domains / research / scientific-claim-appraisal / scientific-critical-thinking

scientific-critical-thinking

Skill card

combined

The K-Dense-AI copy is the most rigorous and current procedure, featuring a specific 6-step 'Appraisal Workflow' (fixing the question, extracting evidence, assessing design, choosing frameworks, synthesizing, and writing a traceable critique) and strict constraints on visual aids. The other three copies (Vibe-Skills, medical-research-skills, and Auto-Empirical-Research-Skills) are largely identical to each other, containing a more verbose 'Core Capabilities' section with detailed checklists for validity and bias, but they lack the structured workflow of the K-Dense version and instead mandate the default generation of schematics. Because the K-Dense version provides a superior procedural framework while the others provide more granular evaluation checklists, and they disagree on the handling of visual aids, they should be combined.

Repositories

Other copies

Combined skill

---
name: scientific-critical-thinking
description: Evaluates scientific claims and evidence quality. Applies to experimental design validity, biases and confounders, statistical interpretation, and evidence grading frameworks (GRADE, Cochrane Risk of Bias).
allowed-tools: Read Write Edit Bash
license: MIT
---

# Scientific Critical Thinking

## Overview

Critical thinking is a systematic process for evaluating scientific rigor. Assess methodology, experimental design, statistical validity, biases, confounding, and evidence quality using GRADE and Cochrane ROB frameworks.

## Appraisal Workflow

1. **Fix the question and unit.** Record population, intervention/exposure, comparator, outcome, time point, effect measure, and target setting. Distinguish descriptive, predictive, and causal claims.
2. **Extract evidence before judging.** Locate the numerical result, uncertainty, denominators, protocol/registration, and analysis plan. Keep missing reporting separate from evidence that a procedure was not performed.
3. **Assess design and analysis.** Check selection, confounding, measurement, attrition, multiplicity, and model assumptions.
4. **Choose the right appraisal framework.** Record the exact tool version and the result being assessed; do not turn checklist counts into a quality score.
5. **Synthesize with scope intact.** Examine independent replication, overlapping samples, and applicability. Apply GRADE per outcome/comparison with explicit domain reasons.
6. **Write a traceable critique.** For each concern give the source location, observation, consequence, uncertainty, and a feasible remedy.

## Core Capabilities

### 1. Methodology Critique
- **Study Design:** Is the design appropriate for the research question? Can it support the causal claims made?
- **Validity Analysis:**
  - **Internal:** Check randomization, confounding control, and selection bias.
  - **External:** Evaluate sample representativeness and ecological validity.
  - **Construct:** Review measurement validation and operational definitions.
  - **Statistical:** Verify power, assumption compliance, and test appropriateness.
- **Control and Blinding:** Assess randomization implementation, blinding feasibility, and control condition appropriateness.

### 2. Bias Detection
- **Cognitive Biases:** Check for confirmation bias, HARKing, and cherry-picking. Verify preregistration.
- **Selection Biases:** Examine sampling, volunteer, attrition, and survivorship biases.
- **Measurement Biases:** Evaluate observer, recall, and social desirability biases.

### 3. Statistical Analysis Evaluation
- Assess power, multiplicity, p-value misuse, and effect sizes.

### 4. Evidence Quality Assessment
- Evaluate study hierarchy, replication, and strength of inference.

## Visual Aids (Optional)

Only add figures when the **user explicitly requests** a diagram (e.g., a GRADE flowchart or bias decision tree).

**How to create figures:**
- **Preferred:** Use the **scientific-schematics** skill for AI-generated diagrams.
- **Alternative:** Build figures in your usual tools (draw.io, matplotlib, etc.).

**Disclosure:** AI schematic generation sends your prompt to a third-party API. Do not include unpublished sensitive details unless appropriate.

## Conflicts
- **K-Dense-AI/scientific-agent-skills vs. foryourhealth111-pixel/Vibe-Skills:** K-Dense-AI requires explicit user request for diagrams, while Vibe-Skills recommends generating them by default for new documents.