Articles

How to Choose a Student Research Project You Can Actually Complete

Written by Icon | Nov 19, 2026, 8:30:31 AM

A research idea can sound impressive and still be a poor project to choose.

“Artificial intelligence”, “cancer treatment” or “climate change” may all be interesting areas, but each is far too broad to investigate meaningfully on its own.

At the other extreme, a project can become so ambitious that it depends on equipment, participants, data or approvals you may never obtain.

For students considering research projects for high school students, the better question is not:

What is the most impressive topic I can choose?

It is:

What question can I investigate well with the time, evidence, resources and supervision I actually have?

A focused literature review may be more appropriate than an experiment you cannot support with the necessary equipment or supervision. Public-data analysis may also be a better fit when access, sample size or approval would make primary data collection unrealistic.

This guide focuses on choosing and narrowing a project you can realistically complete.

The examples below are fictional planning examples. They do not imply guaranteed access to laboratories, participants, publication or any specific research opportunity.

What Makes a Feasible Research Question?

A feasible research question has clear boundaries.

Compare:

How does technology affect education?

with:

What does recent research suggest about the use of generative AI for feedback in secondary education?

The first question leaves almost everything undefined.

The second gives you a clearer topic, evidence source and context.

A useful question should let you answer five practical questions.

A Five-Part Feasibility Test

1. Is the question specific?

Can you identify the population, variable, dataset, comparison or body of literature you want to study?

2. Can you access the evidence?

Do you already know where the articles, datasets or materials might come from?

3. Is the method manageable?

Can you explain what you will actually do with the evidence?

4. Can you finish within the available time?

A question requiring a year of data collection is not a good fit for an eight-week project.

5. Does it require supervision or approval?

Projects involving people, private data, specialized facilities or certain laboratory procedures may create additional requirements.

A project does not become stronger simply because it becomes more complicated.

Good scope means choosing a question that is meaningful enough to investigate, but narrow enough to complete properly.

Question-Narrowing Worksheet

Use this process before you begin collecting evidence.

Step Question to ask Example
Broad interest What area interests me? Adolescent sleep
Specific issue Which part interests me most? Screen exposure before bedtime
Accessible evidence What can I realistically use? Published academic studies
Boundary Which population or context will I focus on? Adolescents aged 13–18
Method What can I do with the evidence? Compare findings and study methods
Final question What can I realistically answer? How is evening screen exposure associated with sleep duration among adolescents in recent research?

The question becomes stronger as the scope becomes clearer.

When choosing a research topic, define the evidence and method at the same time.

If you cannot explain how you would investigate the question, it probably needs another round of narrowing.

What Can You Do With the Time and Resources You Have?

A research project does not automatically require a laboratory, an original survey or access to patients.

Different formats can develop different research skills.

Here are three realistic hypothetical project scopes.

Project Scope 1 — Literature Review

Broad interest: adolescent health

Focused question:
How is evening screen exposure associated with sleep duration among adolescents in recent research?

Possible process:

  • define a recent publication period;
  • identify a manageable set of academic studies;
  • compare populations and methods;
  • examine whether findings agree;
  • identify limitations and unanswered questions;
  • produce a structured literature review.

The project reviews existing evidence. It does not require collecting health information directly from students or making clinical recommendations.

A literature review should also do more than summarize one paper after another.

The research task is to compare evidence:

Where do studies agree? Why might they disagree? Which methods appear stronger? What remains uncertain?

Project Scope 2 — Public-Data Analysis

Broad interest: climate and environment in Singapore

Focused question:
How has monthly mean surface air temperature at Changi Climate Station changed over a defined 20-year period? 

Singapore’s data.gov.sg currently provides an official National Environment Agency dataset containing monthly mean surface air temperature observations from the Changi Climate Station, with records extending from 1982 through 2026.

Possible process:

  • select a clearly defined 20-year period;
  • download the public dataset;
  • organize the monthly observations;
  • calculate annual or seasonal summaries;
  • produce a simple line chart;
  • identify patterns;
  • discuss what the dataset can and cannot explain.

This project does not require collecting new data.

It lets the student focus on:

research question → dataset selection → data handling → visualization → interpretation

That can be more realistic than designing a survey simply because collecting your own data sounds more advanced.

Project Scope 3 — Supervised Experiment

Broad interest: plant biology

Focused question:
How does salt concentration affect mung-bean germination over seven days?

Possible process:

  • choose several safe salt concentrations;
  • keep other growing conditions consistent;
  • define how germination will be measured;
  • collect observations over seven days;
  • compare results across conditions;
  • discuss variation and limitations.

The question is intentionally small.

It does not require sophisticated laboratory equipment and does not claim to represent all plant species.

However, the student should still check the design with the appropriate teacher or mentor before beginning.

The aim is not:

“How advanced can I make this experiment?”

It is:

“Can I carry out the method consistently and interpret the evidence honestly?”

Comparing the Three Approaches

Project type Main resource Main skill Common scope problem
Literature review Academic sources Comparing and synthesizing evidence Trying to review an entire field
Public-data analysis Existing datasets Analysis and interpretation Using too many variables
Supervised experiment Materials + supervision Experimental design and measurement Designing beyond available time or equipment

These are different project formats, not a ranking.

A laboratory experiment is not automatically stronger than a literature review.

The right format depends on the question and the resources available.

When Do You Need Supervision or Approval?

Check this before collecting data.

A project may need additional supervision, institutional review or safety oversight if it involves:

  • human participants;
  • surveys or interviews intended as research;
  • identifiable or non-public personal data;
  • human tissue, cells or biological material;
  • specialized laboratory procedures;
  • potentially hazardous materials;
  • institution-controlled facilities or equipment.

For students exploring research opportunities in Singapore, supervision and approval requirements can vary depending on the host institution and the type of project. 

At NTU, for example, non-NTU students such as JC or IP students conducting human-subject research at NTU through programmes such as the Nanyang Research Programme are required to obtain ethics approval with their NTU supervisors through the relevant ethics route.

NUS guidance also distinguishes between educational exercises and research. At its College of Design and Engineering, a student activity intended mainly to teach research methods without contributing to generalizable knowledge may not require ethics review. If the work is intended to generate generalizable knowledge, however, institutional ethics processes may apply, and supervisors are responsible for determining the appropriate classification.

The practical lesson is not that every student project needs ethics approval.

It is:

Do not assume that having an interesting question automatically gives you permission to collect data from people.

Before Collecting Primary Data

Ask:

Do I need people, private information or restricted resources to answer this question?

If yes, speak to your teacher, mentor or host institution before proceeding.

If access is uncertain, redesigning the question may be more sensible than building the entire project around permission you may not receive.

For example, instead of:

Survey 500 students about stress and academic performance.

you might investigate:

What does recent published research report about academic stress among secondary-school students?

The second question still allows meaningful analysis without assuming access to hundreds of participants.

What Can a Mentor Help With?

Supervision is not only about approval.

A mentor can help you:

  • narrow an overambitious question;
  • identify suitable evidence;
  • choose a realistic method;
  • recognize practical or ethical constraints;
  • determine what evidence would actually answer the question.

Icon Education’s current Research Programme offers one-to-one research mentorship across STEM and Humanities tracks, with students developing a structured research project and research paper.

If your project is specifically STEM-focused, Icon’s STEM admissions guidance also discusses research projects and other forms of technical portfolio evidence.

What Counts as a Useful Outcome?

A useful research project does not need to produce a dramatic result.

It also does not need to end in publication.

What matters is whether the project answers a focused question through an appropriate process and whether you can explain what the evidence supports.

For a literature review, the outcome might include:

  • a structured synthesis;
  • areas of agreement and disagreement;
  • methodological limitations;
  • unanswered questions.

For a public-data project, it might include:

  • a cleaned dataset;
  • appropriate calculations;
  • clear charts;
  • an evidence-based interpretation.

For a supervised experiment, it might include:

  • a defined method;
  • organized observations;
  • analysis of results;
  • limitations and possible improvements.

An inconclusive result can still be useful.

Suppose an experiment shows no clear difference between two conditions.

You can still ask:

  • Was the measurement sensitive enough?
  • Was the observation period too short?
  • Were important variables uncontrolled?
  • What would a follow-up project change?

The research process matters more than forcing the evidence into an exciting conclusion.

Final Feasibility Checklist

Before committing to a project, check:

Question Yes / Needs work
Can I state the research question in one clear sentence?  
Is it narrow enough for the time available?  
Can I access the evidence or data I need?  
Do I understand the method I would use?  
Do I have the necessary materials or software?  
Do I need a teacher, mentor or specialist?  
Do I need safety or ethics approval?  
Can I complete the analysis, not just collect information?  
Do I know what a realistic final output looks like?  
Do I have a simpler backup scope if access changes?  

If several answers are Needs work, narrow the project before beginning.

That is not making the project less ambitious.

It is making it more feasible.

Final Takeaway

The strongest student research project ideas are not necessarily the most complicated ones.

A useful sequence is:

interest → focused question → accessible evidence → realistic method → supervision or approval check → achievable outcome

When comparing research project ideas for high school students, ask:

Can I investigate this question properly with the resources I actually have?

A focused literature review, careful public-data analysis or well-designed supervised experiment can all develop meaningful research skills when the scope is realistic.

Choose the question you can investigate well, not the one that merely sounds the most impressive.

Explore whether a supervised research programme fits your interests.