Add balanced Article Lab research

src/_data/research/claims.yml

@@ -124,3 +124,66 @@
   source_ids:
     - hhs-1979-belmont-report
   limitation: "Protection must not be converted into total erasure of agency, and the report does not resolve every conflict among autonomy, harm, and public duty."
+
+- id: article-one-work-requirements-access
+  article: rights-and-definitions-of-the-way.article-one
+  stance: dissent
+  status: quasi-experimental
+  claim: "Conditioning food assistance on work and reporting requirements can sharply reduce access among clinically and socioeconomically vulnerable recipients without producing a measurable transition off the broader safety net."
+  source_ids:
+    - ndumele-et-al-2025-snap-work-requirements
+  limitation: "The Connecticut study measures safety-net enrollment rather than moral contribution, long-term employment, or resource allocation under acute scarcity."
+
+- id: article-four-teach-back-comprehension
+  article: rights-and-definitions-of-the-way.article-four
+  stance: support
+  status: review-backed
+  claim: "Teaching consequential instructions in usable language and asking the recipient to explain them back can improve comprehension, recall, self-management, and some health outcomes."
+  source_ids:
+    - talevski-et-al-2020-teach-back
+  limitation: "The review concerns heterogeneous health-education studies; it does not prove that teach-back generalizes to all law or that comprehension alone makes enforcement legitimate."
+
+- id: article-four-legislative-complexity
+  article: rights-and-definitions-of-the-way.article-four
+  stance: dissent
+  status: theory-backed
+  claim: "Plain-language publication and teaching cannot by themselves make every legal duty simple or fully knowable because legislation carries irreducible technical and institutional complexity."
+  source_ids:
+    - webb-2020-legislative-complexity
+  limitation: "The legal analysis identifies limits of simplification but does not excuse avoidable opacity, secret rules, inadequate notice, or failure to teach applicable duties."
+
+- id: article-five-legal-capacity
+  article: rights-and-definitions-of-the-way.article-five
+  stance: dissent
+  status: rights-doctrine
+  claim: "Legal capacity and equal civic standing cannot be made contingent on passing a generalized mental-capacity or functional-competence test; decision support should protect the exercise of rights."
+  source_ids:
+    - un-crpd-2014-legal-capacity
+  limitation: "The treaty-body interpretation does not prohibit task-specific competency rules for hazardous professional roles or supported-decision safeguards."
+
+- id: article-six-nonviolent-power
+  article: rights-and-definitions-of-the-way.article-six
+  stance: dissent
+  status: empirical-challenge
+  claim: "Collective political power is not reducible to violence: nonviolent campaigns have often achieved political goals by widening participation, gaining legitimacy, and inducing loyalty shifts."
+  source_ids:
+    - stephan-chenoweth-2008-civil-resistance
+  limitation: "Historical campaign evidence is not a complete theory of authority, does not show violence is never decisive, and does not replace bounded emergency command."
+
+- id: article-seven-graded-categories
+  article: rights-and-definitions-of-the-way.article-seven
+  stance: dissent
+  status: experimental-limit
+  claim: "Many natural-language categories have graded internal structure rather than a single sharp boundary, so imposing a fixed definition before conversation can misrepresent ordinary meaning."
+  source_ids:
+    - rosch-1975-semantic-categories
+  limitation: "Semantic-category experiments do not govern legal drafting or technical interfaces; stipulative definitions remain useful for consequential load-bearing terms."
+
+- id: article-eight-consent-insufficiency
+  article: rights-and-definitions-of-the-way.article-eight
+  stance: dissent
+  status: doctrinal-challenge
+  claim: "Notice and individual consent cannot by themselves legitimize every data practice because cognitive limits, scale, aggregation, downstream uses, and collective harms can defeat meaningful self-management."
+  source_ids:
+    - solove-2013-consent-dilemma
+  limitation: "The analysis concerns information privacy; it does not negate informed consent in medicine or research or establish that consent is irrelevant in other domains."

src/_data/research/sources.yml

@@ -161,3 +161,85 @@
   reviewed_on: "2026-07-21"
   relevance: "Connects respect for persons with information, comprehension, and voluntariness; also requires risk-benefit assessment and protection for diminished autonomy."
   boundary: "Governs human-subject research ethics and supplies a strong consent model; not every civil or private interaction is research."
+
+- id: ndumele-et-al-2025-snap-work-requirements
+  title: "Supplemental Nutrition Assistance Program Work Requirements and Safety-Net Program Participation"
+  authors: "Chima D. Ndumele; Hannah Factor; Matthew Lavallee; Anthony Lollo Jr.; Jacob Wallace"
+  year: 2025
+  publication: "JAMA Internal Medicine 185(1):92-100"
+  type: peer-reviewed quasi-experimental cohort study
+  doi: "10.1001/jamainternmed.2024.5932"
+  url: "https://doi.org/10.1001/jamainternmed.2024.5932"
+  reviewed_on: "2026-07-23"
+  relevance: "Finds that reinstated SNAP work requirements in Connecticut were associated with a substantial enrollment decline concentrated among clinically and socioeconomically vulnerable beneficiaries, without evidence of transition off the broader safety net."
+  boundary: "Studies one state policy and program-enrollment outcomes; it does not determine moral contribution, global food allocation, or the effects of every work requirement."
+
+- id: talevski-et-al-2020-teach-back
+  title: "Teach-back: A systematic review of implementation and impacts"
+  authors: "Jason Talevski; Anna Wong Shee; Bodil Rasmussen; Georgie Kemp; Alison Beauchamp"
+  year: 2020
+  publication: "PLOS ONE 15(4):e0231350"
+  type: peer-reviewed systematic review
+  doi: "10.1371/journal.pone.0231350"
+  url: "https://doi.org/10.1371/journal.pone.0231350"
+  reviewed_on: "2026-07-23"
+  relevance: "Reviews evidence that teach-back can improve health-related knowledge, recall, self-management, and some outcomes by requiring an explanation in the recipient's own words."
+  boundary: "The included studies were heterogeneous and generally moderate quality; health education evidence does not establish that teach-back generalizes to all law or makes enforcement legitimate."
+
+- id: webb-2020-legislative-complexity
+  title: "The Drafters’ Dance: The Complexity of Drafting Legislation and the Limitations of ‘Plain Language’ and ‘Good Law’ Initiatives"
+  authors: "Thomas E. Webb"
+  year: 2020
+  publication: "Statute Law Review 41(2):129-158"
+  type: peer-reviewed legal analysis
+  doi: "10.1093/slr/hmz008"
+  url: "https://doi.org/10.1093/slr/hmz008"
+  reviewed_on: "2026-07-23"
+  relevance: "Argues that legislative drafting contains irreducible technical and institutional complexity that plain-language and good-law initiatives can obscure or displace."
+  boundary: "A theoretical legal analysis of drafting complexity; it does not excuse avoidable opacity, secret rules, inadequate notice, or failure to teach applicable duties."
+
+- id: un-crpd-2014-legal-capacity
+  title: "General comment No. 1 (2014), Article 12: Equal recognition before the law"
+  authors: "United Nations Committee on the Rights of Persons with Disabilities"
+  year: 2014
+  publication: "United Nations document CRPD/C/GC/1"
+  type: treaty-body interpretive guidance
+  url: "https://cambodia.ohchr.org/en/content/general-comment-no-1-committee-rights-persons-disabilities-equal-recognition-law"
+  reviewed_on: "2026-07-23"
+  relevance: "Distinguishes universal legal capacity from variable mental capacity and requires support for exercising rights rather than status removal based on impairment or functional assessment."
+  boundary: "Authoritative human-rights interpretation rather than an empirical outcome study; it does not prohibit task-specific competency requirements for hazardous professional roles or supported-decision safeguards."
+
+- id: stephan-chenoweth-2008-civil-resistance
+  title: "Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict"
+  authors: "Maria J. Stephan; Erica Chenoweth"
+  year: 2008
+  publication: "International Security 33(1):7-44"
+  type: peer-reviewed comparative empirical study
+  doi: "10.1162/isec.2008.33.1.7"
+  url: "https://doi.org/10.1162/isec.2008.33.1.7"
+  reviewed_on: "2026-07-23"
+  relevance: "Compares major violent and nonviolent campaigns and identifies participation, legitimacy, and loyalty shifts as mechanisms through which nonviolent collective power can prevail."
+  boundary: "Campaign-level historical evidence is not a complete theory of authority, does not show that violence is never decisive, and does not replace bounded emergency command."
+
+- id: rosch-1975-semantic-categories
+  title: "Cognitive representations of semantic categories"
+  authors: "Eleanor Rosch"
+  year: 1975
+  publication: "Journal of Experimental Psychology: General 104(3):192-233"
+  type: peer-reviewed experimental study
+  doi: "10.1037/0096-3445.104.3.192"
+  url: "https://doi.org/10.1037/0096-3445.104.3.192"
+  reviewed_on: "2026-07-23"
+  relevance: "Shows that many ordinary semantic categories have graded internal structure and that typicality affects how people learn and process category members."
+  boundary: "Experiments on semantic categories do not govern legal drafting or technical interface standards; stipulative definitions remain useful for consequential load-bearing terms."
+
+- id: solove-2013-consent-dilemma
+  title: "Privacy Self-Management and the Consent Dilemma"
+  authors: "Daniel J. Solove"
+  year: 2013
+  publication: "Harvard Law Review 126:1880-1903"
+  type: legal analysis grounded in empirical and social-science literature
+  url: "https://harvardlawreview.org/print/vol-126/introduction-privacy-self-management-and-the-consent-dilemma/"
+  reviewed_on: "2026-07-23"
+  relevance: "Explains how cognitive limits, information volume, aggregation, downstream uses, and collective effects make notice-and-consent insufficient as the sole governance mechanism for personal data."
+  boundary: "Addresses information privacy; it does not negate informed consent in medicine or research and does not establish that consent is irrelevant in other domains."

src/research/method.njk

@@ -3,7 +3,43 @@ layout: layouts/base.njk
 title: Research method
 description: How The Interdependent Way site distinguishes support, dissent, mixed evidence, implementation, and unresolved research gaps.
 ---
-<header class="page-head"><p class="eyebrow">Research method</p><h1>Evidence is not a decorative citation layer.</h1><p class="lede">Sources are attached to specific claims and units. A metadata service may help locate or normalize a source, but it does not decide whether the source supports, dissents from, limits, or merely contextualizes a claim.</p></header>
-<section><h2>Classification rules</h2><div class="grid"><article class="card"><span class="status status-research">Support</span><h3>Substantive agreement</h3><p>The source supplies evidence or reasoning that materially strengthens the identified claim.</p></article><article class="card"><span class="status status-risk">Dissent</span><h3>Substantive challenge</h3><p>The source disputes a premise, mechanism, prediction, interpretation, or consequence.</p></article><article class="card"><span class="status status-frontier">Mixed</span><h3>Conditional or partial result</h3><p>The source supports one part while limiting another, or shows the claim depends on context.</p></article><article class="card"><span class="status status-hmmm">Gap</span><h3>No qualifying source found</h3><p>The site says so directly. It does not manufacture false balance or cite a weak source to fill visual symmetry.</p></article></div></section>
-<section class="panel"><h2>Publication requirements</h2><ol><li>Identify the exact canonical unit or project claim.</li><li>Record source title, authorship, date, stable identifier, and retrieval date.</li><li>Separate direct findings from editorial inference.</li><li>Label stance through review, not keyword sentiment.</li><li>Preserve contradictory high-quality evidence.</li><li>Mark retractions, corrections, and stale data.</li><li>Keep implementation evidence separate from theoretical validity.</li></ol></section>
-<section class="hmmm"><h2>hmmm</h2><p>The version-controlled research ledger exists as a schema boundary, but the reviewed corpus is not yet broad enough to imply comprehensive coverage.</p></section>
+<header class="page-head">
+  <p class="eyebrow">Research method</p>
+  <h1>Evidence is not a decorative citation layer.</h1>
+  <p class="lede">Sources are attached to specific claims and units. A metadata service may help locate or normalize a source, but it does not decide whether the source supports, dissents from, limits, or merely contextualizes a claim.</p>
+</header>
+<section>
+  <h2>Classification rules</h2>
+  <div class="grid">
+    <article class="card"><span class="status status-research">Support</span><h3>Substantive agreement</h3><p>The source supplies evidence or reasoning that materially strengthens the identified claim.</p></article>
+    <article class="card"><span class="status status-risk">Dissent</span><h3>Substantive challenge</h3><p>The source disputes a premise, mechanism, prediction, interpretation, or consequence.</p></article>
+    <article class="card"><span class="status status-frontier">Mixed</span><h3>Conditional or partial result</h3><p>The source supports one part while limiting another, or shows the claim depends on context.</p></article>
+    <article class="card"><span class="status status-hmmm">Gap</span><h3>No qualifying source found</h3><p>The site says so directly. It does not manufacture false balance or cite a weak source to fill visual symmetry.</p></article>
+  </div>
+</section>
+<section class="panel">
+  <h2>Usage guidance</h2>
+  <ol>
+    <li>Open the exact canonical unit before reading the research cards; the cards test bounded claims and do not replace the source text.</li>
+    <li>Read support and dissent as claim-specific contact. A source may assent to one component while challenging another.</li>
+    <li>Compare source quality, relevance, method, and stated boundary. Never treat citation counts as a vote.</li>
+    <li>Use the two-sided coverage floor to expose each Rights Article to serious challenge, not to manufacture symmetry where evidence quality differs.</li>
+  </ol>
+  <p><strong>Coverage floor:</strong> Every Rights Article carries at least one reviewed support claim and one reviewed dissent claim. This proves contact with both directions of inquiry; it does not prove truth, falsity, balance, or comprehensive coverage.</p>
+</section>
+<section class="panel">
+  <h2>Publication requirements</h2>
+  <ol>
+    <li>Identify the exact canonical unit or project claim.</li>
+    <li>Record source title, authorship, date, stable identifier, and retrieval date.</li>
+    <li>Separate direct findings from editorial inference.</li>
+    <li>Label stance through review, not keyword sentiment.</li>
+    <li>Preserve contradictory high-quality evidence.</li>
+    <li>Mark retractions, corrections, and stale data.</li>
+    <li>Keep implementation evidence separate from theoretical validity.</li>
+  </ol>
+</section>
+<section class="hmmm">
+  <h2>hmmm</h2>
+  <p>The version-controlled research ledger now provides two-sided contact for every Rights Article, but the reviewed corpus remains an opening tranche rather than a comprehensive literature review.</p>
+</section>

tests/research-ledger.test.mjs

@@ -1,4 +1,4 @@
-// Usage: run through `npm test`; add claims only with source ids, limitations, reviewed metadata, and a bounded Article Lab record.
+// Usage: run through `npm test`; every Rights Article must retain at least one reviewed support and dissent claim, while mixed and limit records preserve conditional evidence.
 // Evidence boundary: validates provenance and coverage structure, not the truth of every source claim or application.
 import test from 'node:test';
 import assert from 'node:assert/strict';
@@ -38,6 +38,7 @@ test('research ledgers parse and claims resolve to reviewed sources', async () =
   }
   for (const claim of claims) {
     assert.ok(['support', 'dissent', 'limit', 'mixed'].includes(claim.stance));
+    assert.ok(claim.status && claim.claim);
     assert.ok(labUnitIds.has(claim.article), `claim ${claim.id} is not attached to a Rights Article Lab record`);
     assert.ok(Array.isArray(claim.source_ids) && claim.source_ids.length > 0);
     for (const sourceId of claim.source_ids) assert.ok(sourceIds.has(sourceId), `unknown source ${sourceId}`);
@@ -50,6 +51,9 @@ test('research ledgers parse and claims resolve to reviewed sources', async () =
     assert.ok(Array.isArray(record.worst_practices) && record.worst_practices.length >= 4);
     assert.ok(Array.isArray(record.best_practices) && record.best_practices.length >= 4);
     assert.deepEqual(Object.keys(record.applications), requiredDomains, `${record.unit_id} must preserve the shared domain order`);
-    assert.ok(claims.some(claim => claim.article === record.unit_id), `${record.unit_id} needs reviewed research contact`);
+    const articleClaims = claims.filter(claim => claim.article === record.unit_id);
+    assert.ok(articleClaims.length > 0, `${record.unit_id} needs reviewed research contact`);
+    assert.ok(articleClaims.some(claim => claim.stance === 'support'), `${record.unit_id} needs reviewed support`);
+    assert.ok(articleClaims.some(claim => claim.stance === 'dissent'), `${record.unit_id} needs reviewed dissent`);
   }
 });