A strong mentor does not make a project look advanced on the student's behalf. The mentor helps the student make better scientific decisions—then protects enough space for the student to perform, document, analyze, and defend the work independently.
A mentor should improve decisions, not replace them
Useful mentorship is most visible before and between experiments. A mentor can help a student narrow an unfocused interest, compare prior studies, identify missing controls, assess feasibility, and recognize when a method or claim exceeds the available evidence.
- Turn a broad interest into a testable question
- Teach the student how to compare sources and methods
- Identify safety, approval, and resource constraints early
- Use questions and review to strengthen the student's reasoning
The student must own the core research work
Society for Science states that the student researcher is responsible for all aspects of the project, including experimentation, data collection, engineering, and analysis. The student should also be able to explain why each major decision was made and write the research plan, abstract, poster, and conclusions in their own words.
- Maintain the research log and original files
- Perform or directly participate in the project procedures
- Analyze the data and acknowledge unfavorable results
- Explain the method, limitations, and next steps without a script
Safety supervision is necessary—and different from authorship
An Adult Sponsor or Qualified Scientist may be required to review risk, provide training, supervise procedures, and complete approvals. That oversight does not transfer ownership of the project. The research plan should clearly distinguish the student's contribution from work performed by a mentor or a larger laboratory group.
- Complete required review before experimentation begins
- Document training and professional guidance
- Disclose mentor and institutional contributions
- Keep the student's own portion clearly identifiable
Judges look for understanding and independence
ISEF judging criteria emphasize the interview, including the student's understanding of the science, interpretation, limitations, and degree of independence. A polished result cannot compensate for a student who cannot explain the evidence chain or identify what was done by someone else.
- Can the student explain every figure and axis?
- Can the student identify error and alternative explanations?
- Can the student describe what changed after a failed pilot?
- Can the student separate their work from a mentor's contribution?
Ask these questions before choosing a mentorship
Families should evaluate the working process, not promises of a particular outcome. A credible mentorship model should make student responsibilities visible and explain how the mentor handles access, feedback, safety, documentation, and project scope.
- How often will the student receive direct scientific feedback?
- What records and raw data will the student retain?
- How are mentor contributions disclosed?
- What happens when the first design does not work?