construct-validity-assessment

$npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/construct-validity-assessment

Assesses if a benchmark accurately measures its claimed capability

  • Evaluates alignment between benchmark tasks and intended capability
  • Uses psychometric validity frameworks adapted for AI evaluation
  • Analyzes task examples for content, convergent, and discriminant validity
  • Produces a structured validity verdict with evidence for each dimension

SKILL.md

.github/skills/construct-validity-assessmentView on GitHub ↗
---
name: construct-validity-assessment
description: Evaluate whether benchmark measures its claimed capability
execution: subagent
prompt: ./prompt.md
input: benchmark_name, claimed_capability, task_examples
used-by: benchmark-archaeology
---

# Construct Validity Assessment SOP

Evaluate whether a benchmark actually measures the capability it claims to measure, using psychometric validity frameworks adapted for AI evaluation.

## Input

- **benchmark_name**: Name of the benchmark
- **claimed_capability**: What the benchmark authors claim it measures
- **task_examples**: Representative examples from the benchmark

## Procedure

1. Define the construct (claimed capability) precisely
2. Analyze task requirements — what skills are actually needed to solve examples?
3. Assess content validity — do items representatively sample the construct?
4. Check convergent validity — correlation with other measures of same construct
5. Check discriminant validity — independence from unrelated constructs
6. Identify construct-irrelevant variance (confounds)

## Output

Validity verdict with evidence for each validity dimension.

More from yogsoth-ai/de-anthropocentric-research-engine

SkillDescription
abductive-hypothesis-generationStrategy: 面对异常的最佳解释推理
ablation-brainstormRemove components one by one, observe system changes to reveal hidden dependencies and generate ideas from structural gaps.
ablation-component-mappingMap system architecture to ablatable units for ablation studies
ablation-designDesign ablation studies to isolate component contributions in ML systems
ablation-executionRemove components one by one from a system, record the response/impact of each removal.
abp-vulnerability-classificationClassify assumptions on 2 axes — load-bearing (how much conclusion depends on it) × vulnerable (how likely to be false). Focuses attention on High-Load × High-Vulnerable quadrant.
abstraction-extractionExtract abstract principles from concrete domain cases. Strips domain-specific details to reveal transferable mechanisms.
abstraction-ladderPerform bisociation at multiple abstraction levels
abstraction-ladderingMove between concrete and abstract framings — 3 levels up (Why?) and 3 levels down (How?) to find the most productive research level.
abstraction-to-designAbstract biological principle to design principle. Bridge from biology to engineering.