The "Google Manifesto" which got a Google employee fired (FULL)


#1

Apparently I have been living under a Rock about this issue. Anyway to see what all the fuss is about I hunted down the actual document which has caused an uproar (which was actually be bit tricky to find, most articles just plucked out certain excepts) so anyway here it is.

I found it to be an interesting read. A bit too contentious/political for what I would consider as being accepted as appropriate to send to everyone in a “workplace environment”, but the author seemed to think that Google (being seen as an ultra progressive workplace) could handle it. Some of the points seem solid, some are probably a bit off base (and would need some citations to back up the claim).

My biggest criticism is that the Graph on Page 4 is not clear enough in trying to convey the message. The meaning I get from this is “All genders have the same personality traits, but some traits are statistically more common in one gender than in another”. That sets the tone for the rest of his opinion. Also he seemed to make big guesses on particular attributes that should really be backed by data if he’s going to make them.

Since we are not encumbered by workplace politics, there are probably a few things we could learn from it, especially on the topic of increasing diversity in the party without discriminating.


2017-18 PDC Working Groups
(Andrew Downing) #2

It has it’s own website now… https://diversitymemo.com.
It formats better for mobile.

Anybody reading this could probably benefit from some background on the author. James Damore has a masters in “Systems Biology” from Harvard and was researching in Computational Biology at MIT before being poached by Google 4 years ago via one of their puzzle competitions that he won. So, assume he’s not particularly politically savvy but highly intelligent and writing very objectively.

Much of the terminology is straight from the literature in academic psychology research, so it’s easy to leap to incorrect assumptions based on more colloquial definitions.
For example, “trait neuroticism”. Read https://en.m.wikipedia.org/wiki/Neuroticism to get an idea of the complexity of how that’s really thought of.
I would add, that even though that particular trait is generally considered a negative emotional trait, it actually has it’s value in moderation. More neurotic people are usually the ones calling out for caution in new situations. Google could probably use a bit if that; they tend to just blunder into new things, just because they can.

A useful resource for understanding this is: https://youtu.be/SEDuVF7kiPU, where psychology professor Jordan Peterson interviews James Damore and runs through his memo, discussing the various elements. Also attached to the video are many well organized links to the relevant research literature. There is a lot, and the scope/scale may be unexpected.

If we’re going to discuss this here, please start by assuming positive intent on behalf of other commenters.


#3

There’s certainly space to criticise the memo but it’s solid foundation in current science is undeniable.

It could have been an interesting internal discussion for Google where they could have taken some positive feedback and adjusted it’s approach and pivoted towards deiversity of opinion and equality of opportunity but instead they embodied everything wrong with authoritarian identitarian left and shut him down and his concerns branding him a bigot and a sexist.

By doing so they widen and deepen the chasm between authoritarian far left and racist far right, extinguishing the centrist moderating voices such as James Damore. Pretty poor effort.


(Tim Serong) #4

Not so much. Even the guy behind the research thinks that Googler is wrong – “Using someone’s biological sex to generalise about personality, as James Damore did, is like ‘surgically operating with an axe’”.

I would also encourage anyone interested in this to read Yonatan Zunger’s piece, notably including:

What you just did was incredibly stupid and harmful. You just put out a manifesto inside the company arguing that some large fraction of your colleagues are at root not good enough to do their jobs, and that they’re only being kept in their jobs because of some political ideas. And worse than simply thinking these things or saying them in private, you’ve said them in a way that’s tried to legitimize this kind of thing across the company, causing other people to get up and say “wait, is that right?”

I need to be very clear here: not only was nearly everything you said in that document wrong, the fact that you did that has caused significant harm to people across this company, and to the company’s entire ability to function. And being aware of that kind of consequence is also part of your job, as in fact it would be at pretty much any other job. I am no longer even at the company and I’ve had to spend half of the past day talking to people and cleaning up the mess you’ve made. I can’t even imagine how much time and emotional energy has been sunk into this, not to mention reputational harm more broadly.


#5

It’s quite amazing how much mischaracterising and outright lying there has been in the online media over this. What was written was that the statistical distributions of each gender are different, although overlapping, and that individuals should be judged on their own merits rather than a stereotype based on the mean. It was pro-diversity, anti-discrimination, and suggested ways to improve diversity at Google without affirmative action, aka without discrimination.

I would encourage anyone who hasn’t already to read the memo before reading any articles about it. Quotes like

are obviously wrong, because the memo explicitly argues against that sort of thing. There are diagrams.


#6

“Here’s your point by point refutation of the google memo.” https://medium.com/@adljksbvkj/heres-your-point-by-point-refutation-of-the-google-memo-b7201d0cca04

the real kicker in this is below:

so how about we keep addressing the real problem which is how women are perceived and treated - when it is known they are women. nothing much clearer than that.


#7

I read the whole document, and both those points are wrong.

He wasn’t using someone’s biological sex. They whole thing was general.

He did no such thing. He was talking about genders being predispositioned to certain traits. For example:

More females would prefer to work with other people than men do, as opposed to working on projects alone.

It doesn’t mean that women can’t work on projects alone, or that there aren’t women perfectly capable of working on things alone.

The example he used is coding, which is mainly a male dominated job where coders work alone. His suggestion was to allow coding to be done in pairs, which might be more appealing to people who want to work with other people, as many more women do than men.

There is a long is misconstruction out there. That’s why I criticise his graph on trait overlap for not being clear on what it means, it’s probably the most important part of his memo and it’s getting overlooked because the graph is hard to understand the meaning of.


(Andrew Downing) #8

Did I mention that this is a difficult conversation?

So the quote above totally mischaracterises the memo.
Here’s what Damore actually said about that…

Note, I’m not saying that all men differ from women in the following ways or that these differences are “just.” I’m simply stating that the distribution of preferences and abilities of men and women differ in part due to biological causes and that these differences may explain why we don’t see equal representation of women in tech and leadership. Many of these differences are small and there’s significant overlap between men and women, so you can’t say anything about an individual given these population level distributions.

Later in Wired,

Schmitt told WIRED that while this isn’t his area of expertise, the assumptions made by Damore were unwise. “We should rely on rigorous evidence for making claims in this area. And I believe there is good evidence of both sexism (including sex stereotypes) and real psychological sex differences (some of which may be evolved) to be causes of the gender gaps across occupations,” he said.

Sound familiar?
It should, because that’s also what Damore said.

The problem for someone like Schmitt doing personality research, is that given the rampant PC mindset prevailing in academia today (particularly in the social sciences), Schmitt’s funding and research is in danger if it is perceived to be wrongly aligned with something like this, which is a significant part of the problem here.

To my mind, this is where the left of politics can sometimes go off the rails into authoritarianism, hence why I favour a more libertarian flavour of left, that likes free speech (i.e. Pirate Party)


#9

Jesus Christ this thing is off the charts. The author has gathered a whole bunch of “Sexist assumptions” (at least they admit to them being assumptions) with no explanation of where they got these assumptions from (their arse). In no way did Damore say or imply any of that stuff. The author has completely misinterpreted the whole thing because they don’t understand that he was talking about distributions, not any female specifically, and because of that has conjured up a whole bunch of other allegations of sexism they think he would probably think, because he’s a sexist (according to them).

No wonder why this thing is so controversial, particularly to the far left, because of articles like this which spread misinformation and hate. Absolutely unreal.

Where Gender (, Race, Sexual Orientation, etc.) bias is suspected, especially when it comes to non-face-to-face interactions, anonymise the names.

So for example with GitHub, make everyone use an Alias than their real name and see what happens.
For example with Google, get the recruiters to go over resumes with names like “applicant #12345”.

Damore’s memo seems to think that it’s not the perceptions of females which is the issue, but the predisposition of females and males to have certain personality traits, and by extension those personality traits over a broad population make them more likely to choose jobs which are more suited to those personality traits (if they did and up with those personality traits that are more common).

Examples: More women have a lot of compassion in their personality than men, so that might make them want to be nurses. More men want to do anything get to the top than women, so that might make them want to be CEOs.

It doesn’t mean that men can’t be good nurses nor can women be good CEOs. It doesn’t mean that there aren’t men who want to be nurses and women who want to be CEOs. It’s just general predispositions over a large distribution of people.

It’s not inherently sexist, it’s just that he is coming from the perspective of STEM/Google. The same problems he describes would happen if it was a Hospital who were specifically favouring male applicants to be hired as nurses to balance out the gender ratio (rather than addressing ways to make the job more appealing to their personality so that the pool of good male applicants would be higher and they wouldn’t have to apply any favouritism).

His suggestion is to make the jobs more appealing to the demographic like offering better work/life balance (part time compared to full time+overtime), try reducing stress, allow coders to work in pairs than alone so that there is more social interaction.


(Mark) #10

This has proven to be a hugely fraught and divisive topic just about everywhere it’s been discussed. So please (everyone) avoid inflaming.


(Andrew Downing) #11

Actually, I’m familiar with that study. It flamed out badly, because the study was bogus.
What they did was incredibly duplicitous.

They selected data that showed, as stated in the BBC article you linked, that pull requests (the Github way of submitted code to a group for consideration in their project) that identified the user and specifically showed the gender as female got accepted less than generically anonymous submissions.

Sounds bad right?
Not when you understand that actually the difference was about anonymity vs. identified submissions, regardless of gender. It didn’t matter what gender you showed, anonymous submissions did better. Further, the data showed that amongst non-anonymous submissions, the female submissions fared slightly better.

This is exactly the sort of bad research that comes about when you pollute social research with ideological agendas. Research that sets out to prove a point, and ends up selecting whatever subset of the data it can find to achieve that end, is bad science.

Here’s a more detailed analysis if anybody cares: http://slatestarcodex.com/2016/02/12/before-you-get-too-excited-about-that-github-study/


(Andrew Downing) #12

Actually, one thing that was shown from that Github story, is that the majority of submissions are anonymous, probably for this very reason, so that submissions are gauged on their content rather than the submitter. This is probably even more important for submissions from well known submitters. There would be a tendency to accept submissions based on a history of success, but everybody has a bad day, and one-line changes frequently break the whole system, so anonymous is should be.

As an aside, the Australian public service recently decided to trial the idea of anonymous submissions for positions in the public service, but it didn’t work out as expected.


Note that they quickly backpedalled on the whole anonymous thing, because they discovered that there was actually discrimination going on, but it was in favour of women and minority groups (up to 22% in the case of aboriginal women).
So it because apparent that the real goal was to apply discrimination in the “correct” direction rather than to make a fairer system. I’m surprised they were so open about that.


(Andrew Downing) #13

OK, so there’s quite a lot of assumptions in this, but the most central assumption seems to be about the authors motivation in writing the memo in the first place. Actually, he’s been quite candid about that in personal interviews since.

There were company wide diversity sessions (not bad in concept), however …
Most meetings in Google are video recorded as a part of their openness and transparency (they’re just that sort of organisation usually), but the diversity meetings were explicitly not, and it became apparent why. In these meetings, they discussed discriminatory hiring practices and how these could be pulled off without getting caught out under anti-discrimination laws, in an effort to meet their diversity targets. Things like giving multiple additional interviews to women or minorities who failed earlier interviews, or ensuring that if women or minorities passed the interview process, they would be guaranteed a spot in a project, unlike anybody else that would otherwise fail at the last mark.

It was also discussed, that many managers maintained “black lists” or people they considered to not be sufficiently on board with the ideological agenda. These black lists were reported to be being shared around Google and across other affiliated businesses. They would essentially end your career if you didn’t get on board with their ideology.


(Tim Serong) #14

Which is entirely the problem. He’s positing “that the distribution of preferences and abilities of men and women differ in part due to biological causes and that these differences may explain why we don’t see equal representation of women in tech and leadership.”

Look at what he’s implying:

Women on average show a higher interest in people and men in things

  • We can make software engineering more people-oriented with pair programming and more collaboration

This implies that software engineering is not very people oriented, and so women on average won’t be interested in it.

(This is, incidentally, bollocks – all software engineering involves significant collaboration and cooperation within your team, and between teams)

Women on average are more cooperative

  • Allow those exhibiting cooperative behaviour to thrive […]
  • This doesn’t mean that we should remove all competitiveness from Google […]

There’s a (possibly weak) implication here that tech work requires competitiveness, and again, effectively that it won’t be suited for women in general.

Women on average are more prone to anxiety

  • Make tech leadership less stressful.

Implication: tech leadership is stressful, women in general won’t like it and/or won’t cope.

Women on average look for more work-life balance while men have a higher drive for status on average

  • Unfortunately, as long as tech and leadership remain high status, lucrative careers, men may disproportionately want to be in them. Allowing and truly endorsing (as part of our culture) part time work though can keep more women in tech.

Implication: tech work doesn’t allow a work-life balance which women in general want. Also: women want part time work. (Huh?)

The document reads like it’s well reasoned, but really it’s essentially saying “tech work isn’t people oriented, requires competitiveness, is stressful and doesn’t support a work-life balance, and women on average have traits/characteristics which don’t fit well with this”. So I have no problem with Yonatan characterising it as “a manifesto […] arguing that some large fraction of your colleagues are at root not good enough to do their jobs”.


(Maus) #15

On the contrary. I would interpret it to mean that the job is not good enough for many women.
If the memo hasn’t been scientific enough for my standards, most analyses of it I’ve seen are far worse and/or just hype based on wilful misinterpretation.
The kind of left that assumes that women should assimilate to behave like men and want the exact same things in life, to work full time and over hours for an ungrateful profit-driven company, this is not the left that I allign with. I actually thought some of his ideas for improvement were quite good, even if the rest of the memo was crap. It doesn’t deserve even half the attention it’s been getting, especially with recent political events.


(Ben McGinnes) #16

No, it isn’t. The male vs. female brain nonsense posited on page 3 and onward has been provably shot down in flames many times.


#17

Nope. He did not say abilities.

He’s not implying anything, he’s laying it out factually, you’re assuming

He was referring to software engineering at Google specifically

Nope. There is no implication that they won’t cope, only that most wouldn’t be interested in the first place.

Well not a good work-life balance at Google anyway. More women in general do want work-life balance more than men do, not as many are willing to sacrifice their whole lives to the company.

Part time work is a good example of a way to help with work/life balance.

Yes, well at Google anyway.

Nope. Again, it’s about attracting women into the role to begin with. If a particular woman is interested in the role (despite the conditions that might make most women probably not want it in the first place), there is no inherent reason why they wouldn’t fit

That is the massive insult to suggest such a thing. This was not implied anywhere in the memo. This is the journalist pulling heartstrings and then trying to pin it on someone else for his words! despicable.


(Andrew Downing) #18

I hear rumours to this effect often, and yet never find any substantiation behind them…
but more seriously, if there were no differences, why would a corporation like Google value diversity?

This would seem to be the central dilemma here.
I’m totally down with diversity, but diversity is difference is good right?
So why are we pretending there’s no difference?
Of course there are differences.
Have you not met women?

Like Laura said:

Why not celebrate the differences?
Why not just go crazy and find whole new ways to do business with these differences?
Why do we keep trying to jam them into the same stupid structures?
Could it be because some stupid ideology keeps insisting they’re the same?


(Maus) #19

Diversity is about much more than supposed differences and overlap in the distribution of traits between men and women. The human mind is much less binary than that (pun intended).

The thing about statistical distributions is that they show a range of traits in a population. It’s pretty dumb to only value those that fall somewhere around the middle, when those on the outskirts can bring quite fresh perspectives and ways of approaching a problem. Affirmative action has the potential of enabling these new approaches, so I still disagree with the memo on some points.


(Andrew Downing) #20

Firstly, affirmative action to jam outliers into the work structures designed for middle … seems like you were matrixing (can that be a verb now?) about how crappy that was earlier.

Secondly, affirmative action in favour of one group is always negative discrimination against another, and since we’re already clear that collective stats should never be applied to individuals, why does that logic suddenly become OK in the case of “affirmative action” ?