I'm not sure this is accurate.

Well roughly only 1% of the US population are software developers, so I think you could have a workforce only consisting of autists. Not only that, but if the autism ratio is 2:1 male and you filled all positions with autists, you'd have numbers near what Google has in reality for male/female.

Wikipedia suggests the ratio for autism is actually 4:1. So if a company had 50% autists and 50% non-autists with 50/50 balance, you'd get the ratios like Google has today.

If only 1/3rd of the workforce is autistic (4:1), then you don't need a huge bias in the rest of the hiring to get the 2:1 ratios we see in practise. Systemic bias and any other factors could easily bridge that gap.

1. "Even the most generous estimates of autism prevalence put it at 1-3% of children -- you could hire every man and woman with autism and it would have an entirely negligible effect on gender diversity in the tech workforce at large."

You need to check your math here; on the order of 1% of the US population works in software development.

2. It's called the "autism spectrum" because it's a spectrum, not a binary. People can share some autistic traits without their rising to the level of a diagnosable disorder. If a person is described as nerdy, introverted and eccentric, they probably are further along on the autistic spectrum than most, while not necessarily having full-blown autism.

> It's called the "autism spectrum" because it's a spectrum, not a binary. People can share some autistic traits without their rising to the level of a diagnosable disorder. If a person is described as nerdy, introverted and eccentric, they probably are further along on the autistic spectrum than most, while not necessarily having full-blown autism.

I thought it was called a spectrum because of the variation in severity and symptoms among people with diagnosable autistic disorders. As I understood it, if you don't have such a disorder, you're not on the spectrum.

I think you are getting into a map/territory issue.

We think of it as a spectrum because we observe a spectrum among people with diagnosable autistic disorders.

However, we do not have a fundemental understanding of what autism is, or a binary test to detect it. This would imply that in reality there is a spectrum that extends beyond the diagnosable region(s) we have.

The disorders do not cleanly cleave away from non-autistic. It really can be a judgement call at times.

Absolutely. But that's not, as I understand it, why it's called a spectrum. It's called a spectrum because it covers conditions as different as Asperger syndrome (which often isn't diagnosed until adulthood, and may never get diagnosed at all) and childhood disintegrative disorder, where the onset of symptoms is so sudden and dramatic that the child themselves may notice and ask what is happening to them.

Being "nerdy, introverted and eccentric" doesn't necessarily mean you are on the spectrum. It may do: you may have an ASD, diagnosed or undiagnosed. But it may just mean you are nerdy, introverted or eccentric.

This is a very good point. The only problem would be if people with an ASD are highly self-selecting into the technical field, thus being a significant influence.

Below is what I could find in academic journals. TL;DR it is not significant.

There are 2 studies I could find which found that:

- likelihood of autism (or being on the spectrum) is 3 to 7 times higher among maths students (Cambridge university, 1.85% vs. 0.24% in the control group)[1].

- [2]: People with an ASD (autism spectrum disorder), "have one of the lowest overall college enrollment rates", but "young adults with an ASD who attend college are most likely to pursue STEM majors". It is noted in the analytics section that those with an ASD have a 34.31% enrollment rate in STEM fields with a heavy bias towards maths and CS. This compares to >32.6% in the general population [3] (data from 2012).

So, it looks like the percentage you listed are correct, with a potential of it being quite a bit lower due to hiring practices within the industry.

[1] https://www.ncbi.nlm.nih.gov/pubmed/26181845

[2] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3620841/

[3] https://www.nsf.gov/nsb/sei/edTool/data/college-08.html

The tech industry is only 2% of the US population. [0] The US also imports people from the entire world.

[0] - https://www.comptia.org/about-us/newsroom/press-releases/201...

> And of the ones that do, the women with autism have to deal with both the barriers introduced by their autism and the barriers of sexist treatment from their peers.

With the benefits of diversity hiring.