[sqrt_1]

This reminds me of when the Australian government did blind recruitment.

"Blind recruitment means recruiters cannot tell the gender of candidates because those details are removed from applications."

"The trial found assigning a male name to a candidate made them 3.2 per cent less likely to get a job interview.

Adding a woman's name to a CV made the candidate 2.9 per cent more likely to get a foot in the door."

https://www.abc.net.au/news/2017-06-30/bilnd-recruitment-tri...

[throwawaylinux]

The intersectionality hypothesis is just constantly destroyed by evidence like that, and this voice modulator experiment. And yet it seems to be evangelized more fervently than ever in academic institutions, governments, journalism, etc.

It's really a weird, strange kind of twilight zone. Alleged experts are just spouting totally unfounded (and even contradicted by evidence) assertions and stating them as fact. Some of them really hurtful and hateful even. And few dare to call them out, demand evidence for their unproven claims, or demand metrics and results of improvements they claim their policies would result in. It's like a nutty cult that's being ruled by fear.

In my opinion it is also profoundly damaging to the cause of actually improving inequality and reducing discrimination and reducing division.

[Vt71fcAqt7]

>Alleged experts are just spouting totally unfounded (and even contradicted by evidence) assertions and stating them as fact.

This is because their principal theory of demographics — that all group differences are entirely due to one group's use of power to undermine the other — is wrong. There are many factors involved in group differences, and they are all being completley ignored because of political correctness. Even to say that sexism sometimes only indirectly causes disparity in enmplpyment - as an example, teachers/mentors investing more time in boys - is becoming problematic because it posits that interviewers themselves may possibly not be overtly or implicitly sexist. The idea that not only the "system," but each member in the powerful group is out to get all minorities/women, is ridiculous and, as you say, hateful.

[throwawaylinux]

> The idea that not only the "system," but each member in the powerful group is out to get all minorities/women, is ridiculous and, as you say, hateful.

It is, and you see it in this person's comment. It was actually surprising to them that they did not observe discrimination. They have been indoctrinated to believe without evidence that the society they live in is a sea of hateful and/or ignorant bigots who oppress and discriminate. The oppressed classes are indoctrinated to believe they are under constant attack and oppression by their neighbors and friends and colleagues. And the oppressor classes believe they hold some responsibility and shame for these real or imagined atrocities. It's just a hell of a way to go through life, and I can't see anything much good coming from it.

I'm not saying for a minute that discrimination does not exist, that it should not be addressed and reduced, or that it can not concentrate in organizations and positions of power, before people start accusing and denouncing me.

[faeriechangling]

What is the “intersectionality” hypothesis? I felt intersectionality was a concept so obvious, that intersecting identities create unique experiences that might not be shared by other members of the constituent identities, as to be unremarkable.

The example frequently given is the proverbial workplace that in order to hire enough black people and women, hires many black men on the factory floor and white women in the offices while black women don’t get hired and despite this people declare that equity has been accomplished. People paid attention to women’s issues and black issues, but not the INTERSECTION of those issues.

My criticism is that intersectionality is that it’s blindingly obvious and there is limited value in literally deviding up identity groups into an exponentially expanding amount of subgroups. Once you start exploring the intersection of 3 identities you’ve already kind of lost the plot and may as well start talking about issues on the personal/individual level.

I have no idea what people THINK intersectionality means. I think people who know what that word means keep using it without explaining it and people get really confused and think it’s just a general term for woke ideology and never stop to think what the “intersection” they’re referring to is.

[jwarden]

> Once you start exploring the intersection of 3 identities you’ve already kind of lost the plot and may as well start talking about issues on the personal/individual level.

At the limit, intersectionality approaches individualism.

[SpicyLemonZest]

Most people in my circles understand "intersectionality" to refer to exactly the claim you're disputing, that it's good to identify and focus on an exponentially expanding amount of subgroups. I don't think this is an unreasonable interpretation - the Center for Intersectional Justice, a large nonprofit in the area, says one example of intersectionality (https://www.intersectionaljustice.org/what-is-intersectional...) is that you shouldn't attempt to address the gender pay gap without also thinking about race, socioeconomic status, and immigration status.

[faeriechangling]

I mean it’s cool to propose that and all but how does that work in practice? At some point you need to be a “big tent” to get things done.

How do trans immigrant women feel about the gender wage gap? How do male Irish steelworkers feel about it? So on and so forth… and don’t you dare tell anybody their identity group and their unique intersection of those identities isn’t important.

It’s an ideology I would try to convince my enemies to adopt to encumber them with endless amounts of petty political bickering so they never got anything done.

[dudul]

This last paragraph is totally a conspiracy theory I believe.

This whole intersectionality business is spreading way too fast, being too prominent in the media and front and center everywhere to not be at least encouraged by an hostile nation.

[sidlls]

The use of propaganda and indoctrination to play sub-factions against each other in order to keep the larger group at a disadvantage is a tried and true tactic used by those with power since the beginning of recorded history. This specific implementation may or may not be true, but it’s hardly irrational to suspect.

[hjgjhyuhy]

Either a hostile nation, or economic elite trying to turn people’s attention away from the real issues of income inequality and taxation. Race, gender, whatever, in the end money is what primarily makes things right.

[random314]

There are experiments in the opposite direction too if you are feeling really upset. The claims are not unproven and this was not a properly randomized experiment unlike the far more extensive one I cite below

https://www.wbur.org/hereandnow/2021/08/18/name-discriminati...

[throwawaylinux]

The claim is not that discrimination exists. Of course it does.

The claims basically amount to any inequity between groups is due to discrimination, except in the cases where the inequity "favors" the oppressed class (and where oppressed and oppressor classes have no real objective criteria or definitions, aside from what intersectionality experts claim them to be).

And that institutional policies of discrimination are an the acceptable and effective way to deal with this inequity, but also you must not refer to them as systemic or institutional discrimination.

[reverius42]

What is the intersectionality hypothesis? This is a phrase I'm not familiar with.

[pcthrowaway]

I think the person you're responding to is referencing intersectional feminism (more recently mainstream in social justice and feminism discussions), which is perhaps somewhat related to what's being discussed, but really not the topic at hand.

What is relevant is that the general discourse focuses more on systemic differences in opportunity to disadvantaged groups. In this case, we're talking about women as the ostensibly disadvantaged group.

I think what people fail to understand though, is that marginalization can take many forms. If women are less likely to continue interviewing after rejection, this doesn't mean that they have the same opportunity; instead, it suggests that systemic factors play into this. The "intersectional" take on this is really focused on how different components of identity work together to shape an individual's experience, in subtle ways that compound. Hence, just growing up in a society where young women and men have even mildly different attitudes impressed upon them from a young age, can have more noticeable effects on their outcomes much later in life

[parineum]

>it suggests that systemic factors play into this

No it doesn't. This is a false dichotomy. It can be any number of things, including something systemic, but only one explanation of many is eliminated. All that it suggests is that interviewers don't discriminate (against women).

[adamsb6]

That any disparity between any demographic groups is wholly attributable to bias.

[Spivak]

That's not at all what intersectionality is. It's just a name for the fact that the hardships you face as, say, a black woman isn't the union of the hardships of women and the hardships of black folks. They don't exist independently. This was an important idea because feminist movements focused almost exclusively on white women and it assumed this was fine because "rising tide raises all ships" and that white women had a better chance of having their voices heard -- but it didn't shake out that way and this was the reason.

That's the reason today there's a lot of focus on black trans women because activists are trying not to repeat this mistake.

> any disparity between any demographic groups is wholly attributable to bias

Unless you're literally arguing that some groups are genetically superior or inferior, or that something in the Y chromosome just really draws you to programming this statement is true by definition -- that's what bias means.

If you’re like oh well part of the wage gap is explained by women not being as aggressive in salary negotiations your options are either that there is a bias somewhere or resorting to gender essentialism.

[Der_Einzige]

What the hell is up with HN today? Why is this downvoted so much? Get ahold of yourself HN users!

[BurningFrog]

In other words: correlation is causation.

[GolDDranks]

I haven't heard of it either, but "intersectionality" means (according to Google, but in my experience, sounds accurate.)

> Intersectionality is a theoretical framework rooted in the premise that human experience is jointly shaped by multiple social positions (e.g. race, gender), and cannot be adequately understood by considering social positions independently.

[macrolocal]

This feels like saying that (generalized) linear regressions without interactions are inadequate. And criticisms of intersectionality seem to be that (a) high order interactions are statistically volatile and (b) multivariate averages miss important information.

Is it deeper than that?

[8note]

afaik its a restatement of Simpson's Paradox[0]

If you're a member of two different minority groups, your experience is not described by taking a union of the bad things from each. Eg. If being black results in 3% less <> and being a woman results in 5% less <>, then being a black woman does not result in 8% less <>.

[0]: https://youtu.be/ebEkn-BiW5k

[lioeters]

https://en.wikipedia.org/wiki/Intersectionality

[chitowneats]

The hypothesis is that women, racial minorities, etc, are discriminated against universally in modern society.

[peyton]

You don’t see the behind-the-scenes stuff. There are a lot of threats.

[thrown_22]

Pygmies cast long shadows at sunset.

This is not people interested in truth, this is people who want to control how you think.

No different to the racists of yesteryear, especially since the current wokesters are the grandchildren of klan members.

[faeriechangling]

Citation please, I want to see these wokesters with family ties to the klan.

[tomjakubowski]

> the current wokesters are the grandchildren of klan members

uh, what?

[samplatt]

>racists of yesteryear

Or the racists of currentyear.

>the current wokesters are the grandchildren of klan members

No they're not. I'm definitely not.

[nerdponx]

How do you know that this result is not caused specifically by HR people attempting to counteract discrimination?

[seer-zig]

> And yet it seems to be evangelized more fervently than ever in academic institutions, governments, journalism, etc.

When you know the agendas and world views of those same people, everything starts to make sense. It's quite unfortunate to say the least, but hopefully more people wake up.

[carabiner]

3% more/less likely is a tiny effect. What is the power of this result? Isn't there always noise in this type of thing?

I bet the normal variation in responses for the same group submitted to separate groups of companies exceeds 3%. That is, I could send 1,000 resumes to one batch of companies and 10.0% would get interviews. I could send the same 1,000 resumes to a different batch, and 10.3% would get interviews. Boom 3% difference for the same candidates.

[timr]

> 3% more/less likely is a tiny effect. What is the power of this result? Isn't there always noise in this type of thing?

A result does not have "power". An experiment has power -- the ability to detect a a given effect size a certain percentage of the time -- but a result is either statistically significant, or it is not.

As for "noise", statistical significance takes random noise into account. That is the point of the calculation -- it asks if a given result exceeds the threshold of what you'd expect to find at random some percentage of the time. If it does, the result is deemed significant.

A 3% difference could be enormous, or it could be miniscule. We can't say anything based on this information alone, and certainly can't say it's "likely a tiny effect". On a sample of thousands, a 3% difference is big. On a sample of tens, a 3% difference is small.

[LordDragonfang]

>On a sample of thousands, a 3% difference is big.

Not really. Only if it is many, many thousands. Assuming a totally random acceptance rate of 1/5:

   a = 0; 
   b = 0;
   for (c of Array(1000)) {
       if (Math.random() > .8) 
           a++;
       if (Math.random() > .8) 
           b++;
   }
   console.log(`a=${a}, b=${b}, a is ${(a/b - 1)*100}% more likely than b`)
   > a=209, b=201, a is 3.9800995024875663% more likely than b

literally the first run. And even in absolute terms, I got this on the third run:

   >a=192, b=219, a is -12.328767123287676% more likely than b

That's an absolute difference of 2.7%. Again, 100% random data.

[timr]

> That's an absolute difference of 2.7%. Again, 100% random data.

I think I get what you're going for here -- you're trying to simulate a coin flip? -- but what you've actually done is made successive draws from a uniform random number generator. The software is designed to return numbers that fall along the interval [0,1) with equal probability. Thresholding the numbers and dividing their counts is not a meaningful transformation; the result is still just a uniformly distributed random number. It's like...the ratio of heads in two identical, unfair coins or something.

If all "random numbers" were uniform like this, then no, we wouldn't expect an X% difference to be any more or less likely based on the magnitude of the underlying sample. But when we're talking about something like a a population mean, then the behavior of the errors on estimates is very different indeed, and most estimates cluster around the true (aka population) value:

https://online.stat.psu.edu/stat415/lesson/9/9.4

As the sample size for an experiment of this sort gets larger, the bell curve of expected errors gets sharper and sharper, and it becomes increasingly less likely to see errors >= X, for any value X. In the limit of large N, the distribution of sample errors around a known mean approach a normal distribution:

https://www.jmp.com/en_us/statistics-knowledge-portal/t-test...

For what it's worth, the expected proportion of N heads in M coin flips is modeled using the binomial distribution, which is also bell-shaped and illustrates the same idea:

https://en.wikipedia.org/wiki/Binomial_distribution

[chronial]

> I think I get what you're going for here -- you're trying to simulate a coin flip? -- but what you've actually done is made successive draws from a uniform random number generator. The software is designed to return numbers that fall along the interval [0,1) with equal probability. Thresholding the numbers and dividing their counts is not a meaningful transformation;

This is wrong. That is a very meaningful transformation. It is the standard way (https://stats.stackexchange.com/questions/240338) to turn a uniform distribution into a Bernoulli distribution.

Getting a single value with Bernoulli distribution is called a Bernoulli trial (https://en.wikipedia.org/wiki/Bernoulli_trial). Repeating this gives you a Binomial distribution (see your own wikipedia link).

Long story short: GPs code is a perfectly valid way of sampling the Bernoulli distribution. It is inefficient because it needs so many random values, but it mimics the actual process happening in real life making it easier to understand than generating a Binomial sample from the Binomial distribution's CDF.

[timr]

> This is wrong. That is a very meaningful transformation. It is the standard way (https://stats.stackexchange.com/questions/240338) to turn a uniform distribution into a Bernoulli distribution.

The OP didn't do what was described in the SO post. They did something else -- they calculated the ratio of two binomial random variables, and presented that as a percentage.

Also, no, the SO comment you've cited doesn't describe how to generate a "Bernoulli distribution" (not a thing, btw; it's called a binomial distribution) from a uniform distribution. It tells how to make a single Bernoulli trial...but even that isn't what OP did.

This is how you actually do what you're discussing (draw from the Binomial CDF given a uniform RNG, via a table):

https://math.stackexchange.com/questions/1427288/how-to-samp...

[LordDragonfang]

That's precisely what this is trying to model, yes. The standard computational way to simulate a binary event with probability p is to call rand() and check if rand(0,1) < p (or > 1-p, what I did). Or as you called it, an unfair coin flip.

This model is built on the assumptions that if candidates are actually totally equally likely to be picked (the null hypothesis for the experiment above), any given candidate has a p=.2 chance of being hired (given an arbitrary but reasonable hire vs interview ratio of 1:5). Which is just a weighted coin flip. This is indeed a binomial distribution, and my point is that results ±3% of the mean (p*M), even at M=1000, are still fairly probable. When comparing two such results, it's almost expected.

[timr]

The part where you did rand() < 0.8 ? 1 : 0 is fine. That's a Bernoulli trial with p=0.8

The part where you did this in a loop, with two calls per iteration, and then divided the counts and called it a percentage is wrong. It's certainly not a Binomial distribution. It's just the ratio of two binomial random variables.

[faeriechangling]

Yeah I wouldn’t think anything of this if it were 3% the other direction… that’s a measurement error.

[throwaway27727]

In that case n=1, not n=1000?

[commandlinefan]

My HR department has started blacking out identifying resume details - including dates and durations of employment, so for all I know, I could be looking at somebody with one year of experience who changes jobs every two months.

[patrakov]

This is quite dangerous, because it creates bias against people whose CV is not a traditional CV, but more like a portfolio, with works that uniquely identify the author. My own CV would be way too gray if one excludes my contributions to various open-source projects and articles that I have written - all of which would be identifying.

Another class of extreme examples would be public political figures. There is only one person who has successfully implemented a city-wide smoking ban. There is only one person who was a minister of finance, and then worked on in-game economics.

[intelVISA]

You should fuzz your CV's details before submission

[patrakov]

Easier said than done, but you are welcome to try fuzzing my CV. As I mentioned, this would likely erase half of the CV. https://u.pcloud.link/publink/show?code=wT8otalK (outdated for legal reasons - the current employer does not permit mentioning them, due to social engineering concerns, while I still work for them)

[chii]

did those actions/methods result in any better hiring outcomes than before? Or is it basically the same as security theatre?

[morelisp]

Meanwhile in Germany, our HR department still expresses unease when people don't put their photo on their CV. :(

[dudul]

What was the rationale? That it was a proxy for the candidate's age? That sounds like pushing things a bit far.

[commandlinefan]

They didn’t tell us, but that’s our best guess.

[CamelCaseName]

Perhaps they already screen out people with short job stints? Though you should be able to estimate duration based on the bullet points.

[golemotron]

https://en.wikipedia.org/wiki/Women-are-wonderful_effect

[smeej]

I wondered if it was this, or a difference of expectation.

If you have an expectation that someone won't interview well and they surprise you, you'll be more likely to note their positive characteristics and move them forward.

Likewise, if you expect someone to interview well and they disappoint you, you'll be more likely to note their negative characteristics and won't move them forward.

This will be true even if the interviewees give nearly identical interviews.

Does that explain this? If a woman finds herself being compared to an interviewer's expectations for men instead of the interviewer's expectations for women, she does worse, and vice versa.

[idontwantthis]

This is nothing like that.