discussion¶
Pack: 100xOS shared skills
Category:
draftingField: economics
License:
private (curator-owned)Updated: 2026-05-20
Stages:
paper-draftingCurator-private skill — copy text from 100xOS/shared/skills/writing/discussion.md.
Discussion Section¶
Purpose¶
The Discussion section interprets results — it does not restate them. It connects findings to theory, prior literature, and practice, then honestly addresses limitations and future directions.
5 DOs¶
- Interpret: Explain what results mean, not what numbers are. "The negative coefficient on X implies that agents respond to…"
- Connect to theory: Relate findings back to the theoretical framework or conceptual model. Do predictions hold? Which mechanisms are supported?
- Practical implications: State concrete, proportional implications. "Regulators could … because our evidence shows …" — not "this has important implications for policy."
- Limitations: Be honest but constructive. Frame limitations as scope conditions ("our results apply to X but may not generalize to Y") rather than fatal flaws. Every empirical paper has limitations; the goal is transparency, not self-destruction.
- Future research: Identify specific, answerable follow-up questions that emerge from your findings. Not "more research is needed" — instead, "a natural extension would be to test whether Z holds when…"
5 DON'Ts¶
- Don't repeat results: The reader just read the Results section. No "As shown in Table 3, we found that…" — instead, synthesize.
- Don't overclaim: Match claims to evidence strength. If your design identifies a local average treatment effect, don't claim a universal law.
- Don't be defensive: Limitations are not apologies. State them matter- of-factly and, where possible, explain why they don't undermine the core contribution.
- Don't introduce new results: All empirical findings belong in Results or Robustness. Discussion is for interpretation.
- Don't write a second conclusion: The Discussion is broader and more speculative than the Conclusion. Keep the Conclusion tight and forward- looking.
Structure Template (1.5–2 pages)¶
Text Only
\section{Discussion}
% 1. Summary of main findings (1 paragraph)
% Synthesize — don't restate. "Three key findings emerge…"
% 2. Theoretical implications (2-3 paragraphs)
% How do results advance, refine, or challenge existing theory?
% Which mechanisms are supported by the evidence?
% CRITICAL: What does this finding teach us beyond the specific setting?
% If the finding implies something about a broader market, institution,
% or behavior, this is the paper's most important contribution and must
% be prominently discussed.
% 3. Practical implications (1-2 paragraphs)
% Concrete, proportional to evidence.
% Who should care and what should they do differently?
% 4. Limitations (1-2 short paragraphs, 150-300 words total)
% Scope conditions, data constraints, threats to external validity.
% For each limitation, explain why it does not invalidate the core result.
% MAXIMUM 3-4 limitations. Written as flowing prose, NOT as a numbered
% or bulleted list.
% 5. Future research (1 short paragraph, 100-150 words)
% 2-3 specific research questions that follow from your findings.
% Written as prose, NOT as a list. Keep it tight.
Formatting — MANDATORY¶
The entire discussion section must be written as flowing prose paragraphs. No bullet points. No numbered lists. No enumerated items. This includes limitations and future research subsections.
If a limitation or future research direction deserves mention, it gets woven into a paragraph. If it cannot be woven in, it is probably not important enough to include.
Common Mistakes and Fixes¶
| Mistake | Fix |
|---|---|
| "Our results show that X has a significant effect on Y" (restating) | "The responsiveness of Y to X implies that market participants…" (interpreting) |
| "This has important implications for policy" (generic) | "Exchanges could reduce wash trading by implementing X, since our evidence shows…" (specific) |
| "More research is needed" (vague) | "A natural next step is to test whether this pattern holds in markets with…" (specific) |
| "A limitation is that our data only covers 2020-2023" (defensive) | "Our sample covers 2020-2023, capturing the full NFT boom-bust cycle; extending to later periods would test persistence" (constructive) |
| "Despite these limitations, our results are robust" (hand-waving) | [Delete this sentence — the Robustness section already made this case] |
| Numbered or bulleted limitations list | Write as connected prose paragraphs |
| 500+ word limitations section | Keep to 150-300 words, 3-4 key scope conditions |
| Generic future research wish list | 1 paragraph, 2-3 specific questions, 100-150 words |
IS / Management Science Conventions¶
- Explicit theoretical contribution: State which propositions/hypotheses from your model are supported and which need refinement.
- Design implications: If studying a designed system (marketplace, protocol, mechanism), translate findings into design principles.
- Generalizability: Explicitly discuss which settings your results transfer to and which contextual factors moderate the effects.
- Boundary conditions: State the conditions under which your theory applies. Good theory has clear boundaries.