---

name: hypothesis-gen description: "Structured hypothesis generation workflow. Use when: user needs to formulate testable scientific hypotheses from observations, gaps, or preliminary data. NOT for: testing hypotheses or running experiments." metadata: { "openclaw": { "emoji": "🧪" } }

Hypothesis Generation Skill

Structured workflow for generating testable scientific hypotheses.

When to Use

• "Generate hypotheses for this research question"
• "What could explain this observation?"
• "Propose testable ideas based on this data"
• "Help me formulate H0 and H1"
• "What hypotheses does this gap suggest?"

When NOT to Use

• Testing/verifying hypotheses (use scienceclaw-verification)
• Designing experiments (use experimental-design)
• Literature searching (use literature-search)
• Writing full papers (use paper-writing)

Generation Workflow

Step 1: Observe

Identify the phenomenon, anomaly, or gap:

• What was observed?
• What is unexpected or unexplained?
• What contradicts existing theory?
• What data pattern needs explanation?

Step 2: Contextualize

Ground in existing literature:

• What do current theories predict?
• What related findings exist?
• Where are the knowledge gaps?
• What alternative explanations exist?

Step 3: Formulate

State the hypothesis formally:

Template: "If [independent variable/condition], then [predicted effect on dependent variable], because [proposed mechanism]."

Null Hypothesis (H0): No effect / no difference / no relationship Alternative Hypothesis (H1): The predicted effect exists Directional: Specify direction (increase/decrease) when justified

Step 4: Evaluate

Score each hypothesis on:

Criterion	Score (1-5)	Description
Testability	_	Can be experimentally tested?
Falsifiability	_	Can be proven wrong?
Novelty	_	How new is this idea?
Mechanism	_	Is the proposed mechanism plausible?
Feasibility	_	Can current methods test it?
Impact	_	How significant if confirmed?

Step 5: Prioritize

Rank hypotheses by:

• Total evaluation score
• Risk-reward ratio (impact / feasibility)
• Alignment with available resources
• Potential for publication

Output Format

## Hypothesis [N]: [Short title]

**Statement**: If [condition], then [prediction], because [mechanism].
**H0**: [Null hypothesis]
**H1**: [Alternative hypothesis]

**Variables**:
- Independent: [variable]
- Dependent: [variable]
- Controls: [variables to hold constant]

**Evaluation**: Testability=[X] Falsifiability=[X] Novelty=[X] Mechanism=[X] Feasibility=[X] Impact=[X]
**Priority Score**: [Total/30]

**Key References**: [relevant citations]
**Suggested Test**: [brief experimental approach]

Quality Checklist

• Hypothesis is specific and testable
• Variables are clearly identified
• Mechanism is plausible given known science
• Null hypothesis is properly stated
• Predictions are measurable
• At least one path to falsification exists
• Novel relative to existing literature
• Ethical considerations addressed