Skip to content

scdenney/open-science-skills / plugin/skills/narrative-building

narrative-building

Verified

Drafts or audits a scientific paper’s introduction and framing — the real-world "why", an evidence audit, the funnel from theory to a falsifiable if-then with a named estimand, bridges across experiments, and a contribution list matched to the design. Use when the user asks to write, restructure, or critique an introduction, motivate or frame a paper, repair a gap-in-the-literature opening, sharpen or calibrate a contribution claim, or connect experiments into one argument. Sentence-level flow and wording are out of scope.

FileKindSize
SKILL.mdskill md14.5 KB
reference/example-funnel.mdmarkdown6.2 KB

reference/example-funnel.md

View on GitHub
# Worked example: Why→If-Then funnel

**Topic:** Cross-partisan tolerance under economic threat.

This example shows how the same research topic reads as (a) a weak, gap-driven opening, (b) a strengthened, Why-driven opening, and (c) a funnel that lands on an If-Then hypothesis with a named estimand. The annotation at the end explains what changed.

---

## Weak opening (gap-driven, methods-first)

> A growing literature examines affective polarization in the United States. While many studies document partisan animosity, fewer use experimental methods, and those that do often rely on student samples. To address this gap, we field a survey experiment on a national sample that manipulates economic threat and measures cross-partisan tolerance. Our design extends prior work by using a new conjoint framework. We find evidence that economic threat shapes tolerance, and we discuss implications for the literature.

**Why this is weak.**
- The motivation is the literature's shape ("fewer use experimental methods"), not a real-world stake.
- The "invisible" is never named — what belief, perception, or reasoning is the experiment designed to reveal?
- The target population is implicit ("a national sample"), so the scope of the theoretical claim is unclear.
- "Extends prior work by using a new conjoint framework" is methods-driven: the design is justified by its novelty rather than by a question that requires it.
- The hypothesis is absent. The reader does not know what would falsify the argument.

---

## Strengthened opening (Why-driven)

> When Americans lose jobs and income during economic downturns, do they become more willing to cooperate across party lines because shared hardship is a unifier, or less willing because scarcity sharpens in-group loyalty? Survey data can tell us whether self-reported tolerance correlates with personal economic distress, but cannot reveal the underlying reasoning: whether citizens interpret economic threat as a shared predicament calling for cross-partisan problem-solving, or as a competitive pressure that raises the cost of tolerating opponents. The analytical joint -- where existing theory is silent -- is whether the rally-round-the-flag logic that operates under external threat (Mueller 1970) extends to diffuse economic threat, which lacks an obvious out-group. We field a pre-registered experiment on a probability-based sample of U.S. adults, manipulating the salience of personal economic threat, to adjudicate between these two accounts of how material conditions shape tolerance toward co-nationals of the opposing party.

**Why this is stronger.**
- The opening names a real-world tension (downturns and cross-partisan cooperation) before any method.
- The "invisible" is explicit: the *reasoning* citizens bring to cross-partisan choices under economic pressure, which administrative correlates cannot access.
- The analytical joint is named: whether rally-round-the-flag extends to threats that lack an external out-group.
- The target population is explicit (U.S. adults, probability-based sample), so the PATE interpretation is on the table from the first paragraph.
- The experiment is justified because the question is about reasoning under manipulated conditions, not about observed correlations.

---

## Funnel to If-Then

> Rally-round-the-flag accounts predict that any threat -- including diffuse economic threat -- activates a superordinate national identity that dampens partisan animus (Mueller 1970). Social-identity accounts predict the opposite when the threat lacks an external out-group: scarcity activates the most available in-group, which in a polarized context is the partisan one (Iyengar et al. 2019). These accounts make opposing predictions about how economic threat changes behavior toward co-national partisan opponents. Our target estimand is the average marginal effect of salient personal economic threat on a conjoint-based measure of willingness to cooperate with opposing-party co-nationals, at the U.S. adult population level. **If** rally-round-the-flag extends to diffuse economic threat, **then** respondents assigned to the economic-threat prime will show higher willingness to cooperate with opposing-party co-nationals than control respondents (H1, confirmatory, pre-registered SESOI of 3 percentage points on the cooperation index).

**If-Then specification:**
- **If** salient personal economic threat activates a superordinate identity, **Then** cooperation with opposing-party co-nationals increases relative to control; **If** it activates partisan identity, **Then** cooperation decreases relative to control. The design has power to detect either signed effect at the pre-registered SESOI.
- **Estimand:** PATE of the economic-threat prime on the conjoint cooperation index.
- **Severity:** a null result is informative because both theoretical accounts predict non-null signed effects; the design is therefore a severe test of the "threat activates some identity" family.

---

## What changed, step by step

1. **Why (Step 1).** The opening moved from "there is a gap in the literature" to "there is a real-world tension we do not understand." The invisible (citizens' reasoning under economic pressure) and the analytical joint (whether rally-round-the-flag extends to diffuse threat) are named before any method appears.
2. **Evidence audit (Step 2).** Mueller and Iyengar et al. are cited as *competing predictions*, not as a chronological list. The review sets up a theory-adjudication mission, not a fact-search.
3. **Funnel (Step 3).** The transition explicitly names the target population, the identifying variation (the economic-threat prime), the estimand (PATE on the conjoint cooperation index), and the SESOI before stating the hypothesis.
4. **Multi-experiment bridges (Step 4).** Not applicable in this single-experiment example; a two-experiment version would add a bridging paragraph explaining why Experiment 2 is logically required (e.g., to separate personal from sociotropic threat) rather than a menu option.
5. **Coherence check (Step 5).** The contribution claim is scoped to what the pre-registered design can deliver: a severe test of whether the rally-round-the-flag mechanism operates under diffuse economic threat for U.S. adults. It does not claim to resolve affective polarization writ large.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall