This week has made a case grounded in principle. Today it is worth asking what the available evidence adds to that argument. Across employment, entrepreneurship, collective problem-solving and accessible design, the findings point in a similar direction: organisations often overlook capable people because their systems recognise only a narrow range of backgrounds, behaviours and ways of contributing.

The evidence does not support the claim that every diverse group will outperform every less diverse one, or that every adjustment will benefit everyone equally. It does show that exclusion can conceal valuable capability—and that changing the conditions for participation can produce benefits extending well beyond the people for whom those changes were first made.

Autistic people and employment

Autistic adults experience substantial barriers to employment. UK data from the Office for National Statistics found that, in the year ending June 2020, 21.7 per cent of disabled working-age adults who identified autism as their main impairment were in employment. This was one of the lowest employment rates recorded among the impairment groups in the analysis.

That figure describes employment status, not ability. It also does not tell us that workplace design is the only cause of the gap. Autistic people are a diverse population, and employment outcomes are affected by many factors. Nevertheless, recruitment practices, inaccessible working environments, inadequate support and expectations based on neurotypical communication can all prevent capable candidates from demonstrating what they can do.

Employer programmes offer useful examples of what can happen when some of these barriers are reconsidered. JPMorgan Chase’s Autism at Work programme replaced parts of the conventional recruitment process with approaches designed to recognise candidates’ abilities more effectively. SAP launched its Autism at Work programme in 2013. In 2024, the company reported a 90 per cent retention rate among colleagues hired through the programme; its neuroinclusion hub reports that the initiative has reached more than 240 colleagues across 16 countries.

These company figures should not be treated as controlled scientific comparisons, and they do not prove that autistic employees will be more productive than non-autistic employees. They demonstrate something more grounded: when organisations change how they recruit and support people, they can gain access to talent their conventional processes may have missed.

What immigrant entrepreneurship tells us

The evidence for immigrant contribution to entrepreneurship is substantial, although patterns vary between countries, migrant groups, industries and periods. The clearest figure in this context comes from the American Immigration Council’s analysis of the 2024 Fortune 500. It found that 230 companies—46 per cent of the list—had been founded by immigrants or their children.

This does not prove that an outsider perspective caused those companies to succeed. Nor does it mean that immigrants are uniformly more entrepreneurial than people born in their destination country. Migration is shaped by selection, opportunity, necessity, education, policy and access to capital. What the figure does establish is that immigrants and their children have played a major role in creating some of the largest companies in the United States.

UK evidence is more mixed than a simple claim that immigrants are always more likely to start businesses would suggest. Historical data has found higher self-employment among some migrant groups and lower rates among others, with large differences by nationality and sector. The responsible conclusion is not that every newcomer is an innovator. It is that systems which filter people according to familiarity, accent, background or cultural fluency risk overlooking experience and entrepreneurial capacity that already exist within the population.

What cognitive diversity can contribute

Scott Page’s work helps explain why different perspectives may improve collective problem-solving. In The Difference, Page argues that groups benefit when their members bring different ways of interpreting problems and searching for solutions.

A related mathematical result developed by Lu Hong and Scott Page showed that, under particular modelled conditions, a diverse group of problem-solvers can outperform a group selected only for individual ability. The conditions matter. The result is not a universal law saying that demographic diversity automatically produces better decisions. Participants need relevant knowledge, genuinely different approaches and a process that allows those approaches to influence the outcome.

The practical lesson is therefore more precise than “diversity always wins.” Expertise remains important. But selecting people according to one shared definition of excellence can create a group whose members approach the problem in much the same way. A person who frames the problem differently may contribute a route to a solution that a more cognitively similar group would never consider.

Who benefits from captions?

Captions offer a particularly visible example of an accessibility measure serving a wider audience. Captions remain essential for many deaf and hard-of-hearing people. They are also used by people watching without sound, dealing with unclear audio, learning a language, following an unfamiliar accent or finding that written text supports their concentration and comprehension.

A 2019 online survey commissioned by Verizon Media and Publicis Media questioned 5,616 US adults aged 18 to 54. It found that 80 per cent of respondents who used captions did not identify as deaf or hard of hearing. Because this was a US consumer survey focused partly on online video advertising, it should not be presented as a measure of all caption users everywhere. It nevertheless provides strong evidence that captions have uses extending beyond their original accessibility purpose.

Ofcom’s research into access services provides further insight into why audiences value subtitles, including difficulties with background noise, unclear speech and unfamiliar accents. Together, these findings illustrate a broader design principle: a feature created to remove a specific barrier can also make participation easier for many other people.

What the evidence allows us to say

None of these findings means that inclusion produces an automatic performance advantage. Context, implementation and power all matter. Inviting different people into a room achieves little if the process still prevents them from influencing what happens there.

The evidence supports a more careful and useful conclusion. Conventional systems do not identify talent perfectly. Homogeneous groups can share blind spots. Accessible design can benefit people beyond its intended audience. When organisations widen the ways people can enter, participate and contribute, they improve their chances of recognising capabilities that were previously hidden.

Questions for reflection

  • If your organisation examined who is filtered out during recruitment, what capabilities might it discover that its current process does not recognise?
  • What does the use of captions by wider audiences suggest about adjustments you currently assume benefit only one person or group?
  • Where are the highest-performing people in your organisation also the most cognitively similar? What blind spots might that similarity create?
  • Whose experience or perspective is present in your community but absent from the places where consequential decisions are made?
  • What would it mean to treat this evidence as a prompt for experimentation rather than as proof of a predetermined conclusion?

This post draws on Office for National Statistics employment data, published information from the JPMorgan Chase and SAP neuroinclusion programmes, the American Immigration Council’s New American Fortune 500 in 2024 analysis, Scott E. Page’s The Difference, the Hong and Page problem-solving model, and research into caption use reported by the Advertising Research Foundation and Ofcom.