Skip to content
problem genome

One of 592 unsolved problems in the Problem Genome, a collection written up for student design and engineering teams.

social · education

the seating plan's mixed record

Teachers move seats to help disliked or bullied children, and Dutch classroom trials give mixed results: seating disliked pupils closer for several weeks raised their liking, though relational aggression also rose, and disliked children with behaviour problems became better liked by their seatmates, slightly more beside a well-liked buddy, but a shorter version was followed by more overt aggression class-wide and seating bullied pupils next to their best friend did not improve their wellbeing

Source: Hoekstra et al., “Can classroom seating arrangements help establish a safe…” Open ↗source: primaryLast updated: 2026-11-17

Problem statement

Where a child sits decides whom they talk to all day, so teachers use the seating plan to help children who are disliked, left out or bullied: seat them near someone kind, near a best friend, away from a bully. A Dutch research group has tested these moves in randomised classroom trials with upper primary pupils (aged about 8 to 13), and the results are mixed. Seating children who disliked each other closer together for several weeks raised liking for the most disliked children and reduced reported victimisation in a 2012 field experiment (relational aggression also rose there), and in a 2018 trial children with behaviour problems who were disliked at the start became better liked by their seatmates, slightly more next to a well-liked, prosocial buddy than next to a random classmate; a shorter version of the same move in 59 classrooms was followed by more overt aggression and less perceived cooperation across the whole class, for pupils outside the targeted pairs as well as for those in them. Seating chronically bullied pupils next to their best friend, in a 2024 trial with 250 such pupils in 81 classrooms, did not improve their classroom comfort, internalising problems, engagement or victimisation. Seating does build ties on average (a Hungarian trial in 182 classrooms raised the chance of a mutual friendship between deskmates from 15% to 22%), and software that turns pupils' answers into a seating plan already exists, but none of this says which move helps a child who is disliked or bullied. The unsolved problem is knowing which seating move helps which isolated child, for how long, and at what cost to the classmates around them.

Why this matters

Seating is free, used in every classroom, and fully in the teacher's hands, so it is one of the few levers on children's peer relations that needs no programme or budget. But the trials show it can backfire: a well-meant short rearrangement was followed by more aggression across a whole class, and a support move can cost the helper (seatmates of children with behaviour problems lost a little social status in the 2018 trial). The Radboud announcement notes that one in three bullied pupils tells no one (our translation), so teachers often make these choices without knowing who is isolated.

What’s been tried and why it hasn’t worked

(1) Moving disliked pairs closer (2012, 27 classrooms, 651 pupils). Higher likeability ratings for the children seen most negatively at the start, and less peer-reported victimisation and social withdrawal (full text). The same full text reports that relational aggression (peer nominations for spreading rumours about others, ignoring or excluding them) went up in the experimental classrooms while physical aggression was not affected, and the authors suggest that seating disliked peers closer also gives them more chances to be relationally aggressive to each other. (2) A short version of that move (2020, 59 Grade 5 classrooms, 1,573 pupils). Pupils in intervention classes showed more overt aggression and saw less cooperation among classmates than controls, and the effect held for pupils who were not in a targeted pair; teachers' confidence in managing social dynamics did not change the result. The authors suggest the move may create tension that resolves only later, and say whether it turns positive with time "is to be determined" (full text). (3) Seating children with behaviour problems next to a well-liked, prosocial buddy (2018, 64 classrooms, 221 target children). Target children sat next to a buddy or, in the control arm, a random classmate. Taken together, the target children did not become better liked by their seatmates; those who were disliked at the start did, in both arms and slightly more next to the buddy, and teachers rated the target children's behaviour problems lower over time regardless of arm. The seatmates, buddies and random classmates alike, did not become more aggressive, but became slightly less liked by the class (full text). (4) Seating bullied pupils next to their best friend (2024, 81 classrooms in 35 schools, 1,746 pupils, 250 chronically victimised; 8–12 weeks between measurements). The plan was also to seat victims far from their bullies, but for ethical reasons control classrooms always kept at least two seats between victims and their bullies, and teachers' requested changes were always accepted, so the distance to bullies ended up similar in both arms and the trial in the end tested sitting next to a best friend. No support was found for any outcome, and exploratory analyses of continuous distance from bullies showed no effect either. The authors say teachers "should not expect that merely using victims' placement in the seating arrangement will be sufficient", that it "may be helpful as an add-on to other anti-bullying efforts", and that the null result does not mean it worked for no individual; they suggest the best friend's nomination need not have been mutual and the friend may not have defended the victim (full text). (5) Larger and newer proximity studies. A pre-registered Hungarian trial randomised seating in 182 classrooms (2,966 pupils, one semester) and found that sitting next to each other raised the probability of a mutual friendship from 15% to 22% (full text); a 2026 pre-registered replication in 33 Dutch elementary classrooms, which used existing data from the control classrooms of the 2024 trial's own project (Safe at School), found that rising proximity over time increased mutual liking, but not perceived popularity (full text). Neither targets disliked or bullied pupils. (6) Tools that automate seating plans. The 2024 trial's own seating plans came from an online tool that a software company built for the project, which generated 100 candidate arrangements per class and offered the best fit (full text); so automation was in place for the null trial. A 2026 operations-research paper gives a model and heuristic that assign seats to minimise interpersonal conflicts, tested in computational experiments on real and artificial instances, not in classrooms (abstract). Stoeltjesdans, developed with a Radboud developmental psychologist, asks each pupil who are a classmate's best friends, who is most popular, who is helpful and who gossips most, and produces a seating plan "based on current scientific insights" that the teacher can adjust (our translation). The 2022 announcement presents no outcome data and describes a fundraising campaign to strengthen the platform's scientific basis.

What would unlock progress

Seat-assignment software already exists, so the missing piece is not the optimiser but what it should aim for: a rule set that says, for a given child and class, which move to make and how long to keep it, built from the trials' own contrasts: long versus short exposure (2012 versus 2020), helping a disliked child versus protecting a bullied one (2018 versus 2024), and costs to helpers and bystanders. The 2024 trial also shows a design limit: the most obvious protective move (keeping victims away from bullies) could not be tested for ethical and practical reasons; its authors suggest existing datasets in which victims happened to sit near their bullies, and simulation is another route. The authors also name open variables: whether the friendship is mutual, whether the friend defends the victim, and sight lines to bullies.

How we’d know it’s being solved

(a) For the isolated or bullied child: the trials' own outcomes (likeability ratings, peer-reported victimisation, classroom comfort); (b) for classmates: overt aggression and perceived cooperation, the bystander outcomes that went the wrong way in 2020; (c) for helpers: change in their social status; (d) whether effects hold after the seating changes again. A plan that looks good (best friends together, bullies apart) is a proxy, not an outcome.

Where to start

Facility-free door (build the decision rules): from the four trials' published designs and results, write a teacher-facing decision guide: for each kind of child (disliked, aggressive, bullied), what was tried, what happened to the child and to the class, and what is unknown. Then build a small simulation on invented classroom friendship and dislike networks (a published conflict-minimising seat heuristic can serve as the baseline) to show how often a proposed seating rule moves a buddy's status or puts disliked pairs together. Test the guide for clarity with education students. No partner, school or data beyond the papers is needed. Teacher door: run short design sessions with teachers or student teachers from a local school or the campus teacher-education programme, asking them to make seating decisions for invented classes with and without the guide, and recording what they would do; no pupils are involved, and the team collects no data about real pupils. Design-the-test door: with a local school, design (do not run) a trial that separates exposure length (several weeks versus a short version) and measures bystander outcomes; the participants are children, so the design must include ethics review, active parental consent (the 2024 trial obtained it for 90.4% of pupils), a safety rule such as that trial's minimum of two seats between victims and their bullies, and a plan for checking class-wide aggression during the trial, since that went the wrong way in 2020. Any questions to pupils about whom they like or dislike stay with a trained research team under that review; the student team does not ask them. A physical object does not follow from this problem: the lever is who sits where, not the furniture, and what is missing is evidence on which move to make, not a way to lay out a plan, which existing seating software already does. Relevant skills: education, social network analysis, developmental psychology.

Genes

Related problems

Stuck the same way, in other fields.

field: making & materialssolar glass no furnace wantsSixty to seventy percent of a solar panel is glass, and the antimony added to make it clearer keeps that glass out of float furnaces — so "recycled" glass goes to lower-quality usesbuilt without the people who'd live with itsource: secondaryfield: water & energywhere the sludge goes, nobody counts43% of the world uses pits and septic tanks, but data on whether the sludge is ever treated covers only 1% of the global population — the faecal-sludge accounting gapbuilt without the people who'd live with itsource: primaryfield: health & medicinewhere to put the baby downIn Japan, some parents bathing a baby alone set the baby down on the washing machine or the bath lid; SG-standard bath-lid labels already warn against using a baby bath there, two babies died in neck floats in the bath after repeated warnings, and the appliance makers' 2026 response was a leaflet line, while parents cite narrow changing rooms and conveniencebuilt without the people who'd live with itsource: primaryfield: the built worldthe extra deck nobody calculatedAround 95% of ferry deaths happen on domestic routes that SOLAS never covered — and the boats that capsize have often had decks added and land engines fitted with no naval architect, no stability booklet and no surveyor able to checkbuilt without the people who'd live with itsource: primary

Sources

  • Hoekstra NAH, van den Berg YHM, Lansu TAM, Peetz HK, Mainhard MT, Cillessen AHN. "Can classroom seating arrangements help establish a safe environment for victims? A randomized controlled trial." Aggressive behavior 2024;50(5):e22173, doi:10.1002/ab.22173 (full text, publisher version, CC BY 4.0, Leiden University repository), accessed 2026-11-17. Link ↗
  • Braun SS, van den Berg YHM, Cillessen AHN. "Effects of a seating chart intervention for target and nontarget students." Journal of experimental child psychology 2020;191:104742, doi:10.1016/j.jecp.2019.104742 (full text, publisher version, Radboud Repository), accessed 2026-11-17; van den Berg YHM, Stoltz S. "Enhancing Social Inclusion of Children With Externalizing Problems Through Classroom Seating Arrangements: A Randomized Controlled Trial." Journal of emotional and behavioral disorders 2018, doi:10.1177/1063426617740561, accessed 2026-11-17; van den Berg YH, Segers E, Cillessen AH. "Changing peer perceptions and victimization through classroom arrangements: a field experiment." Journal of abnormal child psychology 2012;40:403-412, doi:10.1007/s10802-011-9567-6 (full text, publisher version, Radboud Repository), accessed 2026-11-17. Link ↗ Link 2 ↗ Link 3 ↗
  • Radboud Universiteit, "'Stoeltjesdans' geeft leerkrachten meer inzicht in eigen klas," 14 April 2022, accessed 2026-11-17 (Dutch, read in the original) Link ↗
  • Rohrer JM, Keller T, Elwert F. "Proximity can induce diverse friendships: A large randomized classroom experiment." PloS one 2021;16(8):e0255097, doi:10.1371/journal.pone.0255097, accessed 2026-11-17. Link ↗
  • Vandemeulebroucke DN, van den Berg YHM, Hoekstra NAH, Burk WJ. "The social implications of classroom seats: replicating and extending previous findings on seating arrangements and peer status." Journal of experimental child psychology 2027;274:106666 (Europe PMC record year; online 2026), doi:10.1016/j.jecp.2026.106666 (full text, publisher version, CC BY 4.0, Radboud Repository), accessed 2026-11-17. Link ↗
  • Charytitsch BCB, Nascimento MCV. "Conflict-aware seat assignment in classroom environments." Computers & Operations Research 195:107613, 2026, doi:10.1016/j.cor.2026.107613 (abstract), accessed 2026-11-17. Link ↗
Revision notes

This brief gets checked against its sources and contains the following revisions. Future revisions get added here.

Worth knowing

  • the 2024 trial reports 250 bullied pupils in 81 classrooms but does not say how many of those classrooms had such a pupil
  • the 2018 paper's own discussion still calls seating these children beside a well-liked buddy beneficial, although its results show only a slight edge over a random classmate

Full check

Sources read

  • van den Berg, Segers and Cillessen, seating field experiment (Journal of Abnormal Child Psychology, 2012): full text
  • Braun, van den Berg and Cillessen, seating chart intervention (Journal of Experimental Child Psychology, 2020): full text
  • Hoekstra et al., seating trial for bullied pupils (Aggressive Behavior, 2024): full text
  • van den Berg and Stoltz, buddy seating trial for children with behaviour problems (Journal of Emotional and Behavioral Disorders, 2018): full text
  • Rohrer, Keller and Elwert, Hungarian classroom seating experiment (PLoS One, 2021): full text
  • Vandemeulebroucke et al., replication on seating and peer status (Journal of Experimental Child Psychology, 2026): full text
  • Radboud University news item on the Stoeltjesdans seating tool (2022): full text (read in Dutch)
  • Charytitsch and Nascimento, conflict-aware seat assignment (Computers & Operations Research, 2026): abstract only (the page claims only what the abstract says)

What changed

  • Stated more strongly than the source: the 2018 trial was described from its summary as making children better liked beside a well-liked buddy; its results show that only children disliked at the start became better liked, in both arms, slightly more beside the buddy, and that buddies and random seatmates alike became a little less liked by the class.
  • Title: the full title said a well-liked buddy helped children with behaviour problems; it now says disliked children with behaviour problems became better liked by their seatmates, slightly more beside a buddy (display title unchanged).
  • Other: the 2026 replication is again described as Dutch, which its full text confirms; the replication is no longer presented as separate evidence: it used data from the control classrooms of the same project as the 2024 trial.

Left open

None noted in this check.

Separate check of specific claims: A later check changed no fact on the page.

Earlier full check: The rise in relational aggression in the 2012 trial was added to the page and the full title, the 2020 class-wide effect is now said of pupils outside the targeted pairs, and two descriptions of newer papers were brought back to what they say.

Separate check of specific claims: Safety lines were added to the teacher and trial-design entry points: the team collects no data on real pupils, questions to pupils stay with a trained research team, and a trial must watch class-wide aggression.

Earlier full check: Both titles were rebalanced (display title The Seating Plan Gamble → The Seating Plan's Mixed Record), the age range was corrected, and a large Hungarian trial, a newer replication and existing seating software were added.