John Stanley Ford, my father, was the first black software engineer in America, hired by IBM in 1946. Passed over for promotions, discriminated against in pay, with many inside IBM working to ensure his failure, he still viewed his job as an opportunity of a lifetime. He refused to give up.
Minority underrepresentation in high tech has been present since the earliest days of the industry. In reflecting upon my father’s career for a new memoir I wrote about him, I saw important lessons about the history and nature of racism in high tech, and about the steps that corporations and individuals can take to bring about much-needed change.
An important and fascinating story – especially since it involves IBM, a company with a long and deep roots in racism, eugenics, and genocide.
Thank you. I appreciate this story. This is a chapter of history I didn’t know.
Looks more like a person with resentment trying to justify his bad experience in work and life. He tries to connect disconnected events to put in the mind of the reader that IBM is a racist company. It was just missing something like “OS/2 was created by IBM to kill all Latinos”
martini,
He’s writing about his father. Idiot.
And they’re not “disconnected events”. They’re connected by the fact that IBM was involved in all of them.
Bloody hell you’re an idiot.
Didn’t have the time yet to read the entire article.
But it’s definitely off to great fals start with:
“The roots of racism in high tech coincide with the advent of the digital age,”
and:
“Eugenics is a pseudoscience that seeks to create a “racially pure” master human race by eliminating those deemed inferior.”
Having worked there years ago, I saw none of that. It was highly racially diverse both of technical and non-technical staff. I can’t speak to decades ago, but I can speak to the early 2000’s. I’d take such an article with a grain of salt.
cacheline,
There’s a similar point to be made about the underrepresentation of females. I’ve had this discussion before and I’ve even been criticized for ignoring discrimination in the tech field, particularly regarding sexism, by people who don’t even work in tech. It’s not that I want to dismiss the harmful effects of discrimination, however when I hear that low representation is evidence a priori of discrimination, I cringe. There is clearly a very high correlation between tech and a predominantly male workforce, but this doesn’t imply a sexist causation. The simplest explanation is that the cause of under-representation (that I’ve witnessed many, many, times) is that few underrepresented groups are applying.
Hypothetically if 5% of qualified applicants are female, and 5% of employees are female, it’s illogical to conclude that the employer discriminated just based on under representation. This is why affirmative action programs are morally controversial. If 5% of qualified applicants are female, and 50% of employees are female, then that might legitimately be seen as reverse discrimination in and of itself.
There were very few females in the CS program at my university, would there have been more if they hadn’t been discriminated against? Or are they underrepresented because they didn’t want to go into CS? Keep in mind that females are actually over-represented at universities overall, so the fact that they are so under-represented in tech seems to strongly suggest they don’t want to be in tech. The question becomes why don’t they show an interest? I have a theory as to why this may be: autism. It makes sense for people with functioning autistic characteristics (all degrees, not just clinically diagnosed) to avoid socializing and not mind so much working on computers for long hours at a time, which is what we do in CS. Those who aren’t at least slightly autistic are quite naturally going to have a harder time in this field. Now here’s the thing, males are 3-4 times more likely to be diagnosed autistic in the general population,
https://www.statista.com/statistics/676697/lifetime-prevelance-of-autism-spectrum-disorder-in-children-in-us-by-gender/
I don’t know if anyone’s done a specific study about this, but I’d expect tech fields to have a significant population of functioning autistic individuals for both sexes, but most of those would naturally be male.
Alfman,
It seems more of a trend to blame lots of things on autism. Autism is not common, and autism does not grant everyone with it with higher intelligence. So even with above average rates in the tech community, it still stands to fact that most males in tech are not autistic. So why isn’t the lack of autism keeping those males out tech, if gender makes no difference at all? Why is there a lack of autistic women in tech (hint: we know that women with autism learn how to act in social situations, so what does it say about what they learn to act like)?
The bad arguments always seems to assume there needs to be some large organized effort to make women not choose tech, but all it really takes is one or two incidents every once in a while. That is more than enough to dwindle the numbers and cause them not to be replenished. People feel comfortable when there are more of them around and when existing numbers dwindle, so do new entrants.
It’s a very simple social dynamic, that happens to any people in some group. Hell, it even happens to jobs and job titles.
kwan_e,
I don’t believe autism has to be a discrete binary characteristic (ie you either have it or you don’t). So I kind of take issue with “it still stands to fact that most males in tech are not autistic” on the basis that it’s not an on/off thing but rather a question of “how much”. And in these analog terms, it seems plausible if not probable that on the whole a majority of people in fields like CS will have above average degrees of autism. Granted, I don’t have any research/facts that credibly proves or dismisses this theory, so if you have more information then I’d be very interested in reading it.
I concede the fact that there will be incidents, but is there clear evidence that more incidents occur in tech companies than with other professions?
Women are underrepresented here on osnews too. Do you think that’s due to discrimination, or are they just uninterested? I’d like to get a female perspective, but I’m not even sure if we have a single female here to discuss this.
I don’t know if it’s true or not, but it sounds plausible: some may choose not enter professions that are male dominated because they are male dominated. Assuming this is a significant factor, this kind of positive feedback loop means a balanced workforce is unlikely to happen naturally.
Alfman,
Nothing I’ve said requires autism to be a discrete binary. Autism is diagnosed, and you are either diagnosed, or you’re not, but that doesn’t mean all autism is the same. But there are people with no autism. And they do significantly outnumber people with autism of all severity.
That’s physically not possible. The number of people diagnosed with autism is in a minority. Then, a significant percentage of people with diagnosed autism have learning disabilities. There are too many male people working in tech to have more than half of them to have autism of any severity, let alone the minority of autistic people with higher than average intelligence.
Trying to attribute any social phenomena to autism is the psychological equivalent of quantum woo, frankly. They’re both not well understood, but portrayed by pop culture to be magical, and so gets shoved into any attempt at explanation.
No more than there is “clear evidence” that all the males in the profession get by on merit, yet people somehow seem to take it as a given that the tech profession is all merit based.
Why is OSAlert somehow singled out as requiring a special explanation? There are fewer females in tech and coming into tech. Therefore, tech websites will get fewer regular female members. You just basically prove my point – it doesn’t take discrimination in every specific category to have an effect. Just decrease the number of females in tech in general and you see the effect.
Yes. And it is profoundly irrational for people to argue that because it’s “natural”, that means it’s the natural order of things.
kwan_e,
You’re still treating it as an on/off characteristic, I’m saying it’s likely more of a continuous spectrum with no concrete mileposts. Even the autistic “tests” recognize this gradient. What we diagnose as autism is really more extreme traits that every one of us has to some extent. To claim some person is 0% autistic would imply that nobody in existence could ever show up even lower on the autistic spectrum, but given infinite possibilities and enough time, people, and precision, you’ll always find someone to beat the previous 0% candidate. The same should be true of 100% autism. In practice our tests may only concern themselves with 1 part per hundred let’s say, but in reality there could be as many buckets as there are people given enough precision.
So again, to me it’s not about whether someone is autistic or not, but where do they fit on the spectrum and what does the distribution look like? Are there clumps towards the low & high ends? Is it a bell curve? Is the distribution linear? IMHO it would be extremely interesting to have these data points. It would obviously convey far more information than the basic “autistic or not autistic” stats we’re familiar with.
This prompted me to search for existing research on the topic…
https://www.telegraph.co.uk/news/science/science-news/11973110/Scientists-and-mathematicians-test-higher-on-autism-spectrum-says-Cambridge-University.html
It’s the exact same idea that I’ve been trying to explain, haha!
No doubt, there’s just as much “it’s not what you know, it’s who you know” in tech as any other profession.
In a way though you’ve made my original point that we see a lot of underrepresentation even without discrimination. The reason this is important to acknowledge is because as a matter of policy (corporate, government, etc) programs aimed at curbing discrimination are clearly NOT going to fix the underrepresentation. Not that we shouldn’t do that anyways, but if you want to address underrepresented groups, then it’s going to require more than just weeding out the bad actors. You’ll need to do something to make the work more appealing to the underrepresented groups, but even so they might not be as interested for the reasons we’ve been talking about.
Alfman,
No I’m not. Otherwise, everyone is autistic. Which makes your argument pointless. If everyone is autistic, then autism CANNOT be an explanation. You may as well be trying to argue that having Neanderthal genes has an effect on people going into tech.
Sorry but that article is FAR FAR from “the majority of people in tech have autism” that you claim. Scientists and mathematicians may have higher rates of autistic traits, but that is far from the “majority has autism”.
And it DOES NOT in any way show that people without “autism” are not suited for tech jobs. Not to mention, science and mathematics are not the only jobs in tech. Hell, if the internet is any indication, people in tech jobs seem to be poor at mathematical and scientific thinking.
Autism is a spectrum. At some point, someone is so low on the spectrum that we can use the shortcut phrase “not autistic” to describe them.
You’re just being pedantic (hyperliteral). I thought it would be obvious, but it seems I have to spell it out because you can’t figure that out on your own.
Correlation is not causation.
And even that article makes a point to differentiate between autistic traits and having autism.
No, you keep using a mental image of discrimination that is like some movie version of discrimination. There is discrimination, but it doesn’t have to be targeted to specific websites, or be in your face, for it to have an effect. Stop basing your world view on cartoons.
The wider point here is that INDIVIDUALS should not be treated as the “underrepresented group” as a whole. Population statistics have no bearing on individuals.
kwan_e,
Not everyone is clinically autistic, but the point is that everyone can be plotted somewhere on the continuous spectrum whether or not they are autistic. You keep speaking as though the spectrum must stop abruptly and everyone who is not considered clinically autistic has exactly zero autistic traits, but that’s not really true. I get that for classification purposes, we assign some threshold to the right of which one is said to be ‘autistic’, but that clearly does not mean the spectrum cannot continue to the left of that same threshold, because logically it can and does.
Please don’t misquote me since the difference is very significant. Here is what I actually said “it’s not an on/off thing but rather a question of ‘how much’. And in these analog terms, it seems plausible if not probable that on the whole a majority of people in fields like CS will have above average degrees of autism.” If it wasn’t clear, then let me clarify by adding that I meant above average across the whole population on the entirety of the spectrum, not above average for people who are on the right side of the threshold used to consider someone autistic.
The way I see it, this lower part of the spectrum is so interesting exactly because it covers most of the population (who are not autistic as you point out). Furthermore based on what we know we can predict that males on average will be higher on the spectrum than females (again even if they are not clinically autistic). And no this isn’t about being pedantic, it goes goes to the point of trying to understand the causes of low female numbers in fields like CS. This very well might be a factor.
We’re both smart people, I welcome your feedback and insight, but there’s no reason to talk down to me. It does not advance the conversation.. Let’s keep it light-hearted and fun
Alfman,
No I don’t. You can have the most gradual spectrum, but AT SOME POINT, that person is not diagnosed as being autistic.
What you’re trying to argue is like saying if I have a thousand dollars, that means I’m technically a millionaire because I’m on the spectrum of having money. As far as I can tell, there’s a gradual spectrum from a thousand dollars to millions of dollars, but AT SOME POINT, I am not a millionaire.
Then by your very own causal argument, we should see more females in tech than we do, because there are more females higher than average on the whole spectra of traits than there are clinically diagnosed autistic females. It’s quite meaningless to talk about the whole spectrum, because at some point, those “autistic traits” blend in with other psychological tendencies, like introversion, or obsessive-compulsiveness. When you set the bar so low, you can make any idea work. It’s quantum woo.
You also have still yet to consider why aren’t people who aren’t above average on a spectrum are not being dissuaded from tech. The fact that the average is only very slightly higher, means there’s actually not that much difference at all, and that there must be a lot of people below the average, of both tech, and the general average. Otherwise the average should be much higher.
That doesn’t mean jack, because not you, and no one, has shown that there’s any link between being on a spectrum, and being good for the job, or being interested in the subject. Correlation is not causation.
What you were being pedantic about was your hyperliteral reading of “not autistic” to mean “0 percent autistic traits”, just so you could continue to:
If that’s me talking down to you, then you’ve been talking down to me almost the whole time, by making it seem like I think autism is an on/off thing.
To me, the portrayal of discrimination that you’re making is a cartoonish version of it. You’re treating discrimination more as an on/off thing than I’m supposedly treating autism as an on/off thing.
kwan_e,
That point is around 3% of the population, however you’ve still got ~97% of the population who are not diagnosed as being autistic yet can still be plotted somewhere on the continuous spectrum of traits associates with autism. Simply plotting someone’s position on a spectrum of autistic traits does not make them autistic! But it is nevertheless possible to do this with everyone in the population even though they don’t pass the threshold.
Put another way, if we break up the spectrum into 1000 buckets, the top 35 buckets might be considered autistic, and the lowest 965 will not be. Most of the population will be in those 965 buckets and not be autistic, yet those buckets still exist anyways and they can still reveal differences between the traits of males and females.
Being a millionaire is a threshold of wealth, but just because people aren’t millionaires does not mean that their wealth is immeasurable or the spectrum ceases to exist below millionares, that makes no sense. A threshold is just that, a threshold.
You’ll need to reword that because I am not sure what you are trying to say. We can split the buckets by male and female, picture two statistical distributions, the male distribution would appear skewed to the right compared the female distribution, meaning males have somewhat more prevalent traits associated with autism, but that doesn’t mean the average male is “autistic”. Only the rightmost 35 buckets fit that classification. (Just as “millionaire” is a classification reserved for people whose wealth is north of a million dollars).
The article I linked to earlier did measure higher percentages for autism traits in STEM compared to the average population in the study. Unfortunately those numbers weren’t broken down by sex. I personally am curious to see what the distributions actually look like, but I have yet to find anyone who’s compiled/published this data.
I never claimed underrepresented groups are not good for the job, but we do know they’re not applying and these personality traits could be one of many factors explaining why. I’m not saying it’s the only possible factor. More data would be helpful though.
I try to be nice, but to the extent that I failed, I offer my apology. Re-read what you said though, you don’t find it was rude and offensive? If you didn’t realize it when you wrote it, fine I’m telling you now that it is. Do you think it’s nice to say things like “I thought it would be obvious, but it seems I have to spell it out because you can’t figure that out on your own.” “Stop basing your world view on cartoons.” Clearly not. Obviously I can take a little sarcasm, but I hope that you don’t speak that way at work because it can make people feel miserable and unwelcome. I won’t touch on this any further,
Alfman,
So what’s the point of it then? Why are you trying to argue that it provides any explanatory power to your theory? You’ve left it so broad that you can set any arbitrary threshold and claims it proves your theory.
“Oh, the empirically determined threshold for autism doesn’t fit my theory, so I’m going to find the threshold that does.”
Like I said, it’s quantum woo. A bit of p-hacking thrown in. A theory is only as good as its contra-examples.
You’re ignoring the thread of conversation. The millionaire analogy was to counter your claim that I treat autism as an on/off thing. A threshold is a threshold. Me simply using a preexisting empirically determined threshold is not treating it as an on/off thing.
If higher incidence of “autistic traits”, and not just “diagnosed autism”, is the criteria for your theory, then we should see more females in tech ALREADY. Because by definition, there are more females that score higher on “autistic traits” than there are females with “diagnosed autism”.
I don’t know, maybe you’ll just keep trying to find that arbitrary threshold until your theory works.
No it’s not nice. But when you keep making the claim about me treating autism as on/off when I have explained to you (twice, at that point) that I didn’t, how else should I call you out on it?
People at work don’t keep making the same claim about what I said when I’ve explained it to them once. Something about being face to face that stops most people from trying to put words into other people’s mouths. Do you put words into other people’s mouths while at work? Do you think you should get a free ride to do that at work, or in general conversation?
And, yes, if someone at work does have a repeated pattern of behaviour, despite being told how to do better, I do say not-nice things to them. Sometimes you need to be clear and get their attention. eg, a few times I had to tell someone to stop treating the compiler like it’s magic, but it was only after many times that person ran themselves into compile issues without understanding how compilers work. But again, people at work tend not to need a talking to. On the other hand, I look back at past jobs and realize there are times I needed to be more firm to avoid things snowballing in the first place.
kwan_e,
The point of looking at the entire spectrum rather than just a piece of it is to recognize the fact that trait differences can exist even below the threshold needed to be diagnosed as autistic. And the main reason to care is that, as you’ve already pointed out, official clinical autistic diagnoses data excludes ~96% of the population, yet there’s no logical reason to assume that there are no male-female trait differences beneath that threshold. I think it’s a safe prediction to say there will be a difference on average, the question is really how much. We ought to have that data is all I’m saying.
To be fair though, I did not ignore your points, I just find it’s tedious to get you to concede anything and I end up having to repeat things a lot. But the fact remains that we can plot data points outside these classification thresholds and I do believe you understand this, despite the argument, so I’ll move on.
That’s a good question, however I’m not actually not looking for an arbitrary threshold at all, quite the contrary I’m interested in the whole population. What we can do is talk about overall averages without regards to any thresholds (just as we can talk about average trends in wealth overall without confining the data to millionares).
You don’t see the irony here? If you do belittle people with ad hominem quips like that at work, then regardless of your intentions don’t you see how language like that contributes the the negative and unwelcoming perceptions of our field?
We don’t have to look hard to find instances of you resorting to ad hominem attacks: “He’s writing about his father. Idiot. And they’re not ‘disconnected events’. They’re connected by the fact that IBM was involved in all of them. Bloody hell you’re an idiot.” I didn’t reply because I agreed with you on the facts, but nobody likes to be spoken to this way. Was it really necessary for you to say it this way? Do you think this person wants to engage more with the community after that sort of interaction? What is your goal in taking such a disparaging tone? Do you see how this can drive people away and even people who weren’t being targeted?
Honestly it’s a dilemma for me even to point this out because I know your natural gut reaction is to deflect and place blame elsewhere. I’m reminded of linus torvalds, who is smart, very well respected and yet has had a troubling history of creating a toxic environment. It took him a very long time to get there, but he eventually apologized over his behavior.
https://www.cnbc.com/2018/09/17/linux-creator-linus-torvalds-takes-time-off-apologizes-for-behavior.html
I don’t like dwelling on this, it’s off topic and if you are not ready to evolve past the use of degrading language then I don’t think it’s necessarily helpful for me to point it out. Inevitably pointing it out is going to sound condescending to you even if that’s not intended. I’m also trying to learn, evolve, and be a nicer person. So say what you need to say, blame me as much as you feel is appropriate and get that out of your system, but after that reflect on how your words effect other people’s feelings and our tech culture. If you don’t care about how people feel or about being welcoming, then so be it, but if you do care then try to consider that this is something you’ll need to work on.
At this point I think we’ve exhausted the original conversation and it’s just becoming boring & redundant, so I don’t think I have anything more to add. Till next time
Alfman,
To make the claim that more than half of the people working in tech place higher on the spectrum of autistic traits enough to make a significant gender imbalance, you have to have an arbitrary threshold. And you’ll probably have to lower the threshold so much that you lose any meaningful trend.
The thing about averages is that half of the people are below it. So it doesn’t matter what differences there are between the averages of male and female, because there are males below their average and ALSO below the female average. If autistic traits mattered at all, then those males below the female average should also be excluded/self-excluded from tech jobs for the same reason.
So differences in autistic traits between males and females can have nothing to do with it.
Economists also use arbitrary thresholds to make their theories fit and tells us nothing about anything. Scientific thinking requires consideration, rather than cherry picking. There’s something strange about people in the tech community when they think they can throw out scientific thinking just because they’re smart.
kwan_e,
This could explain some of the confusion, you are mistaking “average” and “median”. Only with the median, half the people are above and half the people are below.
https://www.mathsisfun.com/median.html
In a skewed distribution, the averages and medians can be very different. Take the wealth example again:
https://www.freemoneyfinance.com/2010/08/big-difference-between-average-and-median-net-worths.html
This difference illustrates how skewed the distribution is.
Again, I have no idea where you get the notion that I am cherry picking data. One normally does NOT cherry pick data to build a distribution or collect averages. We’re not talking about any voodoo statistics, just a bog-standard distribution over the full population.
Many distributions in nature follow a bell curve, and one of my questions earlier was whether autism also naturally follows a bell curve, I think it might. I’d predict the distribution for males to be skewed to right, the question is by how much.
https://www.statisticshowto.datasciencecentral.com/probability-and-statistics/normal-distributions/
Please just finally acknowledge the validity in collecting & plotting distribution data for the full population, this is totally non-controversial and should have been a gimme from the start. It just seems like you are so eager to criticize and argue that understanding what I say first isn’t that important to you. Am I right? I am so tired of this. To me it’s a shame to have an interesting topic to explore only to get completely stuck over such trivial minutia. I don’t know how to encourage you not to dig into a position that stops the conversation in it’s tracks. We can brute force the argument as usual, but man is this boring! In sincerity, do you enjoy these conversations? I don’t come here to argue, I want to have deep interesting conversations, but it seems impossible to get there with you. Maybe it’s time we break up, haha.
I don’t think it’s much of a stretch to say that systematic racism in the US was a lot worse in 1946 than it was in 2006. At least I would sincerely hope so!