Neurodiversity training does not change a system

Think of the manager in your organization who runs the most hiring loops. Not the one with the biggest team. The one whose requisitions close fastest, because speed is what gets noticed. Neurodiversity training reached him in the spring, the feedback scores were good, and the interview scorecard he used the following Monday was the same document it had been for three years.

Most neurodiversity training is bought to change a system and is built to change an audience. Those are different purchases, and only one of them is on the invoice.

What awareness training measurably changes

Awareness training reliably changes what people know. The evaluation literature is fairly consistent on this and fairly consistent on the limit: knowledge moves, attitudes move less, behavior moves least, and the attitudinal effects decay. Bezrukova and colleagues' meta-analytic integration of more than forty years of diversity training evaluation is the cleanest statement of that split, and the decay is the part most procurement decisions do not price.

The sharper evidence is the comparison. Kalev, Dobbin and Kelly's analysis of US private-sector establishments across three decades, published in the American Sociological Review in 2006, found the interventions that moved management diversity were structural: mentoring, diversity task forces and formal accountability. Training was not among them. Chang and colleagues ran online diversity training with a sample in the thousands and reported mixed effects, with attitudinal movement in some groups and little behavioral change.

None of that makes training worthless. It makes training a knowledge intervention being sold as a behavior intervention, which is a procurement problem rather than a moral one.

Why belief is the wrong lever

The reason is mechanical. A manager can hold a sincere belief about inclusion and run a hiring process that filters against it, and experience no contradiction between the two, because the belief and the decision are stored in different places and never consult each other.

I called this compartmentalization in my next book, out in April 2027, and the line I would use here is this one: compartmentalization keeps the training in one box and the behavior in another, and the boxes never touch. The training fills the box it was aimed at. The hiring decision happens in a different box, on a Wednesday, under time pressure, using a scorecard no one has edited since it was written.

So the shortest way to say it is this. The training reaches the people. The decisions run on the instruments, and the instruments were not in the room.

Which is why the instruments are the lever. A manager's beliefs are not available for inspection and not reliably actionable. A job description, an interview scorecard, a promotion criterion and a feedback template are four documents, they are all editable, and they govern the decisions regardless of what anyone in the room believes that week.

There is a sharper version of this on the hiring side, and it is recent. Whelpley and colleagues gave a mental ability test, a personality inventory and a customer-service situational judgment test to 107 autistic and 99 neurotypical employed adults. On the situational judgment test the gap was large, and being autistic still predicted a lower score after controlling for mental ability, personality, gender and age. That result sits inside a test, not inside a person's attitude. No training day in the world reaches it. Changing the test does.

The arithmetic, and it is not a dollar figure

Run this one with the numbers you already have.

Take the number of managers who attended, or are about to. Call it forty. One day each.

Now count the decisions those forty will make in the next twelve months. Résumé screens, phone screens, interview debriefs, offer calls, review cycles, promotion recommendations, the informal decision about who gets the visible project. For forty managers it runs to several thousand decisions.

Then count how many of those decisions pass through a written instrument that was edited this year. For most organizations the answer is a small single-digit number, and often it is one: the job description, reused.

One training day against several thousand decisions governed by documents the training never mentioned. That ratio is the whole argument, and it is calculable from an HRIS export and a calendar.

Where can I find free neurodiversity training for my workplace?

Several organizations publish free material, and the honest answer is that free training is a reasonable thing to use and a poor thing to buy against. If the goal is that managers know what the words mean, free material does that about as well as paid material does. If the goal is that next quarter's hiring produces a different outcome, neither one will, because neither one edits an instrument.

That is the uncomfortable sentence for anyone selling in this category, and it is also the one worth having before a budget cycle. The money is better spent on the four documents, and editing four documents does not require a vendor at all. It requires someone to sit down with each one and decide what it is supposed to be selecting for.

The operating evidence, named as mine

Across the companies I have run I held voluntary retention near ninety-five percent. I did not get there by changing what anyone believed. I got there by making the behaviors I did not want structurally difficult, which is a less inspiring sentence than the alternative and it is the accurate one.

What that looked like in practice: project communication visible to the team rather than routed privately, ideas attributed to the person who had them as a matter of process rather than of courtesy, and escalation paths that did not require permission from the person being escalated past. None of that is a training topic. All of it is a design decision, and each one took an afternoon to write down and years to stop needing to defend.

That is my accounting rather than a published benchmark, and I have no control group. What I can say is that the mechanism is the same one the research above describes from the outside.

Change the structures rather than the hearts is the disability community's design tradition, argued for decades by disabled people who were tired of being the subject of awareness campaigns while the buildings, the forms and the processes stayed exactly as they were. The research caught up. I am applying it to an org chart.

Run it yourself

1. Pull the four governing documents for one function: the job description template, the interview scorecard, the promotion criteria, the performance review form.
2. For each one, record two things. The date it was last edited, and whether a manager is required to complete it or merely encouraged to.
3. For the interview scorecard, mark every criterion that is a judgment about manner rather than an observation about output. Communication skills. Culture fit. Executive presence. Energy.
4. Count the marked criteria as a share of the total. That share is the proportion of your hiring decision that is unwritten, and it is where the training was supposed to land and did not.
5. Edit one document. One. Then re-run the count in two quarters.

An analyst and a hiring manager can do steps one through four in an afternoon. Step five is the one that takes a decision, and it is a decision about a document rather than about a person, which is what makes it survivable.

Four documents, and a count of how many are written and current. Most organizations have never produced that number, and it is the one a budget meeting can act on, because a document has an owner and an edit date and a belief has neither.

How many of your hiring and review instruments exist as a written document, and how many live in a manager's judgment on the day?

---

Forward this to whoever owns your L&D budget before the next cycle closes. The companion piece on what a review is recording when it reaches for a personality word is here, and the three criteria hiding inside the phrase culture fit are unpacked here.

Sources

• Bezrukova, K., Spell, C. S., Perry, J. L., & Jehn, K. A. (2016). A Meta-Analytical Integration of Over 40 Years of Research on Diversity Training Evaluation. Psychological Bulletin, 142(11), 1227–1274. Knowledge gains are the most reliable outcome; attitudinal effects decay over time. Psychological Bulletin, A Meta-Analytical Integration of Over 40 Years of Research on Diversity Training Evaluation
• Kalev, A., Dobbin, F., & Kelly, E. (2006). Best Practices or Best Guesses? Assessing the Efficacy of Corporate Affirmative Action and Diversity Policies. American Sociological Review, 71(4), 589–617. Structural interventions moved management diversity; training did not. American Sociological Review, Best Practices or Best Guesses?
• Dobbin, F., & Kalev, A. (2016). Why Diversity Programs Fail. Harvard Business Review, 94(7–8), 52–60. The authors' own synthesis of the above for a management readership. Harvard Business Review, Why Diversity Programs Fail
• Chang, E. H., Milkman, K. L., Gromet, D. M., Rebele, R. W., Massey, C., Duckworth, A. L., & Grant, A. M. (2019). The Mixed Effects of Online Diversity Training. Proceedings of the National Academy of Sciences, 116(16), 7778–7783. Attitudinal movement in some groups; limited behavioral change. Proceedings of the National Academy of Sciences, The Mixed Effects of Online Diversity Training
• Whelpley, C. E., Fezzey, T., Batchelor, J. H., & Weekley, J. A. (2026). Using selection tests to hire autistic employees: An exploration of methods. PLoS One, 21(8), e0356539. 107 autistic and 99 neurotypical employed adults; being autistic predicted a lower situational judgment test score after controlling for mental ability, personality, gender and age. Open access, data deposited. PLoS One, Using selection tests to hire autistic employees
• Mann, P. A. (2026). The Perception Revolution. Fast Company Press. Chapter 10 on designing environments where the behavior is structurally impossible, and Chapter 13 on SHIELD.
• Mann, P. A. (April 2027). The Projection Fallacy (manuscript). Chapter 2, on compartmentalization. Quoted line marked for checking against the final text.

Peter Allen Mann is the author of The Perception Revolution (Fast Company Press, 2026). He spent thirty-seven years in leadership, including four years as a Navy Surface Warfare Officer and two companies founded and run. He writes about late-recognized autism and perception at peterallenmann.com.