Skills for Real Engineers. Straight from my .agents directory.
science-stats-report
Agent skill fromyzlin286-wq/Deepscient
Use when a statistical result has to leave the analysis tool and enter a manuscript — when writing the statistics paragraph, the table notes, or the figure caption that carries a p-value; when someone asks how to report a test, an effect size, or a confidence interval; when a result is significant but the assumption check failed; when a chi-square had Yates correction or not and the value differs between tools; also use when a reviewer asks for the numbers behind a claim.
How to install science-stats-report
npx skills add yzlin286-wq/Deepscient# or add only this skill to Claude Code: copy the folder
dsh-science/skills/science/science-stats-report/
# into
~/.claude/skills/science-stats-report/ (all projects)
.claude/skills/science-stats-report/ (one project)
1# 把统计结果写成审稿人认的句子23## 一句话标准45一个结果要能独立成立,必须同时给出:**样本量 → 统计量 + 自由度 → p → 效应量 + 置信区间 → 用的是哪个检验、为什么**。6缺任何一项,这一段就会被要求补充;缺效应量最常见。78## 模板(照抄结构,别照抄数字)910```11Welch 双样本 t 检验比较咖啡因组(n = 24,M = 421 ms,SD = 52)与安慰剂组12(n = 24,M = 452 ms,SD = 55)的简单反应时,两组差异未达显著,13t(44.79) = −1.95,p = .058,Cohen's d = −0.57(95% CI −1.15 至 0.02)。14```1516要点:1718- `t(44.79)` 的自由度带小数,因为用的是 Welch 而不是合并方差 t。写成整数说明你用了等方差假设,或者算错了。19- `p = .058` 不写 `p = 0.058`(多数刊要求省略前导零),`p < .001` 才用小于号。20- 效应量后面必须括号给 CI。**CI 跨 0 时不能说"有趋势"**——趋势是没做检验的说法。21- 组统计量(M、SD、n)写在括号里,读者不用回表。2223## 不同检验的报法2425| 检验 | 统计量格式 | 效应量 | 常见错 |26| --- | --- | --- | --- |27| Welch t | `t(df) = x.xx` | Cohen's d + CI | df 写成整数 |28| 配对 t | `t(n−1) = x.xx` | Cohen's dz + CI | 用独立样本 d 代替 dz |29| ANOVA | `F(df1, df2) = x.xx` | η² 或 partial η² | 只报 F 不报 η² |30| 卡方 | `χ²(df) = x.xx` | Cramér's V | 不说明有没有 Yates 校正 |31| Mann-Whitney | `U = x.xx` | rank-biserial r | 报成"中位数检验" |32| 回归 | `β = x.xx, SE = x.xx, t = x.xx` | 调整 R²(整体)/ 半偏 r²(每个系数) | 只报 p 不报系数与 SE |33| Logistic | `OR = x.xx (95% CI …)` | McFadden R² / AUC | OR 与 RR 混用 |34| 生存 | `HR = x.xx (95% CI …)` | 中位生存 / RMST | 不做比例风险检查就报 HR |3536## 卡方:Yates 校正是必须先声明的3738同一个 2×2 表,校正与否统计量差很多(例:7.92 vs 9.09)。`science_stats` 默认用 Yates 并同时给出未校正值,39写在 `uncorrected` 里。正文要说清用的是哪个:4041> `χ²(1) = 7.92,p = .005(Yates 校正);未校正 χ² = 9.09,p = .003。`4243两个都给不丢人,只给一个又不说哪个才是问题。4445## 假设不满足时怎么报4647不能说"虽然不满足正态性,但 t 检验通常稳健"——这句话在有离群点的小样本里是错的。4849正确写法:5051> 反应时分布右偏(Shapiro-Wilk,p < .001),故改用 Mann-Whitney U 检验:52> U = 118,p < .001,rank-biserial r = −0.74。结论方向与 t 检验一致。53…
Excerpt shown under the repository's MIT license.
More skills in Deepscient
- ▸science-agent-scheduling
- ▸science-campaign-design
- ▸science-consensus-report
- ▸science-discipline-methods
- ▸science-experiment-runbook
- ▸science-figure-for-journal
- ▸science-frontier-to-paper
- ▸science-fulltext-hunting
- ▸science-journal-intel
- ▸science-journal-selection
- ▸science-literature-search
- ▸science-manuscript-formatting
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
An AI skill that provides design intelligence for building professional UI/UX across multiple platforms.
