How to Choose a Research Topic in High School (4-Week Method)

Most high school students who want to do research get stuck before they start, because they think they need a brilliant, original topic on day one. You don't. In our program, students arrive with anything from a precise question to "I like biology," and they lock in their research question by the end of week four. Here is the method that gets them there.

Want a PhD mentor to walk you through this? Apply to the STEM Research Program →

Why most students pick the wrong topic

There are two common mistakes, and they're opposites:

• Too broad. "Climate change" or "artificial intelligence" are fields, not topics. You can't finish them, so the paper ends up as a summary of Wikipedia.
• Too narrow, too soon. Picking a hyper-specific question before you've read anything usually means you pick one that's already been answered, or one that needs a lab you don't have.

The fix is to deliberately go broad first, then narrow on purpose. One student described it well: "If you wanted to go broad and then slowly focus in onto a specific thing you want to research, my mentor was completely fine with it."

Week 1: List what you're curious about, not what sounds impressive

Write down 5–10 things you've wondered about in the last year: a documentary that stuck with you, a class topic you wanted more of, a problem in your family or community, a technology you use every day and don't understand.

Then sort them into fields: biology, neuroscience, computer science, physics, chemistry, math, economics. You're looking for the field you return to most. That's your starting point.

Tip: don't filter for "what colleges want to see." Admissions readers can tell the difference between real curiosity and a topic picked for the résumé, and you'll spend weeks reading about this. Pick what you'd read about anyway.

If you need inspiration, browse our list of 45 research topics for high school students.

Weeks 2–3: Read your way into the field

This is the step most students skip, and it's the most important one.

1. Find two or three recent review articles in your field. Search Google Scholar for "[your topic] review" and filter to the last five years. Reviews summarize what's known and, crucially, what isn't.
2. Read the introduction and conclusion first. You don't need to understand every equation or method on the first pass.
3. Keep a "questions" list. Every time you think "wait, why?" or see "it remains unclear whether…", write it down.
4. Follow the citations on the two or three points that interest you most.
5. Use a reference manager like Zotero from day one. You'll thank yourself when you write the paper.

By the end of week three you should have a list of 10–20 open questions. Several will be too big, and a few will be interesting and finishable.

Our mentors are PhDs from MIT, Stanford, Columbia, UC Berkeley and other top-20 universities. See how the program works →

Week 4: Narrow to one research question

Take your best three or four questions and run each through these tests:

• Scope: can I answer this in about 8 weeks of reading and writing? If not, narrow it by one population, method, or time period.
• Resources: can I do it with published papers, public data, simulation, or math? If not, switch to a review or computational angle.
• Openness: is there real disagreement or an unanswered part? If not, look further down your questions list.
• Interest: will I still care about this in two months? If not, drop it.

Then rewrite the survivor as a single sentence ending in a question mark. For example:

• Broad: "Neuroscience and memory."
• Narrower: "How memories are formed."
• Research question: "What does current evidence show about whether a false memory can be created in the hippocampus, and how?"

That last version tells you what to read, what to argue, and when you're done.

What to do if you're still stuck

• Choose a method you want to learn, then find a question it can answer. Wanting to learn machine learning, Python, or a statistics package like R is a perfectly good starting point.
• Pick a comparison. "A vs. B" questions (two treatments, two algorithms, two energy policies) are naturally scoped.
• Ask someone who has read the literature. This is the single biggest accelerator. A mentor who knows the field can tell you in minutes whether a question is already answered, too big, or genuinely open.

Get matched with a PhD mentor in your field. Apply now →

What happens after you have a question

Once the question is set, the work shifts to building the paper: a research proposal, a first draft, and then revisions. We cover that in how to write a research paper in high school.

Students consistently tell us the topic stage was where a mentor made the biggest difference:

"My mentor took the time to assess my pre-existing knowledge and, more importantly, what I didn't know about the chosen topic in order to adapt the program to my level." — learnSTEM student

"The program was very good at helping me determine what I wanted to focus on in my learning…" — learnSTEM student

How learnSTEM helps

In the STEM Research Program, you work one-on-one with a PhD mentor from a top-20 university for 12 weeks, online. You don't need a topic in advance: your mentor helps you explore, and you settle on your research question by week four, then write your paper with one-on-one support. At the end you have a finished research paper and a written evaluation from your mentor.

Apply to the STEM Research Program → · Questions first? Contact us →