What Makes a Project ISEF-Worthy: The Criteria Judges Actually Use

Every spring, thousands of students win their regional fair with a project that was never going to survive at ISEF. The project was good. It was also the wrong kind of good. ISEF judges score something specific, and most regional projects were built to impress a different audience.

This guide explains exactly what ISEF judges score, with the weights, then turns it into a test you can run on your idea before you spend six months on it.

The five ISEF judging criteria and their weights

The Society for Science publishes the rubric. Judges score every project in five categories. The weights below are for science projects; engineering projects swap the second and third categories for design and construction criteria with the same weights.

Two things in that table surprise most students.

First, the interview is the single biggest category. Thirty percent of the score is how you talk about the project, not the project itself. A student who built something brilliant and cannot explain why a control was chosen loses to a student with a smaller project and a complete command of it.

Second, "creativity" is not about the topic. It is about the approach. A creative project on a familiar question outscores a conventional project on an exotic one.

What ISEF-worthy actually means

A project is ISEF-worthy when it can score well in all five categories at once. Here is what each one looks like at the finalist level, compared with the regional-winner level.

Research question: a gap, not a demonstration

Regional-winner level: "Does the type of fertilizer affect plant growth?" The answer is known. The project demonstrates it.

ISEF level: "Does a biochar amendment reduce nitrate leaching from sandy soils at application rates farmers can actually afford?" The question sits inside a real problem, names a specific gap, and has a measurable answer that someone in agriculture would want.

The test: can you name the person or group who would change what they do based on your result? If you cannot, the question is a demonstration.

Methodology: a design that could be published

Regional level: three samples per group, no replication, a single trial.

ISEF level: sample sizes chosen for statistical power, replicated trials, controls for every variable you can name, and a method that a grad student would recognize from the literature. Judges will ask why you chose each parameter. "Because that is what the kit said" ends the conversation.

The test: write the methods section before you start. If you cannot justify every number in it, the design is not ready.

Execution: enough data to support a claim

Regional level: a bar chart of averages.

ISEF level: error bars, a stated statistical test, a p-value or confidence interval, and an honest treatment of outliers. If the project is computational, that means validation on held-out data and a comparison to a baseline. Finalists routinely have more data than they show, and they know which data they left out and why.

The test: can you state your main result as a sentence with a number and an uncertainty in it? "Treatment reduced leaching by 38 percent, with a 95 percent confidence interval of 29 to 47 percent, across six replicated trials."

Creativity: an approach the judge has not seen

Regional level: a standard experiment with a different variable.

ISEF level: a method borrowed from another field, a new way to measure something, an instrument the student built, or a dataset nobody had combined before. One of our Fellows reached the ISEF finals with a project that applied a signal-processing technique from audio engineering to heart-rate data. The question was not new. The approach was.

The test: in one sentence, what did you do that someone with the same question and the same budget would not have thought of?

Presentation: the interview decides it

Regional level: a memorized speech and a tidy poster.

ISEF level: a student who can be interrupted at any point, answer a question about a limitation without flinching, explain what they would do with another six months, and discuss how the work relates to two or three published papers they have actually read. Judges at ISEF are practicing scientists and engineers. They are not grading enthusiasm. They are checking whether the student owns the work.

The test: have a scientist you do not know interview you for fifteen minutes. If you cannot find one, that is exactly the problem a research mentor solves.

The ISEF-worthy checklist

Run your idea through all ten before you build it. Nine out of ten is a project that can compete. Fewer than seven is a regional project.

1. The question has a specific, named gap behind it.
2. Someone outside your school would care about the answer.
3. You can list at least three published papers on the exact question.
4. The method appears in at least one of those papers.
5. Sample size and replication are decided in advance, with a reason.
6. The analysis includes a statistical test, not just a chart.
7. There is a baseline or control the result is compared against.
8. At least one element of the approach is original to you.
9. You can state the main result with a number and an uncertainty.
10. You can explain the biggest limitation and what you would do next.

The categories where ISEF-worthy projects are most common

Some ISEF categories are more open to high school students than others, because the methods are accessible without a university lab:

• Computational biology, bioinformatics and biomedical data science. Public datasets, open-source tools, real questions.
• Environmental engineering and earth science. Field data you can collect yourself, with real local stakeholders.
• Robotics and intelligent machines. A built system is inherently original.
• Systems software and machine learning. Baselines and benchmarks make execution easy to prove.
• Behavioral and social sciences. Survey and experimental designs are feasible, and judges reward rigorous statistics.

Cellular and molecular biology projects can absolutely be ISEF-worthy, but they usually require lab access that most students only get through a mentor or a university program. If that is your field, line up the lab before you commit to the question. Biology Research for High School Students covers the lab and computational routes in detail.

How ISEF-worthy projects actually get built

Finalists almost never work alone. In our experience the pattern looks like this:

1. A mentor with a PhD helps the student find the gap in the literature. This is the step students cannot do alone, because knowing what is already known takes years.
2. The method is lifted from a published paper and adapted, which satisfies the methodology criterion automatically.
3. The student collects more data than seems necessary.
4. The write-up is done as a paper first, then condensed into a poster. Students who write the paper first interview far better, because they have already defended every choice in writing.
5. Mock interviews with scientists, repeatedly.

This is the sequence the YRI Fellowship is built around: a PhD mentor, a publishable paper, then the competition. Fellows have qualified for ISEF and JSHS and placed at the national level, and the paper they wrote along the way is what they put on their college applications.

Frequently Asked Questions

What are the ISEF judging criteria? Research question (15 percent), design and methodology (15 percent), execution including data collection and analysis (20 percent), creativity (20 percent), and presentation including the poster and interview (30 percent). Engineering projects use design and construction criteria in place of the second and third categories with the same weights.

What makes a project ISEF-worthy? A project is ISEF-worthy when it addresses a specific gap that matters to someone outside the student's school, uses a method that could appear in a published paper, has enough replicated data to support a statistical claim, includes an original element in its approach, and can be defended in a fifteen-minute interview with a practicing scientist.

How is the ISEF interview scored? The interview and poster together are 30 percent of the total, the largest single category. Judges assess whether the student understands the work deeply, can discuss its limitations, knows the related literature, and did the work themselves.

Can a simple project win at ISEF? Yes, if it is rigorous and original in approach. Judges reward a focused question with strong methodology and honest analysis over a sprawling project with weak data. Simple is fine. Shallow is not.

Do you need a mentor to qualify for ISEF? There is no rule requiring one, but the overwhelming majority of finalists have an adult scientist involved, because finding a real gap in the literature and designing a publishable method are hard to do without one. How to Find a Research Mentor covers the options, free and paid.

How do I qualify for ISEF? Through an affiliated regional or state fair, which sends its top projects to ISEF. Our ISEF qualification guide explains the fair system and deadlines.