US Applications
Passion Project Ideas That Hold Up
24 research questions a student could genuinely own, grouped by the interest behind them rather than the subject on the timetable. Each one comes with a method and an end product, because that is the difference between a project and a hobby with a name.
Reviewed 6 September 2026. Every idea is drawn from the subject guides, where it was checked for method, level and output.
What a reader can tell, and what they cannot
An admissions reader cannot tell whether a student was "passionate". Nobody can. What they can tell, usually within a paragraph, is whether the project was chosen or assigned. Chosen projects have a specific question, a method the student can describe, a product someone else could read, and, most tellingly, something that went wrong along the way that the student can explain.
That is the standard every idea on this page is held to. None of them is "learn about" a topic. Each asks a question that can be answered, names the method that answers it, and says what the finished thing is. A literature review with a verdict is a project. A data analysis with charts the student made is a project. "Exploring my interest in neuroscience" is not.
One more honest note. The phrase "passion project" has been worn smooth by people selling manufactured profiles, and readers know it. The way to be believed is not to use the word more convincingly. It is to have done the work for 8 to 12 weeks and to be able to talk about it like someone who did.
The ideas
Start from what you already argue about
Pick the group that sounds like you, not the one that sounds impressive. Every question links back to its subject guide, where there are more like it and the method is explained.
You notice unfairness and want to know whether the data agrees
How does bias enter a machine-learning system, and which mitigation methods actually reduce it?
Produces: A review of documented bias cases and which mitigations measurably reduced the gap.
How do socioeconomic inequalities affect early diagnosis rates of a disease such as type 2 diabetes?
Produces: A short data analysis using national health statistics, with charts you made yourself.
What does the evidence show about social mobility trends in a specific country over recent decades?
Produces: A mobility analysis from published cohort or national data, with your own charts.
What drives gentrification in a specific city, and who benefits and who is displaced?
Produces: A study of one neighbourhood pairing price and demographic data with who actually moved.
You like building things and finding out where they break
How do different bridge or truss designs compare for strength against material used?
Produces: A tested comparison of designs with load-to-failure measured and repeated.
How do sorting or pathfinding algorithms compare in practice on real-world data, beyond their theoretical complexity?
Produces: Implementations timed on real inputs, with the gap from theoretical complexity explained.
How do different ways of handling imbalanced data change a classifier’s accuracy and fairness?
Produces: A comparison of resampling methods on one dataset, reporting both overall accuracy and per-group error.
What can a simple model predict about an orbital or projectile system, and how well does it match observation?
Produces: A simulation with predictions compared against measured or published values.
You argue with the news and want to check who is right
How can common statistical claims in the media be checked, and where do they go wrong?
Produces: A set of worked checks on real published claims, with the errors classified.
What explains the rise of populist movements across different countries?
Produces: A review of competing explanations for populism, weighed against the cross-national evidence.
How should the right to free expression be balanced against protection from harm?
Produces: A position paper working from decided cases rather than from abstract principle.
What does behavioural-economics research reveal about how "nudges" change real-world decisions?
Produces: A review of published nudge trials separating those that replicated from those that did not.
You are interested in how people think, and suspect the popular version is wrong
What does the evidence say about how sleep deprivation affects memory and learning in adolescents?
Produces: A review of sleep-and-memory studies in adolescents specifically.
How does the framing of a question measurably change the answers people give?
Produces: A review of framing-effect studies with a verdict on how large the effect really is.
What does research on neuroplasticity imply for how skills are best learned?
Produces: A synthesis distinguishing supported claims from popular overreach.
What does the research on "cognitive load" imply for how students should revise?
Produces: A synthesis translating cognitive-load findings into concrete revision advice.
You care about a place, a species, or a planet, and want more than a poster
What does the data show about how a specific habitat or species population has changed over time?
Produces: A trend analysis of one population or habitat from published monitoring data, with your own charts.
What do public datasets reveal about pollinator populations in a particular region?
Produces: An analysis of pollinator records for one region, honest about recording bias.
What does the data reveal about why the burden of a specific disease falls so unevenly across countries?
Produces: A cross-country analysis of one disease burden using WHO or GBD data, with your own charts.
How can a simple differential-equation model capture the spread of a disease or a rumour?
Produces: An SIR-style model implemented and compared against published outbreak data.
You read for the argument, and want to make one that survives contact
Is CRISPR-based gene editing a proportionate response to hereditary disease, and who should decide?
Produces: A position paper that states a principle and answers the strongest objection.
What is knowledge, and does the Gettier problem show that the standard definition fails?
Produces: A reconstruction of the standard analysis and an assessment of whether Gettier defeats it.
How did a specific group experience a major historical event differently from the mainstream account?
Produces: An essay reconstructing one group’s experience from sources that record it indirectly.
How has the representation of a particular group changed across a specific genre or period?
Produces: A systematic analysis of a defined sample, with the coding scheme stated before analysis.
How to turn one of these into something finished
- 1 Narrow it until it can failA question that cannot come out the other way is not a question. "How does sleep affect memory" cannot fail. "What does the evidence say about sleep deprivation and memory in adolescents specifically" can, and the answer might be "less than the headlines claim". That is a result.
- 2 Choose the method before the readingLiterature review, public dataset, build-and-test, case study, ethics argument. The method decides what "done" looks like, how long it takes, and whether it is honest for a student without a lab. Every idea above names one.
- 3 Keep the thing you produce small and realSix to ten papers reviewed with a verdict. One dataset analysed with three charts the student made. One design built and tested to failure, twice. A reader would rather see a finished small thing than an unfinished ambitious one.
- 4 Write down what went wrongThe dataset had gaps. The papers disagreed. The bridge failed in a way the model did not predict. This is the part students cut and the part readers look for, because it is the part that cannot be faked.
- 5 Put it somewhere it can be readA PDF with a title, an abstract and a method section. A short write-up on a personal page. For the strongest work, a journal that accepts pre-university research. Unread work is indistinguishable from work that was never done.
Frequently asked questions
What actually counts as a passion project for college applications?
Sustained, self-directed work on a question the student chose, with a method and a finished product that someone else could read, use or test. A reader is not looking for the word "passion". They are looking for evidence that the student kept going after the easy part, and can say what they would do differently now.
How long should a passion project take?
Long enough to hit a problem and get past it. Most serious projects on this site are built to run 8 to 12 weeks at a few hours a week. Something finished in a weekend rarely reads as a passion; something that dragged on for a year without an end product reads as a hobby.
Does a passion project need to be original research?
No, and pretending otherwise is the most common mistake. A well-executed literature review that reaches a defended verdict, or a public-dataset analysis with charts the student made themselves, is a research project. Claiming a novel discovery from a school bedroom invites the one question that sinks it.
Can a passion project be about something unrelated to my intended major?
Yes. Readers value sustained curiosity over strategic alignment, and a project chosen to "match" a major is often the one that looks assigned. What matters is that the student can explain why this question and not another.
Is it acceptable to get help from a mentor?
Yes, in the same way a university student gets help from a supervisor: on the question, the sources, the method and the revision. It is not acceptable for a mentor or anyone else to produce the work. Every idea on this page has a method precisely so that the work is the student’s own and can be defended as such.
Free PDF
Take the 50 strongest ideas with you
A free PDF with fifty of these questions tagged by method and difficulty, the output each should produce, and the feasibility checklist we apply before recommending any project. We email it to you, along with occasional guidance worth having. No obligation, and your details are not passed to anyone.
Now placing students for the October start. Places in each field are limited by mentor availability. We review applications on a rolling basis.
ScholarBridge matches students with doctoral-level or equivalent research mentors across six academic fields. Every project is student-led and completed to a standard the student can stand behind in any university interview.
Explore all programmes