We built voice modulation to mask gender in technical interviews • Hacker News Board

Re: We built voice modulation to mask gender in technical interviews

#201

Post by t0mbstone » Wed, Jun 29, 2016, 8:45 PM UTC

leeny wrote: post #39

benkuhn wrote: post #15

I would really have liked to see them check whether the gender-blinding actually worked. Even if someone's voice is modulated, they may have gendered behavior patterns that could influence someone's performance ratings. I wouldn't be surprised if I could guess the gender of a voice-modulated person pretty accurately from other cues. To combat this, they could have asked the interviewer during the performance assessme…

Aline from interviewing.io here. Great idea!

You all should record the audio from the original modulated interviews, then play them in both male and female versions to sample sets of separate observer groups that rank the performance of the person being interviewed.

That way, you would be isolating the thing you are testing to just the voice, and you would be able to establish controls and a baseline for the exact same interview.

The way your test is described right now, there are still way too many factors that could explain the results. Your controls are far too loose.

Re: We built voice modulation to mask gender in technical interviews

#202

Post by stcredzero » Wed, Jun 29, 2016, 8:45 PM UTC

Spivak wrote: post #191

bduerst wrote: post #132

Earlier quoted context omitted.

This is deceptive because it's an opinion piece on Youtube. The BLS study that controlled for occupation, industry, part-time vs. full-time, education levels, and other factors found that there is still a gender pay wage gap of 94 cents to the dollar [1]. It's not-fact-based to keep touting the 77% number, but it's just as much not-fact-based to claim there isn't a gender wage gap because women feel "squeezed". [1] h…

But he brought that up in the video with the "unexplained pay gap" that still remains when you control for known discrepancy causing variables. I don't think people are being deceptive arguing for the non-existence of a gender pay gap since controlling for reasonable variables wiped away 80% of the discrepancy. There's no strong explanation for the remaining 20% so it's really just an argument over whether or not the…

I don't think people are being deceptive arguing for the non-existence of a gender pay gap since controlling for reasonable variables wiped away 80% of the discrepancy.

I would say that the numbers very strongly show there is a 6% pay gap. If we assume the job market is efficient, then either there is some bias against women or women slightly under-perform men or women are worse at negotiating salaries or they are receiving some renumeration in the form of greater flexibility.

There's no strong explanation for the remaining 20%

Yet. I don't find it intellectually satisfying to say something doesn't exist just because we don't have a theory that covers it well. The disparity is smaller, but it does exist. I also observe how hectic the schedules of professional women with children appear to be. I'm not sure we completely understand what's going on there, but I don't think the sentiments of women around this issue are entirely disconnected from objective reality. From what I can see, women are "squeezed" in terms of the resource of time.

Re: We built voice modulation to mask gender in technical interviews

#203

Post by stcredzero » Wed, Jun 29, 2016, 8:50 PM UTC

malandrew wrote: post #139

stcredzero wrote: post #95

Earlier quoted context omitted.

If I were to start a company today, I'd definitely be tempted to only hire women, and pay my entire workforce 80 cents on the dollar, instead of 70 cents on the dollar. Except that correcting for experience, the pay gap shrinks to about 94 cents on the dollar. https://www.youtube.com/watch?v=it0EYBBl5LI I'm with John Green here: Much of the strong negative feeling around the gender pay gap comes from women as a whole…

What is the outsized unpaid work? Any studies and do the studies account for different cohorts by age range?

We already know there is a disparity in the amount of housekeeping between men and women, even in couples where both men and women are working.

Re: We built voice modulation to mask gender in technical interviews

#204

Post by omginternets » Wed, Jun 29, 2016, 8:53 PM UTC

Jemmeh wrote: post #117

jsprogrammer wrote: post #108

Earlier quoted context omitted.

The NYT blog post starts out discussing regional dialects, then presents a study of 23 Southern Californians aged 18-22... The only study the post links to is paywalled and the link to a conference paper is dead. Your second link is also paywalled.

I've found most studies to be paywalled in general aside from the free abstract. Not really sure what to do about that.

Scihub, though I can't find a functioning mirror right now.

Re: We built voice modulation to mask gender in technical interviews

#205

Post by 1_2__3 » Wed, Jun 29, 2016, 8:56 PM UTC

benkuhn wrote: post #15

I would really have liked to see them check whether the gender-blinding actually worked. Even if someone's voice is modulated, they may have gendered behavior patterns that could influence someone's performance ratings. I wouldn't be surprised if I could guess the gender of a voice-modulated person pretty accurately from other cues. To combat this, they could have asked the interviewer during the performance assessme…

Translation: "Hope they keep doing this experiment until it matches my expectations!"

Re: We built voice modulation to mask gender in technical interviews

#206

Post by playerqq » Wed, Jun 29, 2016, 8:57 PM UTC

onion2k wrote: post #168

xienze wrote: post #147

I honestly feel like this sort of thing is a waste of time. Let's pretend someone invents the perfect "gender bias-free" system: no names, no faces, voices are flawlessly transcribed on-the-fly so you can't pick up on speech patterns. What if it works perfectly and you still end up hiring more men than women? I think exercises like this may end up giving people answers they don't want to hear. So, stop wasting time w…

For some companies who believe an equal workplace should be 50% of each gender then you're right, but that doesn't make the system pointless. The idea of a "gender bias-free" hiring process isn't necessarily to make workplaces more equal or to promote diversity. The point is simply to remove a bias - to eliminate something that pushes you to hire someone who isn't the best candidate. Having a bias drives you to make…

But let's say our "gender-bias free" method leads a manager to hire 7 men and 3 women. This leads the manager to draw the conclusion that "My next hire is more likely to be a male than a female". The manager wants to acquire a good candidate pool as quickly as possible, so now he changes his mental search heuristic to favor male applicants, if he can discern their gender.

Haven't we just reinforced his bias further, now cemented by hard fact (and faulty reasoning)?

edit: missed a word

Re: We built voice modulation to mask gender in technical interviews

#207

Post by 1_2__3 » Wed, Jun 29, 2016, 8:58 PM UTC

Jemmeh wrote: post #194

mpweiher wrote: post #176

Earlier quoted context omitted.

Should "we" also try to figure out why large amounts of women have less speed compared to men, and work on that, too?

Sorry, I went back and edited my comment but you had already posted. The reason we need to figure out the confidence gender gap issue is because it leads to less opportunities for women. Anyone who is less confident is going to have less opportunities. But for various reasons, women tend to have a lot less confidence than men. Yes, that's something that is partially internal to each person. But the fact that large gr…

You're not actually answering his question.

Re: We built voice modulation to mask gender in technical interviews

#208

Post by amelius » Wed, Jun 29, 2016, 8:58 PM UTC

xienze wrote: post #147

I honestly feel like this sort of thing is a waste of time. Let's pretend someone invents the perfect "gender bias-free" system: no names, no faces, voices are flawlessly transcribed on-the-fly so you can't pick up on speech patterns. What if it works perfectly and you still end up hiring more men than women? I think exercises like this may end up giving people answers they don't want to hear. So, stop wasting time w…

> What if it works perfectly and you still end up hiring more men than women?

I think the idea is that the company can then say: but we used a gender-mask, so you can't blame us. A form of plausible deniability.

Re: We built voice modulation to mask gender in technical interviews

#209

Post by randyrand » Wed, Jun 29, 2016, 8:59 PM UTC

Someone off-topic but: Laws forbidding sexual and racial discrimination in hiring are a waste for everyone's time involved.

Do I really want to work alongside someone that "isn't racist" or sexist just because the law forbids it? No of course not. I'd rather work next to people that aren't racist and sexists regardless of if the law forbids it. I want to work at companies where people actually want me. How difficult is that to understand? =(

Most people want less racism/sexism. But this is not the way to do it. I'm in favor of removing this law and letting discriminators be upfront about it. So that I can knowingly avoid them without wasting hours and potentially years of my life.

Re: We built voice modulation to mask gender in technical interviews

#210

Post by smadge » Wed, Jun 29, 2016, 9:02 PM UTC

omginternets wrote: post #175

xienze wrote: post #147

I honestly feel like this sort of thing is a waste of time. Let's pretend someone invents the perfect "gender bias-free" system: no names, no faces, voices are flawlessly transcribed on-the-fly so you can't pick up on speech patterns. What if it works perfectly and you still end up hiring more men than women? I think exercises like this may end up giving people answers they don't want to hear. So, stop wasting time w…

I'm always surprised by the strong underlying assumption that removing interviewer/managerial bias will result in a gender ratio of 50%. There's research to suggest that this assumption is flawed: [0] Moreover, the idea that anatomy and physiology don't condition preferences or behaviors flies in the face of everything we know about human biology and (embodied) cognition. Let's recognize gender/feminist theory for wh…

Why are you answering nature vs nurture with (I paraphrase) "100% nature and anyone who says otherwise is an irrational charlatan"? This debate is cliched at this point.

From the study you linked to:

"As suggested by previous literature and the present study, interests stabilize at an early age. Given the essential role of family and other environmental influences in the socialization process of children and their interest development, it may be important for parents, educators, and counselors to be involved early on, when interests seem more malleable."

i.e. the study accepts the theory that environment, family and in general socialization, contribute to a child's interests. This is the exact opposite of what you claim in your comment

"It is important to note that group differences do not speak for every individual within a group,and it is not the purpose of the present study to reify the stereo-types that men and women are interested in their “traditional”areas."

The study is talking about d "it is beyond the scope of the present study to provide a detailed exploration of the environmental, social, and biological factors that have contributed to the development of these sex differences"

The study makes no claim as to the cause of these sex differences. They only point out that there is a difference and suggest "environmental, social, and biological factors" as the cause. Again, the opposite of what you claim.