quant-findings-writer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Quantitative Findings Writer

量化研究结果撰写工具

Draft Results/Findings sections for quantitative sociology articles using structural patterns discovered in 83 Social Problems and Social Forces articles.

基于从83篇《Social Problems》和《Social Forces》期刊文章中总结的结构模式，为量化社会学文章起草结果/发现部分。

Project Integration

项目集成

This skill reads from

project.yaml

when available:

yaml

undefined

本技能会在可用时读取

project.yaml

文件内容：

yaml

undefined

From project.yaml

type: quantitative # This skill is for quantitative projects paths: drafts: drafts/sections/ tables: output/tables/ figures: output/figures/


**Project type:** This skill is designed for **quantitative** projects.

Consumes output from **r-analyst** or **stata-analyst** (tables, figures, interpretation memos from Phase 5).

Updates `progress.yaml` when complete:
```yaml
status:
  results_draft: done
artifacts:
  results_section: drafts/sections/results-section.md

type: quantitative # This skill is for quantitative projects paths: drafts: drafts/sections/ tables: output/tables/ figures: output/figures/


**项目类型**：本技能专为**量化**项目设计。

可接收**r-analyst**或**stata-analyst**的输出内容（第5阶段产出的表格、图表、解读备忘录）。

完成后会更新`progress.yaml`文件：
```yaml
status:
  results_draft: done
artifacts:
  results_section: drafts/sections/results-section.md

Connection to Other Skills

与其他技能的关联

Skill	Relationship	Details
r-analyst	Upstream	Produces tables, figures, interpretation memos (Phase 5 output)
stata-analyst	Upstream	Same as r-analyst but for Stata
article-bookends	Downstream	Takes results section as input for framing
methods-writer	Parallel	Methods section written alongside or before results
lit-synthesis	Upstream	Provides theoretical framework for theory-linking
prose-craft	Craft guide	Sentence/paragraph benchmarks (evaluative mode); tone, anti-LLM rules

技能	关联关系	详情
r-analyst	上游	产出表格、图表、解读备忘录（第5阶段输出）
stata-analyst	上游	和r-analyst功能一致，适用于Stata工具链
article-bookends	下游	接收结果部分作为框架撰写的输入
methods-writer	并行	方法部分可与结果部分同步或提前撰写
lit-synthesis	上游	提供理论关联所需的理论框架
prose-craft	写作指南	句子/段落基准（评估模式）；语气、反LLM规则

File Management

文件管理

This skill uses git to track progress across phases. Before modifying any output file at a new phase:

Stage and commit current state:

git add [files] && git commit -m "quant-findings-writer: Phase N complete"

Then proceed with modifications.

Do NOT create version-suffixed copies (e.g.,

-v2

-final

-working

). The git history serves as the version trail.

本技能使用git追踪各阶段进度。在新阶段修改任何输出文件前：

暂存并提交当前状态：

git add [files] && git commit -m "quant-findings-writer: Phase N complete"

再进行后续修改。

请勿创建带版本后缀的副本（例如

-v2

、

-final

、

-working

），git历史会作为版本追溯依据。

Workflow

工作流

Phase 1: Orient

阶段1：确定方向

Gather from the user:

Method type: secondary-survey-analysis, administrative-data, or content-analysis
Key results: tables, model output, or thematic findings to present
Theoretical predictions: hypotheses or expectations the results address
Target length: typical is 12-25 paragraphs (2,000-5,000 words)

If the user has already written a draft, read it and assess which cluster it most resembles before suggesting revisions.

从用户处收集以下信息：

方法类型：二手调查分析、行政数据分析或内容分析
核心结果：待展示的表格、模型输出或主题发现
理论预测：结果对应验证的假设或预期
目标篇幅：通常为12-25个段落（2000-5000字）

如果用户已写好草稿，先阅读草稿并判断其最符合的聚类，再给出修改建议。

Phase 2: Select Cluster

阶段2：选择聚类

Present the 7 clusters with their canonical arcs. Recommend 1-2 based on method type and analytic strategy:

Cluster	Best for	Arc
Progressive Model Builder	Regression-heavy papers building from simple to complex specs	DESCRIBE → BASELINE → ELABORATE → MECHANISM → ROBUSTNESS
Hypothesis Tester	Papers with numbered H1/H2/H3 predictions	SETUP → BASELINE → ELABORATE → SUBGROUP → SUMMARY
Decomposition Analyst	Gap/disparity papers using Oaxaca-Blinder, mediation	DESCRIBE → BASELINE → DECOMPOSE → MECHANISM → ROBUSTNESS
Subgroup Comparator	Heterogeneity-focused papers (by race, gender, class)	DESCRIBE → BASELINE → SUBGROUP → COMPARISON → ROBUSTNESS
Temporal Tracker	Event studies, trend analysis, periodization	TEMPORAL → BASELINE → TEMPORAL → SUBGROUP → ROBUSTNESS
Thematic Explorer	Content analysis with qualitative themes/frames	THEMATIC → THEMATIC → THEMATIC → SUMMARY
Causal Inference Specialist	DiD, IV, RDD, matching designs	SETUP → BASELINE → ELABORATE → ROBUSTNESS → MECHANISM

Selection heuristics:

Survey data + model progression → Progressive Model Builder
Admin data + quasi-experimental design → Causal Inference Specialist
Admin data + inequality decomposition → Decomposition Analyst
Any method + explicit hypotheses → Hypothesis Tester
Any method + group comparisons as central question → Subgroup Comparator
Content analysis + thematic coding → Thematic Explorer
Panel/longitudinal + change over time → Temporal Tracker

After the user selects a cluster, read the matching guide from

clusters/{cluster-name}.md

for detailed arc, paragraph budget, signature techniques, and exemplar patterns.

展示7种聚类及其标准叙事弧，根据方法类型和分析策略推荐1-2种适配选项：

聚类	适用场景	叙事弧
渐进式模型构建类	回归分析为主、从简单到复杂搭建模型规格的论文	描述→基线→扩展→机制→稳健性
假设验证类	有编号H1/H2/H3预测的论文	铺垫→基线→扩展→子群体→总结
分解分析类	使用Oaxaca-Blinder、中介效应的差距/差异研究论文	描述→基线→分解→机制→稳健性
子群体比较类	聚焦异质性分析的论文（按种族、性别、阶层划分）	描述→基线→子群体→比较→稳健性
时间趋势追踪类	事件研究、趋势分析、分期研究	时间维度→基线→时间维度→子群体→稳健性
主题探索类	带有定性主题/框架的内容分析	主题→主题→主题→总结
因果推断专项类	双重差分、工具变量、断点回归、匹配设计类研究	铺垫→基线→扩展→稳健性→机制

选择启发规则：

调查数据+模型递进→渐进式模型构建类
行政数据+准实验设计→因果推断专项类
行政数据+不平等分解→分解分析类
任意方法+明确假设→假设验证类
任意方法+核心问题为群体比较→子群体比较类
内容分析+主题编码→主题探索类
面板/纵向数据+随时间变化→时间趋势追踪类

用户选择聚类后，读取

clusters/{cluster-name}.md

对应的指南，获取详细的叙事弧、段落分配、特色技巧和范例模式。

Phase 3: Build the Arc

阶段3：搭建叙事弧

Using the cluster guide, construct a section outline:

Map each major finding/table to a MOVE from the standardized vocabulary
Sequence moves following the cluster's canonical arc
Allocate paragraphs using the cluster's paragraph budget
Identify the opening and closing moves

Standardized move vocabulary:

Move	Function
DESCRIBE	Descriptive statistics, sample overview, bivariate patterns
SETUP	Methodological restatement, analytic strategy recap
BASELINE	Initial/simple models, main effects without interactions
ELABORATE	Add complexity: interactions, nonlinearities, mediators
DECOMPOSE	Formal decomposition (Oaxaca-Blinder, mediation, etc.)
SUBGROUP	Heterogeneity by subgroups (race, gender, class)
MECHANISM	Mediation, mechanism tests, process tracing
ROBUSTNESS	Sensitivity analysis, alternative specs, placebo tests
THEMATIC	Substantive theme/topic analysis
TEMPORAL	Over-time patterns, periodization, event studies
COMPARISON	Cross-group or cross-context comparison
VISUAL	Key figure/visualization driving the narrative
SUMMARY	Brief recap paragraph
TRANSITION	Bridge to discussion section

Present the arc to the user as a numbered outline with paragraph counts per move.

使用聚类指南，构建章节大纲：

将每个核心发现/表格匹配到标准化词汇表中的对应布局步骤
按照聚类的标准叙事弧排列布局步骤顺序
参考聚类的段落分配方案规划各部分段落数
确定开篇和收尾的布局步骤

标准化布局步骤词汇表：

步骤	功能
DESCRIBE	描述性统计、样本概览、双变量关联
SETUP	方法重述、分析策略复盘
BASELINE	初始/简单模型、无交互项的主效应
ELABORATE	增加复杂度：交互项、非线性、中介变量
DECOMPOSE	正式分解（Oaxaca-Blinder、中介效应等）
SUBGROUP	子群体异质性分析（种族、性别、阶层）
MECHANISM	中介效应、机制验证、过程追踪
ROBUSTNESS	敏感性分析、替代规格、安慰剂检验
THEMATIC	实质主题/话题分析
TEMPORAL	跨时间模式、分期、事件研究
COMPARISON	跨群体或跨情境比较
VISUAL	支撑叙事的核心图表/可视化内容
SUMMARY	简要回顾段落
TRANSITION	通往讨论部分的过渡

将叙事弧作为带编号的大纲展示给用户，标注每个步骤对应的段落数。

Phase 4: Draft

阶段4：起草内容

Write each move following corpus norms. Consult

techniques/techniques.md

for the full technique catalog.

Opening paragraph (choose one based on cluster):

Table reference (58% of corpus): "Table 2 presents results from..."
Sample description (20%): "Before turning to multivariate models, I describe..."
Hypothesis restatement (14%): "Recall that H1 predicted..."
Methodological setup (5%): "To estimate the causal effect, I use..."

Body paragraphs:

Lead with the finding, not the method
Translate every key coefficient into substantive terms (85% of corpus does this)
Use attenuation tracking when adding controls: "the coefficient falls from .34 to .21"
Connect to theory at moderate density: ~1 theory reference per 3-4 paragraphs for most clusters
Report null findings transparently (45% of corpus does this)

Closing paragraph (choose one):

Robustness cascade (18%): "Results are robust to..."
Strongest finding (18%): save the most important result for the end
Subgroup analysis (17%): end with heterogeneity
Supplemental reference (14%): "Additional specifications in Appendix Table A3..."
Summary (11%): brief recap of all findings

Cross-cutting norms:

Median section length: ~18 paragraphs, 3 tables/figures referenced
75% use hybrid table strategy: tables anchor the narrative but prose interprets
55% link results to theory heavily; 40% moderately; only 5% minimally
Distinguish statistical from practical significance when warranted

遵循语料库规范撰写每个布局步骤的内容，可查阅

techniques/techniques.md

获取完整的技巧目录。

开篇段落（根据聚类选择一种）：

表格引用（占语料库的58%）："表2呈现了来自...的结果"
样本描述（20%）："在进入多变量模型分析前，我先描述..."
假设重述（14%）："回顾可知H1预测..."
方法铺垫（5%）："为估计因果效应，我使用..."

正文段落：

以发现作为开头，而非方法
将每个核心系数转化为实质含义（语料库中85%的文章遵循此规则）
添加控制变量时跟踪系数变化："系数从0.34下降到0.21"
以适中密度关联理论：多数聚类每3-4段约有1次理论引用
透明汇报零结果（语料库中45%的文章遵循此规则）

收尾段落（选择一种）：

稳健性说明（18%）："结果在...情况下依然稳健"
核心发现展示（18%）：将最重要的结果放在末尾
子群体分析（17%）：以异质性分析收尾
补充材料引用（14%）："附录表A3中的额外规格..."
总结（11%）：简要回顾所有发现

通用规范：

章节长度中位数：约18个段落，引用3个表格/图表
75%使用混合表格策略：表格作为叙事支撑，正文内容负责解读
55%的文章紧密关联理论，40%关联度适中，仅5%关联度极低
必要时区分统计显著性和实际显著性

Phase 5: Calibrate

阶段5：校准调整

After drafting, check against cluster norms:

Does the arc match the canonical sequence?
Is the paragraph budget balanced?
Are tables referenced with interpretive guidance, not just pointed at?
Is theory linking at the right density for the cluster?
Are robustness checks present if the cluster expects them?
Are null findings acknowledged rather than buried?

Present the draft with a brief calibration note.

起草完成后，对照聚类规范检查：

叙事弧是否符合标准顺序？
段落分配是否均衡？
引用表格时是否附带解读指导，而非仅提及？
理论关联密度是否符合聚类的要求？
若聚类要求稳健性检验，是否已包含相关内容？
零结果是否被明确提及而非刻意隐藏？

提交草稿时附带简短的校准说明。

Reference Files

参考文件

Cluster guides (read the one matching the selected cluster):

```
clusters/progressive-model-builder.md
```
```
clusters/hypothesis-tester.md
```
```
clusters/decomposition-analyst.md
```
```
clusters/subgroup-comparator.md
```
```
clusters/temporal-tracker.md
```
```
clusters/thematic-explorer.md
```
```
clusters/causal-inference-specialist.md
```

```
techniques/techniques.md
```
— 20 writing techniques with descriptions and frequency data
```
references/corpus-statistics.md
```
— summary statistics from the 83-article analysis corpus

聚类指南（读取与所选聚类匹配的文件）：

```
clusters/progressive-model-builder.md
```
```
clusters/hypothesis-tester.md
```
```
clusters/decomposition-analyst.md
```
```
clusters/subgroup-comparator.md
```
```
clusters/temporal-tracker.md
```
```
clusters/thematic-explorer.md
```
```
clusters/causal-inference-specialist.md
```

```
techniques/techniques.md
```
— 包含20种写作技巧，附带说明和使用频率数据
```
references/corpus-statistics.md
```
— 83篇文章分析语料库的汇总统计数据