reporting one selected analysis can conceal sensitivity and be misleading.","Multiverse analysis improves transparency and robustness assessment but does not itself supply an evidentiary threshold or formal test of selective reporting."],"publisher":"SAGE Publications, Perspectives on Psychological Science","source_id":"SRC2","source_type":"PRIMARY_RESEARCH","title":"Increasing Transparency Through a Multiverse Analysis","url":"https://journals.sagepub.com/doi/10.1177/1745691616658637"},{"claims_supported":["OSF supports general, open-ended, completed-project, secondary-data, replication, and Registered Report registrations.","Registrations can connect project records across the research lifecycle.","Embargoed and anonymized view-only access can support blinded peer review."],"publisher":"Center for Open Science","source_id":"SRC3","source_type":"OFFICIAL_GUIDANCE","title":"Welcome to Registrations & Preregistrations!","url":"https://help.osf.io/article/330-welcome-to-registrations"},{"claims_supported":["Corpus keyness software corrects for multiple testing by default and defines the family from candidate items processed.","Large candidate sets and sample sizes in corpus linguistics can generate many false positives.","The family size may need explicit specification when a corpus is processed in batches."],"publisher":"Comprehensive R Archive Network","source_id":"SRC4","source_type":"FIRST_PARTY_PRODUCT","title":"corpora: Compute best-practice keyness measures","url":"https://search.r-project.org/CRAN/refmans/corpora/html/keyness.html"}],"world_novelty_boundary":"This bounded public-web screen establishes only that the retained sources are adjacent rather than an exact match. It does not establish world novelty, patentability, market size, expert acceptance, implementation feasibility beyond the pilot, or realized scholarly value."}