Capstone: Write a Complete Research Paper

You’ve learned every step: research question design, literature review, methodology, results and discussion, citations, and revision. Now bring it all together by walking through a complete paper from start to finish.

This capstone isn’t about writing an entire paper in one sitting. It’s about demonstrating the full workflow — showing how each technique connects to produce a coherent, publishable manuscript.

🔄 Quick Recall: Across this course you’ve learned: AI-assisted topic exploration (Lesson 2), literature discovery with Semantic Scholar and Elicit (Lesson 3), methodology structuring (Lesson 4), Results and Discussion drafting (Lesson 5), citation management and integrity (Lesson 6), and revision strategies (Lesson 7). The capstone integrates all of them.

The Complete Workflow

Here’s the full AI-assisted research paper process in one view:

Phase 1: Foundation (Lessons 2-3)

Step 1: Research Question

Topic funnel: Broad area → Specific gap → FINER-evaluated question
AI role: Explore subtopics, identify gaps, evaluate question quality
Your role: Choose the question, verify the gap exists

Step 2: Literature Review

Discovery: Semantic Scholar → 30-50 candidate papers
Screening: Elicit → structured extraction table
Synthesis: AI assistant → identify 3-5 themes
Your role: Read key papers yourself, verify all citations

Phase 2: Writing (Lessons 4-5)

Step 3: Methodology (write first)

AI role: Structure subsections, check completeness, suggest stats tests
Your role: Provide all actual study details, verify test assumptions

Step 4: Results (write second)

AI role: Format statistics, create tables, ensure APA reporting
Your role: Run the actual analysis, verify all numbers

Step 5: Discussion (write third)

AI role: Connect findings to literature, identify logical gaps
Your role: Interpret results, form conclusions, acknowledge limitations

Step 6: Introduction (write fourth)

AI role: Structure the narrative arc (broad → specific → gap → contribution)
Your role: Frame the research problem, state your contribution

Step 7: Abstract (write last)

AI role: Condense into word limit, ensure all sections represented
Your role: Verify accuracy, check it matches the actual paper

✅ Quick Check: Why do experienced researchers write the Introduction after the Discussion, not before? (Answer: The Introduction promises what the paper delivers. You can only write that promise accurately once you know what the paper actually delivers — your findings, their interpretation, and their implications. Writing the Introduction last ensures alignment between what you promise and what you provide.)

Capstone Exercise: Write One Complete Section

Choose one section of a paper (real or practice) and apply the full workflow:

Option A: Literature Review Section

1. Define scope: Choose 3-4 themes for a topic you know
2. Discover: Search Semantic Scholar for 10-15 papers across your themes
3. Extract: Create a table with: Author, Year, Method, Key Finding, Theme
4. Synthesize: Use AI to draft thematic paragraphs connecting studies
5. Verify: Check every citation against the original paper
6. Revise: Run the three-pass revision (structure, clarity, tone)

Option B: Methodology + Results

1. Outline: List your study details (design, participants, procedures, analysis)
2. Draft methodology: Use AI for structure, provide all specifics yourself
3. Present results: Format your findings in APA style with proper statistics
4. Review: Use AI as a methodology reviewer to catch missing elements
5. Revise: Ensure replicability — could someone repeat your study from this description?

Option C: Full Paper Outline + Abstract

1. Design: Create a research question using the topic funnel
2. Outline: Generate a complete IMRAD outline with AI
3. Abstract: Write a 250-word abstract covering all sections
4. Cover letter: Draft a cover letter for your target journal
5. Disclosure: Write an AI disclosure statement

Quality Checklist

Before considering any section “done”:

CONTENT:
□ Every claim supported by a citation or your own data
□ No hallucinated references (all verified)
□ Logical flow from paragraph to paragraph
□ Research question clearly addressed

WRITING:
□ Appropriate hedging language ("suggests" not "proves")
□ Consistent tense (past for what you did, present for established facts)
□ Clear topic sentences for each paragraph
□ No unnecessary jargon

INTEGRITY:
□ AI use disclosed per journal requirements
□ AI-generated content properly cited
□ All co-authors reviewed and approved
□ Plagiarism check completed

FORMAT:
□ Matches target journal guidelines
□ Within word count limits
□ Figures and tables referenced in text
□ Reference format consistent throughout

Course Recap

Where to Go Next

Start your paper. The best next step is applying these techniques to a real project. Pick one section and work through the workflow.

Explore specialized tools. Try Elicit for systematic reviews, Connected Papers for citation mapping, or PapersGPT for Zotero integration.

Stay current on policies. Journal AI policies are evolving quickly. Check your target journal’s guidelines before every submission.

Share what works. If you find an AI prompt that produces excellent academic prose, save it. Build a personal library of prompts for each paper section.

Key Takeaways

• Write in this order: Methods → Results → Discussion → Introduction → Abstract
• AI drafts are only as good as the specifics you provide — gather your notes first
• The complete workflow integrates discovery, extraction, synthesis, drafting, verification, and revision
• You are the author — final responsibility for accuracy, integrity, and quality is always yours
• Build a personal prompt library for each section type to speed up future papers