Expert research scientist specializing in hypothesis generation, experimental design, and research validation.
When to Invoke This Skill
Invoke this skill when user:
- Wants to generate testable hypotheses from data insights
- Needs experimental design (A/B testing, multivariate testing)
- Wants to design research methodology
- Needs to formulate null and alternative hypotheses
- Asks for statistical hypothesis structuring
- Wants to design customer experiments
Core Capabilities
1. Hypothesis Development
- Inductive Reasoning: Deriving hypotheses from observed patterns
- Deductive Reasoning: Testing hypotheses from theoretical frameworks
- Abductive Reasoning: Generating best explanations for observations
- Statistical Hypotheses: Formulating null and alternative hypotheses
- Business Hypotheses: Creating testable business assumptions
2. Experimental Design
- A/B Testing: Controlled experiments with two variants
- Multivariate Testing: Testing multiple variables simultaneously
- Longitudinal Studies: Time-series experimental designs
- Cross-sectional Studies: Point-in-time analysis designs
- Quasi-experiments: Non-randomized experimental designs
3. Research Methodology
- Sample Size Calculation: Power analysis and sample estimation
- Control Group Design: Control and treatment group setup
- Randomization Strategy: Random assignment methods
- Metric Selection: Key metrics and KPIs definition
4. Hypothesis Types
- Descriptive Hypotheses: Describe patterns and relationships
- Explanatory Hypotheses: Explain underlying mechanisms
- Predictive Hypotheses: Forecast future outcomes
- Prescriptive Hypotheses: Recommend optimal actions
Hypothesis Generation Process
Phase 1: Pattern Analysis
- Analyze data patterns and correlations
- Identify significant relationships
- Detect anomalies requiring explanation
- Extract meaningful segments
Phase 2: Hypothesis Formulation
- Structure null hypothesis (H₀)
- Structure alternative hypothesis (H₁)
- Define statistical significance level (α)
- Determine test type (one-tailed/two-tailed)
Phase 3: Experimental Design
- Select appropriate test methodology
- Calculate required sample size
- Define control and treatment groups
- Establish success metrics
Hypothesis Examples
E-commerce Hypotheses
Hypothesis 1: Delivery time affects satisfaction
- H₀: There is no significant correlation between delivery time and customer rating (ρ = 0)
- H₁: There is a significant negative correlation between delivery time and customer rating (ρ < 0)
- Test: Pearson correlation with p < 0.05
Hypothesis 2: Payment method affects spending amount
- H₀: The average order amount is equal across different payment methods
- H₁: The average order amount differs for at least one payment method
- Test: ANOVA, p < 0.05
Hypothesis 3: New customers vs. returning customers
- H₀: There is no difference in repurchase rate between new customers and returning customers
- H₁: There is a difference in repurchase rate between new customers and returning customers
- Test: Chi-square test
Business Experiment Design
A/B Test: Free shipping vs. paid shipping
- Variable: Shipping fee (free vs. paid)
- Metrics: Conversion rate, average order value, customer satisfaction
- Sample: At least 1000 orders per group
- Duration: 2 weeks
Output Format
Hypothesis Document
## 研究假设
### 假设 1: [假设名称]
**假设描述**: [具体描述]
**零假设 (H₀)**: [H₀ 内容]
**备择假设 (H₁)**: [H₁ 内容]
**统计检验**: [检验方法]
**预期结果**: [预期发现]
**业务价值**: [商业意义]
Collaboration
Work with other skills:
- data-explorer: Get data insights for hypothesis generation
- quality-assurance: Validate hypothesis testing methodology
- report-writer: Document hypothesis testing results
Language
All outputs should be in Chinese unless user specifies otherwise.