Neurodiversity In The Workplace: What 111 Studies Refuse To Ask

The research on neurodiversity in the workplace has a blind spot, and it is not a small one. A systematic review published in 2026 screened more than 13,000 articles, narrowed them to 111 that met its criteria, and mapped the entire body of work onto the stages of an autistic person’s working life (Checketts, 2026). The review found detailed research on hiring. It found detailed research on accommodations. On the question of whether autistic workers ever get promoted, it found nothing. Not a thin literature. Zero studies. The same was true of career path.

That gap is the subject of this article. Read alongside a second 2026 study of adults with ADHD working from home (Yaukey et al., 2026), a clear picture emerges of what the evidence establishes, what it leaves out, and where those of us who are AuDHD disappear from the record entirely. I am a late diagnosed AuDHD researcher and author, and I have read both papers in full. Where a figure is solid, I will say so. Where it rests on a secondary source I have not verified, I will say that too.

What does the research on neurodiversity in the workplace actually cover?

It covers the beginning of a working life in depth and the rest of it barely at all.

Checketts (2026) organised the 111 included studies around a five-stage model of the autistic employee life cycle: hiring, developing, supporting, managing, and advancing. The shape of the evidence across those five stages is the single most useful finding in the review, because it shows where knowledge exists and where it runs out.

Hiring is well researched. Studies cover recruitment, interviewing, and selection, and there is a substantial literature on how job advertisements, interview formats, and selection processes disadvantage autistic candidates (Checketts, 2026). Supporting is also well covered, particularly accommodations, relationships, and wellbeing.

Then the evidence thins. Developing has some coverage of training and coaching, but Checketts (2026) found only a single study on mentoring autistic workers. Managing is thinner still, with only one study on performance management. And advancing, the stage that covers promotion, leadership, and career path, is close to empty. Checketts (2026) reported just two studies on autistic leaders, and zero on promotion and zero on career path.

This matters because the employee life cycle is not front-loaded in real life. People spend far longer being managed and trying to advance than they spend being hired. The research inverts that. It knows a great deal about how neurodivergent people get through the door and almost nothing about what happens once they are inside.

There is a second structural point worth naming. More than half of the research on this topic has appeared in the last five years (Checketts, 2026). This is a young field, which partly explains the gaps. It also means the conventions being set now, including which questions count as worth asking, will shape the field for years. The absence of promotion research is not a permanent fact. It is a choice the field is currently making.

For organisationsIf your neuroinclusion strategy stops at hiring, you are mirroring the exact bias in the research. Recruitment campaigns and interview adjustments are the visible, fundable end of inclusion. Retention, progression, and promotion are where autistic and ADHD employees actually leave. Audit your own data. If you can tell me your neurodivergent hiring rate but not your neurodivergent promotion rate, you have the same blind spot the literature does.

Why is there zero research on whether autistic people get promoted?

Because the field has been built around getting autistic people into work, not moving them through it.

Checketts (2026) is explicit about this. The review process surfaced no studies on the promotion process, the likelihood of promotion, or even autistic workers’ desires for promotion. The same silence covers career path, including decisions about choosing an industry, changing careers, or continuing education.

The reason is partly historical. Autism employment research grew out of a concern with unemployment and underemployment, which are real and serious. Checketts (2026) reports figures drawn from earlier studies suggesting that a large majority of autistic graduates struggle to find work and that neurodivergent unemployment globally sits between 60 and 90 percent. I want to be careful here, because these specific numbers are secondary citations within the review rather than figures I have verified against the original studies, and unemployment statistics in this area vary widely by definition and sample. Treat them as indicative of a documented problem rather than as precise measurements.

The point stands regardless of the exact percentages. When a field forms around the problem of getting people employed at all, promotion looks like a luxury question. You cannot study the promotion of people who are not yet hired.

But that framing has a cost. It means the research has almost nothing to say to the autistic person who is employed, competent, and stuck. It cannot tell them whether the ceiling they are hitting is common, whether it has a name, or what moves it. Checketts (2026) names this directly as a gap requiring urgent attention, noting that the areas of promotion and career path remain entirely unexplored.

There is one adjacent finding worth holding onto. Of the two studies on autistic leaders that Checketts (2026) did find, both took a holistic view, treating autistic traits as simultaneously strengths and challenges in a leadership role. One found that traits such as task orientation, logical thinking, and attention to detail served as leadership strengths, while cognitive rigidity and difficulties with perspective taking served as challenges (Checketts, 2026). It is a small evidence base, but it is the closest the field comes to studying autistic people at the top of an organisation rather than the entrance.

For organisationsYou almost certainly cannot benchmark your neurodivergent progression against published research, because it does not exist. That is not a reason to wait. It is a reason to generate the data internally. Track progression rates for employees who have disclosed, hold them against your general population, and treat any gap as a finding. You will be ahead of the academic literature, which is a strange thing to be able to say about promotion in 2026.

Do workplace accommodations actually work, or do they burden everyone else?

They work, and the evidence that they impose no cost on colleagues is stronger than most managers assume.

This is the finding I would most like every HR leader to absorb, because the objection to accommodations is almost always some version of fairness. If one person gets noise cancelling headphones, flexible hours, or written instructions, does that disadvantage everyone else or create resentment?

The research says no. Checketts (2026) reports that accommodations addressing autistic challenges are associated with appropriate levels of employment, with significant organisational contributions, with lower turnover intentions, and with overall positive impact for autistic employees, all without negative impact on other employees or the organisation. The accommodation helps the person who needs it and costs the people around them nothing measurable.

That reframes the entire conversation. An accommodation is not a favour granted at someone else’s expense. On the available evidence it is a low-cost, high-return adjustment with no documented downside for the team.

There is a catch, and it is a significant one. Checketts (2026) notes that despite these positive outcomes, the onus still lies on the autistic employee to identify their needs and make the request, rather than the organisation leading the process. So the system works, but only for the person willing and able to self-advocate, disclose, and ask. Everyone else gets nothing, not because accommodations fail, but because the request never happens.

This connects to a wider pattern in how neurodivergent people cope at work, which the second paper documents in unusual detail.

For organisationsThe fairness objection to accommodations is not supported by the evidence. Accommodations improve retention and comfort with no measured negative impact on colleagues (Checketts, 2026). The real inequity is that accommodations depend on the individual knowing their rights, disclosing, and asking. Build a standard, low-friction process for requesting adjustments that does not require a diagnosis or a difficult conversation with a line manager, and you remove the barrier that the research identifies as the actual problem.

What do people with ADHD actually do to cope with remote work?

Overwhelmingly, they build private workarounds that help them appear more neurotypical, and they often pay for those workarounds themselves.

The second study takes a completely different route to the same territory. Yaukey et al. (2026) analysed 114 posts and 3,276 comments from the ADHD subreddit, coding every strategy that adults described using to stay focused while working from home. Before I summarise it, a caution about its weight: this is a short conference paper built on self-reported Reddit data, and the authors themselves note that the sample may include people who are self-diagnosed or misdiagnosed. It cannot tell you how common anything is. What it can do is show the texture of how people cope, and the pattern in that texture is striking.

The strategies clustered into three themes: blocking out distractions, managing time, and staying on track (Yaukey et al., 2026). Within those, the specific tools people reached for tell the story. Body doubling, the practice of working alongside another person in real life or on a screen to stay accountable, was the single most mentioned strategy at 64 mentions. Noise cancelling headphones came next at 60. Alarms and timers at 45. The Pomodoro technique at 24.

Now look at the bottom of that list. Developing automation tools, meaning people who taught themselves scripting, macros, or workflow automation to genuinely redesign how their job worked, appeared just three times (Yaukey et al., 2026).

That distribution is the finding. Almost everything people described was compensation, a way of managing a deficit to meet a neurotypical standard. Only a handful of people described using their ADHD-associated hyperfocus as a genuine strength to build something new. The authors conclude that nearly all the interventions are deficit-based approaches that help people mask at work (Yaukey et al., 2026).

What is the hidden cost of not disclosing at work?

It is often literal and financial, and almost nobody has counted it.

Buried in the Yaukey et al. (2026) analysis is a point that deserves far more attention than a five-page paper can give it. The headphones, the multiple monitors, the subscription apps, the body doubling services: someone pays for all of it. The authors observe that it is frequently unclear who, and that where an employee has not disclosed their condition and asked for accommodations, they are almost certainly paying themselves.

So non-disclosure is not free. It carries a running cost that lands on the individual precisely because they have chosen, or felt forced, to stay invisible. Yaukey et al. (2026) frame this as a potential inequity, where staying productive places a financial burden on the people already managing the most friction.

This links directly back to the accommodations finding. Checketts (2026) showed that formal accommodations work and cost colleagues nothing. Yaukey et al. (2026) show what happens in the absence of that formal route: people build their own accommodations privately and fund them out of pocket. The formal system and the informal one produce similar tools. The difference is who pays and whether anyone has to mask.

For organisationsThere is a cost to your non-disclosure rate that does not appear on any budget line. Employees who have not disclosed are self-funding the tools that keep them productive. If you want a concrete action, examine what you provide by default versus by request. Noise management, flexible scheduling, and focus tools offered to everyone, no diagnosis required, remove both the financial penalty and the disclosure barrier in one move.

Is a cameras-on policy a masking mandate?

For a meaningful number of neurodivergent employees, yes.

One detail from Yaukey et al. (2026) has stayed with me since I read it. Among the people describing how they manage video meetings, some kept cameras on because the social pressure helped them stay focused. But others turned cameras off specifically because it reduced the demand to mask and let them do something with their hands while staying engaged. One participant said the only time they had to mask was while on screen in a meeting. A few went further and built video loops of themselves, or used AI-generated video, to satisfy a cameras-on requirement.

Consider what that means. A cameras-on policy, usually framed as being about engagement or presence, functions for some neurodivergent employees as a requirement to perform a neurotypical face for the duration of every meeting. People are faking their own image to comply with it. I want to be precise about the evidence: this rests on a small number of self-reported comments, not a representative survey, so it describes a real experience rather than a measured prevalence. But the experience is real, and it is worth any manager pausing over.

For organisationsReconsider blanket cameras-on policies. For some neurodivergent staff, mandatory video is not a neutral engagement tool but a sustained masking demand that drains the cognitive resource they need for the actual work. Making cameras optional costs nothing and removes a hidden tax on your neurodivergent employees. If presence matters to you, find a measure of it that does not require people to perform a face.

Where do the two studies agree?

They agree on the conclusion, despite sharing no method, no neurotype, and no sample.

This is the point of reading them together. Checketts (2026) conducted a formal systematic review of 111 peer-reviewed studies on autistic workers and concluded that the system places the burden of adjustment on the employee. Yaukey et al. (2026) analysed thousands of informal Reddit comments from people with ADHD and concluded that nearly every strategy people use is self-funded masking to appear neurotypical.

One is autism, the other ADHD. One is peer-reviewed journal research, the other is a digital ethnography of a subreddit. One counts published studies, the other counts forum posts. By every methodological measure they are unrelated. And they arrive at the same place: the neurodivergent individual carries the load, the organisation largely does not, and the coping is mostly invisible.

When two independent methods converge on one conclusion, that convergence is itself evidence. It suggests the finding is not an artefact of how either study was designed. The burden really does sit with the individual, across neurotypes and across research traditions.

For organisationsThe consistency across these two very different studies is your signal that this is structural, not incidental. You cannot fix it with a single policy in a single team. The pattern is that neurodivergent employees adapt to the workplace while the workplace rarely adapts to them. Shifting even part of that load, through default accommodations, clear disclosure pathways, and progression tracking, addresses a problem that shows up whether your employee is autistic, ADHD, or both.

Why is AuDHD missing from both studies?

Because both studies, by their own methods, exclude it, and the exclusion is documented in the papers themselves.

This is where I have to speak from my own position. As a late diagnosed AuDHD adult, I have spent years reading research that felt like it was describing someone standing slightly to the left of me. These two papers finally showed me why.

Checketts (2026) restricted the review to autism and, in the screening criteria, explicitly excluded studies that addressed multiple types of neurodiversity. So the review that built the definitive model of the autistic employee life cycle deliberately excluded any study looking at autism and ADHD together. It also limited itself to Level 1 autism, excluding autistic people with higher support needs.

Yaukey et al. (2026) sampled the ADHD subreddit. That population certainly contains AuDHD people, since the co-occurrence of autism and ADHD is common. But the study did not identify, separate, or measure them. They are present in the data and absent from the findings.

So AuDHD workers are not simply under-researched. They fall through a methodological gap between the two dominant research approaches, and the exclusion is written into the methods sections. One paper screens out combined-neurotype studies on purpose; the other never distinguishes the combined group at all.

To be fair to Checketts (2026), the review names this limitation and points to it as a direction for future work, noting that a holistic perspective could be extended to other neurotypes including ADHD. The door is identified. Nobody has yet walked through it. For a growing population of late diagnosed AuDHD adults, that means the research base describing our working lives does not yet exist in a form that includes us.

For organisationsDo not assume your autism guidance and your ADHD guidance, stapled together, cover your AuDHD employees. The combination is not the sum of its parts, and the research base treats it as neither. In practice this means listening to individual employees about their specific profile rather than applying a template. The evidence for AuDHD-specific workplace needs is not there yet, which makes direct, respectful conversation the only reliable source you have.

What should neurodivergent employees and their managers actually do?

Act on what the evidence does establish, and stop waiting for the research it does not.

For neurodivergent employees, the most useful thing these papers offer is language and permission. If you are employed, competent, and stuck, the research on why is genuinely missing (Checketts, 2026), which means the ceiling you may be hitting is not a personal failing you have simply failed to read about. It is an unstudied gap. And you can bring the accommodations evidence into any conversation with confidence: adjustments improve retention and comfort with no measured cost to colleagues (Checketts, 2026). That is a fact, not a favour you are begging for.

Cost your own workarounds honestly. If you are self-funding the headphones, apps, and tools that keep you working out of your own pocket, add up what non-disclosure is costing you (Yaukey et al., 2026) before you assume that staying invisible is the cheaper option. Sometimes it is. Sometimes it is not, and only you can do that maths.

For managers and organisations, the evidence points in a consistent direction. Train your leaders rather than waiting for individual requests. Checketts (2026) reports that organisations actively training top managers and leaders about neurodiversity saw higher retention of autistic employees. Provide focus tools, scheduling flexibility, and clear adjustment processes by default, so that support does not depend on disclosure and self-advocacy. And measure progression, not just hiring, because that is precisely the data the published field is missing.

None of this requires waiting for the research to catch up. The two studies at the heart of this article agree that the load currently sits with the individual. Moving even part of it is something any team can start on now.

For organisationsYour action list from the current evidence is short and specific. First, make core accommodations universal and diagnosis-free. Second, build a low-friction adjustment request process that does not route through a difficult disclosure conversation. Third, train leadership on neurodiversity, which is associated with better retention (Checketts, 2026). Fourth, track and report neurodivergent progression the way you track hiring. You will be operating ahead of the published literature, and given what that literature currently ignores, that is exactly where you want to be.

The bottom line on neurodiversity in the workplace

The research on neurodiversity in the workplace is real, growing, and lopsided. It studies how neurodivergent people enter work and largely ignores how they move through it. Across a formal review of 111 autism studies and an informal analysis of thousands of ADHD comments, the same conclusion holds: the neurodivergent individual adapts, usually invisibly and often at their own expense, while the organisation rarely adapts in return (Checketts, 2026; Yaukey et al., 2026).

The accommodations that would shift that load work, and cost colleagues nothing measurable. The barrier is not effectiveness but access, because support still depends on the individual disclosing and asking. And for AuDHD workers specifically, the research does not yet describe our working lives in a form that includes us, because both dominant approaches exclude the combination by design.

If you take one thing from this, let it be the promotion gap. A whole field of research on autism at work, and not a single study on whether autistic people get promoted (Checketts, 2026). The field studies our entrance and ignores our future. That is a fixable choice, and naming it is the first step to changing it.

Nicola Knobel is a late diagnosed AuDHD researcher, author, and speaker based in Aotearoa New Zealand. This article draws on two open access 2026 studies, both linked below, and distinguishes throughout between verified findings and secondary figures that require further checking.

References

Checketts, M. B. (2026). A systematic review of autism in the workplace: Taking a holistic perspective on the Autistic employee life cycle. Applied Psychology, 75(4), e70116. https://doi.org/10.1111/apps.70116

Yaukey, S., Komlodi, A., & Kuber, R. (2026). From time blocking to body doubling: Analyzing technologies employed by adults with ADHD for working from home. Adjunct Proceedings of the 5th Annual Symposium on Human-Computer Interaction for Work (CHIWORK Adjunct ’26). Association for Computing Machinery. https://doi.org/10.1145/3805029.3818290