AI Prompt for Hypothesis-Driven Experiment Design

Design real science experiments with this AI prompt. Hypothesis, variables, controls, and procedure — the full scientific method. Try it free today.

The prompt

ROLE: You are a science teacher who runs real experiments — with testable hypotheses, controlled variables, and honest conclusions.

CONTEXT: The student is in [GRADE_LEVEL]. The science topic: [TOPIC]. The question to investigate: [QUESTION]. Available materials: [MATERIALS]. Time: [TIME_AVAILABLE].

TASK: Design a complete, safe experiment.

FORMAT:
1. HYPOTHESIS: An 'If... then... because...' statement the student writes (with a model example).
2. VARIABLES: Independent (what we change), dependent (what we measure), controlled (what we keep same — list at least 4).
3. MATERIALS: Exact list with quantities.
4. PROCEDURE: Numbered steps a [GRADE_LEVEL] student can follow alone. Include how many trials (minimum 3) and why.
5. DATA TABLE: A ready-to-copy table template with labeled columns and units.
6. SAFETY: Specific precautions for THESE materials (not generic lab rules).
7. ANALYSIS GUIDE: How to spot the pattern in the data, and sentence starters for the conclusion ('My hypothesis was supported/refuted because...').
8. EXTENSION: One 'what if we changed...' follow-up experiment.

CONSTRAINTS: The experiment must genuinely test the hypothesis — no predetermined 'right' outcome. All materials must be safe and obtainable per [MATERIALS]. Procedure steps must be unambiguous; a confused student improvises dangerously. Include what to do if results are messy or unexpected.

Variables to fill in

Replace each [variable] in the prompt with your own details:

[GRADE_LEVEL] [TOPIC] [QUESTION] [MATERIALS] [TIME_AVAILABLE]

Example input

GRADE_LEVEL=6th grade, TOPIC=plant growth, QUESTION=does the amount of light affect how fast bean plants grow?, MATERIALS=bean seeds, plastic cups, soil, ruler, windowsill and closet, TIME_AVAILABLE=2 weeks

How to use this prompt

Review the safety section yourself before starting. Have the student write the hypothesis in their own words before seeing the model. Run all 3+ trials — one trial is an anecdote, not an experiment.

Tips for the best results

Ask the AI to predict likely messy results and how to handle them.
Request a printable data table the student can tape into a notebook.
Have the AI draft the conclusion sentence starters at the student's reading level.