How to Find a Science Fair Project Idea

How to Find a Science Fair Project Idea

The short answer: start with something the student genuinely notices, turn it into a question that changes one measurable factor, and reject any idea that cannot be tested safely, repeatedly and mostly by the student within the available time.

“Climate change,” “cancer,” “AI” and “renewable energy” are important topics. They are not yet project ideas. A science fair project needs a focused question, method, data and conclusion.

The goal is not to find an idea nobody in human history has considered. The goal is to find a question the student can investigate with enough independence and rigour to learn something real.

---

The project-idea funnel

Move through five levels:

1. Interest: What does the student care about or notice?
2. Observation: What pattern, problem or difference have they seen?
3. Variable: What could be changed or compared?
4. Measurement: What data can be collected reliably?
5. Feasibility: Can it be done safely, ethically, repeatedly and on time?

Example:

Plants → houseplants dry at different speeds → compare container material → measure daily mass loss → use identical soil volumes and conditions over several weeks.

That is far stronger than “Which pot is best?” because the variable and measurement are visible.

Topic versus research question

The better questions specify an input, an output and a controlled context.

Start with everyday friction

Good ideas often begin with a small inconvenience:

• a room that overheats;
• food that loses texture;
• a bicycle light that drains quickly;
• noise in a study area;
• packaging that uses too much material;
• a school process that wastes time;
• a garden area where one plant thrives and another does not;
• a sensor that gives inconsistent readings.

Ask: What exactly is happening, and what evidence would help explain or improve it?

Use the seven-question test

Before committing, score the idea against these questions.

1. Is it genuinely interesting to the student?

Research contains repetition and setbacks. Borrowed enthusiasm fades quickly.

2. Is the question focused?

One project should not try to solve a global problem. Narrow the material, setting, age group, variable or mechanism.

3. Is the result measurable?

Useful measurements include time, mass, temperature, voltage, distance, force, pH using safe materials, count, error rate, energy, image features or a clearly defined score.

4. Can it be repeated?

One trial may be an anecdote. Plan enough repetitions to estimate variation.

5. Is it feasible?

Check equipment, cost, time, space, season, data access and adult supervision.

6. Is it safe and ethical?

Rules must be checked before experimentation, especially for people, vertebrate animals, microorganisms, tissues, hazardous chemicals or devices.

7. Does the student own the thinking?

Mentors can teach a method and challenge assumptions. The student should still understand and carry out their own defined contribution.

Science, engineering or computing project?

Science investigation

Asks how or why something behaves. It tests a research question or hypothesis through controlled measurement.

Engineering project

Defines a need, sets criteria, builds a solution, tests it and iterates. The question is often: How well does this design meet the requirements?

Computing/data project

Builds or evaluates an algorithm, model or digital tool. It still needs a defined test set, baseline, metric and limitations.

All three can be strong. Do not force an invention into a fake hypothesis or call an attractive demonstration an experiment.

Ten idea sources that produce better questions

1. A repeated observation at home or school.
2. A local environmental issue with accessible data.
3. A hobby such as sport, music, photography or gardening.
4. A device that could be measured or improved.
5. A public dataset with an unanswered comparison.
6. A surprising result in a classroom practical.
7. A limitation mentioned in a scientific paper.
8. A previous project that suggests a genuinely new variable.
9. A community need that can become an engineering criterion.
10. A failed prototype whose failure can be investigated.

Safe example directions by field

Environmental science

• compare microclimates across measured locations;
• analyse public air-quality or weather data;
• test safe model filtration materials;
• quantify litter patterns with a consistent survey method;
• investigate insulation using model containers.

Physics and engineering

• optimise a model structure's strength-to-mass ratio;
• compare damping methods for vibration;
• test light angle, surface texture or thermal insulation;
• calibrate a low-cost sensor against a reference;
• model drag or rolling resistance safely.

Biology and plant science

• study germination or plant growth using safe species;
• compare environmental conditions with controlled variables;
• use image analysis to measure leaf area;
• analyse an open biological dataset;
• study non-hazardous ecological observations without disturbing wildlife.

Computing and data science

• compare algorithms on speed, memory or accuracy;
• test accessibility improvements with approved methods;
• model a local system using public data;
• evaluate robustness under controlled input changes;
• build a sensor-and-dashboard project with defined performance criteria.

Behavioural and social science

Human-participant projects can require prior approval, consent and careful privacy design. A safer starting point may be an anonymised public dataset. Do not begin surveys, interviews or experiments with people until the competition's rules have been reviewed.

Ideas to avoid or redesign

“Which brand is best?”

This often becomes consumer testing without a scientific mechanism. Improve it by defining one material property and explaining why it matters.

A demonstration with a known outcome

A volcano model shows a reaction but does not answer a new research question. Add a measurable variable only if the resulting method remains safe and meaningful.

Culturing unknown microorganisms

Growing bacteria from phones, hands or public surfaces can create unknown hazards. Many fairs restrict this work and require appropriate facilities and approval. Choose a safer question.

Medical diagnosis or treatment claims

A school project should not present an unvalidated model or product as medical advice. Use carefully defined research language, ethical data and appropriate supervision.

Projects involving people without approval

Surveys, usability tests, fitness measurements and psychological tasks may count as human-participant research. Approval may be required before collecting any data.

A mentor's laboratory project

Access to advanced equipment does not replace student ownership. Judges will ask what the student designed, understood and carried out.

Rules come before data

For an ISEF-affiliated route, official rules require a research plan before experimentation and prior approval for many projects involving human participants, vertebrate animals, potentially hazardous biological agents or hazardous activities.

The official guidance also says that demonstrations, literature reviews and explanation models are not appropriate as ISEF research projects. A project needs original student research or engineering work.

This does not mean every local fair uses identical paperwork. It means students must read the rules of their intended competition before beginning.

Match the idea to the competition

Stripe Young Scientist & Technology Exhibition

Strong for sustained research or technology projects in Ireland with school and teacher involvement. Read the parent-friendly Stripe YSTE guide.

SciFest Ireland

Provides regional and school pathways for science, technology, engineering and maths projects. It can be an excellent first formal fair.

The Big Bang Competition

Suitable for UK science and engineering work where the student can explain the problem, method, evidence, creativity and impact.

Regeneron ISEF

An elite international final reached through affiliated fairs, with detailed rules and approvals. Read how UK and Ireland students reach ISEF before planning around this route.

Breakthrough Junior Challenge

Not a science fair: it is a science-communication video challenge. It may fit a student with a strong explanation rather than original experimental data.

Our comparison of science fairs and Olympiads can help choose the right format.

The one-page idea brief

Before asking a teacher or mentor, write:

1. Observation/problem: what prompted the idea?
2. Question or engineering goal: one sentence.
3. Why it matters: a specific reason, not “this will change the world.”
4. Independent variable/design change: what will vary?
5. Dependent measure: what data will be recorded?
6. Controls/baseline: what makes comparison fair?
7. Method: five to eight steps.
8. Repetitions/sample: how much data is realistic?
9. Safety and ethics: what approvals may be needed?
10. Resources and timeline: equipment, cost and milestones.

If the brief cannot fit on one page, the project may still be too broad.

Pilot before committing

A small pilot can reveal:

• whether the measurement changes enough to detect;
• whether the equipment is precise enough;
• how long each trial takes;
• whether uncontrolled factors dominate;
• whether the question needs narrowing.

Do not collect full data until required approvals are complete. A pilot is not a loophole around ethics or safety rules.

What judges usually value

High-level judging criteria commonly reward:

• a clear and testable question;
• sound design and defined controls;
• systematic data collection;
• appropriate analysis;
• creativity and potential impact;
• student independence;
• understanding of limitations;
• clear presentation and interview responses.

An expensive project with weak ownership can be less convincing than a simple question investigated exceptionally well.

A realistic first-month plan

Week 1

Collect ten observations or problems. Do not choose yet.

Week 2

Turn the best three into testable questions. Check rules and obvious safety issues.

Week 3

Research existing knowledge, define variables and draft the one-page brief.

Week 4

Review with a teacher, narrow scope, obtain required approvals and plan a pilot.

Only then should substantial experimentation or building begin.

Key takeaways

• A broad topic is not a project idea. A project needs a focused question or engineering goal.
• Begin with a real observation, then define a variable and measurement.
• The strongest idea is feasible, repeatable, safe, ethical and mostly owned by the student.
• Simple equipment does not mean a simple project. Rigorous design and analysis matter more than laboratory prestige.
• Read competition rules before collecting data. Some projects require prior approval.
• Avoid demonstrations, brand comparisons without mechanism and unsafe microorganism work.
• Match the project format to the competition. Research fairs, engineering competitions and science-communication challenges reward different outputs.
• Pilot the method and reduce the scope early.
• Judges value understanding, independence and limitations, not only a dramatic result.
• A good project leaves the student with a better question as well as an answer.