TL;DR
Picking a research method for high school students starts with one idea: your question chooses the method. Use an experiment to test cause and effect, a survey to measure what lots of people think or do, and a case study to understand one person, group, place, or event in depth. This page walks you through the comparison table, the “which method fits your question” table, other methods, mixing methods, and common mistakes. Want help matching your question to a method? Apply to Horizon to design your study with a PhD mentor.
What is a research method for a high school student?
A research method is how you collect and analyze evidence to answer your question. Your method isn’t the same as your project type, which is whether you’re doing original research or reviewing existing studies. That choice comes first, and this page starts after you’ve made it.
Data comes in two main kinds. Quantitative data is anything you can count or measure, like quiz scores or minutes of sleep. Qualitative data is words and observations, like interview answers or notes on how a team worked together. Experiments and surveys usually give you numbers. Case studies and interviews usually give you words.
Before you choose, you need a clear question or a hypothesis, which is a testable prediction about what you’ll find. If your question is fuzzy, every method will feel wrong. Our guide to writing and testing a hypothesis shows you how to sharpen it.
Experiments: testing cause and effect
What it is: An experiment is a test where you change one thing, called the independent variable, and measure another, called the dependent variable. A variable is anything that can change in your study. You keep everything else the same and compare your results against a control group, which is a group that doesn’t get the change.
Use it when: Your question asks “Does X cause Y?” or “What happens to Y when I change X?”
Example projects: Test whether background music changes reaction time on a computer task. Compare plant growth under two colors of light. Measure how different filters change the clarity of water samples. Time how fast two sorting algorithms run on the same data. For more ideas, browse STEM research topics. Planning a science fair entry? See what ISEF judges look for.
Data you’ll collect and how to analyze it: You’ll collect measurements, then find averages and graph them. Repeat your trials, because one result could be luck. Replication, repeating your test, makes a result believable. If you can, run a t-test, a statistical test that tells you whether the gap between two groups is probably real or just chance.
Ethics and approvals: Experiments on people, like taste tests, reaction time, sleep, or exercise, count as human participants research. That means an IRB, or institutional review board, a group that checks whether a study is safe and fair, must review it before you start. Read Do High School Researchers Need IRB Approval? for the steps. Projects with vertebrate animals or hazardous materials need extra approvals too. Confirm what applies with your teacher, fair, or program.
- Pros: The strongest evidence for cause and effect.
- Cons: Needs equipment or setup, time for repeated trials, and approvals if people are involved.
Surveys: measuring what people think or do
What it is: A survey is a set of identical questions you ask a group of people. Your sample is the group you actually ask, and your sample size is how many people are in it.
Use it when: Your question asks “how many,” “how often,” or “what do people think about” something, or you want to look for patterns between groups.
Example projects: Screen time and sleep among classmates. Attitudes toward AI tools in school. Commute time and stress. Browse more research topics for high school students, or see psychology programs for high school students if people and behavior interest you.
Data you’ll collect and how to analyze it: Multiple-choice and rating-scale answers become percentages and charts. Open-ended answers get grouped into themes, which are ideas that show up again and again. You can also check for correlation, which means two things tend to move together, like more screen time and less sleep. A survey can show a pattern, but it can’t show that one thing causes the other.
| Tip: how to write survey questions – Ask one idea per question. “How many hours do you sleep?” works. “Do you sleep enough and exercise?” doesn’t. – Avoid leading wording. Ask “How do you feel about the new schedule?” instead of “Don’t you agree the new schedule is awful?” – Balance your answer options, with the same number of positive and negative choices.Test your survey on three to five people first, and fix any question that confuses them. |
Ethics and approvals: Under ISEF rules and Horizon’s IRB guide, surveys count as human participants research no matter how small, and that includes online surveys. You need approval before you recruit anyone, and people under 18 usually need parent permission. The IRB steps in the experiments section above apply here too. Confirm the details with your teacher, fair or program.
- Pros: Reaches many people fast and cheaply.
- Cons: People misremember or answer to look good, and a small or one-school sample limits what you can claim.
Case studies: understanding one example in depth
What it is: A case study closely examines one person, group, organization, place, or event, using multiple sources like documents, interviews, observations, and data.
Use it when: Your question asks “how” or “why” something happened in a specific situation, or the thing you’re studying is rare or can’t be tested in a lab.
Example projects: How one city handled a flood. How a local nonprofit grew. A historical policy decision. A company’s response to a product recall. Looking for more? Try these 45 history research topics.
Data you’ll collect and how to analyze it: You’ll gather notes, documents, and interview transcripts, then look for patterns and themes. Compare what you find with what published research would predict. That comparison is what turns a story into evidence.
Ethics and approvals: A case study built only on public documents usually doesn’t need IRB review. One that includes interviews does. Medical case studies about real patients aren’t a realistic option for a high school student without a supervising institution. Check with your teacher, fair, or program before you begin.
- Pros: Rich detail, works without a lab, and fits social science, history, and business.
- Cons: Findings may not apply to other cases, and it’s easy to let your own opinion steer the story.
Case study vs experiment vs survey: side by side
These three methods answer different kinds of questions, so a side-by-side look helps you see the trade-offs. Read across a row to compare the methods, then read down a column to see how one method behaves. If a row shows a dealbreaker for you, like no access to equipment, cross that method off. No method wins every row, so your question has to do the choosing.
| Factor | Experiment | Survey | Case study |
| Best question type | Does X cause Y? | How many, how often, or what do people think? | How or why did this happen in this situation? |
| Type of data | Numbers | Mostly numbers, some words | Mostly words, some numbers |
| Can it show cause and effect? | Yes, if well designed | No, only patterns | No, but it can suggest explanations |
| Typical sample | Small groups or repeated trials | Many people | One case, or a few |
| Equipment or access needed | Materials, tools, or software | A way to reach people and a survey tool | Documents, records, or interview access |
| Time to collect data | Medium to long | Short to medium | Medium to long |
| Usually needs IRB review? | Yes, if people are involved | Yes | Only if you interview people |
| Common subjects | Biology, chemistry, psychology, engineering, computer science | Psychology, public health, education, social science | History, business, political science, sociology |
Source: Horizon research mentors’ guidance; ISEF human participants rules, October 2026
The method you pick decides what you’re allowed to claim in your conclusion.
Which research method fits your question?
Find the sentence below that sounds most like your question. The middle column shows a method that usually fits, and the right column shows what a real question might look like.
| If your question asks… | Try this method | Example question |
| Does X cause Y? | Experiment | Does listening to music while studying change quiz scores? |
| How common is X, or what do people think about X? | Survey | How many students at my school use AI tools for homework? |
| How or why did X happen in this one situation? | Case study | Why did our town’s recycling program work when a nearby one failed? |
| Are X and Y related in a large group? | Survey or secondary data analysis | Is there a link between hours of sleep and grades? |
| What do people experience when X happens? | Interviews (qualitative) | What is it like to start at a new school in the middle of the year? |
| How does a group behave in its normal setting? | Observational study | How do students use the library during lunch? |
| What would happen if conditions changed, and you can’t test it directly? | Computational model or simulation | How would a flu spread through a school under different vaccination rates? |
| Do existing public numbers show a trend? | Secondary data analysis | Has local air quality changed over the past decade? |
Source: Horizon research mentors’ guidance; ISEF human participants rules, October 2026
Your question comes first, but practical limits matter too: time, cost, access to participants or equipment, and approvals. For most beginners, the easiest options are secondary data analysis or a small survey, as long as you get approval for the survey. If your question still needs work, read how to choose a research topic as a high school student.
Other research methods worth knowing
Observational study
In an observational study, you watch and record what happens without changing anything. That’s the main answer to observational study vs. experiment: in an experiment you change something, and in an observational study you don’t, so you can’t prove cause. Public observation with no interaction and no identifying details is usually exempt from IRB pre-approval under ISEF rules, but confirm with your teacher or fair first.
Secondary data analysis
Secondary data analysis means studying a public dataset that someone else collected, like weather records or school district statistics. You don’t need a lab or participants, and you usually don’t need IRB review when the data is public and de-identified, meaning it has no names or other details that identify a person. Open datasets are covered in 10 Best Ways to Get Research Opportunities in High School.
Interviews and other qualitative methods
Interviews are conversations where you ask a few people open questions and listen closely. You hear from fewer people than a survey reaches, but you get deeper answers, and you analyze them by finding themes. Interviews are human participants research, so they need IRB review before you talk to anyone.
Computational modeling
A computational model is a program that simulates how something works so you can test “what if” questions. You might model how a rumor spreads through a school or how traffic changes with a new light. You adjust the settings in your code and compare the outcomes. Need ideas? Try these 200 AI research topics.
Can you combine research methods?
Yes. Mixed-methods research means using numbers and words together, like a survey followed by a few interviews, or an experiment plus a short questionnaire. Mixing methods helps most when one result surprises you, and you need to explain it. If a survey shows seniors sleep less than freshmen, a few interviews can tell you why.
It hurts when it doubles your work and your approvals on a tight timeline. Surveys need IRB review, and so do interviews, so two methods can mean two sets of paperwork. Pick one main method, and add a second only if it answers a part of your question the first can’t. Whatever you choose, you’ll explain it later in the methods section of your paper, which we cover in the key components of a high school research paper.
Common research method mistakes high school students make
Most method mistakes are easy to avoid once you know to look for them. Two terms help here. Bias tilts your results in one direction, and validity means your method measures what you say it measures. The research terms glossary explains more.
- Your sample is too small to say anything. Fix: collect as much data as your timeline allows, and keep your claims modest. Say “in my sample” instead of “teenagers.”
- Your survey questions are leading or loaded. Fix: read each question out loud and ask whether it nudges people toward an answer. Rewrite it in neutral words and test it on a few people.
- You have no control group, or you changed more than one variable. Fix: always keep a comparison group, and change one thing at a time.
- You treated correlation as cause. Fix: use phrases like “linked to” unless you ran a real experiment.
- You surveyed only friends or one class, then generalized to “teenagers.” Fix: reach a mix of grades and groups, or narrow your claim to the people you actually asked.
- You collected data before getting IRB approval. Fix: get approval first. Data collected too early often can’t be used, so check your fair’s rules.
- You picked the method before the question. Fix: write your question first, then ask which method could actually answer it.
Frequently asked questions
For most beginners, it’s secondary data analysis or a small, well-designed survey. Easy still means careful. You’ll need a clear question, clean data, and, for a survey, approval before you start.
An experiment changes a variable to test cause and effect. A case study examines one situation closely without changing anything, so it can explain how or why something happened but can’t prove cause.
The difference is whether you change anything. In an observational study, you only watch and record. Only an experiment can show cause.
Quantitative research uses numbers and statistics, like test scores or percentages. Qualitative research uses words and themes, like interview answers or field notes.
There’s no magic number. More people usually give you stronger results, and your sample should look like the group you want to talk about. Say honestly in your paper what your sample can and can’t support.
Yes. Under ISEF rules, surveys count as human participants research regardless of size or format, including online surveys. Read our IRB guide and confirm the details with your teacher, fair, or program.
Yes. A mentor can check that your method actually answers your question before you collect anything. At Horizon, you’re paired 1:1 with a PhD scholar or professor for several months of guided research that ends in a roughly 20-page paper. Read What Does a Research Mentor Actually Do? or see Horizon’s programs.
Picking your method
Start with your question, pick the simplest method that can answer it, and plan your approvals before you collect a single answer. Your question picks the method, not the other way around. Worried your method won’t hold up? Apply to Horizon and have a PhD mentor check your plan, or join an info session to see how the program works.
Resources
- Society for Science, ISEF International Rules: Human Participants, https://www.societyforscience.org/isef/international-rules/human-participants/
- Society for Science, ISEF International Rules: Rules for All Projects, https://www.societyforscience.org/isef/international-rules/rules-for-all-projects/
More from Horizon:
- What Is a Hypothesis? How to Write and Test One
- Do High School Researchers Need IRB Approval?
- Research Terms Glossary for High School Researchers
- How to Choose a Research Topic as a High School Student
- What Are the Key Components of a High School Research Paper?
- ISEF: How to Qualify and What Judges Look For
Image source: Horizon Academic Research Program