I wish when people use survey statistics that they would show us the original question. For example I wanted to know if the sexual advances question was specifically AT WORK, or if it could be at any point in their lives in any context.

If they're just asking if women who happen to work in tech now have ever received unwanted sexual advances in their entire life, well, that is worthless without a non-tech statistic to compare/contrast it to.

Also what is an unwanted sexual advance? Someone touching someone? Someone asking someone out once? A boss pressuring their underlings for sex/dates? It has no precise meaning.

Unfortunately I spent some time studying Sociology in my youth and one of the biggest takeaways from that was "you can make people agree with any position you want by changing the wording on the questions." So precisely what you ask is super important for knowing what to do with the conclusion. We actually created surveys where we got completely opposite results from exactly the same democratic just through word-alterations (on purpose).

> "you can make people agree with any position you want by changing the wording on the questions." So precisely what you ask is super important for knowing what to do with the conclusion.

This. Also: any other questions asked which got less news-worthy results. https://en.wikipedia.org/wiki/Publication_bias

I wish there was was something like a wikipedia page which listed every paper / survey with a conclusion on a topic out there. Kind of like examine.com but for social science results rather than nutrition.

> If they're just asking if women who happen to work in tech now have ever received unwanted sexual advances

Given that only 60% of women reported unwanted advances, it's safe to say that it's not "at any point in their lives in any context." That being said, you make a good point about disclosing details of study.

> Given that only 60% of women reported unwanted advances, it's safe to say that it's not "at any point in their lives in any context."

Do you have data on the level of unwanted sexual advances over the course of a woman's life? I just googled it and cannot really see too much.

PS - Although boiling down a definition of an "unwanted sexual advance" is required to compare/contrast different data sets.

Why is the word "unwanted" relevant? Presumably if unwanted advances are not OK in the workplace, wanted ones aren't either, unless the issue is inadequate telepathic abilities among tech workers.

On the internet you can always trawl for a few thousand users that happen to agree with you. When they say sixty percent, they mean sixty percent is their estimate of how many women would report sexual advances, based off of a survey taken by volunteers taking time out of their day to report on gender in the tech industry.

They pull similar tricks with college sexual assault statistics too. They don't ask "Have you been sexually assaulted?" Instead, they ask about a laundry list of scenarios. If you say that one of them happened, you're now a sexual assault victim, even if you don't agree and even if that incident never reached the threshold of illegality. A good analogy is regular assault, which legally includes verbal threatening language. Who hasn't had a heated argument with a friend or family member? But imagine if a survey counted you as an assault victim just for agreeing to the statement "someone has threatened me during a verbal altercation". The law is intentionally defined broadly but applied narrowly. Collecting statistics that count people as victims of violence, when they wouldn't even count themselves, is deceptive and dishonest. It's being done to justify kangaroo courts in colleges, and government mandates in private business.

Disclaimer: Just to preempt the easily offended, I obviously acknowledge that sexual and regular assaults are real crimes that happen way too often, but the statistics being thrown around are ridiculous.

> They don't ask "Have you been sexually assaulted?"

People have different definitions of sexual assault. Asking about specific scenarios is a way to control for those varying definitions, and it's a standard requirement of good survey design.

There may be problems with those surveys, but the solution is not to just ask, "Have you been sexually assaulted?"

Yes, but shouldn't you ask that question too? It's extremely relevant.

The law was written in the context of how it would be applied. It was not designed as a metric for sexual violence in our society. It was designed as the legal rules that judges, just one part of the process, must apply once a case is brought before them. And only then in the context of constitutional law and old common law like de minimis.

It that's true, it would be trivial to design a survey to get whatever sexual assault statistics that you want. Even after the survey has been taken, by having questions ambiguous, and then reclassify them as either sexual assault or not.

The solution is to look at the questions asked.

If you're asking loaded questions, you're getting loaded results. If the results in question are college sexual assault statistics and no one bothered to ask "have you been sexually assaulted" then that's what I would call loaded.

> But imagine if a survey counted you as an assault victim just for agreeing to the statement "someone has threatened me during a verbal altercation".

In England that could meet the threshold for common assault, which is an arrestable offence.

Yes, as I said, the law is defined broadly but applied more narrowly. There's always the question of degree, that's assessed by the victim (in deciding to pursue the issue), the prosecutor, the judge, and the jury. The legal principle is de minimis non curat lex ("the law does not concern itself with trifles").

But that's the point of crime surveys - you ask people what crimes they've experienced and count the results.

That's not because courts interpret the law narrowly, it's because crime is under reported, and under prosecuted.

Plenty of people get prosecuted, and convicted, for common assault that only involved verbal threats.

It also doesn't define what an "unwanted" advance is.

There's a stark difference between an inappropriate advance and an unwanted advance, and it generally comes down to whether or not the recipient has free agency to reject the advance without fear of reprisal in a power-imbalanced situation, workplace or otherwise.

TFA says:

>60% of women in Tech reported unwanted sexual advances.

>

> -- 65% of women who report unwanted sexual advances had received advances from a superior, with half receiving advances more than once

> -- 1 in 3 have felt afraid of their personal safety because of work related circumstances

(emphasis mine)

Power-imbalanced? Yes, for 39% of them, and 30% had some multiple times.

I agree that "unwanted advance" could mean someone asked me out and I didn't like them. I'll be honest, I don't deal with that problem myself because men aren't asked by women in our society, but I think you should define what you think draws the line between unwanted and inappropriate.

I couldn't find that info despite clicking around in "TFA". Thanks.

"24% of women in Tech reported inappropriate sexual advances" is a less scary headline.

It's still way too high, but when you also factor in the selection bias and poor methodology, I have a very hard time taking these numbers at face value.

As for "afraid of their personal safety", that's subjective enough -- and introduces significant response bias in the already selective audience -- to be useless.

Side complaint, it'd be awesome if they put id's on the tags for each subheading, so linking to each would be a pinch.

If the authors are reading this, consider that.

I agree.

And as you've demonstrated, it is always very easy to dismiss thousands of people's experiences as irrelevant in order to be comfortable with the status quo.

I'm just criticizing the methodology that supports the statement that 60 percent of women in tech experience unwanted sexual advances.

This kind of studies are indeed really difficult to make, it's a sensitive subject so not everyone is going to report it the same way, you need to find the right group of people, a sample large enough and compare with across industries which is even harder to do. So I'm convinced personally there's some problems with women in tech which is partially due to the low amount of women going into tech but I think that's really hard to quantify in real numbers.

why do you need to compare across industries, exactly? even if we found the same % in every industry we'd still need to take action.

The link seems to insist there is a problem with women in tech, even if the study is correct, it could just mean that women are not treated the same way in other industries either. It does not stop to take action indeed, I agree with that, I was just reacting to the study itself.

I'm pretty sure nearly all women (and most men) get "unwanted sexual advances" at least a few times in their lives. After all we can't read each other's minds. The crucial question to me is not whether advances happen, but whether they stop when it is made clear that they're unwanted.

And how many in non-tech? Let's say, fashion industry?

Exactly... We can draw exactly zero conclusions regarding gender inequality in the tech industry without a control sample...

Sure we can - let's say we look at the fashion industry and see it's just as bad there - or even worse. So what?

That doesn't mean that as an industry we should put up with gender inequality. Gender inequality is an incredibly irrational historical artifact and in times where it's so hard to hire excellent employees - creating a gender equal workplace gives employers an edge.

This sort of data can give employers the confidence that investing in fixing the problem would indeed help giving them an edge over other workplaces since the problem is so widespread.

I agree with most of your points. Gender equality is an important issue in society at large. But the makers of this site are drawing a tech-industry-specific conclusion, so they should at least attempt to present unbiased data.

Is that serious?

We can draw the conclusion that MOST women in tech are sexually harassed! What comparison is required? Why does it need to be worse than other industries to require examination and action?

It's like saying, "Well, we know 500 people were killed in Chicago last year, but that doesn't tell us anything until we find out how many people were killed in London, Beijing, and Tegucigalpa."

Well if your conclusion is that Chicago is a significantly less safe place to live than other densely populated cities, then yes, you'd need that data.

The conclusion is just that SV has a problem. That's why it's called "Elephant in the Valley". It isn't called, "Elephant in the Valley that doesn't exist in other places".

My wife for example received 100% of sexual advances on the way to and from workplace and none at workplace (yet). The percent of women who received unwanted sexual advances outside industry is definitely much higher than 60% so I think they are quite protected in the industry. That's why I encourage more women to get into tech and also to work more overtime hours; it statistically minimizes the chances of receiving unwanted sexual advances.

Why is this relevant? Why not solve the problem in our industry, instead of pointing fingers?

It is very relevant. I want more women in tech. (I think the problem is they don't want to do the job, but let's put that aside.) Studies like this only reinforce the already existing negative stereotypes about "nerds in tech", in both sexes.

You claim this is a problem. Here's a simple solution: Make tech companies segregated by gender. What about that? The thing is, I presume, people who did the study don't want to solve a problem of "unwanted sex advances" in the workplace. They want to solve a problem of not enough women in tech. And this is not the way to go about it.

It is also an interesting question, and I am genuinely curious. I also doubt women actually choose the career based on amount of unwanted sexual advances, and having the numbers could finally put that stupid theory to rest (if it's false, of course, maybe it's not, I dunno; but based on the above provided link, it probably is false).

I think they want to solve the problem of harassment in tech. "Not enough women in tech" is a trailing indicator for a number of factors, that being one of them.

Yes, "How are other fields doing" is a question that can be asked - but it is one of the least interesting ones. Suppose that every other industry is at 40%. Or suppose they are all at 60%. It doesn't change the fact that ours needs to seriously deal with it. (And as seen by ~1/3rd of the comments in this thread, is more interested in splitting hairs.)

> I think they want to solve the problem of harassment in tech

Fine, why choose tech then? Are they biased? It simply doesn't make sense.

If they want to help women, they could perhaps start in an industry where it's most likely to be a problem (and I'd bet it's somewhere where women are doing low paid job). Here's why - we probably already know at least a partial solution, just look at what are the other industries doing. You don't get either information without a survey.

Because, the authors...

Work in tech.

So yeah, they are biased in wanting to get their industry in order. There is no ulterior motive, here, and no reason to get so defensive.

I don't think they have ulterior motive, I just think they go about it (whatever is it they are actually doing) in a completely wrong way.

I think that these "unwanted sexual advances" may come from the fact there is power differential, i.e. hierarchy and bosses. Maybe we could try to eradicate that first, that is, have more cooperatives (democratically owned and run companies)? In tech, these would be pretty easy to set up due to low investment requirements..

Fallacy of relative privation

http://www.logicallyfallacious.com/index.php/logical-fallaci...

I think you might have a fallacy of missing conclusion here.. i.e. I didn't make any. I just asked a (actually genuine) question.

Quick Google, but see [0]. Moreover, there's no doubt that 60% is a clear majority, which is striking, to be honest.

EDIT: And I forgot the link :)

[0] http://www.huffingtonpost.com/2015/02/19/1-in-3-women-sexual...

Your link says 31% of women in STEM were harassed. 60% is some distance from that, don't you think?

Good point, there are two possibilities here:

1) There is a bias in the cited sample (not the huffpo's, the op's).

2) Silicon Valley is one sleazy place.

I want to be open minded because I as a man am someone oblivious to a lot of these issues, but it might be called for if we had some details on the OP's study.

Agreed. I don't necessarily outright disagree with the OP, I just get the impression that we're missing some of the facts. And not getting a complete picture can do just as much damage as entirely ignoring it.

What are you trying to accomplish by asking that (seemingly rhetorical) question?

Are you implying that 60% is an acceptable number because it may hold in other industries? Are you implying we should only focus on the industry with the highest rate of sexual harassment? Or are you implying that it's unfair to examine tech as a separate industry?

The fact is that tech is a particularly female-unfriendly industry, and it merits examination in isolation. Yes, we should examine all industries, but there's nothing wrong with a subset of people focusing on a subset of different industries.

I think there should be a state agency (employing women only) doing the examination the same way there is IRS for taxes but it should be funded by a tax applied to each woman's salary. They could be called FBI (Feminist Bureau of Investigation) and have their own SWAT (special women attack team) like IRS.

These results might be subject to survivorship bias: all the women who quit the industry before they got 10 years of experience because of these negative experiences are silent. All the women who dropped out of CS in college. All the women who went to a hackathon once and never again.

I can't believe this was the most downvoted comment for a bit.

It's true and incredibly relevant. The responses from women with 10+ year experience paint a shocking enough picture, and this survey doesn't even cover those women who left the industry or never felt welcome in the first place.

200+ is hardly a representative sample for the entire tech industry, especially as most of these reports were taken from the Bay Area. There's also no control group (other gender, other industries). And what specifically is an unwanted advance with regards to these statements?

I'd like to see the peer review on this one, to tell you the truth.

Using a control group here doesn't make sense because we're not trying to benchmark SV against other places, nor are we doing an experiment. This is a survey.

Furthermore, if there are 400,000 tech workers in Silicon Valley, then 200 gives us a less than 8% margin of error (95% confidence level). This still gives us a result of at least 50%.

But we're not surveying the entire tech industry, we're surveying women in Silicon Valley, which is a number much closer to 100,000 (although that doesn't radically change our margin of error).

Honest question: do those confidence levels account for how the participants are chosen? Or is it assumed that it's random?

Randomness is assumed for my number. However, there are lots of ways this survey could have been biased:

- women who were assaulted, harassed, or threatened have already left SV

- women who were assaulted, harassed, or threatened were more likely to respond

- women who don't believe in an anti-female culture in SV were more likely to respond

- etc.

It's up to the designers of the survey to control for these things within the survey and/or disclose them. Unfortunately, they haven't.

Does anybody know of couples successfully formed on the workplace? Seriously curious here. I have witnessed two personally. Maybe your coworkers, you friends or even yourself? For each of these, someone had to first try and make an advance ; Presumably without knowing in advance if wanted or not. Sexual harassment is not seriously characterized by this metric. Or these two couples of mine started with some...

Okay, HN - forget the surveys and methodology. Your actions will be local, so your data will be too. Take a look around you, in your workplace. Are there actually any female developers working alongside you? How about in engineering management, and at the executive level?

The problem is there. It's staring you in the face. The tech industry has not offered the same opportunities to women as it has to men.

Act now and solve it. Don't wait for that perfectly conducted longitudinal study, covering a representative sample of women entering the industry and their experiences ten years later. You don't need it to see what's happening.

"Act now and solve it."

By doing what?

This article was, of course, flagged into oblivion within minutes of being posted, despite having 27 upvotes, because Silicon Valley is quite sure that there's no sexual harassment problem.

It was posted at least 4 times in a 24-hour period; perhaps the latter postings were flagged purely for spam/dupe reasons. Flags aren't delimited by reason.

The original survives:

https://news.ycombinator.com/item?id=10887846

These dupes still survive (as of 0200 GMT):

https://news.ycombinator.com/item?id=10887846

https://news.ycombinator.com/item?id=10895364

Plus the introductory article on Techcrunch:

https://news.ycombinator.com/item?id=10884242

I posted the article. I may be misinformed to what 'flagged' means, but I see no reports or 'flags' on my screens and whatnots. I am not a power user in any sense. As of writing, I see no downvotes, just an upvote total of 35. What does being flagged mean here so that I may avoid it or recognize it in the future?

Yours is the survivor that gained traction. There were earlier (and perhaps later) postings that were flagged (I'm guessing because of duplication rather than nefarious intent.)

Thanks for the clarification.

Every user who has been around for a while has a "flag" option for every story, to be used for stories which should not be on HN, such as spam. Your posting briefly made the front page, then was flagged off of HN by people clicking that link.

There's nothing you can do to avoid it except avoid posting stories which imply Silicon Valley has any sort of sexism problem.

What is a non-disparagement agreement? And I would like to see a similar survey of men.

"A Non-Disparagement clause restricts individuals from taking any action that negatively impacts an organization, its reputation, products, services, management or employees.

Example Definition: Disparaging remarks, comments or statements are those that impugn the character, honesty, integrity, morality or business acumen or abilities in connection with any aspect of the operation of business of the covered individual or entity."

http://www.contractstandards.com/clauses/non-disparagement

[Meta comment.] It is a shame you can't get credibility points for downvoting.

What is "unwanted" sexual advances? Is it a colleague asking out for a drink, ignoring that the woman was already seeing someone or not interested in dating? It is "unwanted", but the colleague may just have been taking a shot. It is said that work is the main environment where people find their partner.

Now, if "unwanted" means "He asked, I declined and he kept asking me out every day after that" then yes it is a problem.