spss-amos-sem
Questionnaire structural equation modeling in the SPSS + AMOS workflow that Chinese management, psychology, education, nursing and tourism theses use — data cleaning of an SPSS .sav file, reverse coding, straight-lining, demographics, Harman common-method-bias test, Cronbach's alpha / CITC reliabili…
Install / Use
npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill spss-amos-semInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
AutomationSupported Platforms
Tags
Our assessment of spss-amos-sem
spss-amos-sem scores 94/100 on our quality scale, 412th of 2,869 Automation skills we index (top 15%).
Its SKILL.md is 8.2 KB long, well organised into 15 sections with 5 code examples: a thorough specification that gives an agent plenty to work with.
With 4,360 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated 40 days ago, so spss-amos-sem is actively maintained.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
spss-amos-sem compared with similar skills
All 4 of these similar skills score higher than spss-amos-sem; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| spss-amos-sem (this skill)by brycewang-stanford | 94 | 4.4k | 40d ago | SKILL.md |
| Agent-Reachby Panniantong | 100 | 108.3k | 2d ago | CLAUDE.md |
| Scraplingby D4Vinci | 100 | 93.4k | today | MCP Server |
| rufloby ruvnet | 100 | 80.7k | today | MCP Server |
| algorithmic-artby anthropics | 100 | 177.9k | 41d ago | SKILL.md |
Frequently asked questions
- How do I install spss-amos-sem?
- Run
npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill spss-amos-sem. The install tabs above show the steps for each supported agent. - Which AI agents does spss-amos-sem work with?
- It is written for Zed, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is spss-amos-sem safe to use?
- It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is spss-amos-sem still maintained?
- The repository was last updated 40 days ago, so spss-amos-sem is actively maintained.
Skill content
View source on GitHubname: spss-amos-sem description: Questionnaire structural equation modeling in the SPSS + AMOS workflow that Chinese management, psychology, education, nursing and tourism theses use — data cleaning of an SPSS .sav file, reverse coding, straight-lining, demographics, Harman common-method-bias test, Cronbach's alpha / CITC reliability, EFA (KMO, Bartlett), CFA convergent validity (standardized loadings, CR, AVE), discriminant validity (Fornell-Larcker, HTMT), model fit (chi2/df, CFI, TLI, RMSEA, SRMR), structural paths and hypothesis tests, bootstrap mediation with specific indirect effects, multi-group measurement invariance, and Hayes PROCESS models 1/4/6/7/8/14. One spec card drives an R/lavaan engine set to AMOS defaults and writes ready-to-run SPSS syntax (.sps), AMOS Program Editor VB.NET code, AMOS user-defined estimands, an AMOS drawing checklist with the numbers AMOS should show, Chinese three-line tables in Word, a results-chapter draft and a transparent model-modification log. Use when the user mentions SPSS, AMOS, SEM, structural equation model, CFA, confirmatory factor analysis, questionnaire / survey / Likert scale data, reliability and validity, CR / AVE, HTMT, common method bias, mediation, moderation, PROCESS, bootstrap indirect effect, multi-group analysis, measurement invariance, or Chinese terms 结构方程, 结构方程模型, 验证性因子分析, 探索性因子分析, 问卷, 量表, 信度, 效度, 收敛效度, 区分效度, 组合信度, 共同方法偏差, 中介效应, 调节效应, 有调节的中介, 多群组, 测量不变性, 模型拟合, 修正指数, 毕业论文问卷分析. license: CC-BY-SA-4.0
问卷数据结构方程全流程(SPSS + AMOS)
一张设定卡(YAML)驱动整条流程:R/lavaan 按 AMOS 的默认设置把所有数字算出来,同时生成用户可以在 SPSS / AMOS 里原样复现的语法和操作步骤,再把结果排成中文论文的三线表和结果段落。
.sav 数据 ──► draft_spec.R ──► 设定卡 spec.yml ──► run_sem.R ──► 输出文件夹
▲ │
└──── 模型修正(写理由)◄─────────┘
先说清楚能做什么、不能做什么
| 能做 | 不能做 / 要说明 |
|---|---|
| 用 lavaan 复算 AMOS 的 ML 估计(第一载荷固定为 1、χ² = (N−1)F、偏差校正 bootstrap),数字和 AMOS 一致到小数点后两三位 | 不能运行 SPSS 或 AMOS 本身。生成的 .sps 和 AMOS 步骤要用户自己在装了软件的电脑上跑 |
| 生成 SPSS 语法:清洗、描述统计、Harman、信度、EFA、相关、VIF、PROCESS | AMOS 的 VB.NET 程序按 IBM Amos 28 用户手册的语法生成,未在真实 AMOS 里跑过;主路线是 Amos Graphics 画图清单 |
| PROCESS 模型 1/4/6/7/8/14(OLS + 百分位 bootstrap,与宏的算法一致) | 不做 PLS-SEM(SmartPLS)、潜变量交互(LMS)、贝叶斯 SEM、纵向 / 多层 SEM |
| 中文三线表(Word)、结果段落初稿、路径图、模型修正记录 | 理论模型(构念、路径、假设)由用户的研究设计决定,不由数据倒推 |
硬规则(每次都要遵守)
- 数字只能来自本次运行的
results.json/results.xlsx。 不要凭印象写 AMOS 输出,不要编造任何一个系数、拟合值或置信区间。用户贴来 SPSS / AMOS 原始输出时,以那份输出为准,并用 references/08-lavaan-amos-parity.md 解释差异。 - 模型修正只通过设定卡的
modifications,每条必须写理由。 引擎会自动核对理由和数据(被删题项在上一步模型里的载荷、误差相关的 MI 和排名),不符时在模型修正记录.md里打 ⚠。不要替用户删题、加误差相关来"凑"拟合;跨构念误差相关和交叉载荷不建议采纳。规则细节见 references/05-fit-and-modification.md。 - 未达标就如实写未达标。 结果段落里的「未达标」「不成立」「不显著」是真实结果,不改写成达标;可以帮用户解释原因、给出改进方向或局限性写法。
- 只引用 references/citations.md 里的文献作为阈值出处。需要别的文献时告诉用户自己查证,不要编。
--quick只用来检查设定卡和数据(bootstrap 只有 200 次),交付前一定用正式模式重跑。- 告诉用户 AMOS 的 C.R. 是 z 值(临界比),不是组合信度 CR——这是问卷论文里最常见的错误之一。
工作流程
0. 环境
需要 R ≥ 4.1 和这些包(第一次用时安装):
install.packages(c("yaml", "lavaan", "semTools", "psych", "GPArotation", "haven", "openxlsx",
"officer", "flextable", "jsonlite", "readxl", "ragg"))
ragg 是可选的(画路径图更清晰);读 .xlsx 需要 openxlsx,读 .xls 需要 readxl。
1. 看数据,起草设定卡
Rscript scripts/draft_spec.R 用户数据.sav --out spec.yml
脚本会打印变量总览(名称、SPSS 标签、取值范围、缺失),按变量名前缀把题项分组(PU1 PU2… → PU),找出与同组负相关的疑似反向题,写一份带 TODO 的设定卡草稿。
2. 和用户确认设定卡
只问数据里看不出来的东西,一次问完:
- 每个构念的中文名和题项归属(草稿按前缀分组,常常要改);
- 结构路径和研究假设(来自用户的理论模型 / 开题报告,不要自己编);
- 反向题是否属实(对照问卷原文);
- 要不要检验具体中介路径、多群组、PROCESS 调节模型。
字段说明见 assets/templates/spec_template.yml,完整示例见 assets/demo/spec_demo.yml。构念代码 ≤ 6 个字母(PROCESS 宏要求变量名不超过 8 个字符,生成的均值变量叫 m_代码)。
3. 快速试跑
Rscript scripts/run_sem.R spec.yml --out 输出文件夹 --quick
读 results.json 和 模型修正记录.md,向用户汇报:样本清洗结果、信度、EFA 是否有归错维度或交叉载荷的题、CFA 载荷 < 0.5 的题、拟合指标、区分效度。
4. 决定是否修正模型
按 references/05-fit-and-modification.md 的顺序:先看低载荷题项(< 0.5,且题干确实有问题才删;每个构念至少保留 3 题),再看同一构念内、题干相似的误差相关。每条修正都要用户同意并写进 modifications(带理由),然后重跑。
5. 正式运行并交付
Rscript scripts/run_sem.R spec.yml --out 输出文件夹
在 demo 数据上,SEM bootstrap 2000 次加 PROCESS 5000 次大约 15 秒。交付时按下表告诉用户每个文件怎么用:
| 文件 | 用途 |
|---|---|
| 结果表格.docx | 中文三线表(宋体 + Times New Roman,五号),直接贴进论文第四章 |
| 结果段落.md | 结果部分初稿,数字全部来自本次运行;用户按学校格式改措辞 |
| 路径图.png | 结构模型标准化路径图(300 dpi),实线显著、虚线不显著 |
| 模型修正记录.md | 每一步修正的内容、理由、数据依据和拟合变化,答辩被问时用 |
| analysis.sps | SPSS 语法:改第一行文件夹路径后全选运行,输出同时导出为 spss_output_tables.xlsx |
| <数据名>_cleaned.sav | 清洗后的数据(反向计分、剔除无效问卷、加了 m_ 均值变量),AMOS 直接读 |
| amos/AMOS操作步骤.md | Amos Graphics 画图与设置清单,附每个 AMOS 表格的预期数值 |
| amos/01_CFA_VB程序.vb、02_SEM_VB程序.vb | AMOS Program Editor 程序(不用手画;未实测,报错时以画图为准) |
| amos/03_自定义估计量.txt | 具体间接效应的 estimands,粘贴到 AMOS 的 Define new estimands |
| results.xlsx / results.json | 全部中间结果;json 供 Claude 读取和核对 |
| lavaan_model.txt | 与 AMOS 设定一致的 lavaan 模型语法 |
6. 用户带回 SPSS / AMOS 输出时
- SPSS:读
spss_output_tables.xlsx(OMS 导出),逐项对照results.xlsx。α、KMO、载荷、相关系数应完全一致;不一致通常是反向题、缺失值处理或题项归属不同。 - AMOS:对照
AMOS操作步骤.md里的预期值。点估计和拟合指标应一致到小数点后两位左右;bootstrap 置信区间因随机数不同会有第三位小数的差异,但显著性判断应一致。差异来源见 references/08-lavaan-amos-parity.md。
设定卡速查
| 字段 | 必填 | 说明 |
|---|---|---|
| data | ✓ | 数据文件,相对设定卡的路径;.sav/.zsav/.dta/.csv/.xlsx/.xls |
| constructs | ✓ | 代码: {label: 中文名, items: [题项...]},代码 ≤ 6 个字母 |
| structural | | 因变量: [自变量...],只支持递归(无环)模型;不写就只做测量模型 |
| hypotheses | | H1: X -> Y,必须对应一条结构路径;不写就自动编号 |
| indirect | | H8: X -> M -> Y(可多级),偏差校正 bootstrap 检验 |
| reverse_items / scale_points | | 反向题与李克特点数(反向 = 点数 + 1 − 原值) |
| missing | | listwise(默认,AMOS 要求完整数据)或 fiml |
| drop_straightliners | | 剔除所有题目同一答案的问卷(在反向计分之前判断) |
| demographics / controls | | 人口统计变量;控制变量(数值型)指向所有内生构念 |
| multigroup | | 2 组分组变量:测量不变性 + 路径组间差异 |
| process | | {model: 4, x: X, m: [M], y: Y, w: W, covariates: [...], center: true} |
| modifications | | {type: drop_item, item: X, reason: ...} 或 {type: error_cov, items: [A, B], reason: ...} |
| bootstrap / process_bootstrap / seed | | 默认 2000 / 5000 / 20261027 |
判断标准速查
| 指标 | 标准 | 出处 | |---|---|---| | Cronbach's α | > 0.7 | Nunnally (1978) | | CITC | > 0.5(常用经验值) | — | | KMO / Bartlett | KMO > 0.7,Bartlett 显著 | Kaiser (1974) | | Harman 第一因子 | < 40% | Podsakoff et al. (2003);周浩、龙立荣 (2004) | | 标准化载荷 | > 0.5(理想 > 0.7) | Hair et al. (2019) | | CR / AVE | CR > 0.7,AVE > 0.5 | Fornell & Larcker (1981) | | 区分效度 | √AVE > 相关系数;HTMT < 0.85(宽松 0.90) | Fornell & Larcker (1981);Henseler et al. (2015) | | 拟合 | χ²/df < 3;CFI、TLI ≥ 0.95(> 0.90 可接受);RMSEA ≤ 0.06(< 0.08 可接受);SRMR < 0.08 | Hu & Bentler (1999);Browne & Cudeck (1993) | | 中介 | bootstrap 95% CI 不含 0;类型按 Zhao et al. (2010) | Preacher & Hayes (2008) | | 测量不变性 | ΔCFI ≤ 0.01,ΔRMSEA ≤ 0.015 | Cheung & Rensvold (2002);Chen (2007) |
参考文档(按需读取)
- references/01-workflow-and-chapters.md:问卷论文第四章的标准结构,每一节报告什么
- references/02-spss-syntax.md:生成的 SPSS 语法逐段说明、怎么读输出、常见报错
- references/03-amos-guide.md:AMOS 画图、设置、读输出、自定义估计量、多群组、常见报错
- references/04-reliability-validity.md:α、CITC、CR、AVE、FL、HTMT 的公式,不达标怎么办
- references/05-fit-and-modification.md:拟合指标含义与模型修正规则
- references/06-mediation-moderation.md:SEM 中介、PROCESS 各模型、选哪个
- references/07-reporting.md:中文写作模板、表格规范、答辩常见问题
- references/08-lavaan-amos-parity.md:lavaan 与 AMOS / SPSS 的设置对应与数值差异
- references/citations.md:全部判断标准的文献出处
演示与自检
assets/demo/questionnaire_demo.sav:模拟的移动支付 App 问卷(430 份,5 个构念,18 题),生成模型已知,并且故意埋了反向题、低载荷题、误差相关、SPSS 缺失代码 9、缺答和规律作答。生成脚本是scripts/make_demo_data.R。assets/demo/example_output/:用 demo 设定卡正式运行的完整输出,可以先给用户看效果。Rscript tests/test_demo.R:端到端自检,共 48 项检查:α / KMO 与 psych 对照,PROCESS 与lm()对照,χ² 的 N−1 规则,能否恢复生成模型,生成的 SPSS / AMOS 文件结构是否正确,设定卡校验,以及 PROCESS 模型 1/6/14、观测调节变量、FIML 的第二套设定。
Related Skills
Agent-Reach
108.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Scrapling
93.4k🕷️ The BEST Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev
ruflo
80.7k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
algorithmic-art
177.9kCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
