How to Do Research in High School: Step-by-Step Guide

Learning how to do research in high school means moving from an interesting subject to an ethical, feasible, and defensible project. It is not about finding the biggest topic or forcing a dramatic conclusion. Strong student research asks a focused question, uses evidence carefully, documents decisions, and explains limits honestly.

This guide explains the research process. If a student still needs to compare programs, labs, internships, virtual work, or science fairs, start with research opportunities for high school students.

What counts as original research?

Original research is a structured effort to answer a focused question by collecting, generating, analyzing, or interpreting evidence in a way that adds something beyond a general summary. The contribution can be modest. It may be a new analysis of public data, a tested engineering design, a documented field observation, a carefully framed archival interpretation, or an experiment that answers a narrow question.

A class report, a copied demonstration, or a prototype without systematic testing can still teach useful skills, but it is not automatically original research. The goal is not to claim novelty before checking the field. It is to define a contribution the student can explain and support.

The student research process: an end-to-end roadmap

1. Choose a broad interest and narrow it to a focused question.
2. Review credible literature and identify what is already known.
3. Define a feasible contribution or gap.
4. Select a method that matches the question.
5. Identify ethics, safety, privacy, and approval requirements.
6. Write a plan with milestones, materials, and evidence needs.
7. Pilot the method when appropriate.
8. Gather data or other evidence carefully.
9. Keep records, versions, and a decision log.
10. Analyze findings without forcing a preferred conclusion.
11. State limitations and report negative or mixed results honestly.
12. Write, revise, and share the work through an appropriate format.

Each step may send the student back to an earlier one. Revising the question after a literature review or changing a method after a pilot is normal when the change is documented and approved where required.

Move from a broad interest to a focused question

A broad subject is not yet a research question. “Improve mental health,” “solve climate change,” and “use AI in medicine” are goals, not manageable studies. A focused question is specific, answerable with available evidence, connected to a defined context, and narrow enough to complete responsibly.

Use this progression:

Broad interest → defined problem → research question → method → measurable evidence

For example, a student interested in urban heat might ask how shade and surface material relate to afternoon temperatures in selected school-courtyard zones using an approved, repeatable protocol. This is hypothetical. It does not claim to solve climate change, and it may still need permission or safety planning.

A narrow question is not less ambitious. It makes accurate evidence and learning possible.

Review literature and identify a gap

Begin with reliable sources: scholarly articles, government reports, university libraries, professional organizations, and primary historical sources where appropriate. Take notes on the question, method, evidence, limitations, and citations. Read beyond an abstract before treating a paper as support for a claim.

A gap can be a limitation in existing work, a new local context, an untested comparison, a needed replication, a data question, or a different interpretation of primary sources. It does not have to be a global breakthrough. Students should avoid calling a project “first” or “novel” unless they can support that statement.

A literature review may itself be a valuable project when it has a clear method and purpose, but it is not automatically empirical research. The chosen format should fit the question and discipline.

Choose a feasible research format and method

Different questions need different evidence.

• Wet-lab research: Uses controlled procedures, materials, and appropriate facilities. It may require trained supervision, safety procedures, and prior review.
• Computational and public-data research: Uses existing datasets, code, models, or reproducible analysis. Students should document data origin, cleaning choices, assumptions, licenses, bias, and limits.
• Engineering design and testing: Defines a need, criteria, and constraints, then builds and tests a design against those criteria. A working-looking prototype is not enough without meaningful testing.
• Environmental or field research: Uses observations or approved sampling with consistent protocols, safety planning, and respect for site rules.
• Surveys and behavioral research: May involve human participants, privacy, consent, and required review before recruitment or data collection.
• Archival or humanities research: Interprets primary sources, texts, artifacts, or records using a transparent method and careful contextualization.

Choose the smallest method that can answer the question. If the method requires unavailable equipment, unsafe work, private data, or an unqualified claim, narrow the question or select a different format.

Ethics, safety, privacy, and approvals

Some approvals must occur before work begins. The Society for Science International Rules are updated annually for ISEF and affiliated fairs and address welfare, safe laboratory practices, environmental protection, documentation, and eligibility. The HHS Office for Human Research Protections provides federal regulations, guidance, and decision resources for protecting human subjects.

Projects involving people, surveys, interviews, identifiable information, vertebrate animals, biological materials, microorganisms, hazardous chemicals, devices, regulated facilities, or some fieldwork may need review, permission, qualified supervision, or training. School, institution, lab, and fair rules may differ or add requirements.

Retrospective approval may not be possible. Self-directed does not mean unsupervised. Students should not conduct unsafe medical, biological, chemical, animal, or human-participant work at home. Parental permission alone may not satisfy every requirement. This article provides general education, not medical, legal, biosafety, or institutional-review advice.

Make a research plan and milestones

Write a short plan before collecting evidence. Include the question, background sources, method, needed materials or data, likely risks, oversight, what will be measured or interpreted, and a recordkeeping method. Add milestones for pilot work, collection, analysis, drafting, feedback, and revision.

Do not use a generic timeline as a promise. A project's duration depends on its scope, approvals, access, setbacks, and method. A smaller project completed carefully is better than an oversized project rushed toward a deadline.

Use this practical checklist before collection begins:

• State one focused research question.
• List credible background sources and open questions.
• Match a feasible method to the question.
• Confirm permissions, review, training, and supervision before starting.
• Define evidence, records, and version-control practices.
• Schedule a pilot, analysis, feedback, and revision.

Pilot test before relying on a method

A pilot is a small test of whether the plan can produce useful evidence. It may reveal unclear instructions, inconsistent measurements, missing controls, impractical equipment, confusing questions, or a need to narrow the scope. A pilot does not justify quietly changing a project after the fact. Record what happened and obtain any needed approval before proceeding with a revised plan.

For an engineering project, pilot testing may compare early designs against defined criteria. For computational work, it may test whether data are complete and whether code produces expected outputs. For archival work, it may test whether selected sources can answer the intended question.

Gather evidence and document the work

Collect evidence according to the plan. Keep a dated notebook or digital log with procedures, source records, raw data, code versions, parameter choices, failed attempts, and changes. Store files in an organized way and protect private information.

Version control can be simple: clear filenames, dated folders, tracked changes, and a record of why a change was made. For code, a repository may help when appropriate. The point is not a particular tool. It is being able to reconstruct what the student did and why.

Students should credit mentors, teammates, laboratories, datasets, tools, and outside feedback. Documentation of support protects student ownership because it makes roles clear.

Analyze without forcing a conclusion

Analysis should follow the question and method, not a hoped-for result. Check calculations, labels, units, assumptions, missing data, alternative explanations, and whether the evidence supports correlation, comparison, or a stronger claim. Do not remove inconvenient evidence without a documented, defensible reason.

Negative and mixed results can still be useful. A project can show that a method did not work as expected, that evidence is insufficient, or that a result applies only in a narrow context. Fabricating data, changing a hypothesis after results appear, or overstating a finding undermines research integrity.

State limitations and write the report

Every project has limits. These may include sample size, measurement precision, missing data, possible bias, short observation periods, source availability, or limits on generalization. Naming limitations does not weaken a project. It shows the student understands what the evidence can and cannot support.

A report may use an abstract, introduction, methods, results, discussion, acknowledgments, and references, but disciplines differ. An engineering report may emphasize criteria, constraints, designs, and testing. A humanities paper may emphasize sources, interpretation, and context. Use the format expected by the intended audience.

Write clearly enough that a reader can understand the question, what the student did, the evidence, the result, and the limits. Cite sources consistently and distinguish the student's work from ideas or words drawn from others.

Get feedback and revise

Ask a qualified teacher, mentor, or reader to question the logic, method, claims, and clarity. Feedback should improve the student's thinking, not replace it. The student must remain able to explain the work and make the final intellectual decisions.

Revise figures, explanations, methods descriptions, citations, and conclusions as needed. A clear revision record can show how the work improved and help the student prepare for a presentation or interview.

Present, compete, report, or publish

A finished project may fit a classroom presentation, poster, science fair, technical report, preprint, student journal, or field-specific journal. These outcomes are not interchangeable. A public PDF is not automatically peer reviewed, and a science-fair result is not a journal publication.

Students considering a fair should read how to prepare for ISEF and locate the appropriate route through the Society-affiliated fair network. Students considering a manuscript should use how to publish research in high school. Neither outcome is required for a project to be meaningful.

Document student ownership

Keep records that clarify the student's work: dated plans, notes, drafts, code, data files, source lists, version histories, test logs, feedback summaries, approvals, and a final reflection. State mentor and teammate roles accurately. A student should not claim a larger team's work as entirely their own, and a mentor should not complete research or writing for the student.

This documentation supports an honest description later, whether the work appears in a portfolio, project page, application, science fair, or conversation.

How Personify supports research execution

Personify works with students in grades 6 through 12 on one standout project through 1-on-1 mentorship one to two times per week. A dedicated admissions expert develops the roadmap, while a field-expert project mentor supports execution. Mentors work alongside students like co-founders, and Personify's internal system tracks progress.

Support may include question scoping, literature strategy, method planning, milestones, analysis feedback, revision, and selecting an appropriate sharing path. Students remain the genuine intellectual owners of their work. Personify does not provide institutional ethics approval, replace qualified safety oversight, complete research for students, or promise lab access, publication, awards, or college admission. Families can explore the research project path and How It Works.