We ran the numbers, and there really is a pipeline problem in engineering hiring • Hacker News Board

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#101

Post by wrs » Tue, Dec 03, 2019, 8:22 PM UTC

btilly wrote: post #83

wrs wrote: post #50

Earlier quoted context omitted.

Of course, if doing more things that involve spatial reasoning makes you better at spatial reasoning (which seems like a very likely assumption), this becomes a circular argument. (Did you read the Gender Differences section of that Wikipedia article?)

Yes, I did read the Gender Differences section of that article. You will note that I stated what research shows about population differences and did not wade into the more controversial question about why those differences exist. That said, I have found that on controversial subjects, people grab every sliver of support for their own ideas, and ignore data that might contradict it. The more heated the debate, the har…

It could be explained by lots of things — my point was that population differences in abilities that can be learned will obviously result if the population doesn’t have a uniform opportunity to practice those abilities uniformly, so it’s hard to avoid explaining the lack of learning by the lack of abilities without becoming circular.

It’s biological fact that brains are extremely plastic, and I’m sure we all know (or are) people who have rendered “innate” talents or lack thereof irrelevant by the application of study and practice. If cultural biases limit practice in a population, that population will certainly skew lower in ability as a result.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#102

Post by sbilstein » Tue, Dec 03, 2019, 8:23 PM UTC

charliesharding wrote: post #79

sbilstein wrote: post #42

Earlier quoted context omitted.

That's not at all what I said. I related a very personal set of anxieties (largely irrational) that I feel as a minority. Plenty of white people are uncomfortable around minorities. Do they all love that feeling? I'm confident most feel intense guilt about it and never act on those impulses in any way. Just having a 'racist' feeling doesn't make someone racist.

> Just having a 'racist' feeling doesn't make someone racist. I would argue that that is exactly what racism is. It's having an irrational feeling towards someone and assuming they have characteristics based solely on their race. Why is it so hard to judge people on a case-by-case basis? By your own admission you say that these personal anxieties are irrational and then you go on to attribute it to being a minority.…

False dichotomy. Some people may feel the weight of group membership more than others. Just cause group membership is the cause does not imply there is no personal solution. It’s both. I’m not gonna solve this for everyone but I can solve it for myself. It’s stupid to ignore my upbringing and blame it instead on just personal failure.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#103

Post by sdnlafkjh34rw » Tue, Dec 03, 2019, 8:29 PM UTC

tasogare wrote: post #52

Collective madness is largely reached at this point: the writer is complaining about a totally self-made issue. If I was trying to recruit men for nurse jobs, or people of color in Korea I would ran in a similar pipeline problem. The real issue is that "diversity" had became a metrics to reach for its own sake by companies instead of simply recruiting people that fit the job. Moreover, you can’t complain about so-cal…

It's weird to use the nursing example since the population of male nurses is exploding. Male nurses have gone from 2% of the population to 13% in just a few decades: https://www.nytimes.com/interactive/2018/01/04/upshot/male-n...

When we see extremes of representation we should take a step back and question why. In a lot of cases you are dealing with something systemic that may be able to change. For nursing, in the 1950s it was perceived as a female profession but as that viewpoint is going away, more men are joining the field.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#104

Post by ajross » Tue, Dec 03, 2019, 8:33 PM UTC

commandlinefan wrote: post #97

Well, the gist of this is: there aren’t enough women in “tech”, so different standards should be applied for women until there are. Setting aside all the other problems I have with this: law and medicine have much stricter barriers for entry. There are no practicing doctors or lawyers without advanced degrees, but there are plenty of women in both professions. So even if you accept the premise, it’s still pretty clea…

Alternative hypothesis: law and medicine started proactively working on their "pipeline problem" many decades earlier, and have had time to reach parity. Computer science, in the sense of the modern job market of elite grads going to bay area employers, is barely into its third decade of existence.

Seriously: go back to the 50's and 60's and you'll find all these same arguments about how women can't be good doctors and law is a man's world. And they were all wrong.

Why is computer science special in a way that medicine isn't? Why can women heal but not code? Isn't that the harder point to prove than "it's a pipeline problem"?

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#105

Post by TeMPOraL » Tue, Dec 03, 2019, 8:35 PM UTC

ajross wrote: post #95

commandlinefan wrote: post #61

Earlier quoted context omitted.

> you cannot expect to make things better than they are at the input of your pipe segment People have been fired from Google for suggesting exactly that.

This isn't a fair characterization. Damore's memo (which I assume is what you're talking about) absolutely did not attribute Google's lack of women candidates to a lack of graduates in the pipeline. It went with a straight on (and largely unsourced) "women are biologically different" argument. Wikipedia's page on the controversy is relatively good: https://en.wikipedia.org/wiki/Google%27s_Ideological_Echo_Ch... And e…

You're misreporting the very link you quoted. Indeed, the Wikipedia description is pretty fair. I'll quote the relevant paragraph directly:

"James Damore was spurred to write the memo when a Google diversity program he attended solicited feedback.[2] The memo was written on a flight to China.[12][13] Calling the culture at Google an "ideological echo chamber", the memo states that while discrimination exists, it is extreme to ascribe all disparities to oppression, and it is authoritarian to try to correct disparities through reverse discrimination. Instead, it argues that male/female disparities can be partly explained by biological differences.[1][14] Damore said that those differences include women generally having a stronger interest in people rather than things, and tending to be more social, artistic, and prone to neuroticism (a higher-order personality trait).[15] Damore's memorandum also suggests ways to adapt the tech workplace to those differences to increase women's representation and comfort, without resorting to discrimination.[1][14]"

In terms of pipeline argument, the memo can be viewed as arguing that the pipeline is biased at every stage, which is "partly explained by biological differences"[0], and that it's wrong to try and solve this by reverse discrimination at the tail end of the pipe (i.e. throwing perfectly good candidates out to improve the ratio). Instead, the memo proposed means to reduce the bias near the end of the pipe without resorting to discrimination.

--

[0] - Note that some of the differences - "stronger interest in people rather than things, and tending to be more social, artistic, and prone to neuroticism" - can be also plausibly explained by social factors (instead of biological), but that's orthogonal to those differences existing and biasing the entire pipeline.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#106

Post by plughs » Tue, Dec 03, 2019, 8:35 PM UTC

majos wrote: post #29

nradov wrote: post #17

My alma mater Harvey Mudd College has had rough gender parity in Computer Science for several years now. It's a small school but other colleges could learn from their practices. https://www.inc.com/kimberly-weisul/how-harvey-mudd-college-...

For anyone who didn't read the linked article (it's short), it suggests that a key element is careful design of intro courses. Students with computer science experience get a more advanced course that is explicitly designed not to make them more competitive in second-year courses (instead, it starts on material that will re-appear in third-year courses). Students with no computer science experience get their own intr…

It's an interesting insight. When applying to colleges I never considered applying for CS as I (female) had no programming experience whatsoever, and seemed to be surrounded by people (male) who'd been working on computers for years,

( I'll leave speculation on that for another day )

It was only because they forced EE majors to take programming courses that I ended up in the field. I wrote my first "hello world" as an 18 year old freshman, but it never really seemed to matter.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#107

Post by gotoeleven » Tue, Dec 03, 2019, 8:35 PM UTC

ajross wrote: post #95

commandlinefan wrote: post #61

Earlier quoted context omitted.

> you cannot expect to make things better than they are at the input of your pipe segment People have been fired from Google for suggesting exactly that.

This isn't a fair characterization. Damore's memo (which I assume is what you're talking about) absolutely did not attribute Google's lack of women candidates to a lack of graduates in the pipeline. It went with a straight on (and largely unsourced) "women are biologically different" argument. Wikipedia's page on the controversy is relatively good: https://en.wikipedia.org/wiki/Google%27s_Ideological_Echo_Ch... And e…

People should look at the memo with all its links to papers to decide if these claims are "largely unsourced."

https://assets.documentcloud.org/documents/3914586/Googles-I...

There was a version initially floating around that had all the links removed.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#108

Post by ajross » Tue, Dec 03, 2019, 8:39 PM UTC

TeMPOraL wrote: post #105

ajross wrote: post #95

Earlier quoted context omitted.

This isn't a fair characterization. Damore's memo (which I assume is what you're talking about) absolutely did not attribute Google's lack of women candidates to a lack of graduates in the pipeline. It went with a straight on (and largely unsourced) "women are biologically different" argument. Wikipedia's page on the controversy is relatively good: https://en.wikipedia.org/wiki/Google%27s_Ideological_Echo_Ch... And e…

You're misreporting the very link you quoted. Indeed, the Wikipedia description is pretty fair. I'll quote the relevant paragraph directly: "James Damore was spurred to write the memo when a Google diversity program he attended solicited feedback.[2] The memo was written on a flight to China.[12][13] Calling the culture at Google an "ideological echo chamber", the memo states that while discrimination exists, it is e…

Sorry, how are these two arguments equivalent in your mind:

1. There are few women engineers at Google because there are few candidates in the pipeline.

2. women generally having a stronger interest in people rather than things, and tending to be more social, artistic, and prone to neuroticism

I don't see that at all. The only way you get there is to assume a common causality. But that assumption (that some people are inherently bad at coding) is offensive if made without good science. You can't start with bad pseudoscience and "prove" that Damore was fired because he pointed out the pipeline problem!

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#109

Post by TeMPOraL » Tue, Dec 03, 2019, 8:40 PM UTC

hinkley wrote: post #85

wayoutthere wrote: post #69

Earlier quoted context omitted.

The author is leaving unstated what we all know to be true: most of these jobs aren't really that hard. You don't really need the people with pedigreed CS backgrounds, but it's easier than actually figuring out how to screen candidates. The "pipeline problem" is the pipeline itself. So you have to go around the pipeline by avoiding the formal education system, which is where the bias is observed.

The problem is that if you give a rocket engineer the job of opening a pickle jar, you are forever in danger of them doing something titanically ornate, instead of just opening the goddamn jar or installing one of those toothy wedges under a cabinet near the fridge, where people can still function normally if they forget it's even there. The one time I worked on a team of all highly senior people you couldn't get any…

> If you aren't constantly trying to make 90% of the tasks simple enough for a boring team member to do then I don't know how you prevent the overall complexity of the system from eventually swamping you all.

One has to keep in mind that the things should be as simple as possible, but no simpler. Otherwise, you risk ending up with the stereotypical low-skill Java Enterprise shop, where inexperienced and not particularly skilled developers end up writing (or autogenerating) code by kilolines, where if someone better trained were to stop and think for an hour, it could all be done in 100-1000x less code, and with similar improvement in maintainability.

Re: We ran the numbers, and there really is a pipeline problem in engineering hiring

#110

Post by ajross » Tue, Dec 03, 2019, 8:43 PM UTC

gotoeleven wrote: post #107

ajross wrote: post #95

Earlier quoted context omitted.

This isn't a fair characterization. Damore's memo (which I assume is what you're talking about) absolutely did not attribute Google's lack of women candidates to a lack of graduates in the pipeline. It went with a straight on (and largely unsourced) "women are biologically different" argument. Wikipedia's page on the controversy is relatively good: https://en.wikipedia.org/wiki/Google%27s_Ideological_Echo_Ch... And e…

People should look at the memo with all its links to papers to decide if these claims are "largely unsourced." https://assets.documentcloud.org/documents/3914586/Googles-I... There was a version initially floating around that had all the links removed.

There is absolutely no sourcing for research showing women being inherently or unfixably uninterested in computer science. None. He had a bunch of links to studies about bland stuff like personality differences, using them to make arguments the study authors never indended and (where asked) refused to support.

Calling that "sourcing" is essentially trying to treat Damore's fundamentally political document as an original piece of social science research.