Loading today’s stories
Science1 distinct publisher3 min readPublished
Manchester's correlates are sharp about which pupils get suspended and quiet about sequence, because the survey answers and the suspension records come from the same academic year, which leaves belonging as plausibly a casualty as a cause.
The Scientist · Science desk
Compiled by The ScientistSomething wrong?How this is made
Odds ratios are worth converting into people before anyone budgets against them. About 6% of the 19,064 pupils were suspended at least once during 2023-24 [9][2], which is roughly 1,140 children [15]. Take that 6% as the comparison point and more than doubled odds arrive at about 11% [16]. The gap deserves attention, and at 11% roughly nine in ten pupils on SEN support finished the year without a suspension [16]. The announcement gives relative comparisons and one overall rate, not rates for the SEN, EHCP and free-school-meal groups separately [18], so read that conversion as an illustration of scale rather than a number from the paper. (One housekeeping oddity: the announcement describes the linked data as covering Hampshire and the Isle of Wight in one sentence and Hampshire and Southampton in its methods paragraph [3][2]. Anyone generalising to another authority should establish which.)
Sequence is the harder problem. Survey responses and suspension records both come from the 2023-24 academic year [19], and the protective correlates are appraisals of school: whether a pupil feels they belong, how they rate their relationships with staff, whether they are satisfied with their attainment [8]. A pupil suspended in the autumn term has reasons to report weaker belonging and cooler staff relationships in a later survey, and that would generate the association reported. Peer pressure and bullying victimisation sit on the same footing [7], as does emotion regulation [8], all of it self-described. A single-year cross-section does not separate the reading where belonging keeps pupils in school from the reading where suspension costs a pupil their belonging.
To its credit, the release names one such problem outright. Pupils reporting any substance use in the past 30 days had nearly three times the odds of suspension, partly because substance use is itself grounds for suspension under Department for Education guidance [6]. That is a predictor overlapping the definition of the outcome, stated plainly rather than left for a reader to notice.
Whether school practice is the lever is a question about variance between the 69 schools [2]. If schools with similar intakes suspend at measurably different rates, practice is implicated. The figure that would show it, the share of suspension variation sitting between schools rather than within them, is not in the announcement [20]. Dr Stephanie Cahill's reading, that suspension is not simply about an individual classroom or a young person's behaviour in isolation [12], is consistent with every correlate reported, and the correlates on their own do not choose between the school and the circumstances the pupil brings to it. What the analysis establishes firmly is who is at risk; the recommendation for early intervention and preventive support [14] rests on that, and on the assumption that the appraisals which track lower suspension odds would still track them if they were measured first.
Ranked by verification strength, evidence, and original report placement.
The reported data cover a single academic year, 2023-24, with the survey responses and the suspension records drawn from that same year.
New research from The University of Manchester on school suspensions is published in the British Educational Research Journal.
The researchers analysed data from 19,064 students in Years 8 and 10 across 69 mainstream secondary schools in Hampshire and Southampton during the 2023-24 academic year.
Students receiving special educational needs (SEN) support were more than twice as likely to be suspended as students without SEN, and students with an Education, Health and Care Plan (EHCP) also had more than twice the odds of suspension.
Students eligible for free school meals were nearly two and a half times more likely to be suspended than those not eligible.
Students reporting any substance use in the past 30 days were nearly three times more likely to be suspended than those reporting no recent substance use, in part because substance use is itself grounds for suspension under DfE guidance.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 4, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
science
Feeds are triaging patients by engagement, and the proposed defence is literacy, not moderation1 distinct publisher
science
Approved airport expansions already push UK aviation past its 2038-42 carbon budget share1 distinct publisher
science
Rewriting 26,000 crowdfunding pitches with AI cut the variation in length and quality by about 80%1 distinct publisher
product
Eight suppliers, 53 bids, £300,000 each: Britain turned AI tutoring into a procurement line1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Peer-reviewed study, single telling
The paper is peer-reviewed, the sample is large and administratively linked, and the citation is given in full. Everything we can inspect, though, is the university's announcement as republished by phys.org: odds ratios arrive without intervals or model detail, no subgroup rate is published, and the text contradicts itself on which authorities supplied the records.
Linkage running, practice unmoved
What is demonstrably in operation is the data infrastructure: 69 mainstream schools, two local authorities, survey answers joined to suspension records for a full year group cohort pairing. Nothing in this reporting shows a school, a council or the Department for Education changing how suspensions are handled because of the findings.
Cautious body, confident frame
The researchers say plainly that these are associations and that mechanisms remain unexplained, and Humphrey's closing quote is about school capacity rather than a fix. The packaging pulls the other way: a headline about the pupils most at risk and what may help them, supported entirely by multiples of other pupils' odds, invites readers to hear prediction where the arithmetic gives about an 11% chance against a 6% baseline.
Research comms with a funding ask
The account originates with the institution that leads the programme supplying the data, alongside a children's mental health charity and a family foundation, and it ends with a professor asking government to fund more school support. That is a legitimate position argued in public, and it also explains which findings lead and which caveats sit near the bottom.
One region, one year, one account
The figures hold together and come from peer-reviewed work, so the direction of the findings is credible. Confidence stops short of high because a single announcement carries every number, the window is one academic year in one corner of southern England, and the unresolved question of which authorities contributed records has not been corrected by anyone.