I'm not sure why you think that your experiences as a youthful nerd make some sort of point about racism and sexism in modern tech-corporate culture, though. I mean, your entire post seems like a non sequitur in the context of the post you're responding to.
The people described in the article worked as computers and mathematicians at NASA because of:
- talent
- born, raised and lived in the right place at the right time (propinquity)
- self interest (they applied for jobs doing work they were interested in performing)
- health (reasonably good enough not to be a distraction)
- education
People who worked as engineers in Silicon Valley since the mini-computer revolution have done so because of:
- talent
- born, raised and lived in the right place at the right time (propinquity)
- self interest (they applied for jobs doing work they were interested in performing)
- health (reasonably good enough not to be a distraction)
- education
Notice that those two lists are largely the same? While I'm not saying discrimination isn't at play or non-existent, it's at best a minor factor relative to lots of other issues. Let's take what is likely to be one of the biggest issues that everyone loves to conveniently ignore, base rates:
In Hampton, Virginia, the birthplace of NASA, the African American population is 49.8% and in Virginia it's 19.7%, whereas in the Bay Area it's about 6.7%, most of which is largely concentrated in Alemeda, Contra Costa and Solano counties. Furthermore, these counties are not the locus of Silicon Valley, San Francisco (5.8%), San Mateo (3.0%) and Santa Clara (2.9%) are. Even the entire state of California is only about 6.5% black. This means, that to have a percentage of the Silicon Valley work force be black at approximately the same rate as the rest of the United States, black people need to move here either for work or for school.
As someone who's moved countries 4 times and states 2 times, this is non-trivial and unpleasant for those without a disposition to tolerate such a jarring change where you're an adult and responsible for supporting yourself and financing your own move. You lose support networks of family and friends. It's a non-trivial cost, especially if you need to move belongings. You need to rebuild and re-purchase the things you didn't bring.
The base rates above are the general population. If you assume that most people in tech have a bachelor degree or better (we no this isn't universally true since many software engineers are self taught), then the base rates gets worse. For people between 25 and 29 years of age, about 20% of black people have a bachelors degree versus 15% for Hispanics, 40% for white people and 58% for Asians.
Furthermore, 52% of college students attend a university within 100 miles of home and only 15.5% attend a university farther than 500 miles away. And a mere ~1.5% of Americans move across state lines each year.
Honestly, most people debating diversity as so incredibly awful at understanding statistics and performing a basic root cause analysis. It's like people start asking the 5 Whys, but the moment diversity is involved people's minds automatically short circuit and jump straight to because #racism or because #patriarchy. I'm not saying neither of those contribute to the problem, but they are far from the biggest factors. There are a LOT of filters that do a lot to modify/distort the available pool of people who would get even as far as applying for a job (which is for the most part a prerequisite for discrimination, active or unconscious)
Basically, if you want to make it so racial diversity in tech approximates the general population in the US, the best solution is not to try to change Silicon Valley, but to help more places like Silicon Valley exist in places with base rates that more closely match the racial distribution you believe is ideal. AnnaLee Saxenian's books are instructive if you're curious how such regions are formed: "Regional Advantage: Culture and Competition in Silicon Valley and Route 128" and "The New Argonauts: Regional Advantage in a Global Economy". A big part of what you need to move is money (i.e. you need a robust angel investor and venture capital industry).
* surface diversity - when you define diversity based on genetics and not content of character and live experiences.
yes, I'm aware that you could argue that recruiting mostly from people who live in the same city, county or state is a form of unconscious bias, but you're going to have a hard time arguing against that one ( http://www.bothsidesofthetable.com/2010/05/28/you-cant-look-... )
Hampton, Virginia http://quickfacts.census.gov/qfd/states/51/51650.html
Santa Clara County http://quickfacts.census.gov/qfd/states/06/06085.html
San Mateo County http://quickfacts.census.gov/qfd/states/06/06081.html
San Francisco County http://quickfacts.census.gov/qfd/states/06/06075.html
https://fivethirtyeight.com/features/race-gap-narrows-in-col...
http://www.thecollegesolution.com/where-most-students-end-up...
http://www.nytimes.com/2013/12/15/magazine/why-are-so-many-a...
Then there's the experiences themselves:
http://www.theguardian.com/technology/2016/jan/12/silicon-va...
Would you work in a place or field like that if all you heard was what's in above link? And people justified it instead of solved it? And hard data, such as pay or hiring w/ performance in context, showed they discriminated in both hiring and advancement? Probably not.
The problems are in the Valley's culture and decision-making, not anything intrinsic. Same as in most other places. Good policies on professionalism, hiring, and promotion which focus on performance instead of identity would go a long way. Have, in few companies that did it.
Some reasons for skepticism about blind auditions making a difference in tech:
1. Github provides a potential blind audition scenario - don't mention your race or age or gender or weight or anything else. Just submit code and prove yourself. Yet, diversity advocates generally see this as not helping minorities in tech.
2. Tech companies constantly publish their diversity stats. They are trying to hire more women and minorities, and are motivated to do so. Blind auditions tend to help in industries where the opposite is true.
Still, I'd love to see a source for your claim, and I'd be happy to be convinced otherwise.
Regarding the one link you do provide: The headline figure is meaningless without some more data on the survey - such as how was the survey conducted, and is it random. If it's not random, then it's very hard to judge how valid it is.
I'm also curious to see a citation for this statement, "Just submit code and prove yourself. Yet, diversity advocates generally see this as not helping minorities in tech.", since I'm curious to read their reasoning.
I actually would expect and hope that such a process doesn't help minorities. I also would expect and hope that it also doesn't help those that are not minorities. After all, isn't a fair, blind process the most just approach we can hope for? However, I think this is where there is a big, largely unrecognized schism, and that is the schism between those that advocate for equality and those that advocate for equity. Those two unfortunately are not the same. I personally am an advocate for equality.
+1 for understanding study methodology and what biases may be present. AFAIK, the study is this one here, http://elephantinthevalley.com/
They state the criteria for including a respondent in the survey, but I didn't find out if the 200+ women included in the study were randomly sampled from all women they "found" who met the criteria or if this represented the entire population of all the women they "found" who met that criteria. An explanation of how they were "found" was also absent.
The final sentence in the introductory block "We encourage you to read, discuss, and add your anonymous stories to this initiative." suggests that self selection bias may influence these figures.
I submitted the following comment on the site: "Could you please add a section on your site describing in detail the study/survey methodologies used? I'm particularly interested in knowing more about how survey participants were selected.". I wish all studies and surveys like this voluntarily shared their methodology along with their results.
I can see a demonstration of 50/50 in a controlled study, but I don't see how you get to 50/50 at a real company in this industry in the Bay Area without actively discriminating against those who meet the hiring criteria and are not women or minority, since 50/50 is so far from the base rate of potential applicants available to hire from.
To answer your question, yes, I would work in a place or field like that if that is all I heard in the above link, because I also know that nothing I read in that article is isolated solely to tech companies in Silicon Valley. AFAIK (and I haven't seen numbers otherwise), everything described in that article happens in many industries and many regions, including other industries I've personally worked in or with previously that weren't tech and weren't in the Bay Area (or necessarily the United States) (including one industry that is mostly female).
I'm not saying that we shouldn't address and change the culture so these unfortunate problems don't happen to anyone, but painting one industry and region as being uniquely a bad apple does a lot to deprive good people interested in that industry from entering it in the first place, leaving those already there further isolated.
I work/worked at one of Silicon Valley's many unicorns on the same team as one of the women who founded the female engineering group at the company and who champions the hiring of more women and one of the greatest challenges I heard her complain regularly about when trying to recruit women to improve our ratios was trying to overcome how the biased tech media likes/liked to portray our company (and any unicorn once it starts experiencing success). I heard of several instances where a woman turned down our offers and went to work elsewhere merely on account of what the media said despite my colleagues claims to the candidate otherwise. The way I see it, the chicken little sky is falling way of talking about things highlights problems but also actively discourages people from joining and making it better. Lots of great female engineers declined a seat on a rocket ship on account of what was being said, and I have no idea how many more didn't even consider applying in the first place.
There are ways to highlight and address the problem without discouraging people from coming to work in tech. I would love to see a comparative analysis of the transgressions in your linked article across multiple industries. Raising awareness for these problems everywhere helps get them solved, but raising awareness for these problems in one particular industry is both productive and counterproductive as it highlights the problem while simultaneously discouraging people from entering that industry to expedite those solutions.
Before I conclude this comment, I want to be very clear that I wholeheartedly agree that everything in that linked article is a problem and needs to be improved, but my particular complaint above and in this comment is about people who seek results without applying logic and rationality to understand the problem and don't also apply logic and rationality before arriving at solutions. Basically, a society that demonstrates logic and rationality is the social cause I most advocate for above all other social causes because it keeps all other social causes and those that advocate for them honest and grounded in reason.
In fact, this colleague who took on the important work of trying to recruit more female candidates started out as one of those women who had a job offer at our unicorn and went to work at another unicorn on a different coast with a better reputation first and ended up back at our unicorn less than a year later finding out that while our reputation was worse, our conditions were better (but still need improvement).
But your numbers are focused on race and ignore gender, which from what I've read is the much bigger problem in modern American tech. GapJumpers' blind auditions, for example, seem to indicate that the "women aren't interested in tech" assertion is less true than people want it to be. Nickpsecurity mentioned GapJumpers by name and you and Brighteyes both asked for citations; it's https://www.gapjumpers.me/ if you don't want to Google.
With gender however, the base rates are more directly influence by society's treatment of gender in the many years between being born and arriving at the point where you start looking for a job. Girls are treated differently by adults and expected and encouraged to take interest in different things. I think that's unfortunate and I hope that changes.
That said, it doesn't do anyone any good when people complain about a problem and only focus on a single contributing factor like racism and sexism instead of scientifically and methodically brainstorming all possible causes and contributing factors, measuring the magnitude of impact of each and then determining which of all the contributory factors provide the most bang for the buck in changing.
There is an entire discipline for improving things you can measure. It's called statistical process control and its a much smarter way of going about improving things than debates we have today where people arguing from either side have no measurements or statistics or a poor understanding about how to measure things and use statistics or some combination of both.
===
As someone with a strong interest in the social history of how the hardware and software industries evolved since WWII, I've provided some other possible explanations, in addition to sexism, that likely contributed to the gender imbalance we see today:
https://news.ycombinator.com/item?id=10928920
https://news.ycombinator.com/item?id=10929042
The big questions that remains in my mind after the observation I made in that second post are:
- "Are there conditions that when afforded to boys and girls equally, generate equal interest in computer science and engineering in people regardless of where they lie on the gender spectrum?"
- "If there are equal conditions that foster equal interest, does that equal condition generate maximal interest from anyone on the gender spectrum, or should we consider a offering a variety so that greater interest if fostered by pairing a person with the environment that is optimal for him her or anything in between?"
- "If no condition exists where you can generate equal interest, how do you design two or more separate set of conditions that approximates equality by being equally equitable without necessarily being truly equal.
===
on race and ignore gender, which from what I've read is
the much bigger problem in modern American tech
===
Re: Gapjumpers
Now that I've googled "gapjumpers" and seen what it is, I'm very inclined to give far less credence to the claims of nickpsecurity because I now know that gapjumpers is a startup selling a product based on their claims and not some non-profit funding independent academic resource with all the hallmarks of good science. I was really hoping for scholarly research and not an organization with a clear financial conflict of interest that is very likely to be hyperbolic or suffer from confirmation bias.
I even went a bit farther and decided to look at some of their jobs and the blind audition task. Here are two engineering ones:
https://www.gapjumpers.me/questions/sendgrid/qs-224/ https://www.gapjumpers.me/questions/bbc-digital/qs-256/
Those have 32 and 36 applicants respectively. While I have no doubt that their process helps create a more equal process, what is see isn't suggestive of a large sample size to make firm claims from like nickpsecurity did.
The figures on their home page are 54%/44% female to male. You can get at these figures with 9 females and 8 males. There is no indication that these results scale to dozens to hundreds of employees. If there were just 17 hired employees for those figures, the margin of error would be mega high (+/- 23.77%) margin. Furthermore, this hiring process is likely to suffer from confirmation bias, since those with the greatest interest in an unbiased process are those that hold the believe that they are most disadvantaged by the bias. This means that the pipeline of applicants into their process isn't representative of the base rates of available candidates in our industry currently.
I'm really not trying to tear their startup apart because I think they are part of the solution, but bad science, bad statistics and unsubstantiated or misleading claims do deserve to be challenged and torn apart all day long.
===
Re: "not all men"
Generalizing ... is the reason that "not all men" has
become a nasty cliche.
Essentially this boils down to the immature playground argument of "Are too!", "Are not!", "Are too!", "Are not!", ad nauseam.
"Hello Pot, this is Kettle calling..."
Those who make generalizations about white people or men are as guilty as those who make generalizations about non-white people or women (or any other discretely identifiable group on the gender spectrum).
And yet you've not cited a single, scholarly treatment of anything you're claiming. You immediately moved to an anecdote to save yourself time and facilitate explanation. Like I did. Upon request and when I found time, I dropped around 3 hours digging up old and some new references that supported my claims including scholarly work and experimental studies. Should I doubt all your future claims because you saved yourself time or didn't drop a Pastebin-sized set of references and commentary in your first comment?
I think that would be a bit unfair, don't you? ;)
"Now that I've googled "gapjumpers" and seen what it is, I'm very inclined to give far less credence to the claims of based on work by gapjumpers"
The reason I made the jump to putting your name in their is because gapjumpers was your primary reference in a few comments, but no link was provided.
I have few doubts that a blind interview process of capable candidates will approximate the base rate of available candidates that participate in the interview process (or mock hiring process for the purpose of demonstrating that a blind interview process eliminates or greatly reduces identity-based biases). I was never debating this claim. Providing references to this claim is tangential to the point I was making. I was debating your original claim.
Your original claim:
It all looks nice until we get numbers from groups like
GapJumpers that institute blind auditions for hiring or
performance evaluations where evaluator can't identify
what group they are. Starting out, these "fair" companies
are usually under 10% minority or women. Once they see
only performance, it's suddenly closer to 50/50. Hmm.
I'm also making a minor assumption that we're talking about engineering positions at these companies since that is where most of the debate regarding the gender ratio is centered at. For non-engineering positions, I've not noticed a huge discrepancy in hiring. At the tech companies where I've worked, the non-technical side of the business is very balanced gender-wise. FWIW the technical recruiting departments responsible for sourcing most of the engineering candidates we interview are mostly women. I don't know if that's important at all, but it's an interesting observation.
With those two fair assumptions, both me and the other commenter considered your claim as written to be a very very bold claim. Your claim was tantamount to "blind auditions are a panacea to the gender (or other identity) diversity problem in hiring". Given that claim, that's was we want a citation for. Where are these mythical 50/50 tech companies and how to they achieve 50/50 without being discriminatory when the base rate they are drawing from is approximately 88/12 (the approximate distribution of men to women with CS degrees)?
Given this, the "scholarly work and experimental studies" you found are interesting, but not really relevant because I never once challenged what blind auditions can achieve in an academic or experimental setting. I wanted a citation that references the "fair" companies you mentioned. because I was challenging that blind auditions can automagically achieve 50/50 without reverse discrimination in an industry where the population from which you draw candidates is far from 50/50. Had you limited your claim to academic and experimental studies, I wouldn't have challenged it at all. I probably wouldn't even had asked for a reference, since that's not even an extraordinary claim.
With respect to your other comments. I'm not sure which claims of mine in particular you're addressing or what anecdote in particular you're talking about. I wrote a lot across several comments. The original claim I made, which was about the relevance of base rates with respect to race didn't require a scholarly citation since I cited the data directly, much of which was US Census data, which is as authoritative as you get for the claim I'm making. For the other references I used, I could try to find the original source material instead of secondary references. Let me know which secondary references you have doubts about and I'd be happy to scrounge up the primary sources so we can debate the likelihood that they are valid.
FWIW, I'm 100% in favor of blind auditions. I would love to be able to reduce a candidate's identity down to a hash. I miss the internet back in 1995/6 when the only identity you had online was the one you created for yourself.