The metric that hasn't moved
Birkbeck’s Centre for Neurodiversity Research at Work, in partnership with the organisation Neurodiversity in Business and sponsored by Sage, published its third annual report on neurodiversity at work in July 2026. It is a substantial piece of work — 605 participants across three strands (428 neurodivergent employees, 122 entrepreneurs, 55 employer representatives), built on validated instruments for psychological safety, wellbeing, burnout, engagement and job crafting, and co-produced with the community it studies. It is not a peer-reviewed study and it does not pretend to be; it is a practitioner report, openly framed to make the business case to employers. That framing matters, because the most telling thing in the report is what happens to that case when you read its own numbers.
The report’s own prior peer-reviewed analysis, of its 2023 data, identified psychological safety as the single strongest driver of wellbeing, career satisfaction and retention for neurodivergent employees. It is, on their account, the load-bearing variable — the thing that, if it moves, moves everything else. So the natural question for a third annual report is: after three years of awareness campaigns, manager neurodiversity training, employee resource groups and inclusion initiatives, has it moved?
It has not. The report states plainly that psychological safety has not improved in 2026, and that the likelihood of burnout is high, particularly for those with multiple neurotypes. Across the board, neurodivergent employees scored significantly worse than both employers and entrepreneurs on psychological safety, wellbeing, and engagement, and significantly higher on burnout. The one variable the field’s own research names as decisive is the one that three years of intervention has failed to shift. Some things did improve — line manager and colleague support, work-life balance, turnover intention — and those are real gains worth acknowledging. But the foundation the whole structure rests on is exactly where it was.
That is a quietly damning finding for a document whose purpose is to demonstrate that the business case is working. And it becomes sharper still when you look at where the talent goes.
The exodus in the data
The genuinely new material in this year’s report is its focus on neurodivergent people who have left conventional employment altogether — the entrepreneurs, the self-employed, the contractors. This is the first UK data of its kind across neurotypes, and it reframes everything around it. The headline numbers: nearly half of these people discovered their neurodivergence before opting out of corporate life, and 61% would not consider returning.
The detail underneath is the part that should trouble any employer reading it. The report is explicit that push factors were as strong as pull factors. People did not, in the main, leave to chase an entrepreneurial dream; a large share left because conventional employment did not fit them and showed no sign of coming to fit them — discrimination and stigma, burnout, a lack of person-environment fit, strengths that went unused. And the entrepreneurs, having left, scored significantly better than the employees who stayed: higher psychological safety, higher wellbeing, higher engagement, lower burnout. On the report’s own measures, the people who left the workplace are doing better than the people still inside it.
Set the two findings side by side and the story writes itself. The metric that determines whether neurodivergent people can thrive at work has not improved in three years. And the neurodivergent people with the means to leave are leaving, discovering their neurotype and then concluding that the environment will not change fast enough to be worth sticking around for. This is not a neurodiversity-gain story, whatever the report calls it. It is a neurodiversity-exodus story — the most able, most self-aware neurodivergent talent voting with its feet, while the organisations running inclusion programmes are the ones being voted against. The report documents an outflow and frames it as an opportunity to learn. Which it is. It is also, read straight, a verdict.
"Neurodiversity gain" and what it obscures
The report’s central coinage is “neurodiversity gain” — borrowed by analogy from “disability gain” in critical disability studies, and defined as the idea that when organisations redesign systems for neurodivergent people, everyone benefits. As an argument, this is sound and I have made versions of it myself: it is the demand-structure case, that fixing the environment rather than the person produces benefits that generalise. The trouble is not the concept. The trouble is what the concept does in a document that has just finished showing the strategy failing on its own terms.
“Neurodiversity gain” is a success frame bolted onto a set of findings about non-improvement and departure. It takes a report whose core data reads “psychological safety flat, burnout high, talent leaving” and gives it a name that says “everyone benefits.” The coinage does not describe what the study found; it describes what the field would like to be true, and supplies a hopeful vocabulary that lets the uncomfortable numbers be reframed as an opportunity rather than a warning based on what’s actually happening. This is a similar move to what I flagged in last week’s piece on affirming mental-health promotion, where a strong structural argument was undercut by an uncritical embrace of Autistic Barbie as a token of progress. The pattern is the same: the affirming frame applied without the critical faculty to notice that the underlying evidence points the other way.
It is worth being precise about why this matters, because the intention behind it is good. A field that narrates its own results in the language of success loses the ability to see when the strategy is not working; and losing that ability is expensive, because the thing that actually helps neurodivergent people — and the employers who could help them — is not a better name for the goal but an honest account of the gap between the goal and the data. “Neurodiversity gain” is a comfortable phrase. The report’s own numbers are an uncomfortable finding. I couldn’t help but see the distance between those two things. When the phrase is used to metabolise the finding, the finding stops doing its job, which was to tell the field that three years of the current approach has not moved the variable that matters.
Leaving the workplace is rational — which is the ultimate indictment
None of this means self-employment is the answer, and the report is honest enough not to claim it is. The entrepreneurs demonstrably traded one set of problems for another. Two-thirds struggle with maintaining a social-media profile; 63% with work-life balance; 57% with the financial and administrative load. Burnout did not disappear when they left — it relocated, into the invisible overwork and isolation of running everything alone. The report flags the likely presence of post-hoc rationalisation in the data, the “I opted out, so I must be doing better” effect, and that caveat should be carried. Leaving is not a cure.
But that is exactly what makes the exodus damning rather than reassuring. These people did not leave because the grass was greener. They left knowing the other side was also hard — trading psychological unsafety inside an organisation for financial precarity outside it — and still, having made that trade, most would not go back. When leaving a stable job for an uncertain one (and finding it equally as difficult, just in different ways) is still the rational choice, the thing being escaped has to be worse than the difficulty of escaping. The report’s own conclusion is that opting out is entirely rational for neurodivergent people who consistently lean into their strengths and meet fewer challenges outside the corporate structure. That sentence is not a finding about entrepreneurship, but a damning indictment of what employment still means for neurodivergent people.
So the honest reading of this report is probably not the one on its cover. Three years of the business case, and the decisive metric has not moved; the most able neurodivergent people are leaving; and their leaving is rational precisely because the environment they are leaving has not changed enough to hold them. The value of the report is not “neurodiversity gain.” It is the evidence, gathered carefully and against its own framing, that awareness and training and lanyards and celebration have not touched the foundation nor fulcrum, and that the people the whole effort is meant to help are increasingly deciding not to wait for it. The business case for neurodiversity does not need a more inspiring name or a new token. It needs us to notice that it is not, on its own terms and in its current ways, working — and to ask what would actually change psychological safety rather than what would more warmly describe its absence.
Citations
McDowall, A. & Srinivasan, A. (2026) — Neurodiversity in Business: Research Report 2026 — Neurodiversity Gain, Inclusive Practice and Self-employment — Birkbeck, University of London / Neurodiversity in Business
Edmondson, A. (1999) — Psychological Safety and Learning Behavior in Work Teams — Administrative Science Quarterly
Doyle, N. & McDowall, A. (2022) — Diamond in the rough? An empirical test of the strengths-based approach to neurodiversity at work — Equality, Diversity and Inclusion
Branicki, L. J., Brammer, S., Brosnan, M. et al. (2024) — Factors shaping the employment outcomes of neurodivergent and neurotypical people — Human Resource Management
