DSA-Sec group rounds · Checked 8 August 2026
Several Singapore secondary schools publish that a group round is part of their Direct School Admission selection. Almost nobody publishes what happens inside one, and no preparation provider trains it properly, because a group task needs other children. This page covers what is known, marks clearly what is only assumed, and gets practical about the part that actually trips children up.
Read this bit first
The evidence on this page is deliberately uneven, and we have marked it that way throughout. That group rounds exist, and at which schools, is confirmed from schools’ own documents. What panels actually score inside one is not published by anyone. Where this page gets specific about behaviour, it is describing skills worth having — not a rubric we have seen.
Definitions
A stage where your child is assessed alongside other candidates rather than alone. It takes several shapes: a group interview where everyone answers in front of each other, a group discussion, a group task with something to solve or build, or a selection camp containing several of these across hours or days.
The distinction matters because the skills differ. In a group interview, the adults are asking and the difficulty is social: your child hears three other answers before giving their own. In a group task, the adults mostly watch, and the difficulty is collaborative: a group of children who have never met has to produce something within a time limit while every one of them is being observed.
Children are almost never told which they are walking into. Some schools name the component; most do not describe it. So the useful preparation is not format-specific drilling — it is being comfortable enough with all four shapes that none of them is the surprise.
Several candidates are interviewed together by the same staff. CHIJ Secondary publishes this as a formal second stage applied to every talent area, and states the purpose in its own words: to assess commitment and motivation level. Anderson Secondary publishes a small group interview following its sports trials.
A structured discussion or short exercise, observed rather than led. Raffles Institution publishes it for Leadership and Character, in a sequence of personal statement, group activity, then interview. SOTA lists group activities as one of four mandatory Talent Academy components.
A problem the group must actually solve or build. The School of Science and Technology publishes a Phase 2 STEAM Challenge in which candidates work in groups on real-world issues, following a written phase of about two hours.
A longer event for shortlisted candidates, made up of several group activities. NUS High publishes a Selection Camp as phase two of its process, after Selection Tests. SOTA's Talent Academy runs across two to three days and mixes group and individual components.
Not usually called a group round, but it poses the same problem in physical form: play well with children you have never met. One parent described her daughter struggling at a trial because it was difficult to receive passes from girls she was not familiar with — which is the group-task difficulty exactly.
Primary sources
CHIJ Secondary publishes a group interview for every talent area. SOTA publishes group activities as one of four mandatory components. Raffles Institution publishes a group activity for Leadership and Character. The School of Science and Technology runs a group STEAM Challenge. NUS High runs a Selection Camp.
Publishes a group interview as Selection Process 2, applied identically across every talent area, conducted with key personnel in the school, to assess commitment and motivation level. The strongest single piece of evidence that group rounds exist as a formal gate.
Source: chijsec.edu.sg (PDF)Publishes group activities as one of four mandatory Talent Academy components, alongside the individual audition or portfolio presentation, a separate panel interview, and online questionnaires. Also states plainly that it keeps its procedures and criteria confidential and gives no feedback.
Source: sota.edu.sgPublishes a per-domain process table in which Leadership and Character runs as personal statement, then group activity, then interview. The only major school we found that names a group stage against a specific talent area in a public table.
Source: ri.edu.sgPublishes a two-phase process ending in a STEAM Challenge — an activity-based assessment where candidates work in groups on real-world problems. Read via a search extract rather than the live page, so corroborated rather than confirmed.
Source: sst.edu.sgPublishes a per-talent-area table stating that sports trials are followed by a small group interview in August, and that trials assess sport-specific skills, fitness and in-game decision making.
Source: andersonsec.moe.edu.sgPublishes a Selection Camp for shortlisted candidates as phase two of its process, with dates. One parent's account of a 2024 camp describes candidates split into sessions of about thirty-five and then into groups of five, with sessions kept apart from one another — a single account, and marked as such.
Source: nushigh.edu.sgA group task and an interview on the same day for its STEM talent area. This comes from a single parent post about one cycle, not from anything the school publishes — the school's DSA page carries a talent list and no process at all.
Source: mgs.moe.edu.sgIts published matrix shows Service Leadership assessed by personal statement plus interview. A parent separately described a leadership round one run as a group session with staggered timeslots, followed by a further round for those who progressed — reported, not published.
Source: en.nygh.moe.edu.sgThis list is what we could verify, not the full set. Plenty of schools publish no process at all, so the absence of a school here says nothing about whether it runs a group round. PSLEPrep is not affiliated with, endorsed by, or connected to any school named above; each entry summarises that school’s own published material as at the date on this page.
The purpose
Because a group shows something a one-to-one interview cannot: how a child behaves around peers when nobody is managing them. A prepared answer survives a solo interview. It does not survive twenty minutes of being disagreed with, ignored, talked over and asked to share a decision.
A secondary school taking a child through DSA-Sec is committing to them for four to six years, and committing them in return: an accepted offer removes a child from Secondary 1 Posting and binds them to that talent for years. Schools are choosing a member of a team, an ensemble, a squad or a committee — and the cheapest way to see whether a child can function in one is to put them in one.
There is also a practical reason. Group rounds are efficient. Where a school shortlists more than a hundred candidates, seeing five at a time is the difference between a week of interviews and a month of them.
And there is a third reason, which is worth naming plainly: the group format is harder to coach. Task content is often withheld until late — one school publishes that its sports tasks are made known only after the application period, and that its debating topic is released near the trial date. A round that requires other children, in real time, is a round in which a scripted child has less advantage.
One correction to a widely repeated assumption. It is often said that group rounds are there to assess collaboration and leadership. The one school-published purpose we could find says something narrower: CHIJ Secondary states that its group interview assesses commitment and motivation level. That may be specific to that school — but it is the only stated purpose in the public record, and it does not say what the prep industry says it says.
Reported versus inferred
Nobody outside the schools knows. The only school-published criterion we found is commitment and motivation level. Everything else in circulation — collaboration, listening, leading without dominating, composure under observation — comes from preparation providers rather than from schools, and is a reasonable guess rather than a rubric.
We have written the rest of this page against those assumptions anyway, for one reason: a child who can hold a view, take a challenge, credit somebody else and stop talking on time is better equipped for secondary school whether or not any panel is scoring it. What we will not do is tell you those behaviours are on a scoresheet, because we have not seen one — and neither has anybody charging you for it.
The central tension
By being useful to the group rather than visible in it. Start the group and then hand over. Ask a question you do not know the answer to. Turn somebody else’s idea into a next step. Children fail this in both directions — the loud one taking every turn, the quiet one waiting to be invited.
This is the piece of advice every DSA guide repeats and none of them makes actionable. Told to lead without dominating, an eleven-year-old hears two contradictory instructions and picks one. The child who picks lead talks over everybody for fifteen minutes. The child who picks do not dominate says almost nothing and leaves the room having contributed one sentence.
The way out is to stop framing it as a volume problem. Leading is not the amount of talking; it is whether the group is better off for your child having been in it. That reframing is usable by a child, and it converts into concrete moves.
The most useful thing a child can do in the first thirty seconds is get the group moving — suggest how to begin, or ask what everyone thinks the task is actually asking. The move that turns that from dominating into leading is the second half: name someone and pass it on. Shall we each say one idea first? You go ahead. The child has organised the group without taking the floor.
Groups of eleven-year-olds who have never met tend to jump straight to answers. A child who says something like, we have got ten minutes, so maybe five minutes of ideas and five minutes deciding, has contributed something the whole group can use, and it does not require them to have the best idea in the room.
Questions are the most underrated group move because they cost nothing to make and they cannot be faked. Why would that not work? and what would we need to make that work? both push the task forward. A child asking a genuine question is visibly thinking rather than performing.
Leading without dominating does not mean never disagreeing. It means disagreeing with an idea and then genuinely changing your mind when the answer is good. Say what you think, hear the reply, and be capable of saying, that is better than mine, let us do that. Children who cannot lose an argument gracefully are the ones who look difficult.
A workable private rule for a child: after I have spoken twice, somebody else speaks before I do again, unless I am asked directly. It is crude, and it is far better than the two failure modes it prevents. It also stops the child who has been told to speak up from doing nothing but speak.
Long, unfinished contributions do more damage than short ones. A child who makes one point in fifteen seconds and stops has said something. A child who is still qualifying it forty-five seconds later has taken the floor and given nobody anything to respond to.
A caution about the quiet child
A naturally quiet child should not be turned into a loud one for a selection round. It rarely holds under pressure, and coaches describe over-rehearsed candidates as visibly afraid to make a mistake. The achievable goal is narrower and kinder: two or three real contributions, one question asked, one other child brought in. That is a full contribution to a group, and it is reachable in a few weeks of practice.
The most common panic
Respond to it rather than abandon it. Agree and extend it, test it with a question, apply it to a next step, or name a real trade-off — and credit the person who said it. Repeating the point in different words is the one move that costs something, because it reads as not listening.
This is close to guaranteed in a group round, and it is the moment children describe afterwards as the one that threw them. A child arrives with one good idea, someone else says it in the first thirty seconds, and the child spends the rest of the session searching for something new instead of listening to what is being said.
Rehearsing the recovery is more valuable than rehearsing the idea. Five moves cover almost every instance.
Agree explicitly, then add the piece nobody has said. I think Aisyah is right that we should ask the younger students first — and we could do it during recess so more of them can come. The child has credited someone else and still moved the task forward.
Take the idea seriously enough to poke at it. That would work if the hall is free. What do we do if it is not? This is difficult to fake and it is the clearest evidence a child is thinking about the task rather than about their turn.
Move the group from idea to consequence: so if we are doing it that way, one of us has to ask the teacher. Groups often generate five ideas and implement none. The child who converts an idea into a next step is doing work nobody else is.
Name a real trade-off. It is a good idea but it costs money, so maybe we need a smaller version. Disagreement done on the idea rather than on the person is exactly the behaviour prep guides say panels notice — and it is the most useful thing a child can practise regardless.
The move nobody teaches. If the point is made, well made, and nothing needs adding, letting it stand is a legitimate contribution. A child who speaks only when they have something is not invisible, provided they have spoken at all — and the child who fills every silence to prove engagement is the one panels are reported to find tiring.
The move adults notice
By naming a person and asking them something easy to answer. Adults watching a group notice it immediately, because it is the one behaviour that cannot be produced by a child who is only managing their own performance — it requires having noticed somebody else in the room.
Be careful with the claim, though. That panels reward this comes from preparation providers, not from any school document we could find. What is defensible is simpler: in any group of children, the one who makes space for a quieter member is doing something observably generous, and adults in the room are human.
There is also a reason to teach it that has nothing to do with selection. A child who has practised drawing somebody else out has a tool for the first week of Secondary 1, which is a room full of strangers with no adult managing the conversation.
What do you think, Wei Ming? works. Anyone else? does not — it invites the loudest child to go again. Naming a person and making the question easy to answer is the single highest-value social move in a group round.
Ask a quiet child which of two options they prefer, rather than asking them to invent something. A closed question is a doorway; an open one can feel like being put on the spot in front of strangers.
When a quieter child gets talked over, saying sorry, I think Nadia was still going hands the floor back. It takes three seconds and it is the clearest single signal of the behaviour the whole format is thought to be looking for.
That was Rohan's idea. Children rarely do this, adults notice it immediately, and it costs a child nothing except the chance to look like they thought of it themselves.
One warning worth passing on. Done as a technique, this backfires. A child who asks what do you think in a bright voice and then talks straight over the answer has performed inclusion rather than done it. If your child brings somebody in, they have to actually listen to what comes back — which is why practising it as a habit at a dinner table works better than learning it as a line.
The other failure mode
Not as confidence. From outside the group, the tell is not volume but repetition — the same point restated, answers given to questions asked of somebody else, and commentary that fills time without moving the task. The group gets less done and one child has used most of the airtime.
Children who over-talk are almost never trying to. They are usually anxious, and often have been told to make sure you speak up, which for a nervous child converts into fill every silence. Naming the specific behaviours turns a vague instruction into something a child can actually monitor.
The most common form of over-talking is not new content — it is the same point restated because the child is not sure it landed. From outside, it reads as not listening.
When an adult names a child and another child answers, everyone in the room notices, including the adults writing things down.
Long running commentary on what the group is doing, without adding a decision or an idea, fills the time without moving the task. The group ends up with less done and one child has used most of the airtime.
Children are often surprised that enthusiastic agreement can be an interruption. Cutting across someone to say yes exactly still takes the floor from them.
Prepared lines land differently in a group, because everyone else is speaking spontaneously. Coaches interviewed about DSA preparation independently describe over-rehearsed candidates as visibly afraid to make a mistake, and describe responsiveness to correction as the thing that cannot be scripted. Their commercial interest is worth noting; so is the fact that a parent who had been through it said much the same thing without selling anything.
The counterweight is worth saying too, because parents of talkative children read pages like this one and over-correct. A child who contributes a lot and is genuinely responsive is not doing anything wrong. The problem is never the number of turns on its own — it is turns that take something from other people and give nothing back to the task.
The practical problem
Build the group out of the family and play badly on purpose. Give the table a real question and eight minutes, have one adult talk too much and another say nothing, and require a shared decision at the end. Rotate your child’s role each round, and count who spoke afterwards.
This is the reason the group format is under-served commercially. Tuition centres sell one-to-one mock interviews because they can staff them. A group task needs four children who do not know each other, which is expensive to arrange and impossible to guarantee — so the market mostly sells the format it can deliver and calls it preparation.
Families can do better than that at home than they expect, because the goal is not to simulate the room. It is to make three specific situations familiar: being disagreed with, having your idea taken, and sharing a decision with people who are not managing you.
Give the table a real, unresolvable question — how should our block handle people not clearing the lift lobby, or what should the school do about food waste — set eight minutes, and require a group decision at the end. Adults deliberately play badly: one adult talks too much, one says almost nothing. The child's job is not to win the argument. It is to reach a decision that everyone had a hand in.
Run the same eight minutes three times with your child in a different role each time — the one who starts the group, the one who brings in the quiet person, the one who has to disagree with something. Naming the role beforehand and reviewing it afterwards is what turns a conversation into practice.
Everybody gets two turns before anybody gets a third. Children feel the constraint immediately: it forces the talkative child to choose what is worth saying, and it gives the quiet child a turn nobody has to fight for. Drop the rule in later rounds and see whether the behaviour survives without it.
Deliberately steal your child's idea before they can say it, and require them to respond anyway using one of the five moves above — extend, test, apply, complicate, or pass. This is the single most transferable drill on this page, and the situation is close to guaranteed in a real group round.
Not to critique the content. Count who spoke, for roughly how long, and how many times somebody was interrupted. Children are usually surprised — the child who felt silent often spoke more than they thought, and the child who felt helpful often spoke over people. The count does the teaching.
One evening with three classmates and a task beats a month of family practice, because the social risk is real. Failing that, existing group settings do the same work: CCA committee discussions, a class project, church or community youth groups, a debate club. The skill being trained is not exam technique, and it does not have to be practised in exam conditions.
Much of what makes a group round hard is not the group. It is speaking a complete thought out loud, in front of adults, when nobody tells you whether it was any good. That part can be rehearsed alone: take a position, give one reason, give one real example, and stop.
One honest limitation. Nothing here simulates a real group round, and any product that claims to would be over-claiming — including ours. The inside of these rooms is genuinely undocumented, and schools intend it that way. What can be trained is the underlying behaviour, and the spoken half of it is the part PSLEPrep already works on for the PSLE oral papers: taking a position, giving a reason, giving a real example, and stopping on time.
Transparent sourcing
Thinner than for one-to-one interviews, and we would rather say so. That group rounds exist, and at which schools, is confirmed from schools’ own documents. What panels score inside one is not published by anybody, and the familiar advice about collaboration comes from providers who sell preparation.
CHIJ Secondary publishes a group interview for every talent area, with its stated purpose. SOTA publishes group activities as one of four mandatory components. Raffles Institution publishes a group activity for Leadership and Character. NUS High publishes a Selection Camp. Anderson Secondary publishes a small group interview. Each is linked above.
Group rounds, group tasks and camps appear consistently across independent parent accounts and prep-industry descriptions from multiple schools. The breadth is well supported even where the individual instance is not.
The widely repeated claim that panels reward a child who leads without dominating comes from prep-industry sources only. No first-hand account describing an actual observed rubric exists in our research at all, and CHIJ's own published criterion — commitment and motivation — is narrower than the market's framing and arguably contradicts it. The internal structure of the 2024 NUS High camp, the Methodist Girls' group task, and the Nanyang Girls' leadership round each trace to a single post.
The gap is structural rather than accidental, and it will not close. SOTA states that it keeps its procedures and criteria confidential and gives no feedback; one parent described a selection camp arranged so that sessions could not communicate with each other. That is defensible on the schools’ part — a published rubric would be drilled within a season — but it means the honest offer here is narrow. We can tell you the format exists and where, from schools’ own documents, and describe behaviours worth having. We cannot tell you what is being scored.
Free preparation pack
The DSA-Sec Interview Preparation Pack includes the round-table exercises on this page, question archetypes by talent area, worked answer structures, and the format-by-school reference. Written for parents, sent to parents.
It is free, and nothing about it expires. We will also let you know when our spoken-answer practice for DSA-Sec interviews opens. Unsubscribe whenever you like.
We are still writing it, so leave your email and we will send it the day it is done rather than send you something half-finished. A free guide for parents: how the DSA-Sec selection process works, the kinds of question interviews tend to explore, and how to help your child practise talking about their own interests and work — drawn from MOE’s published DSA-Sec framework.
These are not real questions from any school, and we neither publish nor ask for actual interview content. DSA-Sec selection decisions are made entirely by each school, on its own criteria, and no preparation can guarantee an interview or an offer. What preparation can do is narrower: help a child speak clearly and honestly about work that is genuinely their own. PSLEPrep is independent and is not affiliated with, endorsed by, or connected to the Ministry of Education or any school.
FAQ
A stage where several candidates are interviewed together by the same staff, rather than one at a time. CHIJ Secondary publishes one as a second stage for every talent area, conducted with key personnel in the school to assess commitment and motivation level.
In a group interview, adults ask questions and candidates answer in front of each other. In a group task, the group is given something to solve or build and the adults mostly watch. Selection camps usually contain several of both.
CHIJ Secondary, SOTA, Raffles Institution for Leadership and Character, the School of Science and Technology, Anderson Secondary, and NUS High through its Selection Camp all publish a group component. Others are reported by parents but not published by the school.
Honestly, nobody outside the schools knows. The only school-published criterion we found is CHIJ Secondary's commitment and motivation level. The familiar advice about collaboration and leading without dominating comes from preparation providers, not from schools, and should be treated as a reasonable guess.
Try to be useful to the group, which sometimes means starting it and often means not. Getting the group moving and then handing over, asking a real question, or converting an idea into a next step are all leadership without taking the floor.
Respond to it rather than abandon it. Agree and extend it, test it with a question, apply it to a next step, or name a trade-off. Credit the person who said it. This situation is close to guaranteed in a group round and is worth rehearsing directly.
Run a four-person family round table on a real, unresolvable question, with adults deliberately playing a talkative and a silent role. Rotate the role your child plays, use a two-turn rule, and record eight minutes and count who spoke. Borrowing three classmates for one evening beats all of it.
Parents ask this repeatedly on forums and nobody answers it, including us. At CHIJ Secondary the group interview is published as a formal stage of the process; whether any individual school treats it as a cut is not published anywhere we could find.
Keep reading
DSA-Sec explained
What Direct School Admission is, how selection works, and the rules that decide what happens after PSLE.
The DSA interview
What actually happens in the room — formats, length, who is on the panel, and how to prepare for it.
Interview questions and worked answers
Reported questions with honest sourcing, and full spoken answers a Primary 6 child could actually say out loud.
Should your child apply?
The honest version. What DSA costs a family in time and commitment, and when the right answer is no.
Interview formats
One-to-one, panel, group task, trial, written test — compiled from what schools publish themselves.
About this page
PSLEPrep is not affiliated with, endorsed by, or connected to the Ministry of Education or any school named on this page. Scheme details are summarised from MOE's published pages and were last checked on 8 August 2026. Always confirm dates and requirements against MOE and the school's own website before acting on them. Reported interview questions are parent- and student-submitted accounts of past experience, not predictions, and are labelled with how well each is sourced.