brycewang-stanford/paper-submission

Paper Submission Evaluator

Overview

This skill evaluates an academic paper and produces a comprehensive submission target report. It performs four assessments:

1. Contribution Novelty Score (0-100): How much the paper advances existing literature, assessed via web search for related work.
2. Best-Fit SSCI Journal Fields: Which ABS field categories best match the paper.
3. Appropriate ABS Star Rating: What star level (1, 2, 3, 4, 4*) the paper's quality warrants.
4. Top 20 Journal Recommendations: From the 2 best-fitting fields, at the recommended ABS star level, list 20 SSCI-indexed journals with rationale.

The ABS journal list is read from C:\Users\Admin\.claude\skills\paper-submission\asset\Business School Journal List 2023.pdf (referred to as AJG2025). This file is bundled with the skill in the asset/ folder, so it is always available regardless of changes to the user's desktop.

The final report is saved as target.pdf in the paper's directory.

Workflow

Phase 1: Initialization

1. Receive paper path from user (via $ARGUMENTS or ask). Accept either:
   • A PDF file containing the full paper
   • A project folder containing LaTeX files (main.tex, tables, etc.)
2. Read the paper thoroughly:
   • For PDF: Read all pages using the Read tool with page ranges.
   • For LaTeX project: Read main.tex and key result files.
3. Extract key information and record internally:
   • Research question / hypothesis
   • Methodology (empirical strategy, identification, data)
   • Data source and sample (country, market, time period)
   • Key findings (baseline results, mechanisms, heterogeneity)
   • Stated or implied contributions
   • Keywords and JEL codes (if present)
4. Summarize the paper in ~200 words for use in subsequent phases. Present this summary to the user for confirmation:

   我已阅读论文，以下是摘要：
   
   研究问题：[...]
   方法：[...]
   数据：[...]
   主要发现：[...]
   贡献方向：[...]
   
   请确认以上理解是否正确，或进行调整。
   

5. Wait for user confirmation before proceeding.

---

Phase 2: Literature Novelty Assessment (Web Search Required)

This phase evaluates how novel the paper's contribution is relative to existing literature. Web search is mandatory.

Step 2a: Identify Search Dimensions

Based on the paper summary, identify 3-5 search dimensions that capture the paper's core novelty claims. Each dimension represents a facet of the paper's contribution. Examples:

• Topic novelty: "Has [X effect on Y] been studied before?"
• Methodological novelty: "Has [identification strategy Z] been applied to [this question]?"
• Data novelty: "Has [this data source / market / country] been used for [this question]?"
• Mechanism novelty: "Have [these channels] been documented?"
• Setting novelty: "Has [this policy / institutional context] been exploited?"

Step 2b: Conduct Web Searches

For each dimension, conduct at least 2 targeted web searches using WebSearch. Search queries should be in English and target academic literature. Example queries:

• "[dependent variable]" AND "[independent variable]" site:ssrn.com OR site:nber.org
• "[key mechanism]" AND "[research context]" journal article
• [topic keywords] survey OR review OR meta-analysis

For each search:

1. Execute the WebSearch query.
2. Read the top results using WebFetch to check abstracts and findings.
3. Record: (a) how many closely related papers exist, (b) how the current paper differs from them, (c) whether the core finding has been documented before.

Step 2c: Score the Contribution

Assign a novelty score out of 100 based on these criteria:

| Score Range | Meaning | Criteria | |-------------|---------|----------| | 85-100 | Highly novel | No prior paper addresses this exact question with this approach. Opens a new research direction. | | 70-84 | Substantially novel | Few prior papers on a similar topic, but this paper offers a clearly distinct angle (new data, new mechanism, new identification). | | 55-69 | Moderately novel | Topic has been studied, but this paper contributes incremental insights (new setting, additional robustness, extension of known results). | | 40-54 | Limited novelty | Multiple papers have addressed similar questions with similar methods. Contribution is primarily confirmatory or extends to a new sample. | | 0-39 | Low novelty | The main findings have been well-documented. Contribution is marginal. |

For each dimension, assign a sub-score and weight. The final score is the weighted average.

Present the assessment to the user:

文献创新性评估结果：

维度1: [dimension name] — 子分 [X]/100
  已有文献：[list 2-3 most relevant prior papers with year]
  本文区别：[how this paper differs]

维度2: [dimension name] — 子分 [X]/100
  ...

综合创新性得分：[SCORE]/100
评级：[Highly novel / Substantially novel / Moderately novel / Limited novelty / Low novelty]

主要创新点：
1. [innovation point 1]
2. [innovation point 2]
3. [innovation point 3]

潜在风险：
- [e.g., "Reviewer may argue that [X] has been shown by [Author, Year]"]

---

Phase 3: Field Matching and ABS Star Rating

Step 3a: Identify Best-Fit Fields

The ABS journal list uses these field categories (22 fields total):

• ACCOUNT (Accounting)
• FINANCE (Finance)
• ECON (Economics)
• STRAT (Strategy)
• MKT (Marketing)
• OPS&TECH (Operations and Technology)
• OR&MANSCI (Operations Research and Management Science)
• ORG STUD (Organization Studies)
• HRM&EMP (Human Resource Management and Employment)
• IB&AREA (International Business and Area Studies)
• INNOV (Innovation)
• PUB SEC (Public Sector Management)
• SOC SCI (Social Sciences)
• SECTOR (Sector Studies)
• BUS & ECON HIST (Business and Economic History)
• MDEV&EDU (Management Development and Education)
• REGIONAL (Regional Studies)
• ENT-SBM (Entrepreneurship and Small Business Management)
• ETHICS-CSR-MAN (Ethics, CSR and Management)
• INFO MAN (Information Management)
• PSYCH (GENERAL) (Psychology - General)
• PSYCH (WOP-OB) (Psychology - Work and Organizational)

Based on the paper's topic, methodology, and data, identify the 2 most suitable fields. Consider:

• What is the paper's primary disciplinary home? (e.g., corporate finance paper → FINANCE)
• What is a strong secondary field? (e.g., uses accounting data → ACCOUNT; studies innovation → INNOV)
• Where would the paper's contribution resonate most?

Step 3b: Determine Appropriate ABS Star Rating

Assess the paper's quality level to determine the appropriate ABS star tier for targeting:

| Star Level | Criteria | |------------|----------| | 4* | World-leading journals. Paper must have: exceptional novelty (score 85+), rigorous identification, clean causal story, broad implications, polished writing. Very selective — only recommend if the paper is truly outstanding. | | 4 | Top field journals. Paper should have: high novelty (score 70+), solid identification strategy, clear contribution, well-executed empirics. | | 3 | Highly regarded journals. Paper should have: moderate-to-high novelty (score 55+), reasonable identification, clear results, good execution. | | 2 | Well-recognized journals. Paper with: some novelty (score 40+), standard methodology, sound results. | | 1 | Recognized journals. Paper with: limited novelty, basic methodology, narrow contribution. |

Decision rules:

• Novelty score alone does not determine the star level — also consider methodology rigor, data quality, writing quality, and scope of implications.
• If the paper uses a novel identification strategy (natural experiment, RDD, etc.), upgrade by 0.5 star.
• If the paper uses Chinese data targeting international journals, be realistic: Chinese-market papers rarely appear in 4* journals unless the finding has universal implications.
• Be honest and calibrated. Over-optimistic recommendations waste the author's time.

Present the assessment:

领域匹配与星级评估：

最佳匹配领域：
  1. [Field 1] — [rationale]
  2. [Field 2] — [rationale]

建议投稿星级：ABS [N] 星
理由：
- 创新性：[novelty score] 分，[assessment]
- 方法论：[methodology assessment]
- 数据质量：[data assessment]
- 贡献范围：[scope assessment]

是否同意以上评估？如需调整星级，请告知。

Wait for user confirmation before proceeding.

---

Phase 4: Journal Recommendations

Step 4a: Extract Journal Data from ABS PDF

Read the ABS journal list PDF (C:\Users\Admin\.claude\skills\paper-submission\asset\Business School Journal List 2023.pdf) using a Python script to extract all journals matching:

• Field = one of the 2 identified fields
• ABS star rating = the recommended level (also include one level above and one level below for reference)
• SSCI indexed = Yes (prioritize SSCI journals, but include non-SSCI journals as backup)

Use the following Python approach via Bash:

import fitz
doc = fitz.open(r'C:\Users\Admin\.claude\skills\paper-submission\asset\Business School Journal List 2023.pdf')
# Parse the tabular data from each page
# Extract: ISSN, Field, Journal Title, ABS rating, ABDC rating, SSCI status, JCR quartile, JIF

Step 4b: Rank and Select 20 Journals

From the extracted journals, select the top 20 recommendations across the 2 fields. Ranking criteria:

1. Field relevance: How well the journal's scope matches the paper's topic
2. Star level match: Journals at the recommended star level ranked first
3. SSCI status: SSCI-indexed journals preferred
4. JIF: Higher impact factor preferred (as tiebreaker)
5. Publication precedent: Journals that have published similar topics (based on your knowledge)

Organize the list as:

• Field 1: 10 journals (ranked by fit)
• Field 2: 10 journals (ranked by fit)

For each journal, provide:

• Journal name
• ABS star rating
• SSCI status and JCR quartile
• JIF (if available)
• 1-sentence rationale for why this journal fits the paper

---

Phase 5: Report Generation

Generate the final report as target.pdf saved in the paper's directory (or user-specified location).

Report Structure

The report should contain:

═══════════════════════════════════════════
        论文投稿目标评估报告
        Paper Submission Target Report
═══════════════════════════════════════════

生成日期：[YYYY-MM-DD]
论文标题：[Paper title if available]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

一、论文概要
[200-word paper summary]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

二、文献创新性评估

综合得分：[SCORE]/100 — [Rating]

[For each dimension:]
维度 [N]: [Name] — [Sub-score]/100
  相关文献：[2-3 papers]
  本文创新：[How this paper differs]

主要创新点：
1. [...]
2. [...]
3. [...]

潜在审稿风险：
- [...]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

三、目标领域与星级

最佳领域：[Field 1], [Field 2]
建议星级：ABS [N] 星

评估维度：
- 创新性：[...]
- 方法论：[...]
- 数据质量：[...]
- 贡献范围：[...]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

四、推荐期刊（共20本）

[Field 1 Name]（10本）:
┌────┬──────────────────────┬──────┬──────┬──────┬─────────────────────────┐
│ #  │ Journal              │ ABS  │ SSCI │ JIF  │ 推荐理由                │
├────┼──────────────────────┼──────┼──────┼──────┼─────────────────────────┤
│ 1  │ ...                  │ ...  │ ...  │ ...  │ ...                     │
│ ...│                      │      │      │      │                         │
└────┴──────────────────────┴──────┴──────┴──────┴─────────────────────────┘

[Field 2 Name]（10本）:
┌────┬──────────────────────┬──────┬──────┬──────┬─────────────────────────┐
│ #  │ Journal              │ ABS  │ SSCI │ JIF  │ 推荐理由                │
├────┼──────────────────────┼──────┼──────┼──────┼─────────────────────────┤
│ 1  │ ...                  │ ...  │ ...  │ ...  │ ...                     │
│ ...│                      │      │      │      │                         │
└────┴──────────────────────┴──────┴──────┴──────┴─────────────────────────┘

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

五、投稿建议

[2-3 paragraphs of strategic advice:]
- Which journal to try first and why
- Backup strategy if rejected
- Any adjustments to the paper that would improve chances at higher-tier journals

PDF Generation Method

Use the Python script scripts/generate_report.py to produce the PDF. The script uses fpdf2 with Chinese font support (SimSun from C:\Windows\Fonts\simsun.ttc).

Interaction pattern:

报告已生成并保存至：[path]/target.pdf

报告包含：
- 创新性评估：[SCORE]/100
- 推荐领域：[Field 1], [Field 2]
- 推荐星级：ABS [N] 星
- 推荐期刊：20本（每个领域10本）

---

Important Notes

• Web search is mandatory for Phase 2. Do not skip novelty assessment.
• Be calibrated and honest in scoring. An inflated score wastes the author's time on unrealistic targets.
• ABS journal list PDF is the authoritative source for journal data. Extract data programmatically from the PDF — do not rely on memory alone.
• SSCI indexing is strongly preferred but not strictly required. If fewer than 10 SSCI journals exist in a field at the target star level, supplement with non-SSCI journals and mark them clearly.
• Star level flexibility: The 20 journal recommendations should primarily be at the recommended star level. If fewer than 10 journals exist at that level in a field, include journals from one level above or below, clearly marked.
• The report language is Chinese for headings and explanations, English for journal names and academic content.
• If the user provides a specific target journal or field preference, adjust recommendations accordingly.
• The output file is always named target.pdf unless the user specifies otherwise.