We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Diversity, Finally

00:00

Formal Metadata

Title
Diversity, Finally
Title of Series
Number of Parts
490
Author
License
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
What if we decided to solve, once and for all, the problem of underrepresentation in the Go community of women, gender minorities, people of color, or any other group the same way we handle our problems in production, by identifying "bugs" and then fixing them? Can it even be done? What if we took the engineering approach? Ronna is planning to convince you it's not a matter of if, but a matter of how, and we are going to analyze some of the statistics, find where the problems actually lay, and build a Trello card full of achievable tasks to address them.
Data conversionBitSet (mathematics)Computer animation
Computer programComputer programmingComputer animation
Group actionNumberSystem callComputer animation
MathematicianPosition operatorGraph (mathematics)GenderLevel (video gaming)Line (geometry)Computer animationDiagram
Archaeological field surveyChemical equationComputer animation
Variable (mathematics)Group actionSystem callComputer programmingVariable (mathematics)Group actionProgramming paradigmFormal languageComputer animation
GenderBinary fileArchaeological field surveyBuffer overflowArchaeological field surveyGenderState of matterStack (abstract data type)Ocean currentDependent and independent variablesStatisticsComputer animation
Archaeological field surveyArchaeological field surveyProcess (computing)InformationGroup actionPlastikkarteToken ringConfidence intervalEqualiser (mathematics)Product (business)Position operatorResultantSoftware testingCartesian coordinate systemMultiplication signNumberComputer configurationCore dumpCASE <Informatik>Observational studyFood energyStatisticsBitComputer animation
ResultantComputer animation
Electric currentFraction (mathematics)Profil (magazine)MultiplicationAutomatic differentiationCartesian coordinate systemFraction (mathematics)Computer animation
Process (computing)Degree (graph theory)Process (computing)Physical systemMetropolitan area networkComplete metric spacePrime idealObservational studyCartesian coordinate systemMultiplication signElectronic mailing listMarginal distributionPlastikkarteExpected valueFunctional (mathematics)Social classGroup actionStatisticsSpring (hydrology)Computer animation
CodeCodeComputer programmingConstructor (object-oriented programming)Group actionSet (mathematics)Scheduling (computing)Performance appraisalResource allocationSelf-organization
Binary fileFigurate numberThumbnailProcess (computing)Computer animation
Menu (computing)FeedbackJava appletComputer scienceMathematicsUniverse (mathematics)Data managementQuicksortProcess (computing)Decision theoryFigurate numberObservational studyShared memoryModal logicParameter (computer programming)Online helpTwitterGradientPlastikkartePoint (geometry)Multiplication signSelectivity (electronic)Source code
Point cloudFacebookOpen sourceComputer animation
Transcript: English(auto-generated)
Hi everybody. More to this side, okay. Wait, I'll move my computer then. Hi everybody. I challenge you, the Go community, to end, fix, solve, once and for all,
the discrimination and underrepresentation of women, people of color, people with disabilities, LGBTQIA, anybody, within one year.
That's all you get. One year. Is it achievable? I think so. I intend to show you that it probably is if we moved the conversation to a more pragmatic,
to more pragmatic solutions. Discuss the how instead of the if. Why a year? Because when you want to get something done, you set a deadline. Don't like my deadline? Set your own, but then you have to make it.
Because we're not likely to discover anything new next year about this topic, so why not just do something now? So a little bit about me.
I'm Rona, and I organize women who go Berlin. I also organize various scholarship programs for GoBridge. I run my own mentoring program, and I'm mainly an engineer, and I love people, opinions, code, and data.
So naturally, before giving this talk, I check to see what kind of impact I can expect from my own call to action. According to science, only 10% of you are likely to take action after hearing my talk. To those 10%, thank you. You're actually all I need.
If you implement some of the solutions that I'm going to lay out today, or even come up with your own, other companies will take notice, so thank you very much in advance. To the remaining 90%, I made that number up, so you have no excuse.
Don't listen to people who say according to science. So why do I do what I do? In my entire career, it's been 16 years, I reported to a woman exactly once for a few months directly, for a few months in 2004. It's been almost 16 years.
And as I become more and more senior, it's likely to never happen for me again, unless I go work for a very large company. Anybody hiring? And get very lucky. And this is what that looks like in a graph. I call this work of art the flat line.
I can therefore summarize... mathematicians. I can therefore summarize my biggest concern with the tech world, especially startups, back end, and ops with, really, that's where the glass ceiling is. Tech lead, team lead, the most junior leadership positions.
Well, it turns out that the glass ceiling is even lower than that. According to HackerRank's women's report from 2018, women are by far likelier to be in junior position than men.
Women as a gender are statistically stuck in junior positions. Some are truly stuck as juniors, are not promoted, not trained, or promoted very slowly, combined with a very, very serious retention problem where women across all levels are constantly leaving tech. In fact, according to a survey from November 2018, from Indeed,
the leading reasons for women to leave tech were lack of career growth, poor management, slow salary growth, lack of challenge, if you can believe that, and lack of work-life balance. Now, this is in multiple selections, as you can see, it adds to more than 100%.
So, lack of work-life balance was actually expected to be higher on the list, and this is surprising. Now, if you take all of those items, we can do something about everything on it. So, I can share countless stories to explain how we got to this point,
and I have researches coming out of my ears, but I only have 25 minutes. Sorry, call for papers committee for having to review my paper with all the data that I have on this topic. So, I want to focus instead today on one message, that like you, I'm in the business of solving problems,
and this is one that we can solve. I am, after all, an engineer, and this is what I do for a living. So, we can solve it using the same methodologies that we used to solve everything else. It's simple to understand, it's simple to break down. We can experiment and monitor our progress, and what else does a good troubleshooter actually need?
So, why am I coming to you with this? Simple. The Go community is modern. As a community, we've set aside paradigms that have shaped programming for three plus decades, changed our perspective, and adapted to a new way of doing things, to a new way of doing things.
We are not here to serve the language. The language is there to serve us. So, that makes me wonder why our variables deserve a safe space,
but our underrepresented groups don't. And what would happen if we treated ourselves as seriously as we treat our code? So, let's examine the current state of diversity in the Go community.
This statistic I took from the Stack Overflows survey of 2019. I selected the responders that worked with Go in the past year, and this is their gender breakdown. And this is the breakdown of their ethnicities, and I remind you, this is a global survey.
In Europe, the situation is a lot more dire. I couldn't bring myself to bring you the situation in the EU because I couldn't take away United Kingdom just yet. I just decided not to. So, a disclaimer. Most of my talk is going to be based on data about women,
about the underrepresentation of women in tech. I'd love to be able to talk more about the realities that other groups face, but sadly, there's just not enough data on that. I believe that some of the information that I'm going to lay out today is relevant to them as well. So, on a side note, I also want to point out something that I think is not addressed enough.
Diversity is not equality. And a diverse team is not necessarily an equal opportunity team. And I think it goes to the core of the problem where people do not want to be the tokens of our diversity. So, let's break down the problem.
Women leave tech while we're pursuing engineers. Whose problem is it? Well, it's very easy to think that it's my problem or people who care, you know, the ideologists, the activists. But I argue it's our problem as a community. First, because we're business-driven
and currently we're paying recruiters a fortune when there's probably available talent. How do we know there's available talent? First, depending on where you live, 60 to 80% of the population is not represented, is not participating. It's the vast majority.
Of course, there's talent. And second, because diversity is reported to increase individual success while improving creativity and eventual growth. So, at the very least, we're not benefiting there. So, now that we understand the problem, let's fix it. What were to happen if we were to create a Trello card
for diversity and add it to our tasks? What would that Trello card look like? The reason I'm turning this into a Trello card is because if any of our products was failing with women and minorities or with the vast majority of the population,
we would have solutions with AAB tests in production within a month. There is no way that we would let this go. Except we do have a failing product. It just happens to be the positions that we cannot fill. So, we should be very incentivized to fix this once and for all.
So, let's start with the biggest funnel. Everybody complains they don't get CVS from women. They are simply not applying. And you've probably heard about this one. A highly quoted HPE internal research has reported that women apply to jobs only if they meet 100% of the criteria
versus men with only 60% of the criteria. I'm happy to call BS. First, a little bit of background. This is known as the confidence gap between men and women. And it's amazing because HPE apparently were able to put
a number on the difference, on the confidence gap, 40%. 40% is massive, right? Might as well be 100%. So, I'm happy to inform you that this statistics is probably not scientific. First, there is no 100% in any real serious research.
It just doesn't exist. Two, if the HPE research even exists, it has never been published anywhere. We don't know who conducted it. We don't know what they asked. We don't know who they asked.
And we don't know if it was peer reviewed. We know nothing about it except it went viral. So, please do not continue telling everybody about this, okay? However, there seems to be some consensus in the industry that women truly don't apply as easily to jobs as men.
And there have been real studies. For instance, the one that was conducted by Dennings and Erlinger, the same Dunning from the Krueger-Dunning effect, if you're familiar. And they asked men and women to guess how much they scored in a test, and women did underestimate their scores more than men.
So, a woman named Tara Sophia Moore decided to ask men and women, if you decided not to apply for a job because you didn't meet all the qualifications, why didn't you apply? And her results were published in Harvard Business Review, and they are interesting.
She gave them five options to choose from. Not all of them are important, so I'm just going to read them quickly. I didn't think they would hire me, and I didn't want to waste my time and energy. I didn't think they would hire me, and I didn't want to put myself out there in case I fail. I was being respectful of the time and preferences of the person reviewing applications.
They had already made clear who they were looking for, and I didn't think I could do the job well. I was following the guidelines about who should apply. So, very good news. The least common answer for women was that they didn't think they could do the job well.
That is not why they didn't apply, or at least not the leading reason why they didn't apply. It was similar to men, 12.4%, but lower, we can rejoice in that. So, Moore explains, people who weren't applying believed they needed the qualifications not to do the job well,
but to be hired in the first place. They thought that the required qualifications were, well, required. Now, to be completely fair, we cannot say for sure that they thought they could do, that were confident they could do the job well, but it is not the reason why they are not applying. And in the troubleshooting world, Moore is right.
This is a lead, and we can try and poke it and see what kind of results we can get from it. So, I propose a few solutions. Eliminate everything that can be learned in two days or even a week. It probably doesn't matter. And if the applicants don't know the technology, they don't know how easy it is to learn,
why not experiment? Post multiple ads with multiple profiles and see who applies to what. At some point, you've accumulated a bunch of CVs. Not all of them meet 100% of your criteria.
Look at the least common denominator of the applicants that you're currently considering. It's probably the profile you're seeking anyway, and repost the ad. And you should probably rethink all the superlatives. Killer programmers, pretty obvious, but what about Ninja or Rockstar? Instead of, we're hiring the best, why not, we'll turn you into the best.
And while we're on the topic of confidence, it's being addressed disproportionately. Imposter syndrome, the self-doubt, all of these are well documented and discussed very often.
But I have not found a single research showing that if they applied, they'd actually get the job. So what if women statistically are not likely to get the job with the infamous 60% of the criteria? What if women, what if when a woman asks for a salary similar to a man's,
she wouldn't get the job in the first place? So I know we all think that we're the one good employer that doesn't discriminate, but how is she supposed to know that in advance? And I'm not the only one saying this. Several studies have already shown that women's self-promotion is tied to fear of backlash
for overconfidence, overconfidence. So what about degrees? Most that begin with a bachelor's degree qualification. And there's a real problem with that. We know there's discrimination in education system towards all underrepresented groups,
mainly people of color, women, LGBTQ, everybody. Campuses are not famous for being particularly safe, and young girls are silenced in classes worldwide.
This is a worldwide phenomenon. So if we were to consider everything that we're doing, we might want to look into that one as well. And finally, since women leave tech in masses, if we're actively looking for a senior, we're not likely to even get CVs from underrepresented groups.
And we're not likely to recruit them in the end. So the statistics is just not on our side. And it's impossible that everybody only hires seniors 100% of the time. So I challenge you then to do one additional thing. Every time you say go is simple, hire a junior.
So next on our Trello card is defining goals to our recruiters, especially if you're using external recruiters, you must set some expectations. I'll explain. External recruiters are paid a margin.
If they think that a man is likely to get a job, they will send you men. If they think that a man is likely to be paid better, they will send you men. So we must create an incentive for them to encourage diversity.
And whether you're using internal or external recruiters, you can define some quotas on the amounts of CVs that you're expecting. And this will allow you to also see what kind of CVs were being passed over so far. So back to our Trello card, go to meetups, find your applicants where they are.
When I started organizing Women Who Go Berlin, I didn't wait for people to come. I went everywhere. Women Who Code, Women Techmakers, Geekyard girls. And I, unlike you, didn't have money to offer them to learn Go. So you can be more convincing. And I got this question a lot.
Rona, we love what you're doing. How can we support you better? Hire my goddamn members, that's how. So next on our list is coding challenges. And they started with the thought that you might overlook someone
because they don't have the experience you're looking for. So there was supposed to be a way to allow applicants to surprise you, and it's a fine idea, it really is. However, companies are now sending out challenges that take days or even weeks to complete, and we need to revise this process.
Here's why. And this is true to probably all parents, but especially women and especially to mothers with babies. A young working mother can maybe allocate four to five hours a week to complete a challenge, and she's probably applying to other places as well. So if your challenge cannot be complete within four to five hours,
and I mean four to five hours, not spring points, not t-shirt sizes, four to five hours, you've probably no business sending it to anyone. But I'll take it a step further because now you're telling yourself, whew, hours takes only two hours to complete. If you're sending out the challenge with the expectation that you're going to hire
the person whose solution stands out, that is a function of time and resources. And our young working mother, again, cannot compete. So this is where we implicitly prefer young people, younger people with no commitments.
So consider this. People are likely to have children after spending a few years in the industry, after have been built a career. So we might be eliminating people at the primes of their careers. Our next item is, of course, retention. Let's retain our hires because otherwise we didn't achieve anything.
Code reviews are a great way to build and destroy trust between colleagues. So make sure to have constructive code review process. And this one is really important, training, education, budget, and conferences. So we see this everywhere. Conferences constantly have to allocate diversity tickets to underrepresented groups.
And then people like me distribute them. Set a clear budget and let your employees decide how they are going to spend it. Finally, define a clear path to promotion. Give your employees a sense of opportunity. Create a mentoring program in your organization.
Scheduled performance evaluation. Scheduled raises. You don't have to wait to be asked if you value your employees. And consider also anonymizing the CV process and your coding challenges, if you choose to keep them. And let's talk about the thumbs up process. Over the past 15 years, companies have been making interview processes harder and longer.
And if before you would interview with two, maybe three people, today you're likely to be interviewed by at least four to six people. And they all have to give the thumbs up. But we just doubled the chances that one of your interviewers is biased.
So finally, people should feel safe at work, physically and emotionally. If you have any kind of bullying at work, on any background, any background, end it. It is inexcusable. Figure it out.
And here's the promised Trello card with all of the ideas I just explained. This is not everything. You might come up with your own ideas. And if you do, please share them back with me. Share a tweet, tell your friends, your recruiters, and your managers. Be brave. And thank you very much.
Cool. So we have some time for questions. So any questions in the audience? You're really far.
Thank you for coming this way, yes. So first a short disclaimer. I'm white, cis, and a man. So I'm obviously very privileged to even talk about anything in this subject. I don't consider myself a particular defender of the cause, but I try to help whenever I can.
And one thing that comes up pretty often for me is that whenever I start any discussion in this area, it's very hard to point out what problem you're trying to solve. Like, why is it a problem? I kind of see it for me, but it's really hard for me to try to convince someone that just shrugs and says,
yeah, why is it my problem? Why should I even care? So do you have any short and to-the-point arguments that could help me when I'm in this sort of discussion to say, hey, look, this is something that we should try to solve, even if we're in a very privileged position and don't see it as a day-to-day problem? Well, one there is the recruitment.
So I can tell you that I mentored about 100 people to date, and they're all incredibly talented. And I'm enjoying what's called the selection bias, where all the people who are brave enough to come to me and ask for help are incredibly talented and probably could make it on their own. So think about this.
If you started your own mentoring program, you could probably hire 100 people tomorrow that would be incredibly talented and very, very good. That is one. Two, my first job was actually, and this was 2003, I was doing C++, and at some point we were a majority of women. To me, it's a no-brainer, because I know what that workplace looked like, and it was quite amazing.
I mean, I understand that people speak from experience and do not understand what a diverse workplace is and therefore do not understand necessarily the necessity. They will just figure it out along the way, because it is nicer. It is just nicer.
For the people in the back, we can hear you perfectly from here. Thanks. So most of the things you mentioned are about the hiring process, but you didn't say a lot about actually the education problem, right? There are a lot less women actually studying computer science and such.
Can you elaborate about that, please? So this changes from country to country, but we see worldwide, a worldwide trend that women do sign up less. Those studies are also hard to get to, and if you are – okay, a curious thing.
Do I have – how long do I have? You have one minute. I have one minute? Yes. I have one minute. Okay, I'll try to make it short. In my country, in my home country, I come from Israel. There's something called the sacrometry. It's an – when you try to enter the university, you have to do that.
It weighs 50% of the score that it's actually going to send you to the university. Women do pretty well for 12 years in my own country, do pretty well and do way better and have better grades than men during the 12 years. Men do better on the sacrometric exams.
Why 50%? I'm always puzzled. Like, why not 40%? Why not 80% if you're sure that it's really that important or it's more important or it's as important as 12 years of grades, why not that? Those things are so arbitrary that we don't even see them. That's the answer, basically.
We need to poke at every step of the way and understand what decisions we've made. And nobody's ever going to change it, you know. Nobody. This is not even – nobody's thinking about these things. So that's the problem. And I'm just going to go away now because I'm being kicked out. Thank you. Thanks so much.