flonat-research 0.1.0

package/.claude/agents/references/journal-referee-profiles.md

@@ -0,0 +1,354 @@
+# Journal Referee Profiles
+
+> Used by `referee2-reviewer` and `peer-reviewer` agents to calibrate review intensity and focus.
+> When a journal is specified, the reviewer adopts that journal's typical referee perspective.
+> Each profile includes a **Referee pool** that weights disposition draws (see `referee-config.md`).
+
+## How to Use
+
+When reviewing with a target journal:
+1. Look up the journal below
+2. Adjust **domain** focus (what matters most substantively)
+3. Adjust **methods** focus (rigour expectations)
+4. Ask the journal's **typical concerns** as explicit review questions
+5. Weight disposition draws using the **Referee pool** field
+
+When no journal is specified, use generic top-field behaviour with equal disposition weights.
+
+---
+
+## Economics
+
+**Top-5 General Interest**
+
+### American Economic Review (AER)
+**Focus:** All fields — broadest audience
+**Bar:** Must interest economists outside your subfield. Big question, clean execution, clear contribution.
+**Domain focus:** "Would a labour economist care about this health paper?" Contribution must be broad. Literature positioning against the *general* frontier, not just subfield. Policy implications welcome but not required — insight is enough.
+**Methods focus:** Identification must be convincing to non-specialists. Clean, transparent design preferred over technically complex one. Standard errors and robustness should be thorough but not excessive.
+**Typical concerns:** "Why should economists outside this field care?" "Is the contribution big enough for AER?" "Is this too narrow/specialised?"
+**Referee pool:** CREDIBILITY (high), POLICY (medium), STRUCTURAL (medium), MEASUREMENT (low), THEORY (low), SKEPTIC (low)
+
+### Econometrica (ECMA)
+**Focus:** Theoretical and empirical economics with formal rigour
+**Bar:** Methodological innovation or empirical work with exceptional identification and formal results.
+**Domain focus:** Theoretical contribution valued highly. If empirical, the design must be near-airtight. Formal welfare analysis expected. Less emphasis on policy narrative, more on economic theory and mechanisms.
+**Methods focus:** Formal proofs or near-formal arguments expected for key results. Asymptotic properties discussed. Novel estimators should have theoretical justification. Simulation evidence for finite-sample properties.
+**Typical concerns:** "Where's the formal result?" "What are the asymptotic properties?" "Is this a methods contribution or an applied contribution?"
+**Referee pool:** THEORY (high), STRUCTURAL (high), SKEPTIC (medium), CREDIBILITY (low), MEASUREMENT (low), POLICY (low)
+
+### Journal of Political Economy (JPE)
+**Focus:** All fields — strong emphasis on economic mechanisms and structural thinking
+**Bar:** Deep economic insight. Values understanding *why* something happens, not just *that* it happens.
+**Domain focus:** Mechanism is king. Reduced-form results alone insufficient — need to explain the economics. Structural models or mechanism tests expected. Theoretical framework (even informal) valued.
+**Methods focus:** Identification strong, but mechanism evidence equally important. Heterogeneity that illuminates the mechanism. Willing to accept some identification imperfection if the economic insight is deep enough.
+**Typical concerns:** "What's the mechanism?" "Can you decompose the effect?" "What does this tell us about economic behaviour?"
+**Referee pool:** STRUCTURAL (high), THEORY (high), CREDIBILITY (medium), SKEPTIC (medium), POLICY (low), MEASUREMENT (low)
+
+### Quarterly Journal of Economics (QJE)
+**Focus:** All fields — prizes compelling narrative and important questions
+**Bar:** The question must be important and the answer must surprise.
+**Domain focus:** Narrative matters enormously. The paper should read like a story with a punchline. Broad implications. Creative use of data or setting. "Clever" identification valued.
+**Methods focus:** Identification must be clean and intuitive — easy to explain. Transparency and simplicity over complexity. Visual evidence (event studies, RD plots) highly valued.
+**Typical concerns:** "Is this surprising?" "Does this change how we think about X?" "Can you explain the identification in one sentence?"
+**Referee pool:** CREDIBILITY (high), POLICY (high), SKEPTIC (medium), STRUCTURAL (low), THEORY (low), MEASUREMENT (low)
+
+### Review of Economic Studies (REStud)
+**Focus:** All fields — technically excellent empirical and theoretical work
+**Bar:** Technical quality must be top-tier. Values precision and completeness over narrative.
+**Domain focus:** Thoroughness expected — address every possible objection. Complete set of robustness checks. Careful literature review. Less emphasis on storytelling than QJE, more on completeness.
+**Methods focus:** Every specification must be justified. Full battery of robustness checks expected. Sensitivity analysis (Oster bounds, etc.). Careful treatment of inference. Multiple testing corrections if applicable.
+**Typical concerns:** "Have you checked robustness to X?" "What about specification Y?" "The inference needs more care."
+**Referee pool:** SKEPTIC (high), MEASUREMENT (high), CREDIBILITY (medium), THEORY (medium), STRUCTURAL (low), POLICY (low)
+
+---
+
+**Top Field Journals**
+
+### AEJ: Applied Economics
+**Focus:** Empirical microeconomics — labour, health, education, development, public
+**Bar:** Clean applied micro with credible identification. Same rigour as top-5 but contribution can be more subfield-specific.
+**Domain focus:** Meaningful subfield contribution. Practical policy relevance appreciated. Literature positioning within the subfield, not the general field.
+**Methods focus:** Modern estimators expected (no naive TWFE for staggered). Replication package expected.
+**Typical concerns:** "Is this incremental relative to [closely related paper]?" "Would this be better in a field journal?"
+**Referee pool:** CREDIBILITY (high), POLICY (medium), MEASUREMENT (medium), SKEPTIC (low), STRUCTURAL (low), THEORY (low)
+
+### AEJ: Economic Policy
+**Focus:** Policy evaluation and design
+**Bar:** Direct policy relevance. Natural experiments from actual policy changes preferred.
+**Domain focus:** Policy implications front and centre. Cost-benefit or welfare discussion expected. Institutional details must be well-documented. Generalisability to other policy contexts.
+**Methods focus:** Identification from actual policy variation. Pre-trends must be clean. Heterogeneity by policy-relevant subgroups expected. Back-of-envelope welfare calculations.
+**Typical concerns:** "What should policymakers do with this?" "Does this generalise?" "What's the cost-benefit?"
+**Referee pool:** POLICY (high), CREDIBILITY (high), MEASUREMENT (medium), STRUCTURAL (low), THEORY (low), SKEPTIC (low)
+
+### Journal of Human Resources (JHR)
+**Focus:** Labour economics, education, health, demography
+**Bar:** Strong empirical contribution with clear policy relevance and careful identification.
+**Domain focus:** Policy relevance matters more than theoretical novelty. External validity. Heterogeneity by policy-relevant subgroups expected.
+**Methods focus:** Modern staggered DiD estimators if applicable. Clean pre-trends. Replication package expected at acceptance.
+**Typical concerns:** "What's the policy implication?" "Does this generalise beyond your sample?" "Have you considered heterogeneity by [race/gender/income]?"
+**Referee pool:** CREDIBILITY (high), POLICY (high), MEASUREMENT (medium), SKEPTIC (low), STRUCTURAL (low), THEORY (low)
+
+### Journal of Health Economics (JHE)
+**Focus:** Health economics — insurance, utilisation, provider behaviour, public health
+**Bar:** Sound health economics with credible identification. Institutional knowledge expected.
+**Domain focus:** Deep understanding of health care institutions. Moral hazard vs adverse selection distinction. Welfare implications. Connection to health policy debates.
+**Methods focus:** Health-specific threats: selection into insurance, Ashenfelter dip, moral hazard confounding. GLM for cost outcomes alongside OLS.
+**Typical concerns:** "Is this moral hazard or adverse selection?" "Have you addressed selection into treatment?"
+**Referee pool:** STRUCTURAL (high), MEASUREMENT (high), POLICY (medium), CREDIBILITY (medium), THEORY (low), SKEPTIC (low)
+
+### RAND Journal of Economics
+**Focus:** Industrial organisation, regulation, antitrust, health care markets
+**Bar:** IO-flavoured analysis with market structure or firm behaviour component.
+**Domain focus:** Market structure and competition implications. Welfare analysis (consumer/total surplus). Regulatory implications.
+**Methods focus:** Structural models valued alongside reduced-form. Demand estimation methods (BLP, discrete choice).
+**Typical concerns:** "What does this imply for market structure?" "Consumer welfare impact?" "Can you do a structural analysis?"
+**Referee pool:** STRUCTURAL (high), THEORY (high), SKEPTIC (medium), POLICY (medium), CREDIBILITY (low), MEASUREMENT (low)
+
+### Journal of Public Economics (JPubE)
+**Focus:** Tax policy, public goods, redistribution, government programmes
+**Bar:** Public finance question with clean identification. Knowledge of tax/transfer system mechanics.
+**Domain focus:** Tax incidence, deadweight loss, behavioural responses. Programme evaluation. Fiscal federalism.
+**Methods focus:** Bunching estimators for kinks/notches. RDD at eligibility thresholds. DiD around policy changes. Extensive vs intensive margin effects.
+**Typical concerns:** "What's the elasticity?" "Extensive or intensive margin?" "Welfare implications?"
+**Referee pool:** STRUCTURAL (high), POLICY (high), CREDIBILITY (medium), THEORY (medium), MEASUREMENT (low), SKEPTIC (low)
+
+### Journal of Labour Economics (JLE)
+**Focus:** Labour markets — wages, employment, human capital, discrimination, immigration
+**Bar:** Clean labour economics with careful identification.
+**Domain focus:** Wage determination, employment effects, human capital, discrimination. Monopsony and market power.
+**Methods focus:** Selection correction when relevant. Decomposition methods for wage gaps. Event study designs around job transitions or policy changes.
+**Typical concerns:** "Is this a supply or demand effect?" "Selection into employment?" "General equilibrium effects?"
+**Referee pool:** CREDIBILITY (high), STRUCTURAL (medium), MEASUREMENT (medium), THEORY (medium), POLICY (low), SKEPTIC (low)
+
+### Journal of Development Economics (JDE)
+**Focus:** Development economics — poverty, institutions, agriculture, trade in developing countries
+**Bar:** Credible empirical evidence on development questions. RCTs or strong quasi-experimental designs.
+**Domain focus:** Deep country/region knowledge. External validity. Gender and equity dimensions. Cost-effectiveness.
+**Methods focus:** Randomisation checks, attrition, compliance, spillovers, pre-analysis plan for RCTs.
+**Typical concerns:** "Does this generalise beyond this context?" "What about attrition?" "Cost-effectiveness?"
+**Referee pool:** CREDIBILITY (high), POLICY (high), MEASUREMENT (high), SKEPTIC (medium), STRUCTURAL (low), THEORY (low)
+
+---
+
+**Short Format**
+
+### AER: Insights
+**Focus:** Same breadth as AER but shorter format
+**Bar:** AER-quality insight in a shorter paper. Must be self-contained and punchy.
+**Domain focus:** Brevity is a feature. One clean result is enough.
+**Methods focus:** Core identification must be clean. Fewer robustness checks acceptable given format.
+**Typical concerns:** "Can this be communicated in 10 pages?" "Is the single result compelling enough?"
+**Referee pool:** CREDIBILITY (high), POLICY (medium), SKEPTIC (medium), STRUCTURAL (low), THEORY (low), MEASUREMENT (low)
+
+### Economics Letters
+**Focus:** Short papers across all fields — theoretical and empirical
+**Bar:** One clear result in 5-8 pages. Speed of publication valued.
+**Domain focus:** Contribution in the first paragraph. One key finding, cleanly presented.
+**Methods focus:** Clean identification but extensive robustness not expected given format.
+**Typical concerns:** "Can this be said in 5 pages?" "Is the single result robust?"
+**Referee pool:** CREDIBILITY (high), SKEPTIC (medium), THEORY (medium), MEASUREMENT (low), STRUCTURAL (low), POLICY (low)
+
+---
+
+**Econometrics**
+
+### Journal of Econometrics
+**Focus:** Econometric theory and methods
+**Bar:** Methodological contribution with formal results. The method must be the contribution.
+**Domain focus:** Theoretical novelty paramount. Monte Carlo simulations expected. Empirical illustration should showcase the method.
+**Methods focus:** Formal proofs required (consistency, asymptotic normality, convergence rates). Comparison with existing estimators analytically and via simulation.
+**Typical concerns:** "What are the asymptotic properties?" "How does this compare to [existing method]?" "Finite-sample performance?"
+**Referee pool:** THEORY (high), SKEPTIC (high), MEASUREMENT (high), STRUCTURAL (medium), CREDIBILITY (low), POLICY (low)
+
+### Review of Economics and Statistics (REStat)
+**Focus:** Empirical economics — all fields, emphasis on careful measurement and methods
+**Bar:** Technically excellent empirical work. Values careful econometrics.
+**Domain focus:** Measurement quality paramount. Novel data or measurement approaches valued.
+**Methods focus:** Highest econometric standards short of Econometrica. Full sensitivity analysis.
+**Typical concerns:** "Is the measurement precise enough?" "Have you tested every assumption?"
+**Referee pool:** MEASUREMENT (high), SKEPTIC (high), CREDIBILITY (high), THEORY (medium), STRUCTURAL (low), POLICY (low)
+
+---
+
+**Health Economics**
+
+### Health Economics
+**Focus:** Health economics with emphasis on empirical applications
+**Bar:** Sound empirical health economics. Good outlet for well-executed applied work.
+**Domain focus:** Same institutional knowledge as JHE. Behavioural health economics increasingly accepted.
+**Methods focus:** Clean identification. Both OLS and GLM for cost outcomes.
+**Typical concerns:** "Have you addressed selection?" "What about moral hazard?" "Policy implications?"
+**Referee pool:** CREDIBILITY (high), MEASUREMENT (high), POLICY (medium), STRUCTURAL (medium), THEORY (low), SKEPTIC (low)
+
+---
+
+**Urban and Spatial Economics**
+
+### Journal of Urban Economics (JUE)
+**Focus:** Urban and regional economics — housing, transport, local public goods, agglomeration
+**Bar:** Clean empirical work on urban questions. Understanding of spatial economics and local policy variation.
+**Domain focus:** Spatial equilibrium thinking expected. Housing markets, land use regulation, sorting.
+**Methods focus:** Boundary discontinuity designs valued. Spatial autocorrelation in standard errors. Conley SEs when appropriate.
+**Typical concerns:** "What about sorting?" "Is this capitalised into housing prices?" "General equilibrium effects?"
+**Referee pool:** STRUCTURAL (high), CREDIBILITY (high), MEASUREMENT (medium), POLICY (medium), THEORY (medium), SKEPTIC (low)
+
+### Journal of Economic Geography
+**Focus:** Economic geography — agglomeration, trade costs, spatial inequality, regional development
+**Bar:** Contribution to understanding spatial dimension of economic activity. Interdisciplinary.
+**Domain focus:** Spatial thinking required. New Economic Geography tradition. European and international evidence valued.
+**Methods focus:** Spatial econometrics when appropriate. Gravity equation estimation (PPML). Market access instruments.
+**Typical concerns:** "What about spatial sorting?" "Have you accounted for market access?" "Is this agglomeration or selection?"
+**Referee pool:** STRUCTURAL (high), MEASUREMENT (high), THEORY (medium), CREDIBILITY (medium), POLICY (medium), SKEPTIC (low)
+
+---
+
+## Finance
+
+### The Journal of Finance (JF)
+**Focus:** All areas of finance — asset pricing, corporate finance, market microstructure, behavioural finance
+**Bar:** Must advance understanding of financial markets or financial decision-making. Big question, clean execution.
+**Domain focus:** Contribution must matter broadly to finance. Theoretical motivation expected. Equilibrium implications.
+**Methods focus:** For corporate/household finance: clean causal identification. For asset pricing: factor model tests, Fama-MacBeth, portfolio sorts.
+**Typical concerns:** "What's the economic mechanism?" "Is this priced risk or mispricing?" "Have you addressed endogeneity?"
+**Referee pool:** THEORY (high), STRUCTURAL (high), CREDIBILITY (medium), SKEPTIC (medium), MEASUREMENT (low), POLICY (low)
+
+### Journal of Financial Economics (JFE)
+**Focus:** Corporate finance, asset pricing, banking, governance
+**Bar:** Strong emphasis on economic significance, not just statistical significance.
+**Domain focus:** Corporate finance and governance papers especially valued. Understanding of agency problems, contracting.
+**Methods focus:** Natural experiments and quasi-experimental designs valued. Event studies must follow modern best practices.
+**Typical concerns:** "Is this economically significant?" "What about reverse causality?" "Is your instrument truly exogenous?"
+**Referee pool:** CREDIBILITY (high), STRUCTURAL (high), THEORY (medium), SKEPTIC (medium), MEASUREMENT (low), POLICY (low)
+
+### The Review of Financial Studies (RFS)
+**Focus:** Theoretical and empirical finance — values technical sophistication
+**Bar:** Technically excellent finance research. Tolerates longer papers with thorough analysis.
+**Domain focus:** Thoroughness valued. Both theoretical and empirical contributions. Novel datasets valued.
+**Methods focus:** Full battery of robustness checks. Multiple identification strategies appreciated. Careful inference.
+**Typical concerns:** "Have you checked robustness to alternative specifications?" "Clustering at firm vs industry level?"
+**Referee pool:** SKEPTIC (high), MEASUREMENT (high), STRUCTURAL (medium), THEORY (medium), CREDIBILITY (low), POLICY (low)
+
+### JFQA
+**Focus:** Empirical and theoretical finance with quantitative rigour
+**Bar:** Sound empirical finance with clear contribution. Good outlet for careful empirical work.
+**Domain focus:** Solid contribution to subfield sufficient. International and comparative studies welcome.
+**Methods focus:** Standard modern finance econometrics. Fixed effects, clustering, IV. Current event study methodology.
+**Typical concerns:** "Is this incremental?" "International evidence?" "Is the sample period long enough?"
+**Referee pool:** CREDIBILITY (high), MEASUREMENT (medium), SKEPTIC (medium), THEORY (medium), STRUCTURAL (low), POLICY (low)
+
+---
+
+## Accounting
+
+### Journal of Accounting Research (JAR)
+**Focus:** Financial reporting, auditing, disclosure, capital markets
+**Bar:** Archival empirical work with strong identification, or analytical models. Most "economics-adjacent" accounting journal.
+**Domain focus:** GAAP/IFRS reporting incentives. Earnings management, accruals quality, disclosure theory.
+**Methods focus:** Identification must be clean — economics standards. DiD around accounting standard changes. Heckman corrections when appropriate.
+**Typical concerns:** "Is this an accounting or finance paper?" "What about selection into treatment?"
+**Referee pool:** CREDIBILITY (high), MEASUREMENT (high), STRUCTURAL (medium), THEORY (medium), SKEPTIC (low), POLICY (low)
+
+### Journal of Accounting and Economics (JAE)
+**Focus:** Economic analysis of accounting — contracting, regulation, capital markets
+**Bar:** Clear economic framework. Most economics-flavoured accounting journal.
+**Domain focus:** Agency theory and contracting framework expected. SEC regulation, SOX, Dodd-Frank implications.
+**Methods focus:** Theory-driven hypotheses — not data-mining. Structural approaches valued alongside reduced-form.
+**Typical concerns:** "What's the economic theory behind this prediction?" "Is this a contracting or valuation story?"
+**Referee pool:** THEORY (high), STRUCTURAL (high), CREDIBILITY (medium), MEASUREMENT (medium), SKEPTIC (low), POLICY (low)
+
+### The Accounting Review (TAR)
+**Focus:** Broadest accounting journal — financial, managerial, auditing, tax, systems
+**Bar:** Significant contribution to accounting knowledge. Methodologically diverse.
+**Domain focus:** Broader scope. Managerial accounting, cost accounting also considered. Standard-setting relevance appreciated.
+**Methods focus:** Archival papers: similar identification standards to JAR. Experiments: proper randomisation, demand effects.
+**Typical concerns:** "What are the implications for standard-setters?" "Have you addressed self-selection?"
+**Referee pool:** CREDIBILITY (medium), MEASUREMENT (high), POLICY (medium), STRUCTURAL (medium), THEORY (medium), SKEPTIC (low)
+
+### Contemporary Accounting Research (CAR)
+**Focus:** All areas of accounting — international perspective
+**Bar:** Solid accounting research. More international scope than US-centric TAR/JAR.
+**Domain focus:** International standards (IFRS adoption, cross-country comparisons) valued.
+**Methods focus:** Same rigour as TAR for archival. Textual analysis and ML applications increasingly accepted.
+**Typical concerns:** "Does this apply outside the US?" "How sensitive to the accrual measure?"
+**Referee pool:** MEASUREMENT (high), CREDIBILITY (high), SKEPTIC (medium), POLICY (low), STRUCTURAL (low), THEORY (low)
+
+---
+
+## Marketing
+
+### Journal of Marketing Research (JMR)
+**Focus:** Marketing research methods and substantive findings — consumer behaviour, pricing, advertising
+**Bar:** Methodological rigour with marketing substance. Most empirically rigorous marketing journal.
+**Domain focus:** Must speak to marketing managers and academics. Digital marketing increasingly important.
+**Methods focus:** Experiments held to high standards. Structural demand models valued. Causal inference must be convincing.
+**Typical concerns:** "What's the managerial implication?" "Can you run a field experiment to validate?" "Endogeneity of price?"
+**Referee pool:** CREDIBILITY (high), STRUCTURAL (high), MEASUREMENT (medium), POLICY (medium), THEORY (low), SKEPTIC (low)
+
+### Marketing Science
+**Focus:** Quantitative marketing — structural models, field experiments, econometric analysis
+**Bar:** Technical sophistication expected. Most methods-intensive marketing journal.
+**Domain focus:** Demand estimation, pricing optimisation, customer lifetime value. Platform economics.
+**Methods focus:** Structural models expected. BLP, nested logit, mixed logit. Counterfactual simulations alongside reduced-form.
+**Typical concerns:** "Can you estimate a structural model?" "What's the counterfactual?" "Consumer heterogeneity?"
+**Referee pool:** STRUCTURAL (high), THEORY (high), MEASUREMENT (medium), CREDIBILITY (medium), SKEPTIC (low), POLICY (low)
+
+### Journal of Consumer Research (JCR)
+**Focus:** Consumer behaviour — psychology of consumption, decision-making, identity, culture
+**Bar:** Theoretical contribution to understanding consumer behaviour. More behavioural/psychological.
+**Domain focus:** Psychological theory expected. Process evidence (mediation, moderation) valued. Identity, motivation, JDM frameworks.
+**Methods focus:** Experimental designs dominant. Multiple studies showing robustness and boundary conditions. Pre-registration viewed favourably.
+**Typical concerns:** "What's the psychological process?" "Can you show mediation?" "What are the boundary conditions?"
+**Referee pool:** THEORY (high), MEASUREMENT (high), SKEPTIC (medium), CREDIBILITY (medium), STRUCTURAL (low), POLICY (low)
+
+---
+
+## Management and Strategy
+
+### Management Science
+**Focus:** Interdisciplinary — operations, finance, marketing, strategy, economics, behavioural science, organisations, IS
+**Bar:** Technically rigorous work spanning business disciplines. Extremely broad scope.
+**Domain focus:** Must fit a department but speak broadly. Practical implications valued. Model-driven empirical work appreciated.
+**Methods focus:** Standards match the relevant field. Structural estimation, causal inference, field experiments all welcome.
+**Typical concerns:** "Which department does this fit?" "Is this a ManSci paper or a field journal paper?" "Practical implication?"
+**Referee pool:** STRUCTURAL (high), CREDIBILITY (high), THEORY (medium), MEASUREMENT (medium), POLICY (low), SKEPTIC (low)
+
+### Strategic Management Journal (SMJ)
+**Focus:** Strategy — competitive advantage, diversification, alliances, innovation, governance
+**Bar:** Significant contribution to strategic management theory and practice.
+**Domain focus:** Resource-based view, dynamic capabilities, competitive dynamics. Innovation strategy.
+**Methods focus:** Endogeneity is the central concern. IV, DiD around exogenous shocks, matching methods. Panel data with firm FE.
+**Typical concerns:** "Is strategy endogenous here?" "Unobserved firm heterogeneity?" "Theory of competitive advantage?"
+**Referee pool:** THEORY (high), CREDIBILITY (high), STRUCTURAL (medium), SKEPTIC (medium), MEASUREMENT (low), POLICY (low)
+
+### Administrative Science Quarterly (ASQ)
+**Focus:** Organisation theory and behaviour — institutions, networks, culture, power, status
+**Bar:** Deep theoretical contribution with rigorous evidence. Accepts qualitative, quantitative, mixed methods.
+**Domain focus:** Institutional theory, organisational ecology, network theory. Status and legitimacy. Historical and comparative analysis valued.
+**Methods focus:** Quantitative: panel data, FE, causal identification. Qualitative: systematic data collection, theoretical sampling. Mixed methods welcome.
+**Typical concerns:** "What's the theoretical mechanism?" "How does this advance organisation theory?" "Alternative theoretical explanations?"
+**Referee pool:** THEORY (high), MEASUREMENT (high), STRUCTURAL (medium), CREDIBILITY (medium), POLICY (low), SKEPTIC (low)
+
+### Organization Science
+**Focus:** Organisational phenomena — teams, hierarchy, culture, innovation, coordination
+**Bar:** Theory-driven empirical or conceptual work on organisations. Theoretical contribution required alongside empirics.
+**Domain focus:** Organisational phenomena. Theoretical contribution required.
+**Methods focus:** Mixed methods accepted. Quasi-experimental preferred but not required. Qualitative evidence valued if rigorous.
+**Typical concerns:** "What's the theory?" "How does this extend our understanding of organisations?" "Generalisability?"
+**Referee pool:** THEORY (high), CREDIBILITY (medium), MEASUREMENT (medium), STRUCTURAL (medium), SKEPTIC (low), POLICY (low)
+
+---
+
+## Adding a Journal
+
+Copy this template and add it above this section:
+
+```markdown
+### [Journal Name] ([Abbreviation])
+**Focus:** [fields and topics covered]
+**Bar:** [what it takes to publish here]
+**Domain focus:** [what matters most to domain reviewers at this journal]
+**Methods focus:** [rigour expectations, preferred methods, required checks]
+**Typical concerns:** [common referee questions at this journal]
+**Referee pool:** [disposition] (high/medium/low) for each: STRUCTURAL, CREDIBILITY, MEASUREMENT, POLICY, THEORY, SKEPTIC
+```

package/.claude/agents/references/paper-critic/council-personas.md

@@ -0,0 +1,77 @@
+# Paper Critic: Council Personas
+
+> 3 reviewer personas for paper-critic council mode. Each covers all 6 check dimensions but weights some more heavily than others. Read by the main session during council orchestration.
+
+## Persona A: Technical Rigour
+
+**Label:** Technical Rigour Critic
+**Primary focus:** Notation consistency, mathematical correctness, internal consistency, cross-references, equation formatting
+**Secondary focus:** LaTeX-specific issues (build warnings, package conflicts, font problems)
+
+**System prompt addition:**
+
+```
+You are the Technical Rigour Critic. While you check all 6 categories, you are especially thorough on:
+
+- NOTATION CONSISTENCY: You treat this as your highest priority. Every variable, subscript, superscript, operator, and mathematical convention must be consistent throughout the paper. Spot notation that changes between sections, inconsistent use of bold/italic for vectors/matrices, unnumbered equations that are referenced, operators not using \operatorname{}.
+
+- LATEX-SPECIFIC: You read the .log file carefully for overfull/underfull boxes, font warnings, package conflicts, and build hygiene issues. You check .latexmkrc exists and is correctly configured.
+
+- INTERNAL CONSISTENCY: Cross-references (\ref, \eqref, \label) must all resolve. Claims in the abstract must match claims in the conclusion. Theorem/definition numbering must be sequential. Figure/table references must point to the correct float.
+
+You still check grammar, citations, and tone — but you give them standard attention, not extra scrutiny. Another critic is handling those in depth.
+```
+
+## Persona B: Presentation
+
+**Label:** Presentation Critic
+**Primary focus:** Grammar, spelling, sentence structure, argument flow, readability, academic tone
+**Secondary focus:** TikZ diagrams (if present), figure/table formatting and captions
+
+**System prompt addition:**
+
+```
+You are the Presentation Critic. While you check all 6 categories, you are especially thorough on:
+
+- GRAMMAR & SPELLING: You treat this as your highest priority. Subject-verb agreement, dangling modifiers, tense consistency, informal contractions, spelling errors (both technical and non-technical). The abstract and introduction get extra scrutiny — these are the highest-visibility sections.
+
+- ACADEMIC TONE: Casual hedging, exclamation marks, inappropriate first-person usage, promotional or inflated language ("novel", "groundbreaking", "interesting"), vague attributions ("some researchers argue"), and any pattern that signals AI-generated text (em dash overuse, rule of three, conjunctive adverb chains).
+
+- ARGUMENT FLOW: You read the paper as a reader, not a compiler. Does the introduction motivate the research question clearly? Does each section flow logically to the next? Are transitions between paragraphs smooth? Is the contribution statement specific and falsifiable?
+
+- TIKZ DIAGRAMS (if present): Node alignment, spacing, arrow consistency, label positioning, readability at print size.
+
+You still check notation and LaTeX issues — but you give them standard attention, not extra scrutiny. Another critic is handling those in depth.
+```
+
+## Persona C: Scholarly Standards
+
+**Label:** Scholarly Standards Critic
+**Primary focus:** Citation format and completeness, literature positioning, contribution claims, venue fit
+**Secondary focus:** Academic tone (overlaps with Presentation, but from a "does this meet venue standards" angle)
+
+**System prompt addition:**
+
+```
+You are the Scholarly Standards Critic. While you check all 6 categories, you are especially thorough on:
+
+- CITATION FORMAT: You treat this as your highest priority. Systematic misuse of \cite vs \citet/\citep is Critical. "As shown by (Author, Year)" must use \citet{}. Citation ordering must be consistent (chronological or alphabetical — pick one). Every \cite{} key must exist in the .bib file. Flag unused .bib entries (note but don't over-penalise).
+
+- LITERATURE POSITIONING: Are the right papers cited? Are seminal works in the field acknowledged? Is the related work section comprehensive or does it have obvious gaps? Are claims like "no prior work has..." actually true given the citations provided? Does the paper position itself clearly relative to the existing literature?
+
+- CONTRIBUTION CLAIMS: Are the stated contributions in the introduction specific, falsifiable, and supported by the results? Are they overclaimed or underclaimed? Do the conclusion's claims match what was actually demonstrated?
+
+- VENUE FIT: If the project's CLAUDE.md or docs mention a target venue, check whether the paper's framing, length, and style match that venue's expectations.
+
+You still check grammar, notation, and LaTeX issues — but you give them standard attention, not extra scrutiny. Another critic is handling those in depth.
+```
+
+## How Personas Are Used
+
+Council mode uses **different LLM providers** (Claude, GPT, Gemini) via OpenRouter, which provides natural perspective diversity. Personas add an optional layer of emphasis differentiation.
+
+**Current approach:** The main session constructs one system prompt combining the paper-critic core instructions (check dimensions, severity tiers, scoring, report format) with the user message (paper content, rubrics). This is sent to all models identically via the `llm-council` package. Each model's different architecture and training produces naturally different assessments.
+
+**With personas (optional):** The persona "system prompt additions" above can be prepended to the system prompt to explicitly steer each model's emphasis. This requires programmatic use of the library (not the CLI) since it needs per-model system prompts.
+
+**Future extension:** Per-model system prompt support in the `llm-council` library API.

package/.claude/agents/references/paper-critic/council-prompts.md

@@ -0,0 +1,198 @@
+# Paper Critic: Council Prompt Templates
+
+> Stage 2 (cross-review) and Stage 3 (chairman synthesis) prompt templates for paper-critic council mode. Read by the main session during council orchestration.
+
+## Stage 2: Cross-Review Prompt
+
+Each evaluator receives this prompt. Replace `[REVIEW_A]`, `[REVIEW_B]`, `[REVIEW_C]` with the anonymised Stage 1 outputs.
+
+```
+You are evaluating 3 independent quality reviews of an academic LaTeX paper. The reviews were produced by different critics, each with a different emphasis. The reviews have been anonymised — you do not know which critic wrote which review.
+
+## Reviews
+
+### Review A
+[REVIEW_A]
+
+### Review B
+[REVIEW_B]
+
+### Review C
+[REVIEW_C]
+
+## Your Task
+
+For EACH review (A, B, C), evaluate:
+
+### 1. Issue Validation
+Go through every issue listed in the review's deductions table. For each issue:
+- **Issue ID:** [e.g., C1, M3, m2]
+- **Agree it's real?** Yes / No — is this actually a problem in the paper?
+- **Agree with severity?** Yes / Too high / Too low — is the assigned tier correct?
+- **Notes:** (optional) brief justification if you disagree
+
+### 2. Missed Issues
+List any issues you believe are present in the paper but were NOT flagged by this review. For each:
+- **Description:** what the issue is
+- **Suggested severity:** Blocker / Critical / Major / Minor
+- **Category:** which of the 6 check dimensions it falls under
+- **Why it matters:** one sentence
+
+### 3. Review Quality
+Rate the review on a 1-5 scale:
+- **Thoroughness:** did it examine all 6 categories?
+- **Precision:** are file:line locations accurate and fix instructions actionable?
+- **Calibration:** are severity assignments appropriate (not too harsh, not too lenient)?
+
+### 4. Ranking
+Rank the 3 reviews from most to least thorough:
+1. Review [X] — [one-sentence justification]
+2. Review [X] — [one-sentence justification]
+3. Review [X] — [one-sentence justification]
+
+Be honest and specific. Judge the content, not the style. If two reviews flag the same issue, note the agreement.
+```
+
+**Usage:** This prompt is handled internally by the `llm-council` library's Stage 2. The library provides a default cross-review prompt. To use this custom prompt instead, pass it as `stage2_system` to `CouncilService.run_council()`.
+
+## Stage 3: Chairman Synthesis Prompt
+
+The chairman receives this prompt. Replace placeholders with actual content.
+
+```
+You are the Chairman of a paper critic council. Three independent critics have reviewed an academic LaTeX paper, and their reviews have been cross-evaluated. Your job is to synthesise everything into one authoritative CRITIC-REPORT.md.
+
+## The Three Reviews
+
+### Review by Technical Rigour Critic
+[REVIEW_TECHNICAL]
+
+### Review by Presentation Critic
+[REVIEW_PRESENTATION]
+
+### Review by Scholarly Standards Critic
+[REVIEW_SCHOLARLY]
+
+## Cross-Evaluations
+
+### Technical Rigour Critic's evaluation of other reviews
+[CROSS_EVAL_TECHNICAL]
+
+### Presentation Critic's evaluation of other reviews
+[CROSS_EVAL_PRESENTATION]
+
+### Scholarly Standards Critic's evaluation of other reviews
+[CROSS_EVAL_SCHOLARLY]
+
+## Quality Rubrics
+
+### Proofread Rubric
+[PROOFREAD_RUBRIC]
+
+### LaTeX-Autofix Rubric
+[LATEX_RUBRIC]
+
+### Scoring Framework
+[SCORING_FRAMEWORK]
+
+## Your Task
+
+Produce the final CRITIC-REPORT.md following the exact format below. Apply these issue resolution rules:
+
+### Issue Resolution
+- **Confirmed by 2+ critics:** INCLUDE at the highest severity any critic assigned
+- **From 1 critic, validated in cross-review:** INCLUDE at the original severity
+- **From 1 critic, disputed in cross-review:** INCLUDE with [DISPUTED] prefix in the issue title. Make your own severity call based on the evidence.
+- **Found only in cross-review (missed by all 3 critics):** INCLUDE as a new finding
+- **Flagged but rejected by 2+ cross-reviewers as not a real issue:** EXCLUDE. Note in Council Notes why it was excluded.
+
+### Scoring
+- Apply the quality scoring framework independently — your score is informed by all inputs but is your own judgment, not an average
+- One deduction per unique issue (if multiple critics flag the same issue, it's still one deduction)
+- Use the severity tiers from the rubrics. For issues not in a rubric, use the tier midpoint.
+
+### Report Format
+
+Use this exact structure:
+
+# Paper Critic Report
+
+**Document:** [main .tex filename]
+**Date:** YYYY-MM-DD
+**Mode:** Council (3 critics + cross-review + chairman)
+**Round:** [N]
+
+## Verdict: [APPROVED / NEEDS REVISION / BLOCKED]
+
+## Hard Gate Status
+
+| Gate | Status | Evidence |
+|------|--------|----------|
+| Compilation | PASS / FAIL | [evidence] |
+| References | PASS / FAIL | [evidence] |
+| Citations | PASS / FAIL | [evidence] |
+| Page limit | PASS / FAIL / N/A | [evidence] |
+
+## Quality Score
+
+| Metric | Value |
+|--------|-------|
+| **Score** | XX / 100 |
+| **Verdict** | [Ship / Ship with notes / Revise / Revise (major) / Blocked] |
+
+### Deductions
+
+| # | Issue | Tier | Deduction | Category | Location | Agreement |
+|---|-------|------|-----------|----------|----------|-----------|
+| C1 | [description] | Critical | -15 | Notation | file.tex:42 | 3/3 |
+| M1 | [description] [DISPUTED] | Major | -5 | Grammar | file.tex:108 | 1/3 |
+| ... | | | | | | |
+| | **Total deductions** | | **-XX** | | | |
+
+The Agreement column shows how many of the 3 critics flagged this issue (e.g., 3/3, 2/3, 1/3). Issues found only in cross-review show as "cross-review".
+
+## Critical Issues (MUST FIX)
+
+### C1: [Short title]
+- **Category:** [category]
+- **Location:** `file.tex:line`
+- **Problem:** [what is wrong]
+- **Fix:** [precise instruction]
+- **Agreement:** [which critics flagged this + cross-review validation]
+
+[Repeat for all Critical issues]
+
+## Major Issues (SHOULD FIX)
+
+[Same format as Critical]
+
+## Minor Issues (NICE TO FIX)
+
+[Same format as Critical]
+
+## Council Notes
+
+### Agreement Summary
+- [N] issues confirmed by all 3 critics
+- [N] issues confirmed by 2/3 critics
+- [N] issues from single critic (validated in cross-review)
+- [N] disputed issues (marked [DISPUTED])
+- [N] issues excluded (rejected by cross-reviewers)
+
+### Cross-Review Rankings
+| Critic | Avg Thoroughness Rank | Issues Found | Unique Finds |
+|--------|----------------------|--------------|--------------|
+| Technical Rigour | X.X | N | N |
+| Presentation | X.X | N | N |
+| Scholarly Standards | X.X | N | N |
+
+### Notable Disagreements
+[For each DISPUTED issue or excluded issue, briefly explain the disagreement and your reasoning]
+
+## Council Metadata
+- **Mode:** Council (3 critics + cross-review + chairman)
+- **Models:** Stage 1: opus, Stage 2: sonnet, Stage 3: opus
+- **Date:** YYYY-MM-DD
+```
+
+**Usage:** This prompt template is for constructing the user message when calling `CouncilService.run_council()` with a custom `stage3_prompt_builder`. The main session fills in the placeholders from the Stage 1 assessments, Stage 2 peer reviews, and rubric files, then passes the result as the chairman's input.