Hypothesis to Experiment Design

Formulates a testable hypothesis, designs a rigorous experiment, then creates an analysis plan before any data is collected.

Category: research Difficulty: beginner

Platforms: chatgpt claude perplexity

Tags: hypothesis testing experiment design scientific method analysis plan research

Prompt Template

You are a research scientist. Help me formulate a testable hypothesis:

**What I want to investigate:** {{investigation}}
**Context/background:** {{context}}
**What I think might be true:** {{intuition}}
**Domain:** {{domain: science/business/product/social/health}}

Develop:
1. OBSERVATION: State the observation or pattern that sparked the question
2. RESEARCH QUESTION: Formulate as a clear, answerable question
3. HYPOTHESIS VERSIONS:
   - Null hypothesis (H0): The default — no effect/relationship
   - Alternative hypothesis (H1): What you predict, stated precisely
   - Directional version: Specifying the direction of effect
4. VARIABLE DEFINITION:
   - Independent variable(s): What you manipulate or compare
   - Dependent variable(s): What you measure
   - Control variables: What you hold constant
   - Confounding variables: What could invalidate your results
5. FALSIFIABILITY CHECK: What specific result would disprove your hypothesis? (If nothing can disprove it, it is not scientific)

End with "EXPERIMENT BRIEF" — a one-paragraph summary of what needs to be tested and the minimum evidence needed to accept or reject H1.

Tips

Step 1 hypothesis formulation determines everything that follows — invest time here
Pre-registering your analysis plan (step 3) before collecting data prevents p-hacking
The power analysis from step 2 tells you if your experiment is worth running
Pilot test with 5-10 participants before full deployment
This chain works for academic research, product experiments, and business decisions