draft-section

name	draft-section
description	Drafts academic prose for a manuscript section from bullet points or an outline. Use when writing or expanding a section.
argument-hint	<section>: <notes>
allowed-tools	Bash, Read, Write, Edit, Glob, Grep
version	1.1.0
workflow_stage	writing
tags	["manuscript","prose","IMRAD","writing"]

Draft Manuscript Section

Draft academic prose for a manuscript section from bullet points or an outline.

Arguments

• $ARGUMENTS — section name and content notes (e.g., "Introduction: regional GDP disparities, panel data from 2000-2020, spatial econometrics, contributes to convergence literature")

Steps

1. Parse the section name and content bullets/notes from the arguments.

2. Read index.qmd to understand:

   • The manuscript's existing tone and writing style
   • What sections already exist and their content
   • Citation conventions used (narrative @key vs parenthetical [@key])
   • What figures and tables are embedded (to reference them)

3. Read references.bib to know which citations are available for use.

4. Draft 2–5 paragraphs of academic prose:

   • Write in the register of empirical economics journals (AER, QJE, ReStud style)
   • Use formal but accessible language
   • Structure paragraphs logically: general → specific, or claim → evidence → implication
   • Include Quarto cross-references where appropriate (@sec-, @fig-, @tbl-)
   • Insert citations from references.bib where they strengthen the argument
   • Where a citation would be helpful but none exists in .bib, insert a placeholder: [CITE: description of needed reference]

5. Apply section-specific templates depending on the target section:

   • Introduction: Hook (relevance or puzzle) → Research question statement → Main result preview (one sentence) → Brief methodology description → Literature positioning (identify 2–3 strands the paper contributes to) → Roadmap paragraph ("The remainder of this paper is organized as follows...")
   • Data / Empirical Strategy: Data source and sample construction → Variable definitions → Summary statistics reference → Estimation equation with clear notation → Identification assumptions and threats
   • Results: Lead with the main coefficient and its interpretation → Progress from simple to complex specifications → Discuss economic significance alongside statistical significance → Heterogeneity results → Brief robustness summary (point to appendix or robustness section)
   • Conclusion: Restate research question and answer → Policy implications → Theoretical implications → Honest limitations → Future research directions. A conclusion should NOT merely summarize — it should interpret.

6. Follow economics writing conventions:

   • Use present tense for established facts and theory; past tense for own empirical analysis ("We find..." → "We found..." or "Column (2) shows...")
   • Use precise causal vs. associational language: "is associated with" for correlations, "causes" or "leads to" only with credible identification
   • Always interpret economic magnitude alongside statistical significance (e.g., "a one standard deviation increase in X is associated with a Y-unit change in the outcome, equivalent to Z% of the sample mean")
   • Lead with results, not methodology
   • Use hedging language: "results suggest", "we find evidence consistent with", "the estimates imply"

7. If the section matches an existing section in index.qmd (e.g., "Introduction" matches ## Introduction {#sec-introduction}):

   • Show how the draft would replace the current [FILL:] placeholders or extend existing content
   • Preserve any {{< embed >}} shortcodes already in that section

8. Present the draft to the user for review. On approval, insert or replace the content in index.qmd.

Error handling

• If no section name is provided, ask the user which section to draft.
• If the arguments are too vague, ask for more specific content points.

Common Pitfalls

• Burying the main result deep in the paper — lead with the finding, then build support
• Confusing statistical significance with economic significance — always report magnitudes in interpretable units
• Overclaiming causality without proper identification — match language to research design
• Writing a conclusion that merely summarizes results — a conclusion should add interpretation, implications, and limitations
• Excessive passive voice — prefer active constructions ("We estimate..." over "It is estimated that...")

References

• Cochrane (2005) "Writing Tips for Ph.D. Students"
• Shapiro (2022) "How to Give an Applied Micro Talk"
• Thomson (2011) "A Guide for the Young Economist"
• Nikolov (2013) "Writing Tips for Economics Research Papers"