operationalization
$
npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/operationalization将假设中的抽象概念转化为具体可测量的指标,并论证测量有效性。
SKILL.md
.github/skills/operationalizationView on GitHub ↗
---
name: operationalization
description: "SOP: 将抽象概念操作化为可测量的指标和方法"
version: 1.0.0
category: hypothesis-formation
type: sop
campaign: hypothesis-formulation
input: "抽象变量描述(来自 variable-identification 输出)"
output: "操作定义 + 测量方法 + 效度论证(内容/构念/标准)"
dependencies:
skills:
- subagent-spawning
---
# Operationalization
将假设中的抽象概念转化为具体可测量的指标,并论证测量有效性。
## HARD-GATE
<HARD-GATE>
前置条件(全部满足才能开始):
1. 已有至少 1 个需要操作化的变量(operationalizable ≠ "high" 时尤其需要)
2. 变量的理论定义已提供(来自 variable-identification 的 description 字段)
不满足 → 停止,返回错误:需要先完成 variable-identification。
</HARD-GATE>
## Pipeline
1. 前置检查:验证变量描述完整性
2. 概念分析:分解变量的核心属性(概念维度)
3. 指标选择:为每个维度选择 1-2 个可测量指标
4. 测量方法确定:指定数据收集方式(survey/experiment/observation/archival/computational)
5. 效度论证:
- 内容效度:指标是否覆盖概念的全部关键维度?
- 构念效度:指标是否与相关构念收敛、与不相关构念区分?
- 标准效度:指标是否与已验证的标准测量相关?
6. 输出操作定义
## Output Format
```json
[
{
"variable": "Variable name",
"theoretical_definition": "Abstract definition",
"dimensions": ["Dimension 1", "Dimension 2"],
"indicators": [
{
"indicator": "Indicator name",
"measurement_method": "How to collect/measure",
"scale": "nominal | ordinal | interval | ratio",
"validity": {
"content": "Justification",
"construct": "Justification",
"criterion": "Justification or null if not applicable"
}
}
],
"operationalization_notes": "Any remaining challenges or alternatives"
}
]
```