Hacker News Reader

Comment by dx034 | Hacker News Reader

> He was fired because he basically said women are underrepresented in IT because they are biologically unsuited for it.

I don't understand why people keep repeating that statement. He said that biological differences explain the different ratios of men and women, NOT that women are not suited for tech jobs. He explicitely did not say (nor mean) that a woman can't be a good software engineer. Just that on average, more men than women will show the features needed for software engineers.

That is a huge difference. It doesn't disqualify any woman working at Google. It just means that the goal to reach 50% women could lead to a worse outcome.

Disclaimer: I don't necessarily support his claims, nor do I support his memo as such. I just think that it's worth thinking about and that there should be more scientific studies to explore this. Diversity campaigns should be based on science, not on populism.

View on HN

I think it has to do with low knowledge of statistics.

The funny thing is that all the following sentences can be true at the same time:

1. Across the whole population, fewer women than men have the biological/psychological traits needed to become software engineers.

2. Female engineers are underrepresented in Silicon Valley and face many different forms of discrimination.

3. Female engineers at Google are on average better than their male colleagues.

I'm not suggesting any of these sentences are necessarily true, just that they can be. This is not intuitive, unless you know a bit of statistics.

> I don't understand why people keep repeating that statement. He said that biological differences explain the different ratios of men and women, NOT that women are not suited for tech jobs. He explicitely did not say (nor mean) that a woman can't be a good software engineer. Just that on average, more men than women will show the features needed for software engineers.

The problem is that his whole argument is based around biological determinism and those are the same arguments that true misogynists often use to discriminate against women. He even stated 'Neuroticism' as a trait women are predisposed to, which could've been pulled right out of an argument for why women shouldn't be allowed to vote.

Damore devotes zero attention to the impact of culture on what fields women go into. And in a nutshell, that's a major reason this manifesto is problematic and got him fired. Women in IT sometimes have to contend with attitudes that 'women are bad at math' or 'women are irrational' or 'women only care about aesthetics'. These are the attitudes that make some technical fields unwelcoming to women.

He hedges a little bit by saying 'distribution in traits between genders may be a part of why the gender gap exists' but then he proceeds to discuss that as the only reason the gender gap exists.

> He even stated 'Neuroticism' as a trait women are predisposed to, which could've been pulled right out of an argument for why women shouldn't be allowed to vote.

Did you read this article? Or references linked by the original memo? Even this article clearly points out that "neuroticism" thing is pretty much established, accepted, and uncontroversial science.

No, I don't think it is. Not even the definition of neuroticism is uncontroversial.

> The problem is that his whole argument is based around biological determinism

It's not based on biological determinism, he provides references to the current state of psychological research and simply wants a discussion about the policy

> this manifesto

it's not a manifesto by any means unless you want to call everything a manifesto including your post.

> Women in IT sometimes have to contend with attitudes that 'women are bad at math' or 'women are irrational' or 'women only care about aesthetics'.

There were no such things in his text, not even remotely close

> then he proceeds to discuss that as the only reason the gender gap exists.

He explicitly questions those causes, please stop projecting your ideology on things you read

The term "neuroticism" stems from the five factor model of psychological traits. The memo author doesn't really mention this too much, but if you know about the model, he uses the other factors throughout his work, which means he has read and understood the model.

There have been several studies showing that women scores higher in the agreeableness and neuroticism factors, on average (and with a quite large variance and a rather small effect I might add). The memo author makes it clear by writing "in part" in multiple places: a moderate effect can't explain the whole story: either there is another factor which truly dominates, or we are just looking at one piece of the puzzle as a whole.

Neuroticism has a very specific meaning in this case and it pertains to "negative emotion" in the sense that extremely high scores make you more likely to be stressed, anxious and so on. Likewise, extremely low scores make you overconfident and so stable you are unlikely to venture out into new ideas.

From an evolutionary phsych. standpoint, any score which exist tend to serve a purpose some way or the other. There are no desirable or undesirable traits, because too much or too little of a trait will be challenging in some situations, but a great advantage in others. So the question to ask is: if women scores higher in neuroticism, what advantage does that buy them in what situations? More anxiety could for instance have you second-guessing yourself in more situations, and that could lead to you finding more bugs in programs up front, which is an excellent trait to have in software development.

The other factor where there is a measurable gender difference, agreeableness, is a two-edged sword: High agreeableness scores are highly desirable for building coherent and working team/group structures. Low agreeableness allows people to pick an individualist/isolationist path when needed and make something great by ignoring the group consensus. Good group dynamic requires both traits to some extent.

My hunch is that if you scored the women at Google on the five factor model, they'd be likely to fall somewhat outside the gender norms. There is a considerable selection bias if you only consider tech people who work for Google. You are easily looking at a small fraction of the population as a whole, and the hiring selection process is likely to skew the scores quite a lot from the average norm.

The thing about pointing at the 5 factor psych model then applying it to something like gender disparity in tech, is that it's a pretty big leap to take scores from research studies to then draw conclusions about gender disparities in a given profession. Especially because at the core, it's very difficult to disentangle gender differences in psychological traits from learned social behaviors.

You bring up a lot of questions about what effects predispositions have in this case and that is exactly why it creates problems when people take findings on predispositions and run with them.

It's no greater than the leap required to attribute gender disparity to internalised misogyny.

Except we do have direct evidence that gender bias leads to worse callback rates, worse initial evaluation of qualifications, worse promotion results, greater punishment for negotiating salaries, and expectations that are outside of job requirements.

But we also have direct evidence that the distribution of genders across the disciplines increases as opportunities between the genders become more equal. That strongly implies that biology does have an affect on life choices.

It also has to be said that not all evidence is equal.

I was reading an article about this just today, as it happens, complete with citations: https://www.psychologytoday.com/blog/rabble-rouser/201707/wh...

I think the key point the author is trying to make is: "Does biology play a role in what choices women make when it comes to selecting a career path?"

Clearly, this is not attributable to a single argument or factor. Rather, you have to look at several different reasons. Some are likely cultural, some are likely genetic and the interplay is going to be vastly complex (see e.g., epigenetics).

To make the argument, you have to survey a larger number of areas and show that they all point in the same direction, which is what the author is trying to do in the first place. And in particular, you have to demonstrate how the areas contribute to women making a choice against tech in general.

Hence, the studies showing how men are more often systemizing than women and so on are far better markers/arguments from a pure biological standpoint, than a difference in the 5 factor model.

Personally, I think most diversity programs fail to achieve any kind of success in improving the gender balance. In two prongs, you have to retain women who are already in tech, and you have to get more women into tech. The latter of those two is a hard sell if women simply don't want to be in the tech world, no matter how much you try to seduce and coerce them. And I think it would be too authoritarian to force them into a role they don't want to belong in.

I think the author is spot on with the observation that more competitive fields are more likely to have men employed over women. And more cooperative fields are better at retaining women. E.g., Law in Denmark has an almost 50/50 gender split, but women are far more likely to be employed in the public sector as lawyers. The law firms, where 80+ work weeks are the norm, are usually employing men. Jordan B. Peterson, whether you like him or not, posed the question as to why men are willing to fucking do that, and I happen to find that question quite interesting.

I also think tech in general would benefit from having some ventures which are more cooperative, but this is not likely to come from a highly competitive venture capitalist private sector, I'm afraid. Nor do I think Google can pull that one off.

> Some are likely cultural, some are likely genetic and the interplay is going to be vastly complex (see e.g., epigenetics).

OK, let's say that some segment of it is cultural. Now what?

Let's say that I'm female, and I have some talent for engineering/programming, but I'm raised in a culture that says "Women don't do that. Women marry, stay at home, and raise children." And I look at my mother, who is doing exactly that, and I see how trapped she feels, how stifling it is, and I really don't want to wind up in the same place, but because my culture is what it is, I don't think I have any other choice. Clearly, that's bad.

But let's say instead I look at my mother, and see that she's fairly content and happy, and I think, "That looks pretty good." It's still the same cultural constraint, but now I am willingly assenting to it. Now, is that still bad? (Yes, it's still bad as for the general culture, because it's still constraining others who don't want that. Is it bad for me specifically?)

If you say "Yes, it's still bad", you're saying that you know better than I am what decisions I ought to make. (This presumes that I genuinely understand the alternative. But don't say that, if I really understood, I wouldn't want to choose that. That's cheating. It's also insulting.)

Note that, in my example, the differences are neither genetic nor cultural. The difference is in the example that I see growing up, whether or not people are happy fitting into their culturally-assigned place.

> He explicitely did not say (nor mean) that a woman can't be a good software engineer.

> Just that on average, more men than women will show the features needed for software engineers.

Read those two sentences you wrote again. You'll be shocked when you see it.

I'm not seeing it. Help me out? It's not a contradiction, if that's what you're implying. Have you ever taken a statistics class?

Oh, throwaway257157, if I thought you were posting in good faith, I'd certainly engage with you but using a newly created account means you aren't.

They are publically defending an argument that got someone fired, so using a throwaway account is totally reasonable here.

No.

Assuming someone is posting in bad faith for this reason is an act of bad faith.

Your statement reveals hypocrisy.

A throwaway account created specifically to post one comment is a sign of bad faith. It's entirely possible that I am wrong but the lack of subsequent engagement suggests otherwise.

"using a newly created account means you aren't."

This is wrong. Using a newly created account does not mean the communicator is approaching this in bad faith. That's a complete non sequitor.

"lack of subsequent engagement suggests otherwise."

Not at all. There are numerous reason they might not engage further within this time frame - or ever.

> Using a newly created account does not mean [...] bad faith

Specifically to post one (1) comment strongly implies it though. You disagree, fine; we'll agree to differ on this.

The point the throwaway raises is valid, though.

Where is the contradiction?

If you can't see it already, I doubt anything I could say to you would help that.

What I see is that you don't understand statistics. And, rather than offer a snide comment, I will try to show you why. Here are the two statements you quoted:

> He explicitely did not say (nor mean) that a woman can't be a good software engineer. > Just that on average, more men than women will show the features needed for software engineers.

Rephrased:

a) a woman can be a good software engineer b) on average, more men possess the skill set required for a software engineer

There is no contradiction between a) and b) whatsoever, since a) is about an individual and b) is about average features of different groups.

Of course, one may argue about the veracity of either a) or b). But, regardless of their factual accuracy, you did not make the point you think you did.

I took the liberty of labeling one of the diagrams in the document with two points labeled A and B. How would you interpret the relationship of these points if A was a man and B was a woman?

http://imgur.com/a/kWyTi

I would definitely agree that A was to the left of B. I would also concede that B is to the right of A. But since the only label on the graph is an unidentified "Trait", I couldn't really comment any further.

What point are you trying to make?

That the value on the y axis, for those two points taken from presented distributions, is the fucking same...

It's hard to make that point even more obvious.

It's pretty clear from the totality of this person's comments that they are here to obfuscate and troll. Not sure why this is allowed on HN.

Would you care to revise your statement given https://news.ycombinator.com/item?id=14981300 ?

What? It's a normal distribution graph. The height of the curve denotes how many of a particular group (group is denoted by the blue and green curves) score with regard to some trait x. The y-values for A and B on the graph are irrelevant.

You are right, of course. A dumb mistake of mine.

The point he's making is that, on that graph, it is clear that many women are better at software engineering than many men. The women closer to the tip of their graph will easily beat men at the tail of theirs. Women can be very competent software engineers. That is clear to all (Sandy Metz is a great example).

But if, in general women prefer different roles, is that unimportant to you?

Why does it matter which trait this describes? Surely this demonstrates that averages don't matter for how a trait is expressed in individuals, no matter how it's distributed in different groups?

Does anyone know if this graph is to scale?

Seems to me it would have radically different interpretations depending on scale, and I'm not sure if it's just illustrating the concept of overlap or if it's claiming to represent "software engineer ability" or any other real measurement.

Reply on news.ycombinator.com