That's actually quite clever, wrt. generating everyone who applies a nonsense pseudonym so there's no unconscious bias happening in the pre-selection process.
It is, but what I've found a bit funny about the entire recent 'diversity' discussion is that such things aren't already done as a matter of course.
We're in an industry where modeling, A/B testing, logging, trending and engineering for redundancy is implicit - except when it comes to hiring apparently.
If a load balancer assumed to split 50/50 was actually operating at 80/20 someone would notice. If a particular demographic were bouncing from the landing page at disproportionate rates, someone would figure out why.
Both problems would be broken down empirically and solved with a thorough post-mortem published on the company blog.
To this point, those same sort of issues applied to hiring have been largely addressed with "We've got to do better" and "The problem is upstream" with barely a word about how things managed to creep so far out of whack to begin with.
It's really refreshing to see admission of an actual (potential) problem and some of brains being applied to mitigate it directly.
Good point. One possible explanation is that in many cases it's currently seen as sufficeint to say the right things about diversity and do enough to show that you're making an investment. Actually improving diversity requires significant changes that can be very uncomfortable; and is typically fiercely resisted by many (although certainly not all!) of the people who have the advantage in non-diverse environments. So the problems don't get attacked in the same way a company would approach things they truly see as business priorities.
Several years ago I had the charter for "game-changing strategies" at Microsoft, and some of what we came up was diversity-related. When I give the quick thumbnail sketch of my career, I sum this stage up by saying "the company didn't really want to change". And everybody nods sadly.
Just about anything when it comes to social equality is about as far divorced from objectivity as it's humanly possible to be. Too many emotions running hot, too many knee-jerk decisions made.
It is refreshing to see an actually objective, fact-based, and mostly well thought out approach to solving the problem.
As an industry, we used to be in favor of approaching meritocracy as a goal. We'd take pride in using github in lieu of a resume, difficult technical tests, and general attempts at objectivity.
Strangely, the pushback against meritocracy as a goal and objective measurement as a method doesn't come from the alleged "good old boy" network.
- A has a created small number of functional and elegant projects
- B has created a large number of functional but sloppy projects
Which is objectively better? The candidate who is more productive, or the candidate whose products have higher quality?
One may be a better fit for your organization, based on your group's needs and existing makeup, but that doesn't mean they are objectively "better" than the other candidate. Do you believe one of these candidates has more "merit" than the other?
Some people who believe a person has objective merit (as opposed to suitability for a particular organization or role) are actually expressing an aesthetic preference (selecting for tidy code, or work ethic, etc) while claiming/believing that their preference is "objective" truth. People opposed to using "objective" or "meritocracy" as labels are arguing for more nuance in understanding the needs and biases of people and teams.
Wait, I can go one better. Candidate A) has a large number of elegant projects. Candidate B) is a ninja with a large number of high profile assassinations under his belt.
Obviously the question of "objectively better" comes with an implicit "for the role in question".
People opposed to using "objective" or "meritocracy" as labels are arguing for more nuance in understanding the needs and biases of people and teams.
That sounds pretty but it's totally vacuous - it doesn't give me any information on how the decision process would be altered.
People opposed to using "objective" or "meritocracy" as labels are arguing for more nuance
This sentence is an oxymoron. An objective hiring process is one that looks at facts, more specifically those related to job performance and past performance with related tasks. Those are, IMO, quite literally the only things that should factor into the process since at the end of the day they're the only things that matter.
Coincidentally, this is described as a meritocracy. A process where you're judged solely on what you do, not what you are.
That's completely non-responsive to the question of hiring. Salary negotiations usually come after acceptability as a candidate has been established. (Indeed, it's considered a faux pas to bring it up before then)
The goal of any hiring process would seem to be to acquire the best possible candidate for the position in question, where "best" is usually determined by factors such as job performance and culture fit.
Note that if you're trying to come up with what the definition of "best" is as a rule that can be applied generally, you lose already, since that will vary based on what the job is, where the job is located, who you're working with, and many many other factors.
"The word meritocracy comes from a political satire. It was never meant to be something we should aspire to. It was the opposite, actually, a warning about how we rationalize what we believe we’ve 'earned'. If that sentence doesn’t seem to you applicable to the tech industry and our cyclical discussions about sexism, racism, and even occasionally classism, go get yourself another cup of coffee."
Thanks! I'm actually planning on writing more about what we've learned from anonymizing our applications and from our application process in general in the next few days. I think some of the things we've learned could probably also be applied to hiring at a company with some modification (it's harder to anonymize resumes and emails than it is to anonymize a standard application).
Another side effect of anonymized applications is that we get to read a lot of funny application names (Beet Manager, Pastry Magician, Bean Pain).
Since there is a well documented hiring[1] bias against people with both female-sounding-names and non-white-sounding names, I think anonymized applications is an amazing idea! So many people never make it to even a phone screen because of insidious biases.
Honestly, I thought it was embarrassing. The idea that you can't control your feelings enough to judge someone based on merit - even when you're conscious of your own bias - reveals an incredible lack of emotional intelligence. It seems like a gimmick created to soothe the nerves of social justice activists.
It was very strange to me that anyone would focus so much on diversity as much as Hacker School has. So I was looking around their site, and I suddenly found the quote that explains their near-obsession: "We make money by helping companies hire our alumni." Suddenly it all makes sense! Companies need to fill diversity quotas, so they turn to Hacker School to get as many women, hispanic/latino and black people as they can.
There's plenty of research pointing to the fact that some racial bias is actually unconscious. Conscious or not, this seem like a perfectly reasonable way to avoid it.
Your second conjecture is a bit tin-foil-hat to even address, but here goes: why does it seem strange to you that a prominent organization in the technology community would focus on diversity? Diversity breeds innovation - the more diverse the tech industry, the better it is for all of us (and by all of us, I'm referring to the entirety of the species). Diversity is a worthy goal in and of itself. There doesn't need to be an ulterior motive.
Comments
That's actually quite clever, wrt. generating everyone who applies a nonsense pseudonym so there's no unconscious bias happening in the pre-selection process.
It is, but what I've found a bit funny about the entire recent 'diversity' discussion is that such things aren't already done as a matter of course.
We're in an industry where modeling, A/B testing, logging, trending and engineering for redundancy is implicit - except when it comes to hiring apparently.
If a load balancer assumed to split 50/50 was actually operating at 80/20 someone would notice. If a particular demographic were bouncing from the landing page at disproportionate rates, someone would figure out why.
Both problems would be broken down empirically and solved with a thorough post-mortem published on the company blog.
To this point, those same sort of issues applied to hiring have been largely addressed with "We've got to do better" and "The problem is upstream" with barely a word about how things managed to creep so far out of whack to begin with.
It's really refreshing to see admission of an actual (potential) problem and some of brains being applied to mitigate it directly.
Good point. One possible explanation is that in many cases it's currently seen as sufficeint to say the right things about diversity and do enough to show that you're making an investment. Actually improving diversity requires significant changes that can be very uncomfortable; and is typically fiercely resisted by many (although certainly not all!) of the people who have the advantage in non-diverse environments. So the problems don't get attacked in the same way a company would approach things they truly see as business priorities.
Several years ago I had the charter for "game-changing strategies" at Microsoft, and some of what we came up was diversity-related. When I give the quick thumbnail sketch of my career, I sum this stage up by saying "the company didn't really want to change". And everybody nods sadly.
Just about anything when it comes to social equality is about as far divorced from objectivity as it's humanly possible to be. Too many emotions running hot, too many knee-jerk decisions made.
It is refreshing to see an actually objective, fact-based, and mostly well thought out approach to solving the problem.
As an industry, we used to be in favor of approaching meritocracy as a goal. We'd take pride in using github in lieu of a resume, difficult technical tests, and general attempts at objectivity.
Strangely, the pushback against meritocracy as a goal and objective measurement as a method doesn't come from the alleged "good old boy" network.
http://readwrite.com/2014/01/24/github-meritocracy-rug
http://www.ashedryden.com/blog/the-ethics-of-unpaid-labor-an...
Consider 2 candidates with github profiles:
- A has a created small number of functional and elegant projects
- B has created a large number of functional but sloppy projects
Which is objectively better? The candidate who is more productive, or the candidate whose products have higher quality?
One may be a better fit for your organization, based on your group's needs and existing makeup, but that doesn't mean they are objectively "better" than the other candidate. Do you believe one of these candidates has more "merit" than the other?
Some people who believe a person has objective merit (as opposed to suitability for a particular organization or role) are actually expressing an aesthetic preference (selecting for tidy code, or work ethic, etc) while claiming/believing that their preference is "objective" truth. People opposed to using "objective" or "meritocracy" as labels are arguing for more nuance in understanding the needs and biases of people and teams.
Wait, I can go one better. Candidate A) has a large number of elegant projects. Candidate B) is a ninja with a large number of high profile assassinations under his belt.
Obviously the question of "objectively better" comes with an implicit "for the role in question".
People opposed to using "objective" or "meritocracy" as labels are arguing for more nuance in understanding the needs and biases of people and teams.
That sounds pretty but it's totally vacuous - it doesn't give me any information on how the decision process would be altered.
This sentence is an oxymoron. An objective hiring process is one that looks at facts, more specifically those related to job performance and past performance with related tasks. Those are, IMO, quite literally the only things that should factor into the process since at the end of the day they're the only things that matter.
Coincidentally, this is described as a meritocracy. A process where you're judged solely on what you do, not what you are.
The problem with "meritocracy" is that the judging mechanism is inherently vulnerable to bias.
So therefore we shouldn't try? What should our goal be instead?
http://www.ncsociology.org/sociationtoday/v21/merit.htm http://valleywag.gawker.com/the-myth-of-silicon-valley-merit...
As for what our goal should be instead: equal pay for equal work.
That's completely non-responsive to the question of hiring. Salary negotiations usually come after acceptability as a candidate has been established. (Indeed, it's considered a faux pas to bring it up before then)
The goal of any hiring process would seem to be to acquire the best possible candidate for the position in question, where "best" is usually determined by factors such as job performance and culture fit.
Note that if you're trying to come up with what the definition of "best" is as a rule that can be applied generally, you lose already, since that will vary based on what the job is, where the job is located, who you're working with, and many many other factors.
"The word meritocracy comes from a political satire. It was never meant to be something we should aspire to. It was the opposite, actually, a warning about how we rationalize what we believe we’ve 'earned'. If that sentence doesn’t seem to you applicable to the tech industry and our cyclical discussions about sexism, racism, and even occasionally classism, go get yourself another cup of coffee."
from http://www.garann.com/dev/2012/you-keep-using-that-word/
Thanks! I'm actually planning on writing more about what we've learned from anonymizing our applications and from our application process in general in the next few days. I think some of the things we've learned could probably also be applied to hiring at a company with some modification (it's harder to anonymize resumes and emails than it is to anonymize a standard application).
Another side effect of anonymized applications is that we get to read a lot of funny application names (Beet Manager, Pastry Magician, Bean Pain).
Since there is a well documented hiring[1] bias against people with both female-sounding-names and non-white-sounding names, I think anonymized applications is an amazing idea! So many people never make it to even a phone screen because of insidious biases.
[1] seriously, so many examples, https://www.google.com/search?q=research+bias+hiring+names
Honestly, I thought it was embarrassing. The idea that you can't control your feelings enough to judge someone based on merit - even when you're conscious of your own bias - reveals an incredible lack of emotional intelligence. It seems like a gimmick created to soothe the nerves of social justice activists.
It was very strange to me that anyone would focus so much on diversity as much as Hacker School has. So I was looking around their site, and I suddenly found the quote that explains their near-obsession: "We make money by helping companies hire our alumni." Suddenly it all makes sense! Companies need to fill diversity quotas, so they turn to Hacker School to get as many women, hispanic/latino and black people as they can.
There's plenty of research pointing to the fact that some racial bias is actually unconscious. Conscious or not, this seem like a perfectly reasonable way to avoid it.
Your second conjecture is a bit tin-foil-hat to even address, but here goes: why does it seem strange to you that a prominent organization in the technology community would focus on diversity? Diversity breeds innovation - the more diverse the tech industry, the better it is for all of us (and by all of us, I'm referring to the entirety of the species). Diversity is a worthy goal in and of itself. There doesn't need to be an ulterior motive.