Q: What should a STEM portfolio contain for a Singapore application? A: Start with the exact target's current form and selection criteria. Record a small number of verifiable activities with the student's role, dates, work, evidence, outcome, and reflection. DSA-Sec and university admissions are different processes; neither publishes a universal competition ladder or points table.
Last reviewed: 23 October 2026 against MOE's DSA-Sec pages and the current NUS and NTU aptitude-based admissions pages.
First separate the two pathways
Pathway
Applicant stage
Published central process
DSA-Sec
Primary 6
Apply for a school's published talent area through the DSA-Sec Portal; the school uses its own criteria and selection process
NUS holistic admissions
Pre-university or other qualifying applicant
All applicants list up to four achievements, co-curricular activities, or non-academic activities and answer short response questions
NTU aptitude-based admissions
Pre-university or other qualifying applicant
Applicants may describe achievements, answer short responses of up to 200 words each, and submit one or two appraisals
DSA-Sec is a Primary 6 application for Secondary 1 admission, so a Secondary 1 to JC2 plan does not feed into it. Work done in secondary school is evidence for later applications, such as university admissions.
DSA-Sec: follow the target school
MOE lists Science, Mathematics and Engineering as a broad DSA-Sec talent area. Schools do not all offer the same area or assess it in the same way.
MOE says applicants generally demonstrate talent, potential, passion, and character through the school track record and online application. It also says a child without prior experience may apply; the school may assess potential. Shortlisted applicants may be asked to attend interviews, auditions, or trials, depending on the school and category.
check the target school's current talent-area page;
use the wording and evidence types that school requests;
do not assume a medal is mandatory;
do not assume one organiser's award guarantees shortlisting; and
prepare for the actual selection task rather than an invented national rubric.
PSLE still matters. MOE says a successful DSA student must qualify for the Posting Group offered by the school. A student admitted through DSA-Sec also commits to that school for the programme and cannot submit ordinary S1 choices or transfer after PSLE results.
University applications: the form is not a scrapbook
NUS's central page says every applicant may list up to four achievements or activities and answer short response questions. It gives examples such as International Olympiad medals, relevant internships, entrepreneurship, leadership, and software or open-source contributions used by many people.
Those are examples, not a checklist or ranking. NUS states that each College, Faculty, or School considers different factors and may request additional information, a test, or an interview if an applicant is shortlisted.
NTU's central page says aptitude-based admissions consider passions, interests, and strengths, subject to a minimum level of academic competence. It asks for achievement details, short responses of no more than 200 words each, and one or two online appraisals. Shortlisted applicants may need an interview and supporting documents.
Neither central page publishes:
an expected floor such as an SMO, NOI, or science-olympiad award;
a three-tier competition ranking;
a fixed number of research attachments;
a guaranteed advantage for a GitHub profile;
one score assigned to every activity; or
a universal STEM interview question list.
Build an evidence record, not an admissions claim
For each activity, record:
Field
Useful evidence
Activity
Official name, organiser, programme, or project title
Dates
Start, end, and time commitment
Role
What the student personally did
Work
Problem, method, design, code, experiment, or analysis
Verification
Certificate, organiser result, supervisor record, repository, report, or permitted media
Outcome
Exact award, result, users, measurement, or honest non-result
Reflection
What changed after testing, feedback, or failure
Attribution
Team members, mentor contribution, reused code, data, and sources
An activity is not stronger because its description sounds grand. It is stronger when the claims are precise and can be verified.
Evidence patterns by activity type
Competition
Record the organiser, year, division, round, exact award, and official result or certificate. Do not convert participation into a placing or add a percentile the organiser did not issue.
Research or science project
Record the question, safety and supervision, the student's own method, data provenance, controls, result, limitations, and mentor role. A null result is acceptable. Fabricated precision is not.
Software or data project
Record the user problem, the student's contribution, repository history, dependencies, data permissions, testing, deployment status, and actual usage evidence. A public repository alone does not prove authorship, quality, or impact.
Robotics or engineering build
Record design constraints, versions, the student's subsystem, test evidence, failures, and changes. Credit the team and adult support.
Internship or attachment
Record the host, dates, assigned work, supervision, and permitted output. Do not publish confidential code, personal data, laboratory records, or a host's intellectual property to make the portfolio look complete.
A defensible project summary
Use four parts:
Problem: What specific question or need did the work address?
Contribution: What did the student personally design, test, analyse, or build?
Evidence: What happened, and how was it measured or verified?
Reflection: What limitation, failure, or feedback changed the student's next step?
Example structure:
Our team tested whether a low-cost sensor could track classroom carbon-dioxide changes. I wrote the calibration and missing-data checks, then compared readings against the reference device. The sensor followed broad changes but drifted after long runs, so I added a recalibration alert and documented why the prototype should not be used for health decisions.
This does not claim that an admissions office prefers the project. It makes the work understandable and bounded.
Select evidence for relevance and truth
When a form permits only a few entries:
choose activities that the student can explain without a script;
prefer a clear personal contribution over a prestigious team name with no defined role;
retain exact organiser language for awards;
show learning or iteration without inventing impact;
include non-STEM activity where it is genuinely among the applicant's strongest evidence; and
follow the programme's current subject prerequisites and supporting-document rules.
NUS expressly allows non-academic activities in the central achievement list. A coherent application does not require every entry to be STEM-branded.
Safety and integrity boundaries
Do not create an application project by:
conducting unsupervised work with biological, chemical, electrical, medical, or other hazardous materials;
collecting personal data without permission and a legitimate purpose;
copying code or prose without attribution;
buying authorship, a publication, an award, or a place in a project;
overstating a mentor's work as the student's work;
inventing users, measurements, awards, or research results; or
disclosing confidential material from a school, employer, laboratory, or participant.
For school or laboratory projects, follow the owner institution's safety, ethics, data, and publication rules.
Application-day checklist
Open the current official form and programme criteria.
Confirm the number and format of entries permitted.
Match every claim to a verification source.
Check dates, award names, links, and team attribution.
Remove claims about admissions value that the institution does not publish.
Keep originals of requested supporting documents.
Prepare to explain the work, limitations, and personal contribution honestly.
