SSC Survey Results: Sexual Harassment Levels By Field

[content note: sexual harassment]


Recent discussion of sexual harassment at work has focused on a few high-profile industries. But there has been relatively little credible research as to how rates really differ by occupation type.

There are many surveys of harassment rates in specific industries, but they can’t be credibly compared with one another. The percent of people who report sexual harassment varies wildly from survey to survey – thus studies finding that anywhere from 12 percent to 48 percent to 60 percent to 85 percent of women have been harassed at work. If a survey shows that 60% of female nurses get sexually harassed at work, does that mean nurses are victimized particularly often (because more than 12%) or are unusually safe (because less than 85%)? It doesn’t matter, because another study says only 19% of nurses get harassed.

Why do all these numbers differ so dramatically? The most important issue seems to be how you ask the question. “Have you ever been harassed?” gets numbers more like 12%; giving a long list of specific behaviors and asking “Have you ever experienced any of these?” gets numbers closer to 85%, depending on what the behaviors are. Surveys also differ on whether they ask all employees or just women, whether they include a time frame (eg “…in the past two years”), whether they specify that it had to be at work vs. work-related events, and whether they include witnessing someone else’s harassment. Taking these surveys entirely seriously would lead to the conclusion that Uber has the lowest sexual harassment rate of any company or industry in the world; I choose not to take them seriously.

This means we need investigations that use the same methodology across multiple fields. Whenever the media talks about this – see eg the Washington Post’s The Industries With The Worst Sexual Harassment Problem – they’re working off of the Equal Employment Opportunity Commission’s records. But these are totally unsuitable for the task – they just report raw number of claims per industry. The industries that rank lowest in EEOC’s data tend to be small industries with very few women – for example, taken seriously the WaPo’s graph shows that mining has the least problem with sexual harassment of any industry in the world. Is this thanks to their uniquely progressive culture – or because there are practically no female miners? I’m going to say the second one. The takeaway that most real researchers take from the EEOC claims is that the lowest-paying and most mundane occupations – retail, restaurant work, hotel work, etc – have much higher sexual harassment rates than the prestigious occupations people generally talk about. Eyeballing the data, this looks basically true. But trying to get anything more fine-grained than that out of EEOC is basically hopeless.

I only know of two surveys that have even attempted to compare different fields in a principled way, and neither really inspires confidence.

First is a survey by Cosmopolitan magazine, that asked 2235 women in different fields about sexual harassment in their industry. About 33% reported being harassed at work. Women in the retail and restaurant industries reported the highest rates of sexual harassment; despite media coverage’s focus on STEM and entertainment, both reported average to just-below-average harassment rates.

Second is the Project XX Survey, where the automotive industry decided to survey their workers using methodology previously used in Silicon Valley, which made their results at least somewhat comparable. The advertising and market research industries seem to have joined in later. They find the automotive industry has the most harassment, followed by Silicon Valley, advertising, and market research – but there aren’t really big differences and industry rankings frequently reverse across questions.

I interpret existing data in this area as being basically useless, but at least suggestive that the media focus on a few prestigious industries is mistargeted.


The Slate Star Codex survey is an online poll of readers of this blog about various aspects of their life and personality. This year 8,077 people participated. The survey asked a set of questions on sexual harassment, providing a unique dataset with which to investigate the question.

Due to the blog’s tech focus, readers skew well-off, male, and STEM-focused. This made it impossible to investigate the restaurant and retail industries that previous research suggests have the most problems. But there were still enough responses to get a wide cross-section of white-collar occupations and academic fields.

Respondents were asked four questions:

1. Have you ever been sexually harassed or assaulted at work?

2. Have you ever been sexually harassed or assaulted outside of work?

3. Have you ever made unwanted sexual advances to someone at work that you think they interpreted as sexual harassment or assault?

4. Have you ever made unwanted sexual advances to someone outside of work that you think they interpreted as sexual harassment or assault?

Respondents were given a “prefer not to answer” option in case the questions made them upset; in order to make people entirely comfortable using the option, they were asked to select it automatically if the last digit of the time was a “1”.

Data were analyzed about female victims, male victims, and male perpetrators. Although some women admitted to perpetration, the sample size across fields was too small to be useful. Industries were included if they had a sample size greater than thirty; because there were more male respondents and so all industries had a greater male sample size, some industries are included on the male graphs but not the female graphs.


Overall, 21% of women and 6% of men reported being sexually harassed at work. 1% of women and 2.2% of men reported committing sexual harassment at work.

Fields listed do not exactly match fields on the survey as some were combined in order to increase sample size, eg “physics” and “chemistry” into “hard sciences”. There’s a discussion involving pre-registered categories below.

Here’s all three measures combined into my best guess for the relative rates in each field:

Made by taking each of the three previous graphs, standardizing each field to percent of the worst field in that category, and then adding them up and dividing by three to get an average for each field

But this combines three different things. First, the actual rate of harassment. Second, a measure of how willing people are to report harassment. And third, a measure of how willing people are to consider some specific act to be harassing. This is a major problem. People from traditionalist cultures and subcultures may have a higher threshold for calling something harassing; if (for example), more traditional people go into Business, and more socially liberal people go into Art, that could skew the numbers.

In order to control for this and other factors, I asked subjects how often they are harassed outside of work. Of note, men, women, victims, perpetrators, everybody in every industry, reported much more sexual harassment outside of work than in it. Although there were enough female perpetrators to do statistics on this time, I left them out to keep it as similar to the in-work data as possible.

Here’s an overall summary, made through the same combination process as Figure 4.

Levels of at-work harassment and out-of-work harassment correlate at 0.81 across fields. This suggests that most of the differences observed here are differences related to the people in a field, rather than to any sort of field culture. Given how strong this effect is, it’s unclear whether it’s possible to adjust for it and get a measure of field culture. If we try (by dividing the at-work harassment levels by at-home harassment levels), we get this:

The most striking finding I find on all these graphs is that “nerdy” / STEM / traditionally-male jobs have the least harassment. The less nerdy / more verbal-personal-skills / traditionally-gender-balanced jobs have the most harassment.

Here’s Figure 4 again, but this time with nerdy-STEM-male occupations in red, verbal-personal-balanced occupations in blue, and edge cases in grey.

If categories like nerdy/STEM/traditionally-male seem unnatural, that’s fine – these are always a judgment call. But in December, I preregistered which clusters I would use, to prevent me from making them up to get the result I wanted. It’s not exactly the same as the clusters above – I ended out having to combine a few to get a big enough sample size to be meaningful. But using the original pre-registered clusters and the original pre-registered endpoint: in STEM/male-type fields, 12% of women (n = 212) reported being harassed at work. In non-STEM/balanced-type fields, 25% of women (n = 161) reported being harassed at work. p = 0.002. The preregistered analysis confirms the finding above.

I preregistered the hypotheses that STEM would have more female victimization (because there’s a higher ratio of heterosexual men to heterosexual women), but the same amount of male perpetration (because there is not an inherent difference in the sort of people who go into the field). Both preregistered hypotheses were wrong; female victimization and male perpetration were less.


Overall victimization rates were comparable to past surveys. These were on the low end of reported results, likely for several reasons. First, we strongly encouraged participants who were at all uncomfortable to say “prefer not to answer”, including asking 10% to say this outright regardless of experiences; this clearly depressed results and makes absolute numbers inaccurate. Second, participants were generally higher-income, and this group usually reports less harassment. Finally, we asked participants only about “sexual harassment” in general, and not about specific misbehaviors. Given these issues, the modest differences between these and past results (eg our 21% average female harassment rate compared to Cosmo’s 33%) are not surprising.

More surprising were the lower harassment rates in STEM and male-dominated fields, for two reasons. First, if most harassers are heterosexual, then a high percent male workforce in an industry should create a situation where a large number of harassers are focused on a small number of potential victims, raising the average harassment rate per victim. But this was not observed in our survey, nor in the Cosmo survey, nor (as far as can be eyeballed from the not-really-inter-comparable data) in the EEOC survey.

I don’t know of a simple explanation for the discrepancy. A friend suggested that it might be a customer effect. That is, lawyers / health-care-workers / etc often have to deal with sexist customers, whereas female programmers might not encounter anyone on the job except their coworkers. But this seems unlikely to make much difference; this survey finds that customers represent only 9% of sexual harassment incidents. Also, remember that the rates at which men admit to perpetrating sexual harassment match the rates at which women admit to experiencing it, and there’s no reason why the male-perpetration rates should have anything to do with customers. Also, social scientists don’t have customers any more than hard scientists do, but the social science harassment rate is higher. Perhaps the sort of people who go into STEM are socially oblivious in a way that prevents them from both noticing if they are sexually harassing someone else, and from noticing if they are sexually harassed. This would result in somewhat lower self-reported perpetration rates, and self-reported victimization rates that are even lower than what the low perpetration rates would predict. It would also match the high correlation between in-work and out-of-work harassment, which suggests that demographic factors are making most of the difference.

A second reason the low harassment rates in STEM were surprising is conventional wisdom claiming the opposite. In media coverage, STEM fields are portrayed as full of “techbros”, and uniquely unwelcoming to women. How do we square this with the data?

Some people who reviewed these data suggested that the media focuses more on a narrative where STEM workers are dismissive of women, rather than one where they harass them. I don’t find this accurate – a quick glance at any media source will show they focus on both narratives. But it seems possible that the dismissiveness narrative is correct, and the harassment narrative is a side effect of this. But this is somewhat contradicted by the auto industry survey mentioned above, which found tech was unexceptional either in harassment or in general dismissiveness. However, the other industries participating in the survey may themselves be unusually bad.

Another possibility is that the respondents to the survey may be skewed for some reason. For example, they may have been selected by this blog’s previous discussion of these issues, which have also argued that tech is not a uniquely bad industry.

But a final possibility is that the media coverage is inaccurate. The minimal amount of previous research in this area has all shown that the highest rates of sexual harassment are in the retail and restaurant industries. The media has ignored this. Before last year, the media almost never mentioned sexual harassment in Hollywood; now they hardly mention anything else. Given that already know media coverage here is generally inaccurate, maybe the data don’t need to be squared with it. Maybe the media is just wrong.

Every industry has enough sexual harassment to produce horror stories. If the media is disproportionately interested in the horror stories of relatively well-off people, and in stories that confirm their existing prejudices, we might hear many genuinely horrible stories about harassment in technology, but fewer about harassment in eg health care, philosophy, or law – even if objectively those fields are much worse. That might further encourage people in technology to come forward (consider what happened in Hollywood after the Weinstein revelations) and people in other fields to stay silent, contributing to a vicious cycle.

Compare all the anecdotes and popular lore about how immigrants are criminals. It’s totally false – immigrants have crime rates well below native-born citizens. But we only know that because there have been really good studies. If the studies hadn’t been done, and all we had to go on was the daily lurid stories about a Mexican guy knifing someone, who would believe it? In the absence of real studies, the media’s ability to spin a compelling narrative casting some people as monsters feeds on itself forever. We know that happens relatively often. Are we sure this time is different?

There’s only been one strong previous survey on this subject – the one in Cosmo I referenced earlier. They found the tech industry had an around-average to slightly-less-than-average rate of harassment. Anything else you’ve heard – any news story, any anecdote – is at the same level of evidence as the stories of immigrant crime.

It seems possible that this narrative grew up to explain why there are so few women in technology – maybe they’re all being harassed out by lecherous professors and bosses. But this explanation doesn’t hold water – the rate at which men vs. women choose tech courses in high school exactly parallels the rate at which they go into tech occupations. And gender balance in fields isn’t predicted by harassment levels in those fields anyway – otherwise you’d never see women working in restaurants, retail, or Hollywood.

I have been expecting results like these even before I did this survey, and this calls my data into question. Whenever someone with an unusual opinion produces an experiment that confirms their opinion, you should always be skeptical until it is verified by other sources.

So aside from appealing to the unimpeachable authority of Cosmopolitan as partially confirming my results, I urge anyone with the relevant skills to download the raw survey data themselves and see if they can replicate my conclusions. You can get them as .xlsx or .csv. Some people have complained of weird problems in the csv format and I recommend the xlsx if at all possible. I have removed the data of a few people who did not want their answers to be public, so you may not get exactly the same numbers I did, but they should be pretty close.

And more important, I appeal to anyone with an interest in this topic to do larger and more formal surveys comparing different fields. This may be the single topic where the extent of public interest is most disproportionate to the minimal amount of good research done. If people more prestigious than me or Cosmo were working on this, we would be in a much better position to know what to believe.

316 Responses to SSC Survey Results: Sexual Harassment Levels By Field

  1. SolveIt says:

    Red and blue are switched in the main text but fine in the in-picture caption.

  2. tlwest says:

    Is there anything tagged ‘race/gender/etc’ that isn’t also tagged ‘things I will regret writing’? :-).

    Anyway, reading the tag gave me a good laugh. Thanks.

  3. Steve Sailer says:

    I offered the opposite hypothesis last November:

    “For years the press has been telling us that industries that hire mostly men—such as computer programming, defense, and the military—must be bad for women. No doubt, it is explained, all those horrible, evil male engineers must be teaming up to exploit the handful of female employees. After all, men and women are enemy genders. I mean, that’s what every lesbian women’s-studies professor says, and they wouldn’t have any incentive to lie, would they?

    “Therefore, women must be given much more in the way of affirmative-action quotas in technology companies. …

    “Instead, however, we see that careers where women are most abundant and most ambitious, such as television and movies, are where they are most exploited.

    “Why? It’s simple supply and demand.”

    • CthulhuChild says:

      I think the fundamental problem is the conception of how sexual harassment works. The prototypical model is the idea that you have a harrasser-type male, representing some non-zero percentage of the population, who will satiate himself upon a set number of victims, with various levels of “need” in terms of number of victims and intensity of harassment. This concept would lead to Scott’s hypothesis of higher-than-normal victimization in STEM, where the gender imbalance creates a larger (absolute) number of abusers and a smaller pool of potential victims.

      What I’d like suggest is controversial and will inevitably be misunderstood as victim-blaming, but here goes. Is it not possible that the victims have a role in the prevalence of sexual harassment? IE, that a certain percentage of men are potential harassers, but are in some sense waiting for the right victim. It’s not a question of satiation or meeting a quota, but rather that harasser-types will probably harass victim-types wherever they meet in a model that’s sort of like stoichiometry only for terrible workplace outcomes. The point being, overall harassment will go down in any industry that selects for for low numbers of either types.

      This would actually explain the apparently low sexual harassment rate in mining industry. As Scott points out, there aren’t many women there, but if the male-harasser satiation model were true this just means that nearly all of them SHOULD have been subject to abuse (their low absolute numbers shouldn’t have altered the percentage). But if the harasser/victim/stoichiometric model holds, then it’s also possible that there are no victim type women in the mining industry. Rather, the culture simply selects for women who will not attract harassment (possibly by immediately harassing any woman who would so badly that they immediately quit).

      Edit: To be really, really, really clear, I am not suggesting that a causative role in attracting harassment equates in any way to a moral assignation with respect to that harassment. Take a hypothetical white supremecist who is very respectful to women and would never attack someone on those grounds, but flies into a violent rage in the presence of blacks of either gender. A black woman who is attacked by this person is not chosen at random and inherent attributes have contributed to them being attacked. Identifying this fact should not be construed as suggesting the victim is somehow responsible. But to follow the metaphor, our radical white supremacist would be invisible (both as a self-reporting harasser and a generator of harassed women) if they worked at a place where everyone was white. Again, this has nothing to do with the progressive culture of this workplace, and everything to do with the fact that the type of victim our harasser prefers is not available.

      This theory partly comes from my own school experience, where I was badly bullied by people who had never been bullies before they met me. Similarly, there were other kids who were bullied by basically everyone, even by kids who didn’t bully me. This phenomenon doesn’t make sense if we assume the decision to engage in harassing/abusive behavior is entirely dependent on the disposition of the perpetrators, rather it was the interaction between their disposition and some intangible characteristic of their victims.

      • Aapje says:


        My perception is that women in IT tend to be quite masculine, so they may, like men typically do, accept a more rough culture without feeling harassed & perhaps also attract less sexual attention.

      • taradinoc says:

        One of the recurring themes on Loveline was that people who are victimized early in life end up being disproportionately victimized throughout the rest of their lives: they become “good victims”, attracting abusers like an injured gazelle attracts lions.

        • Education Hero says:

          What evidence is there that they become “good victims” after early victimization, as opposed to already having traits which attract abusers, including before the first instance?

          • Simon_Jester says:

            The first most obvious evidence is that he first-instance abuser tends not to be ‘drawn’ to the child in the first place; they tend to be an adult caregiver or authority figure who already had contact with the child for other reasons. Convenience and ease of access is generally more important than some kind of careful selection for traits that ‘invite’ abuse.

            The second most obvious evidence is that when abusers can get away with it they tend to abuse everyone around them, or nearly everyone. This strongly suggests that if there is an inborn “abuse me!” trait, it must be either super-duper-heritable (so that all children in a family consistently get it from their parents) or super-duper-environmental (strongly caused by some consistent environmental factor, such as, y’know, being abused in the first place.

            The third most obvious evidence is that victimization may begin before a child’s personality is fully formed, in many cases before it is even *differentiated* clearly (e.g. a toddler). It seems a priori unlikely that developing children would spontaneously and independently develop an “abuse me!” tag that expresses itself just as strongly in toddlerhood, drawing abuse from caretaker adults, as it does in adulthood, drawing abuse from romantic partners. Most character traits evolve much more significantly over time.

            Much more plausible that a consistent environmental factor present in a developing mind’s environment (child abuse received as a child) causes a consistent alteration in the mind of the child in question (predisposition to tolerate abuse and/or to subtly signal that predisposition).

            Much as PTSD is a consistent family of reactions to a consistent family of stimuli. We would not for a moment seriously entertain the notion that PTSD is actually a natural inborn trait, whose victims just happen to live much more traumatic and stressful lives and find themselves attracted to war zones and so on.

            We might also examine the behaviors of people with no prior history of committing abuse, when they come into contact with people who have chronic histories of being abused.

            If “abuse me” signals are frequently the cause of abusive behavior by people who would otherwise not commit abuses, then if Alice has an “abuse me” signal so strong that of the four boyfriends in her life, Bob, Charlie, and Earl were ALL abusive to her but Dale was not… Well, the obvious interpretation is that her signal is so strong it can turn 75% of randomly chosen men into domestic abusers. This is a prediction that can be tested when she next gets together with Frank, because even if Frank has no prior history of domestic abuse, IF the hypothesis is true, it’s likely that he will acquire one.

            If, by contrast, our hypothesis is that Alice is sending out a signal that systematically attracts would-be abusers, then if Frank has no prior history of domestic abuse, it is unlikely that he will suddenly sprout one. He must have been attracted to Alice for unrelated reasons.

          • Calion says:

            Er…except that we know that certain types of people are significantly more likely to experience PTSD than others, given similar experiences.

          • Nancy Lebovitz says:

            Calion, what types of people are more likely to experience PTSD? The only category I’ve heard of is previous history of trauma increases the odds of PTSD.

          • Simon_Jester says:

            Even stipulating that there are personality types predisposed to PTSD, that is not the same thing as saying that such a predisposition causes the stressors that in turn cause PTSD.

            In the absence of traumatic stress, PTSD will not occur. To be sure, some people may be more likely to get PTSD from traumatic stress than others. But their greater susceptibility is not (normally) a causative factor explaining how they came to experience the traumatic stress.

            Now, it is likely that there are certain factors that make a person both more susceptible to domestic abuse (such as willingness to comply with an abuser) and more likely to attract abusers (such as the abusers noticing how compliant a prospective victim is). Since domestic abuse tends to be, for lack of a better term, predatory, it is unsurprising that the predators in question would seek out and exploit weak ‘prey.’

            But all of the above statements about domestic abuse are easily explained by the idea that the first instance of abuse shapes or distorts personality, creating reactions and behaviors that future would-be abusers can detect and exploit.

            So sure, there are probably traits of abuse victims that increase the likelihood that they, out of all people in the world, will be targeted for abuse. This is very far from saying “I think the traits that make a person an easy target for domestic abuse are innate, and a causal factor in bringing about the abuse, including the first instance of abuse.”

        • Aapje says:


          The question is whether these people have trait(s) that attract abusers, whether the abuse causes people to develop trait(s) that attract abusers or whether seeing others get away with abuse causes people to join in.

          I suspect that all of these are true to some extent.

          [EDIT] Ninja’d

          • taradinoc says:

            Simon_Jester responded to most of that above, but regarding this:

            or whether seeing others get away with abuse causes people to join in

            The archetypal call went something like this:

            Someone who’s recently been assaulted by (a partner / a stranger at a party) calls in with a question about (emergency contraception / STDs / how to find partners who won’t do this / whether smoking pot will hurt the baby / when the guests, the Insane Clown Posse, are releasing their next album).

            It turns out she’s been sexually assaulted in similar situations before, and each time, she was (“frozen” / asleep) and (she felt like she couldn’t react / she didn’t wake up until too late and then couldn’t react). She shows no interest in reporting any of them.

            The hosts ask if she was molested as a child. Her response is “yes, when I was (a number less than 10), but (I’ve never told anyone / no one believed me / he went to jail after molesting another kid / that’s not what I’m calling about, what do you mean, I’m fine)”.

            Or, occasionally, “I don’t think so, but I was taken away from my birth parents when I was (a number less than 5) and I don’t remember anything before then”.

            So, it’s not that abusers were emboldened by seeing other people abuse her, because the initial abuse happened out of sight at a young age, and the pattern continues up to the present day with strangers.

        • moscanarius says:


          I recall reading something about a study on PTSD where people who had a more emotional remembrance of the details of a plane crash (“I was wearing my favorite shirt that day”) had a greater chance of suffering from PTSD than the more straightforward people. Unfortunately, I can’t find it anywhere anymore. Maybe this is what Calion is referring to.

      • albatross11 says:

        One obvious victim-side issue w.r.t. sexual harassment is how many options the victim has for other jobs. A woman who’s tired of getting hit on by her boss and whose skills are in high demand can just take another job somewhere else. And her boss knows this, and has an incentive to moderate his annoying passes/risqué comments/etc., because he doesn’t want to drive away his high-value hard-to-replace talent. And also, *she* knows she has options, which will make her more willing to say “Knock it off, Jerry. I don’t have to put up with this shit.”

        A woman whose skills aren’t in much demand, who is in a position where she can’t easily find an equally-good job somewhere else, is *way* more vulnerable to ongoing harassment. She still has to pay the rent, after all, even if that means putting up with being hit on by the boss all the time, and having him “accidentally” brush against her in the elevator and such. He knows she has few options, so he can push it, and she also knows she has few options, and has to put up with it. She can quit or go to HR, but those are both high-cost actions, and maybe the harassment is less trouble than either of those would be.

        • Mary says:

          One notes that nerdy/STEM jobs also have more objective standards of measurement. If your code runs, it runs. If it throws hard errors every time you “fix” the most trivial of errors, it throws them.

          1. makes your skills more objective, too
          2. makes it harder for someone to retaliate by claiming your work is substandard
          3. makes it harder for someone to reward you for compliance.

          A “coding couch” would be, though possible, manifestly a bad idea.

          • moonfirestorm says:

            Coding can objectively differentiate horrible coders from not-horrible coders. It’s not really accurate beyond that: the difference between a good coder and a mediocre coder is really subjective, or at least objective-but-incredibly-difficult-to-spell-out.

          • Mary says:

            True, but that’s more than many other jobs have.

      • vV_Vv says:

        What I’d like suggest is controversial and will inevitably be misunderstood as victim-blaming, but here goes. Is it not possible that the victims have a role in the prevalence of sexual harassment? IE, that a certain percentage of men are potential harassers, but are in some sense waiting for the right victim. It’s not a question of satiation or meeting a quota, but rather that harasser-types will probably harass victim-types wherever they meet in a model that’s sort of like stoichiometry only for terrible workplace outcomes. The point being, overall harassment will go down in any industry that selects for for low numbers of either types.

        This is consistent with something I’ve read before about sexual harassment and rape being largely crimes of opportunity.


        This would actually explain the apparently low sexual harassment rate in mining industry. As Scott points out, there aren’t many women there, but if the male-harasser satiation model were true this just means that nearly all of them SHOULD have been subject to abuse (their low absolute numbers shouldn’t have altered the percentage). But if the harasser/victim/stoichiometric model holds, then it’s also possible that there are no victim type women in the mining industry. Rather, the culture simply selects for women who will not attract harassment (possibly by immediately harassing any woman who would so badly that they immediately quit).

        By why would women in mining or computers be less victimizable than those in healthcare or media?

        • Nate the Albatross says:


          “By why would women in mining or computers be less victimizable than those in healthcare or media?”

          Modern mining is mostly strip mining. If a woman is driving a bulldozer or operating an excavator she can’t hear you and you might get squished if you approach her while she is working. On coffee break or at lunch is an extremely public environment, where all of the other men will see your attempt. Most harassment happens in retail, restaurants, media and health care – where there are long periods of slow times and there are private areas away from prying eyes. Mining happens on a mountain top and doesn’t have long periods of time without a customer or interview.

          In tech the environment is a lot different, but we are still seeing an environment much different than the high harassment categories. Work is typically completed in cubicles, in a factory, or lab – there isn’t a kitchen or stockroom or dressing room or empty hospital room where harassers can have long private conversations with their victim. Even conference rooms and offices often have glass walls so anyone walking by can see.

          Women in tech and mining are also be selected a bit more stringently than retail/restaurant workers and are more likely to have completed some kind of formal education (truck driving, coding, often more) where they learned that there are compliance officers and regulators and human resource professionals that will help them deal with harassment. And the potential harassers also say through those same presentations.

          Restaurant workers, etc are often “trained” by their manager, and may or may not have received corporate info. Mining has some serious safety speeches, and I’m sure HR dumps the rest of the policies into the introductory training too. Tech is also often dominated by company policies. Some of the restaurant and retail companies are owned by the manager, and staffed by only a couple of other people. And the Screen Actors Guild just recently moved to crack down on private auditions.

          Opportunity, policies and training matter at lot.

          • vV_Vv says:

            Restaurant workers, etc are often “trained” by their manager, and may or may not have received corporate info. Mining has some serious safety speeches, and I’m sure HR dumps the rest of the policies into the introductory training too. Tech is also often dominated by company policies. Some of the restaurant and retail companies are owned by the manager, and staffed by only a couple of other people. And the Screen Actors Guild just recently moved to crack down on private auditions.

            This may explain harassment in restaurants, retail and maybe media, but what about healthcare? Hospitals are highly public places, run by organized bureaucracies with HR and legal staff, and female employees above the level of janitor or cook are all highly trained staff.

            And it does not explain why social sciences have higher harassment than hard sciences, math or philosophy.

          • mdet says:

            My initial thought when i saw the high rate in health care was that it’s an unusually intimate field. Fairly common to have a doctor alone with a patient behind closed doors, possibly with the patient partially undressed. All it takes is one off comment (from the doctor or patient) to turn this from “routine” into “uncomfortable”.

            This matches the fact that Health Care is the second highest for at-work harassment, but looks to be fifth for outside-work harassment. It doesn’t match the idea that better educated white collar employees have more recourse. Or maybe it does, if you consider that a doctor is unlikely to report their own patients to HR.

            Edit: “Doctors & nurses are frequently sexually harsssed by their own patients” sounds like something we would’ve heard about by now if it were the case. Seems like an easy and low-stakes thing to vent about outside work. So I’m not sure this is exactly the case, but I think the intimacy of healthcare is a point in the right direction

        • dvr says:

          By why would women in mining or computers be less victimizable than those in healthcare or media?

          I can’t tell if this is a rhetorical question or not, but if we’re accepting the premise that victim attributes impact harassment it could just be that the women in mining and computers are similar to the men in those industries.

          That is, those in mining might often have the general build and attitude of a blue-collar manual laborer, with broad shoulders, callused hands, beer belly, and zero hesistation telling people to fuck off. Those in tech could tend towards the stereotypical disastrous fashion sense, flabby figure, and aversion to human interaction often attributed to the men they work with.

          Lower physical attractiveness means lower marginal temptation to harass, and lower agreeableness means lower probability of a male co-worker convincing himself that a woman is receptive to his clumsy and line-crossing advances.

          On the other hand, the job description for many roles in media involves both being physically attractive and acting friendly regardless of your actual feelings. There isn’t a clear case for healthcare, but its unlikely that victim-attribute differences explain 100% of the variance in harassment rates.

          We could test this by checking whether these stereotypes are accurate by measuring average attractiveness and agreeableness by industry, and checking whether those traits correlate with harassment on either a personal or industry level.

      • Forge the Sky says:

        I mean, not to put too fine a point on it, but do we know if victims of sexual harassment tend to be more sexually attractive on average? Seems like it would be a plausible ‘inviting’ factor which the victim holds no culpability for, but also something that people would tend to not like to talk about.

        It could also explain some difference across industries, with Hollywood being the obvious visible example currently.

        • Mary says:

          Hollywood also has other things that make it a harassment prone industry.

          For instance, there is a vast oversupply of attractive young women for the roles on offer. This makes it more unlikely that she will complain, because she’s easily replaced.

          For another, the criteria for actresses are subjective. Therefore it’s much harder to prove retaliation or reward.

          • albatross11 says:

            It seems likely that there’s some institutional culture thing going on, too. There is a really long history of attractive young women and men getting the casting-room couch treatment. Some of that’s probably the incentives within that industry, but I bet another big part is the culture that has grown up in the industry over decades of its history.

          • Aapje says:

            It is also normal in Hollywood to push actors into being nude with strangers and to grope and kiss them. It seems like a relatively small step from doing that to pressuring people into sex.

            Also, from what I’ve heard, acting schools seem to specifically set out to break down the norms of students, to make them comfortable with doing things that most people consider transgressive. This is akin to grooming.

    • Lambert says:

      Heinlein’s Moon Is a Harsh Mistress put forward that hypothesis, too.

      • Simon_Jester says:

        I suspect Heinlein’s thoughts on this subject were inspired by Wild West settlements, to which the lunar colonies of his novel bear some resemblance. Wild West boom towns could easily have male:female ratios far more imbalanced than the 2:1 ratio he attributes to Luna.

        If Heinlein’s social hypothesis is correct, then in such a society, any given woman will have little difficulty finding men who will cheerfully punish a male who offends against her, and (this is important) this will give her control over how men behave towards her.

        An alternate hypothesis is that his will tend to result in women being ‘parceled out’ by what amounts to a buyers’ cartel among the men, who will enforce their own system of punishing any male who tries to violate the rules of the cartel. While the outcome (men who intrude upon women are harshly punished) is in some ways similar, the process is very different and affords women very little agency.

        One results in ‘free Luna,’ the other results in harems and purdah. The fact that the former is a fictional setting and the latter is a pair of real terms for real things suggests that maaaaybe Heinlein got it wrong, or at least partly-wrong.

        • Calion says:

          Clearly we need an additional element to move to “Free Luna.” I wonder what that would be, and if Heinlein included it?

        • Ben Thompson says:

          You could just as well have said “one results in chivalry, the other results in moon harems and lunar purdah.” Your prose style does not constrain reality.

          • Simon_Jester says:

            I could have, but then I wouldn’t be talking about the novel I was trying to talk about.

            In the fictitious context of that specific novel, a high male:female ratio resulted in women enjoying an extreme degree of freedom to kick aside obnoxious men and rely on other men to protect hem from hostile male behaviors. This is the (fictional) society I described as ‘free Luna’

            In reality, this doesn’t seem to happen very consistently. Thus real social phenomena involving patriarchal family leader figures forming a cartel and imposing coercive control over the scarce ‘resource’ of access to women is… pretty common.

        • littskad says:

          But did men outnumber women 2 to 1 in the places that gave us the terms “harem” and “purdah”? If not, how is it relevant?

          • Simon_Jester says:

            Because if women’s access to the outside world (and men’s access to women) can be locked down by powerful men under conditions where scarcity has to be artificially engineered, it is alarmingly plausible that it could be locked down under conditions where scarcity occurs naturally.

          • littskad says:

            That doesn’t seem right to me. If women are actually scarce, it seems unlikely that some men would be permitted to effectively hoard the few there are. In what sense is scarcity of women “artificially engineered” by harems? It’s my understanding that only a very small minority of men ever had harems, not enough to actually create any effective artificial scarcity. If it had actually created an artificial scarcity, it would likely not have been permitted by the masses of men.

          • Aapje says:


            If those masses of men were oppressed by a tiny number of men, they might have been powerless.

          • Mary says:

            You can’t just “oppress” someone. you have to have the means to do it. Remember that they didn’t even have machine guns at the time.

          • Nornagest says:

            It’s my understanding that only a very small minority of men ever had harems, not enough to actually create any effective artificial scarcity.

            Yes and no. It was rare — low single digits in most of the sources I’ve seen — but if, say, 2% of men each keep four wives, then all the other men are looking at a sex ratio that’s not huge, but enough to change behavior, if my experiences at a college with a similar sex ratio (going in the other direction, though) are anything to go by.

          • Aapje says:


            By your strict definition of oppression, it seems that almost no one counts as having been oppressed, including most slaves. Roman slaves were not kept in line using machine guns, nor African-American slaves, nor the Israelites in Egypt, etc, etc.

            If we use the same kind of standards that are used to claim that most women were oppressed in hierarchical societies of the past, like them having no or very limited political power, then most men were oppressed in these societies as well.

          • Notsocrazy 24 says:


            I think you’re taking Mary’s comment a little too literally. I think the point was that masses of men don’t get oppressed by <1% of the population without some means of doing so. Slaves in various times and places weren't kept in line with machine guns, but there were other structures (and a much more significant portion of the population!) in place to keep them in line.

          • Aapje says:

            @Notsocrazy 24

            Sure, but that is not exactly a remarkable insight.

        • Simon_Jester says:

          The secret to oppression in a pre-machine-gun society is decentralized quasi-official terrorism by an elite class against a servant class.

          If you have a society where a noble can disembowel a peasant just to see if his sword is sharp and get away with it, then that society will quite effectively ‘oppress’ the peasants.

          Sure, your army might run into trouble if fifty thousand peasants outraged by this kind of treatment rose up as one, but that rarely happens, because the vast majority of the people likely to lead such a revolt, or join one in the early phases, are hanging on gibbets on the main road out of town.

          In short, they key is to exploit the coordination problems implicit in ‘revolt for better working conditions.’ And to create a manager/aristocrat caste (~10% of the population) whose job is to exploit and whip the peasants.

          Men are the manager-caste applied to women in a patriarchal society, complete with the state terrorism. Those men may themselves be peasants ruled by a whip-wielding aristocrat, though, who is in turn ruled by a whip-wielding duke.

          • Aapje says:

            Men are the manager-caste applied to women in a patriarchal society, complete with the state terrorism.

            This is very simplistic. There is a lot of evidence that women did and do a lot of that manager/kapo work too.

            For example, women seem to be in charge of the female circumcision apparatus in countries where that is the cultural norm.

      • Steve Sailer says:

        American women got the vote earliest in Western territories like Wyoming because Wild West towns might live or die depending on whether they could attract Respectable School Marms and the like. Three years before women’s suffrage on a national scale, Wyoming elected a woman, Jeannette Rankin, to the US House of Representatives. She was no shrinking violet politically: she was the only member of the House to vote against the 1917 Declaration of War.

    • Speaker To Animals says:

      Is it possible that, precisely because women are under-represented in STEM fields, sexual harassment is more actively discouraged – as, no doubt haters of STEM will argue, a form of benevolent sexism?

      The tech industry seems particularly keen on self-policing/prone to virtue signalling.

      The Damore memo incident would support this. It’s hard to believe factory workers would openly advocate for violence against co-workers they considered to hold conservative views.

      • Notsocrazy 24 says:

        I don’t know that the Damore incident supports that or not, as someone who works in your purportedly different industry of manufacturing. The main reason Damore got fired seems to be the public scrutiny that resulted from the memo getting leaked, not the fact that the internal memo happened in the first place (although I guess I may have the timeline mixed up there, correct me if I’m wrong). I can say that one of my coworkers would have been fired immediately if the incidents made headlines with a Gizmodo leak among other things, yet the actual situation was that he made comments about black people that no one outside the factory knew about, so he basically got the leniency of a three-strike rule before being fired.

        So although there is a contrast between my anecdote and Damore’s, I think that the difference is the publicity of the story outside of what the company had control over, not a difference in virtue-signalling between Google and Nestle.

        • Speaker To Animals says:

          Damore’s lawsuit against Google alleged a hostile workplace existed for anyone with conservative or even classical liberal views long before he published his memo.

          If you can get dogpiled for questioning the belief that the gender imbalance might reflect innate predispositions it’s hardly surprising nobody is going around calling anyone Sugar Tits.

          • Speaker To Animals says:

            Sorry, mangled that last paragraph.

            Should be ‘suggesting’ rather than ‘questioning the belief’.

        • moscanarius says:

          I think you are partially correctly; he was fired mostly because of publicity. The first news I found on the matter are from 5 August 2017, and Damore was fired two days latter.

          But I think it’s important to notice that the memo became an incident outside Google exactly because it was an incident inside the company. Many of Damore’s coworkers were very incensed by what he wrote; even after the memo hit the fan and Google had to do something about it, we kept reading article after article by tech people condemning Damore for his supposed sexism.

          • Steve Sailer says:

            Damore got fired because Youtube CEO Susan Wojcicki, who used to be Sergey Brin’s landlord and then sister-in-law, felt personally aggrieved by his memo — How would she answer its arguments to her five children?

            So she spoke Power to Truth.

          • Steve Sailer says:

            The funny thing is that James Damore was explaining the rational reasons behind what Susan Wojcicki and the other Google executives actually do: hire mostly males for tech jobs.

          • moscanarius says:

            I remember you writing this at the time, but I’ll insist that Wojcicki only knew about it because quite a number of Damore’s colleagues were outraged by the memo and willing to make sure EVERYONE ELSE heard about his sins.

        • The Nybbler says:

          IMO, speaking as an ex-Googler, he almost certainly would have been fired even if it hadn’t leaked but only gone viral within the company. It just would have taken longer.

    • dick says:

      > No doubt, it is explained, all those horrible, evil male engineers must be teaming up to exploit the handful of female employees. After all, men and women are enemy genders. I mean, that’s what every lesbian women’s-studies professor says…

      I’ve been linked to dozens of your essays over the years, and you seem like a smart guy with interesting things to say, but it also seems like you have a congenital inability to paraphrase positions you disagree with without exaggerating them into a caricature.

      • One Name May Hide Another says:

        Actually, I’d guess most of us are here because Scott is really good at it. When he tries. Sometimes he doesn’t try and exaggerates for comedic effect. Like in this example. It’s meant to be humorous.

        • dick says:

          The bit I quoted was from Steve Sailer, not Scott Alexander. I agree that Scott generally does make a good-faith effort to restate the views he disagrees with in a way that the holders of those views would agree with, and that that’s rare and part of what makes his political posts so interesting.

      • Steve Sailer says:

        It’s called satire.

        • dick says:

          I don’t disagree, mockery and exaggeration are classic elements of satire, but they’re also anathema to useful discourse. I get the impression that you believe you’re engaging with liberal views but doing so in a forceful and entertaining way; I would argue that satirizing a position is the opposite, a way to avoid engaging with it.

        • Aapje says:

          @Steve Sailer

          This kind of satire can be entertaining to the ingroup, by making them laugh at a stereotype of the outgroup. Doing that is not appropriate at SSC.

    • Scott Alexander says:

      Banned for one month. Usually content like “men and women are enemy genders. I mean, that’s what every lesbian women’s-studies professor says, and they wouldn’t have any incentive to lie, would they?” will result in a permaban, but quoting yourself is a sufficiently weird edge case that I’m not sure what to do with it.

      • Aapje says:

        That seems a little excessive. I agree that it is type of argumentation that has no place here, but he is a regular commenter who doesn’t usually do this, so I would suggest a slap on the wrist (one week ban).

        • Scott Alexander says:

          Oh, fine.

        • Notsocrazy 24 says:

          This also doesn’t seem that out of line considering the obvious sarcasm. It’s like one point above the amount of sarcasm that Scott allows himself in a lot of #ThingsIWillRegretWriting posts.

          • Aapje says:

            Perhaps, but it was just not very amusing or insightful, so it comes across as baiting or outgroup-bashing. Scott tries to be generous to the outgroup as well, so it’s more love/hate than hate/hate.

  4. shakeddown says:

    Another factor could be the number of people/number of interactions you have at work. Women in tech may interact with a higher percentage of men, but tech workers interact with fewer co-workers in the course of their jobs than in interaction-heavy jobs like media or law, so they still have fewer overall interactions with men, hence fewer interactions with potential harassers.

    • bayesianinvestor says:

      The number of interactions I have doesn’t just influence my risk of meeting a harasser. It also affects how I react to people. If I spent 90% of my work time interacting with people, I’d be pretty quick to label additional interaction as unwanted. But since I currently spend close to 0% of my work time interacting with people, I’m relatively open to wanting additional interactions with people. I.e. the data may be saying mainly that STEM workers are lonelier than average.

    • Squirrel of Doom says:

      That’s another way of saying that people in tech work with the same people day after day.

      I think it’s much easier, and lower stakes, to harass people you only encounter occasionally.

    • zzzzort says:

      I also wanted (somewhat seriously) to see the out of workplace complaints normalized to the amount of time/interaction people had outside of the workplace. Maybe people in the arts are not more sensitive to harassment, they’re just more social and apt to spend time around potential harassers outside of work (my initial crotchety scientist objection, that maybe artists just spend more total hours not at work, does not fit with the high correlation across several fields, nor with anecdotal evidence for how much time people in the arts spend working).

  5. doubleunplussed says:

    One remaining confounder is that people might hang around the same sorts of people outside of work as they do at work. Filter bubbles and all that. So you can’t necessarily normalise by outside-work harassment rates, since these may be signal, not measurement bias. Measuring actual differences in harassment rates between cultures, not just their propensity to identify an act as harassment. But I understand this normalisation is the best you can do in the absence of more objective questions about specific acts.

    • pjs says:

      The EEOC data seems (if I’m looking at it correctly) to have number of filings plus a breakdown by all sorts of possible resolutions (including many ending in ‘no reasonable cause’: “EEOC’s determination of no reasonable cause to believe that discrimination occurred based upon evidence obtained in investigation.”)

      Couldn’t this be used to (try to!) normalize the culture-gap across industries? The EEOC sees complaints from all of society, so hopefully is somewhat industry independent, and so if industry A has 30% N.R.C. vs industry B’s 90% it might be meaningful (and the finer disposition statistics could be used to make this cleverer). This use of the EEOC numbers would not be affected by Scott’s absolute number complaint.

    • Φ says:

      Well said, doubleunplussed. A better question than “Have you ever been sexually harassed or assaulted at work?” for next time might be “Have you ever been sexually harassed or assaulted by a co-worker?”

      I personally have almost no interaction beyond the incidental with people to whom I don’t have either a work or church related connection (not counting family members). I suspect that I am not alone in this regard. Which causes me to marvel at rates of harassment “outside of work” so much higher than “at work”. Seriously, where are you meeting people?

      • Evan Þ says:

        At church, at book clubs, at hobby groups, on public transit… If I went to bars or talked to my neighbors, those’d be two other huge opportunities.

        (Myself, I haven’t come close to being sexually harassed on any of those occasions. Your results may vary.)

        • Φ says:

          Evan: to clarify, public transit and bars would have gone in that “incidental contact” category. Otherwise, I guess I’m lucky from a harassment-avoidance perspective that my hobbies are solitary and I read without a club . . .

      • Aapje says:

        You have to keep in mind that “Have you ever been sexually harassed or assaulted outside of work?” includes childhood and student life.

  6. Scott says:

    Perhaps the sort of people who go into STEM are socially oblivious in a way that prevents them from both noticing if they are sexually harassing someone else, and from noticing if they are sexually harassed.

    This is close to my personal intuition – introversion is overrepresented in these fields, and I would hazard that introversion is negatively correlated with being sexually forward. My other thought is that both men and women in the stem fields tend to be less physically attractive (because of lack of effort and because attractiveness is less of an advantage than in other fields), but I’m not sure that this matters much if at all to harassers.

    • Aapje says:

      I also suspect that scrupulosity is higher among more nerdy types. I certainly feel obliged to not help create a ’10 guys trying to woo one woman’ situation.

  7. atribecallederek says:

    Has there been any hypothesis about combining the ever-trustworthy Myers-Briggs personality types to the job occupations?

    It seems like the types of occupations with higher harassment rates might also attract the “F” type personalities; it also seems more likely the “F” type personalities are more likely to act off gut emotion and either a) commit an act of harassment on a whim, or b) take an unintended act from someone else as an act of harassment, and report it as such.

    This might explain the difference between the “feminine” job types, and the more “T” personality/analytical job types.

    • BlindKungFuMaster says:

      Yeah, as a recovering mathematician I’m not surprised that maths has the lowest harassment rates, just because of the personality of people who study maths. There should be some decent correlation with how agreeable and introverted people in the respective fields are.

    • mdet says:

      This SSC post tries to connect what’s more or less T vs F Myers-Briggs to gender imbalances in a field, but not harassment specifically.

      “Feeling” Myers-Briggs in my understanding doesn’t exactly match your description of seemingly impulsive or moody people though.

  8. Rolaran says:

    Another possibility is that the respondents to the survey may be skewed for some reason. For example, they may have been selected by this blog’s previous discussion of these issues, which have also argued that tech is not a uniquely bad industry.

    I do feel like this is likely to be a factor. I feel like a person who has had bad experiences in the tech industry is less likely to have continued reading a blog that argues (among other things) that problems in that field are overemphasized, and a person who has had nothing but good experiences in tech is more likely to have begun hanging around a blog where they felt like the author was sticking up for them. This is useful data for determining the experiences, preferences and demographics of SSC readers, but I would be considerably surprised if that could be generalized without some serious legwork. (Which, to be fair, it looks like you’ve implemented – I can’t even pretend to be a statistician but it looks like you found some relevant control sets.)

    Surprised you didn’t link this, since this looks to me like you putting your money where your mouth is and offering the data that was missing in one case of that fallacy? And just saw that you did in fact do that.

  9. drunkfish says:

    I don’t think not-at-work harassment is a very good thing to control with, and I think the high correlation supports this. People’s out of work social lives are highly correlated with their at-work social lives, not to mention workplace culture will be influenced by local culture. I think you’d expect to see correlating changes if you could, for example, track a single person who moved regions+jobs.

    I don’t know what a good proxy for personal interpretations of harrassment is, but one that comes to mind is something about seeing harassment around you? Still biased by local culture though…

    • BlindKungFuMaster says:

      Seconded. I would control with “have you ever been harassed by a stranger” or something like that.

      • Aapje says:

        ‘Stranger’ seems like a very ambiguous term. Is the person who you just met at the bar, who gave you her name and who talked with you for half an hour, a stranger?

  10. jw says:

    I am not surprised at all about your results regarding STEM harassment.

    My working theory would be that men who went through school being harassed for being nerds and geeks, actually don’t want to harass others the same way.

    I think that the media’s attacks on STEM are partially fueled by that same sentiment and not by real evidence (i.e. the cool kids still jeer at the nerds).

  11. Chris says:

    Relevant reading:

    Dating isn’t the only reason high school seniors should consider gender ratios when selecting a college. Given the epidemic of campus rape, teenage girls and their parents are justifiably concerned about safety, just as teenage boys and their parents are worried about false accusations. What does any of that have to do with gender ratios? Well, there have been multiple studies showing a correlation between gender ratios and rates of sexual assault. As counterintuitive as it may sound, elevated rates of sexual assault are a predictable feature of communities with oversupplies of women, according to studies by sociologists Nigel Barber and Robert O’Brien.

    The opposite is true of communities with oversupplies of men. Columbia University economics professor Lena Edlund investigated the impact of lopsided sex ratios in China, where young men now outnumber women by 20% due to sex selection, abortion, female infanticide, and other outgrowths of China’s old “One Child” policy. Edlund and her co-authors discovered that although overall crime rates went up in China as the gender ratio skewed more male—not surprising given that men are more prone to criminality—there was a precipitous decline in rape. It seems that men treat women better, and protect them more, when women are in shorter supply.

    Can I prove beyond all doubt that Edlund’s and Barber’s findings also apply to college campuses—i.e. that rape is less common at schools that are at least half male? No, because the available data on campus rape tends to reveal as much about how forthright colleges are in handling sexual assaults—and how comfortable women feel reporting them—as it does about the actual frequency of assaults on a particular campus.

    That said, I was intrigued by a recent Washington Post story on the topic. The article ranked 27 top colleges by their sexual assault rates, and I couldn’t help but notice which college had the lowest rate.

    It was CalTech, a school that is 59% male.

    • tjohnson314 says:

      As a Caltech alum (class of 2013), I’m proud that we have such a low rate of sexual assault. But Caltech is an outlier in so many ways that I don’t know what that proves.

      In particular, the Honor Code is very strong at Caltech. For example, we were allowed to have practically every exam as take-home, including the ones with time limits, which we enforced for ourselves.

      And the Honor Code is widely understood to prohibit taking advantage of other students in non-academic ways as well. So it doesn’t surprise me that our rate of sexual assault would be so low, apart from any other gender ratio effects.

    • Edward Scizorhands says:


      If there are few women, they have strong power in the dating market, and can afford to make waves (and everyone knows this). If there are many women, they have less power, and feel the need to put up with crap (and everyone knows this).

      • vV_Vv says:

        Expanding this, women tend to compete with each other in terms of attractiveness, both to actually attract males and as a status game played among themselves.

        In places where there are few women, they will tend to wear plain clothing, no makeup and they don’t flirt with men, hence men will read them as sexually unavailable. In places where there are many women, they will tend to wear makeup, provocative clothing, and to flirt with men. Men will read them as sexually available and make advances, a fraction of which will end up being perceived as sexual harassment.

        • Watchman says:

          This fits with the oft-repeated observation that girls at single sex schools tend to dress more revealingly (short skirts against trousers being the most obvious manifestation) than at mixed schools.

          • Simon_Jester says:

            Remember the question “is a zebra a black horse with white stripes, or a white horse with black stripes?”

            I think this raises a similar question.

            Is it that women are naturally inclined to dress more revealingly, but cover up more when more males are around, to avoid unwanted attention or the obnoxious behavior of specific controlling males?

            Or is it that women are naturally inclined to dress less revealingly, but cover up less when more females are around, out of a desire to compete with them for status and wanted attention?

            I suspect that which hypothesis people gravitate towards says a lot about their priors regarding gender behavior. I wonder if there’s a way to experiment on this question?

      • Forge the Sky says:

        I would add: where there are fewer women, men will also tend to defend them more against the advances of other men, both altruistically and as a way of improving their own chances.

      • eccdogg says:


        My home state split the majors allowed between the two major state universities. One school was not allowed to offer engineering and land grant stuff and the other was. This resulted in one university having a more applied focus and the other more liberal arts.

        Consequently the gender balances at the two schools was very different. When I went to the more applied school it was something on the order of 60-40 male. The other school was roughly 60-40 female.

        Going to one and having friends at the other as well as visiting the other to party, I found the treatment of women to be very different. Women were treated with vastly more respect at the engineering school, they were treated as a commodity at the other. My guy friends at the other school were admittedly hound dogs, but who know if the situation caused that or not. They were not like that in HS.

    • Nyx says:

      In Mormon fundamentalist cults like FLDS, expulsion of men is commonplace and goes hand in hand with subjugation of women (polygyny is practiced enthusiastically). Anecdotally, the low number of men in black American communities (due to mass incarceration) leads to black women having a weaker bargaining position and being left to raise families on their own.

      • dndnrsn says:

        Anecdotally, at least, it would appear that women in a very different environment from poor black communities – university-educated bubbles, and especially undergraduate university itself – might face a similarly unfriendly market due to the shifting ratio of women to men getting degrees. I was at a school with a nearing-2-to-1 female-male ratio, and it was not unusual to see guys with women who were more attractive than they were, and frequently also both superior both in intellect and in character.

        That was a relatively benign manifestation, but it seems fairly obvious how the imbalance could lead to women putting up with worse behaviour from men, especially higher-status men.

    • Eric L says:

      I think you guys are overthinking this one. As noted in the article, in environments with more men, women find long term boyfriends, spend less time single, and have few partners in total, whereas where there are more women they spend more time single and have casual relationships with more men. Most rapes are committed by acquaintances, especially dates or guys you met at a party/bar. So a big rape risk factor for women is looking for new partners, which is something women spend more time doing where the dating environment is not working in their favor.

  12. meh says:

    This may be the single topic where the extent of public interest is most disproportionate to the minimal amount of good research done.

    I would guess there is a fair amount of reputation risk in studying this topic.

  13. Chris says:

    I’d like to see what the last graph would look like if you excluded male victimization. Let’s be honest, nobody really cares about that and it might be confounding the results to find more harassment in gender-balanced fields if both men and women are more likely to be harassed in more fields with more women or men (respectively).

  14. Tracy W says:

    The low rate of harassment in STEM calls for this story:

    An engineer was crossing a road one day when a frog called out to him and said, “If you kiss me, I’ll turn into a beautiful princess.” He bent over, picked up the frog, and put it in his pocket. The frog spoke up again and said, “If you kiss me and turn me back into a beautiful princess, I will stay with you for one week.” The engineer took the frog out of his pocket, smiled at it, and returned it to the pocket. The frog then cried out, “If you kiss me and turn me back into a princess, I’ll stay with you and do ANYTHING you want.” Again the engineer took the frog out, smiled at it, and put it back into his pocket. Finally, the frog asked, “What is the matter? I’ve told you I’m a beautiful princess, that I’ll stay with you for a week, and do anything you want. Why won’t you kiss me?” The engineer said, “Look, I’m an engineer. I don’t have time for a girlfriend, but a talking frog – now, that’s cool!


    • gbdub says:

      There are of course many along these lines but one of my favorites is:

      An engineer sees his colleague riding by on a brand new bicycle and flags him down. “Sweet ride! Where’d you get that?”
      “Well,” says the second engineer, “I was walking and this beautiful woman pulls up next to me on this bike. She jumped off the bike, tore off all her clothes, and said ‘Take whatever you want!’”
      “Good choice taking the bike,” the second engineer replied approvingly, “the clothes probably wouldn’t have fit”

      • Edward Scizorhands says:

        A lawyer, a doctor, and a mathematician are debating the merits of a girlfriend versus a wife.

        The doctor says to have a wife. There is more fidelity and you are exposed to less STDs.

        The lawyer says that’s dumb. You can wear a condom and it’s cheaper to get rid of a girlfriend.

        The mathematician says they are both dumb. You should have both a wife and a girlfriend. Because while your wife thinks you are with you girlfriend, and your girlfriend thinks you are with your wife, you can get some math done.

  15. Emma the trust fund baby says:

    One explanation jumps out to me: bargaining power of employees. In fields where employees have low bargaining power (the blue ones), sexual harassment is a cost they have to bear. In fields where employees have high bargaining power (the red ones), employees can easily move to other organizations if they encounter harassment. This incentivizes organizations to reduce sexual harassment.

    • stevemckay says:

      I don’t believe this is the case. That sounds a lot like saying that companies won’t discriminate against women because discrimination is inefficient, which sounds reasonable, but the data disagree. Additionally, I think it’s reasonable to view ignoring complaints of sexual harassment, which predominantly from women, as a form of sex discrimination.

      And labor mobility, at least in tech, isn’t as large as you might think given the amount of complaining that employers do about their inability to hire talented people. The labor market is incredibly disfunctional.

      • Aapje says:

        That sounds a lot like saying that companies won’t discriminate against women because discrimination is inefficient, which sounds reasonable, but the data disagree.

        It depends on what you count as ‘discrimination.’ Discrimination merely by gender/race/etc seems obviously inefficient. Treating people differently based on factors that impact their job performance & that happen to correlate with gender/race/etc (and thus have disparate impact) is quite efficient. Activists very often conflate the two.

        Additionally, I think it’s reasonable to view ignoring complaints of sexual harassment, which predominantly from women, as a form of sex discrimination.

        Such a collectivist ‘disparate impact’-based argument can lead to strange conclusions, where people who get treated better individually are portrayed as worse off.

        For example, imagine a company where 40 women and 10 men consider themselves to be sexually harassed. Let’s assume that each of these complaints deserve strong scrutiny. The company has two people who deal with sexual assault allegations. One of these, Mary Koss, believes that men cannot be harmed by sexual assault and only takes sexual assault of women seriously. The other, Bob Lawyer, is an ass-covering bureaucrat who tries to suppress all allegations.

        So if each handles half the allegations, then 20 women and 10 men have their complaint not taken seriously. So if I understand you correctly, you believe that this company has sex discrimination against women.

        However, only men’s allegations are actually suppressed for them being men, given the motivations of Mary and Bob. There is no actual misogyny here. Furthermore, the individual experience of a complaining woman is that she has 50% chance of being taken seriously, while it is 0% for a complaining man.

        Of course, in practice men who figure out that they have 0% chance of being treated fairly may not come forward at all, so the gender disparity in sexual assault that you use as a basis for concluding that sex discrimination against women exists, may actually (partly) be caused by sex discrimination against men.

      • Emma the trust fund baby says:

        I maybe could describe my position more accurately. In my model of companies, there is a cost X to enforcing rules against sexual harassment (takes manager time, expensive to fire and retrain people, others), and a cost Y to sexual harassment (people who get harassed don’t like it, affecting their performance, or they leave). They will enforce the rules if X is lower than Y.

        When employees have high bargaining power, Y is high. They have options, so they can leave, or go to the media, etc. When they have low bargaining power, they can’t leave as easily, they will be blacklisted if they go to the media, and so on.

        • albatross11 says:


          If this explanation is part of the picture, it should show up in surveys of other non-sexual kinds of bad treatment/harassment/bullying at work.

      • Darwin says:

        I think it matters here that the argument is marginal – not that they won’t discriminate/harass *at all* for this reason, but that they might do so on average *less* than another industry without such a structure.

    • peterispaikens says:

      “fields where employees have high bargaining power (the red ones)” seems fishy to me – this would be a plausible explanation for the Hollywood scenario, but I’d assume that employees in law and medicine have much more bargaining power than, say, in non-finance math (which has the lowest harassment rates), as evidenced by the disparity of average income in these fields.

      • Watchman says:

        I think you need to consider the supply of potential workers. There are a lot more trained lawyers and doctors than professional mathematicians for various reasons. And lawyers and doctors are to some extent interchangeable (at least when young and malleable) whilst mathematicians come with specialisms already loaded so are less easily swapped.

        • albatross11 says:

          I suppose it would be interesting to see whether a lot of the harrassment targets in medicine were residents/fellows who are kind-of chained to their jobs if they want to be able to eventually practice medicine in their chosen area.

  16. stevemckay says:

    The CSV is definitely broken. Don’t use it. Looks like multi-line free text entries got CSV’d brokenly. It may not be fixable.

    P.S. is the source data available in some more…usable…format? CSV and XLSX are possibly the two worst data interchange formats on the planet.

    P.P.S People who like SQL can go at the data with BigQuery:

    I got a clean CSV export from the XLSX so here are links to alternate formats:

    CSV (gzip):
    JSON (gzip):

    • Scott Alexander says:


    • Error says:

      For tabular data, I find tsv makes a strictly better format than csv. Most datasets will have commas in them somewhere, but very few will have hard tabs.

      Still has an issue with newlines, though.

      • stevemckay says:

        TSV still suffers from having no types. And when tabs do occur, how should they be represented? How do you distinguish NULL from empty? If you do have a special NULL value, how do you handle that value appearing in the data? How are dates and times represented?

        Of course those problems all occur to some extent in any format in common usage. It’s where a large part of my job comes from. Hell, recently my job is setting up exports for data scientists and I’ve been finding that even the “clean”, already-loaded-and-processed data we use in production, needs fixing in flight to make it usable for sciencing.

        • Error says:

          A lot depends on what you’re trying to do with it. If you want to represent a complete self-describing object hierarchy or something, to be processed in a fully generic way, obviously tsv is the wrong tool. I like yaml for that. Cue holy wars.

          But for survey results where all records are flat, you know their types in advance, and mostly you just want to view them as ordinary spreadsheets? tsv works fine. Represent literal tabs as \t or something. Nulls, where applicable, are empty strings — there aren’t that many cases where an empty string and NULL make sense as distinct values in the same field, and I’d be surprised if any of them applied here. Date format doesn’t matter as long as it’s consistent.

          Sure, it makes processing it more complex, in that metadata not given in the file format must be specified in the code. But it makes *reading* it with stock tools much easier.

          Edit: I think what I’m actually plumping for is something like DSV, i.e. escape your delimiter and special characters, but don’t do MS-CSV’s retarded quoting garbage. TSV with quoting turned off is just an easy way to accomplish something close to that from spreadsheet tools. Tabs are not the ideal separator, but they’re less bad than commas, and easy to use in Excel/Calc.

    • ASparklingViking says:

      The bigquery link is broken for me: “Unable to find dataset: ssc-survery:survey_data”

    • terran says:

      I’m getting “access denied” errors for your csv and csv.gz links, unfortunately.

    • dcwuser says:

      I also get 403 Forbidden. Any other download site?

  17. userfriendlyyy says:

    With the men reporting victimization do you have data on the gender of the victimizer? I wonder just how many of the male victims might more accurately be described as a gay guy made a pass at them because they weren’t sure about their sexual preference and promptly moved on once rejected. I also wonder how many of the women were sexually harassed by other women; separating out ‘slut shaming’ type vs. lesbian interest. Maybe questions for next time.

    • Scott Alexander says:

      No, I don’t have such data.

      I do have data on whether offenders report themselves as being gay or straight in general. IIRC there wasn’t a huge difference in at-work harassment between gays and straights.

      • albatross11 says:

        Any self-report data w.r.t. sexual harassment probably has some element of this issue, though. Different people have different notions of when some interaction crosses the line into harassment, and there are definitely kinds of banter which seem harmless to one person and really uncomfortable to another.

        Actually, this is the place where I’d expect people with fewer people skills to get higher levels of sexual harassment. Andy (with some level of social skills) makes some lewd jokes and sexual references around a female coworker, recognizes that she’s not uncomfortable with it, and stops. Bob (with fewer social skills) makes some lewd jokes and sexual references, and doesn’t pick up that her uncomfortable insincere laugh and eye-rolling mean she’s actually pretty uncomfortable with it and wants him to stop.

        • SaiNushi says:

          “Actually, this is the place where I’d expect people with fewer people skills to get higher levels of sexual harassment.”

          Your example which follows is about the person who is doing the harassing not having social skills, but your conjecture is about the person being harassed not having social skills.

          As someone who spent my teenage years completely oblivious to almost every sexual overture… I can verify that a woman who doesn’t have social skills might not realize she’s being hit on, let alone feel harassed. I wouldn’t be surprised if that applies to men without social skills as well. Put a man who can’t read social cues with a woman who can’t read social cues, and the man’s attempts to hit on her go right past her head, and he never gets any indication that anything is wrong because she didn’t even notice, so his attempts continue and possibly escalate until she DOES notice, but by then she’s just flattered that a guy would put that much effort into getting her attention. Put a man who can read social cues with a woman who can’t, and his attempts to hit on her are met with indifference, which causes him to give up long before she notices that he’s hitting on her.

  18. dark orchid says:

    In media coverage, STEM fields are portrayed as full of “techbros”, and uniquely unwelcoming to women. How do we square this with the data?

    My hypothesis is that they are mixing up two things: (a) for want of a better name, “high-systematising behaviour and perhaps somewhat disregarding social norms” and (b) harassment.

    For an anecdote of a woman leaving tech due to (a), I give you As a Woman In Tech, These Are Not My People:

    No, the reason I left is that I came into work one Monday morning and joined the guys at our work table, and one of them said “What did you do this weekend?”

    I was in the throes of a brief, doomed romance. I had attended a concert that Saturday night. I answered the question with an account of both. The guys stared blankly. Then silence. Then one of them said: “I built a fiber-channel network in my basement,” and our co-workers fell all over themselves asking him to describe every step in loving detail.

    At that moment I realized that fundamentally, these are not my people. I liked the work. But I was never going to like it enough to blow a weekend doing more of it for free. Which meant that I was never going to be as good at that job as the guys around me.

    If you ask people “have any of your work colleagues ever made you feel uncomfortable?” or anything like that where you’re not careful to separate (a) and (b), it sounds reasonable to me that you’ll pick up a mix of both.

    I would really love to have good and reliable numbers on sexual harassment in tech because I can see three cases:

    If there really is more harassment in tech than in other industries, then we have a huge problem and it should be one of our highest priorities to do something about it.

    If the harassment levels are broadly similar to other industries, we internally should be doing everything we can to reduce the levels but externally, the media going on about tech is about as useful as the claim in Breitbart that illegal immigrants commit 3.8% of federal sex abuse cases – without mentioning that they’re also 3.8% of the population. (source: lies, damned lies and the media part 6)

    If harassment levels in tech are actually lower than other industries, we should still be working internally on “aiming for zero” but we need something a lot better than media reports to figure out how to do that.

    • samuelthefifth says:

      Exactly. I’ve been in tech for over 15 years across many jobs and have not once met a techbro. Those who get called techbros seem very nerdy, spergy types who’ve never thrown a football in their life. I’ve observed tech feminists of the Shanley Kane variety use this epiteth in vindictive and malicious ways. And I hear Silicon Valley is full of asshole business types.

      My somewhat shitlordy thesis is that queen bee types are coming into tech, find themselves both tittilated and offended by the high status men who are close to the big money, and then transfer that discomfort onto the low status engineers who actually keep the place running. “Angry men in my mentions” is code for “your feelings don’t matter and you’re supposed to keep your opinions to yourself”, and I see this play out daily on tech twitter. It was also the script behind 2013s donglegate, where an anti harassment activist harassed two guys on Twitter and managed to paint herself the victim.

      It is particularly glaring that they’ve all collectively forgotten that being a computer aficionado in the 80s and 90s made you an outcast. We didn’t need social approval to get into this, and social shaming isn’t going to work now either. The result is a frantic moral panic of foot stomping and melodrama, where a minority of women run to the media to monopolize the entire industry’s image.

      • cryptoshill says:

        I wonder if “disproportionate cultural status” (Engineering has always been a highly paid profession, but only in the last 20 ish years or so has it enjoyed the level of cultural cachet that used to be for Doctors, Lawyers, and Wall St. Traders) attracts “people who are likely to overreport the incidence of sexual harassment”. I will call these people the “victim adjacent” in my model. My model for the broader conversation about this is that there are really six kinds of people:
        Sexists: People who think women are inherently worse at certain professions and want them out.
        Sexist-Adjacent: People who behave in a sexist way for non-sexist reasons. Here is where I place the MGTOW types who refuse to interact with women because they truly believe that false accusations of misogyny are life-ruining.
        Abusers: People who are actually perpetrating the sexual harassment that we are attempting to reduce.
        Abuser-Adjacent: People who frequently engage in behavior that if totally oblivious (or malicious) *could* be construed as harassment (donglegate specifically comes to mind).
        Victims: This is pretty obvious.
        Victim-Adjacent: People who take on the mantle of preventing sexual harassment and abuse to further their own agenda.
        I would like to point out that the majority of the “culture war” conversations seem to be between Abuser-Adjacent and Victim-Adjacent types, and thus not really about preventing sexual abuse. I would like to see good data on whether or not previous “high status” occupations exhibited more or less sexual harassment as well as their incidence rate of “massively overblown scandals”.

        • Aapje says:

          Engineering has always been a highly paid profession, but only in the last 20 ish years or so has it enjoyed the level of cultural cachet that used to be for Doctors, Lawyers, and Wall St. Traders

          Really? It seems to me that those first two professions get a level of respect that Wall St. Traders and engineers don’t get.

          I also consider your ‘adjacent’ types to be very uncharitable and well-poisoning. People who worry about men being victimized are not sexist-adjacent, in that they seek to help sexists. They seek to help people they perceive as being victimized. Victimization is not just something that happens to women or that happens on one dimension.

          I also don’t think that the ‘victim-adjacent’ people don’t actually care about sexual abuse. If anything, they care so much that they are willing to do very harmful things which they think are necessary to solve the problem.

      • herculesorion says:

        It is particularly glaring that they’ve all collectively forgotten that being a computer aficionado in the 80s and 90s made you an outcast.

        The new hotness is to suggest that it wasn’t interest in nerdy topics, it was just your personality in general that made you an outcast, and that you aren’t actually any smarter or better than the jocks; they just didn’t care enough about computers to bother learning anything about them. (This is Freddie de Boer’s feeling on the matter.)

        • Freddie deBoer says:


          • herculesorion says:

            well, you deleted the post where you said this so I guess we can’t talk about it anymore :V

        • Conrad Honcho says:

          This does not fit my school experience. If interest in computers and social skills are not correlated, then why were no socially adept people (i.e., the popular kids) interested in computers?

          • Freddie deBoer says:

            I’m guessing there’s a ton of confirmation bias in how you perceive the social standing of the people who do computers.

          • Aapje says:

            @Conrad Honcho

            The obvious answer is that if being into computers is stigmatized, then socially adept people who have other options would shun the field.

            This is similar to how many gay men used to get married to women in the past, because of stigmatization. This doesn’t reflect a lack of interest in having a intimate relationship with men on their part.

      • The Nybbler says:


        I’ve heard that one, and I guess it’s supposed to be a devastating put-down. But mostly it evokes feelings of contemptuous amusement and a desire to say “Show me the code, genius”.

      • arlie says:

        I’ve been in tech since 1978. I’m not sure what constitutes a techbro. But I’ve certainly encountered assholes whose behaviour had a strong gendered element. And I’ve certainly seen the experience played out in real life – females getting more invitations than they could cope with, and males getting none (and in some cases, whining about that to any female who didn’t run fast enough).

        I’m having trouble describing the gendered misbehaviour. Examples are all over the map. One gentleman wanted to discuss his marriage with his only female peer at work – who wanted to talk about linux kernel issues, not absolve the coworker of his fear that he was preventing his wife from achieving her full potential, by relegating her to homemaking. Several favoured flirtatious humour. One cornered a female coworker in a printer room, blocking the way out, so he could explain his position in whatever problem they were having. A large number favoured lunchtime conversation complaining about women in general, but with emphasis on their wives; in one case their female coworker eventually changed the direction of the overgeneralization by rolling her eyes and talking about “spouses” and/or the art of compromise/staying married. In one workplace, most women were clerical, and they would include young female engineers in discussions of managers who “couldn’t keep their hands to themselves”. And in one workplace there were young women insisting on discussing male attractiveness – making things uncomfortable for other women who did not wish to participate. (I don’t recall them doing this in mixed company, except perhaps by accident.)

        Other things were generally more problematic, such as the too common presumption that females weren’t interested in advancement, or would quietly do the scutwork while males became the “stars”.

        I’m of two minds about complaints about interpersonal style in tech. In many cases, “normal” behaviour involves things women are strongly socialized never to do – such as fairly blatant aggression. I believe the stories that if women conform to the style, it will often show up on their performance reviews as excess aggression – and if they don’t, then it will be “failure to fit in” instead. Without the performance review effect, I’d be very ambivalent about this. (It’s reasonable to expect people to learn to speak the local interpersonal language, even if it’s easier for some than others. But at the same time it’s unfair if there’s only one such language, and it’s one only some people speak naturally.) *With* the double bind though, I’m not ambivalent at all.

        And on the third hand, I’m a nerd. At 60, I’ve learned to speak and understand “normal” fairly well. But I had the social skills of a brick early in my career, and would have been happy to be left alone with only a computer to talk to. I want modern 20 year olds of the same type to have the same chance I had.

        • mtl1882 says:

          I think you pinpointed a lot of common issues in this post – things that aren’t egregious and may seem harmless to the person crossing the line, but that definitely affect people. I definitely agree regarding the aggression issue. I’m a 5 foot tall female, pretty low key and reserved. I never got accused of any aggression growing up. After going to law school and working a bit, I became more assertive, and have a direct, somewhat blunt style, but I’m in no way angry or competitive. Yet I sometimes get called nice euphemisms for aggressive, and my manner gets compared to really blunt, aggressive men who are way older and more advanced than me. Meanwhile, women who actually are cutthroat adopt a really agreeable persona, and everyone buys it. People just interpret female assertiveness bizarrely. There’s a quote along the lines of a man has to be Hitler to be called aggressive – all a woman has to do is hang up on you. I saw an article on Barbara Bush today that mentioned her “acid wit” and used as an example a very benign joke, not the least bit offensive. How often does a guy get accused of having an acid wit?

          • taradinoc says:

            A couple competing hypotheses:

            – Women are being compared to their peers, and people saying “yikes, she’s aggressive” really mean “she’s aggressive for a woman”, even if it’s a level of aggression they wouldn’t comment on for a man.

            – When women make a point to be assertive in a way that gets results in a male environment, they’re in unfamiliar territory, so they overdo it and that’s when they get noticed.

            Regarding the second one, this article compares the behavior of people interrupting each other at work by gender. In addition to the expected result (men interrupt more, and they interrupt women more than other men), it also finds that the highest-ranking women interrupt everyone and do more interrupting than anyone else.

            The correlation between seniority and this sort of aggression suggests it’s an adaptive behavior, but the fact that women don’t do it much among themselves suggests that, if there’s any nuance involved in deploying it without seeming aggressive, they’re less likely to have the practice to do so.

          • Aapje says:


            A possibility is also that benevolent sexism makes it less acceptable for women to be assertive/aggressive, because it is seen as ‘double dipping.’

  19. realwelder says:

    Another possible explanation for there being less harrassment in “thing” professions compared to professions with a large “people” component:

    Introverts (by aptitude and preference) are likely overrepresented in “thing” work and underrepresented in “people” work.

    After all, harassing someone would require voluntarily interacting with them. Which of course, introverts are less inclined towards.

    • Aapje says:

      I also wonder if there is a difference in the kind of socially inept behavior, in the sense that thing-oriented people may commit generic social faux pas’s more often in their normal interactions, but may often actually be very wary of going too far in a sexual context; while people-oriented persons may tend to be very confident of their natural social skills & may often be unaware of the possibility of making errors in a sexual context.

      Some natural social skills that work pretty well for in non-sexual interactions may not work at all well in sexual situations.

  20. sty_silver says:

    My intuition is that working in nerdy fields correlates (A) with introversion and (B) with low self-deception, both of which correlate negatively with comitting sexual harassment.

    I feel like I missed an opportunity to make a prediction beforehand.

  21. Aqua says:

    Another possible confounder:

    Does attractiveness correlate with being harassed?
    It’s possible certain fields have higher, either proportional or absolute, numbers of attractive women. If these are more likely to be victims, this would result in more harassment in more “attractive” fields, or also in fields that have more even gender ratios, as there would be more attractive women.

    • Aapje says:

      Perhaps the same goes for attractive harassers too. Studies show that people tend to be very positively inclined towards pretty people, which may give these people too much confidence.

      • Forge the Sky says:

        Complicated, of course, by the possibility that sexual behavior coming from an attractive person is less likely to be perceived negatively than the same behavior from an unattractive person.

        • Aapje says:

          True, although if that perception is inconsistent with the amount of harm done, you may have psychological damage that people deny exists, because the scenario or perpetrator doesn’t fit with their narrative about what they should feel.

          It’s something that I’ve seen in some male victims of female perpetrators: anger or trust issues that they didn’t attribute to the rape/assault at the time, because it doesn’t fit their mental model of how they should feel about having a sexual experience, even an unwanted one.

          • Forge the Sky says:

            Hot people: the Stealth Rapists

            But hey, I’ve actually been there, so it is a thing.

            On the flip side, I’ve also been the hot dude a girl has a fling with that she freaks out about the next morning ’cause she just cheated on her boyfriend. Imagine my surprise. That’s a weird feeling, I’m not sure if I should feel flattered, guilty, or just kinda sad because I liked this one.

          • Aapje says:

            I’m not sure if I should feel flattered, guilty, or just kinda sad because I liked this one.

            If you didn’t know, there seems to be no reason to feel guilty, unless you cannot distinguish between what you should have done with 20/20 hindsight and what you did with the knowledge of the time.

            The other two seem reasonable emotions, that one could have simultaneously.

  22. Rachael says:

    I agree with most of your conclusions, but there’s still a major flaw in the survey methodology. It asked what industry people currently work in, and whether they’ve *ever* been harassed at work. Plenty of people have switched industries over the course of their career, and nearly everyone worked in retail or food service or similar as a teenager or young adult. So we have no idea how much of the reported harassment took place in industries other than the ones people currently work in, and probably skewed towards the industries that hire teenagers.

    • briguybrn says:

      A really good point. If asked, I would have said that I have been harassed, but it was as a 17 year old at my first job by an older co-worker at a restaurant, and not in my current career.

    • brmic says:

      As you said, ‘neary everyone worked in retail or food service or similar as a teenager or young adult’. That means, the base rate for ‘ever at work’ should be the same for pretty much everyone, and any differences can then be reasonably attributed to the field worked in since then.
      But yes, you’re right, it should ideally have been ‘current field’ and also number of years in the field as a question.

    • Rosemary7391 says:

      Actually I don’t think that’s true – I’d be amazed if “worked in retail/food service as a teen” wasn’t at least somewhat correlated with parental income. Which probably correlates with a bunch of things that might affect entry into certain careers… enough to make a difference? Not sure.

      Agree with your general point though.

    • Deiseach says:

      Yeah, my example of at-work harassment happened thirty-five years and about two career changes ago.

    • Notsocrazy 24 says:

      I’m not sure “nearly everyone” worked in retail or food service as teenagers or young adults, but perhaps I’m biased by my personal experience and the fact that I go to a pretty stringently upper middle-class school where I’m one of the few who did a significant amount of self-support financially before completing undergrad.

    • cuke says:

      I’m glad you raised this, I was just about to post the same. My experience of harassment was two careers ago and so my current industry cannot be tagged to my experience of harassment, and it also was not in retail or service work, but in academia.

  23. Tatterdemalion says:

    SSC survey commentators are a rather weird demographic. I’m not sure

    “How likely are high-IQ ASD rationalist types to self-report being sexually harassed in this field”

    is a meaningful proxy for

    “how likely is the average worker in this field to be sexually harassed”.

    • Darwin says:

      There’s that, and then there’s who has stuck around to become a fan of this blog, which I was glad to see Scott point out as it was my first thought as to confounds.

      I don’t disagree with Scott on what he’s said or how he’s phrased his beliefs/opinions about the narrative around women in tech over the years, but I imagine that if I were a woman in tech who felt I had experienced harassment and that the field had a problem with it, I probably would have gotten pretty annoyed with this blog at some point in its history and maybe given up on it.

      Maybe we cold test this with an anonymous surveymonkey test… gather several of this blog’s posts about women and the tech industry, ask random women in the tech industry to read them and then say whether or not they would choose to read more by this author, then ask them if they’re ever been harassed in a tech industry job, and look for a correlation. Would that make sense at all?

  24. nameless1 says:

    >People from traditionalist cultures and subcultures may have a higher threshold for calling something harassing; if (for example), more traditional people go into Business

    I don’t know if this is a wrong view or one that is true for America but certainly not true for Europe. But the idea that trads may go into Business gives actually a clue. The kind of trads I know would generally avoid that Gomorrah of superficial Mammon worship that business means and become school music teachers directing church choirs on the side. The stereotypical business school type may be a libertarian. And yes I get it why a libertarian why would not consider pushing an offer harassing – not saying libertarians are like that, but they tend to be free from both trad religious inhibitions and modern feminist inhibitions. But why call libertarians trad? I get it that they are often called conservatives in the US but the whole point is that they are the non-trad-conservatives, while the trad-conservatives (Dreher) being an entirely different animal.

    • Calion says:

      I think we’ve got a whole definitional muddle here. First off, “libertarians” are a minority of the population (about 16% by some measures) to begin with, and tech fields skew libertarian-heavy, leaving not too many left for business and other fields. Second, American conservatives are not libertarians. Sure, there are libertarians who call themselves conservative, but most of those we call conservative are “trads,” though probably not nearly as traditionalist as the people you’re talking about.

      Here’s the thing: Due to the Puritan ethic (read Weber’s *The Protestant Ethic and the Spirit of Capitalism*), making money has been associated with being a good Christian, so business is not really an anathema occupation for “trads.”

  25. rlms says:

    I was suspicious that varying rates of transness across professions would be a confounder, but from taking a quick look it doesn’t appear to be (87% of women who responded were cis, 87% again for women in media, 82% for women in maths).

    • Reasoner says:

      Wait, 13% of the women who read SSC are trans? That seems really high.

      Am I an out-of-touch old fogey already?

      • Steve Sailer says:

        Extremely logical high IQ males on the autism spectrum have high rates of transgenderism. For example, the smartest, most arrogant guy I knew at MBA school in 1981 was recently listed by the Washington Post as the highest paid female CEO in America.

        I suspect Robert Heinlein noticed this pattern among his fans way back in the 1950s, judging by his short story “All You Zombies.”

      • rlms says:

        SSC readers are not drawn uniformly from the general population, where the rate of transness is at least an order of magnitude lower.

    • Scott Alexander says:

      I might have forgotten to mention it in the post, but to avoid exactly this possibility the analysis was limited to cis people. Adding trans people doesn’t change much, though.

  26. amoeba says:

    Why do you say that you “have been expecting results like these even before I did this survey” if your preregistered hypothesis was that “STEM would have more female victimization” which as you say turned out to be wrong?

    • Darwin says:

      Yeah, I noticed that and was confused, too.

    • Freddie deBoer says:

      Hypothesis selection is not always congruent with a researcher’s prior belief. Sometimes a hypothesis can be chosen in order to set the null in one particular direction.

      • Forge the Sky says:

        Yes. In the undergrad research I participated in, it was actually more common to start with a hypothesis you expected to be disproven, than one you expected to be proven.

        In this case, Scott is setting out to disprove a hypothesis he believes is likely to be inaccurate. It’s easier and more accurate than trying to prove a specific other hypothesis you hold, since you would likely need to prove multiple elements simultaneously, and you still wouldn’t be directly disproving the hypothesis you disagree with.

    • C_B says:

      I think these results are even farther from the media narrative than Scott was predicting, but in the same direction.

      That is, Scott’s hypothesis was that STEM would be weakly “vindicated” (i.e., shown not to be disproportionately full of harassers) – the consensus that women in STEM are more frequently harassed would be true, but it would turn out to be only the small effect predicted by the over-representation of men in those fields.

      Instead, these results suggest STEM is strongly vindicated – women in STEM aren’t more frequently harassed at all, they’re less frequently harassed.

    • Scott Alexander says:

      Yeah, good point.

      I think what I meant was that I’ve been saying for a while that STEM is not uniquely harassing. A while ago I sort of softened that based on the gender-ratio argument that even if STEM people were less harassy the ratio alone might cause more harassment, but you’re right that my (new, softened) opinion was disproven and my (original, intuitive) opinion was supported.

  27. e.samedi says:

    It seems to me that a scientifically rigorous definition of sexual harassment is needed to compare data across studies. Without such a definition how do you know they are measuring the same phenomena?

    Also, with respect to surveys, there is some legitimate criticism of psychometrics. University of Sydney professor Joel Michell is probably the most prominent critic (see his article “Is Psychometrics Pathological Science?”). Consideration of his argument would make an excellent SSC post in my view.

    • Darwin says:

      To be fair, if we are comparing across fields and don’t care so much about absolute counts, we should be fine as long as the nebulous definitions that respondents are applying to the phrase ‘sexual harassment’ are evenly distributed across fields. In this case we’d just be adding noise, and hopefully a sample of 7000+ would give us enough power to still see the effects.

      Of course, those definitions might *not* be evenly distributed across industries, for a few imaginable reasons. In that case we’d definitely have a problem.

      • arlie says:

        My impression is that the definition is not evenly distributed.

      • Steve Sailer says:

        What are the odds that, say, cocktail waitresses and professors of gender studies have identical definitions of “sexual harassment” in mind?

      • e.samedi says:

        The article Using Surveys to Assess the Prevalence of Sexual Harassment: Some Methodological Problems (Richard D. Arvey; Marcie A. Cavanaugh) discusses problems with surveys such as these.

        For example,
        “There are problems with simply the basic definition of sexual harassment. Researchers differ considerably in terms of the kinds of events and behaviors that constitute sexual harassment.”

        “Another problem is the minimal consideration of the severity of the offense when labeling it as sexual harassment. Researchers often do not attend much to the severity of a behavior or practice in designating whether sexual harassment has occurred.”

        I do not think the usual methodology warrants any confidence in the results. The problems are: vague and inconsistent definitions, potentially flawed and biased survey instruments, the well-known issues with self-reported data, and (possibly) the mathematical foundation of psychometrics itself. In light of the replication crisis, it seems to me that social scientists would be well-served by a critical re-evaluation of their core methodology.

    • Steve Sailer says:

      The definition of sexual harassment as “unwanted sexual advances” is a logical morass. As I pointed out in 1992 when forecasting that the incoming Clinton administration would eventually be rocked by a sexual-harassment scandal (which indeed happened in 1998):

      “What self-respecting woman would admit that no man had ever made an unwanted sexual advance toward her? She’d be admitting either that no man’s ever made her a sexual advance or that she’s never met a sexual advance she didn’t like.”

      • Steve Sailer says:

        Has anybody ever surveyed women about how many wanted sexual advances they’ve gotten at work?

        For example, my dad was a bachelor engineer at Lockheed and my mom was a young widowed secretary at Lockheed whose Marine first husband had been killed on Iwo Jima.

        I realize than in the Current Year our culture’s highest priority is preventing and punishing unwanted heterosexual advances. But I kind of think unwanted heterosexual advances are more or less the price our species has to pay for the wanted heterosexual advances that keep homo sapiens out of extinction.

        • Simon_Jester says:

          Women in your parents’ generation had pretty limited practical options to do well in life unless they could “catch a man.” So they perforce sought out opportunities to do it whether they wanted to or not, for the same reason that people of both sexes today apply for jobs whether they want the job or not.

          [Note that a phrase like “catch a man” would fit far more readily into conversation between two women in a 1950 period piece than it would in 2015, for this exact reason.]

          The thing is, it is not at all difficult for modern women to see too it that their rate of wanted sexual attention rises. They are if anything greatly oversupplied with options for doing so.

          By contrast, it is rather more difficult for women to throttle down the level of unwanted sexual attention. If a woman wants to meet twice as many men who will proposition her, she knows or can figure out what to do. If a woman wants to meet half as many men who will proposition her, she may have a rather harder problem on her hands. So she puts more effort into making sure she has the ability to slam on the brakes than she does into making sure she can slam on the gas pedal.

          The species is unlikely to be in danger of extinction from this status quo.

        • Mary says:

          They are if anything greatly oversupplied with options for doing so.

          Are they? And if so, do the options yield any men they would be interested in? A woman might be far more interested in the engineer she could meet at work than the guy she could pick up at the bar.

          • Simon_Jester says:

            As this becomes a more widespread problem, we would expect to see more and more women who were no longer more interested in the ‘brakes’ on the dating circuit than they were in the ‘gas pedal.’

            Informally, that does not seem to be happening yet, so I infer that the gas pedal is still signficantly stronger than the brakes.

          • The Nybbler says:

            As this becomes a more widespread problem, we would expect to see more and more women who were no longer more interested in the ‘brakes’ on the dating circuit than they were in the ‘gas pedal.’

            Except the tendency to put on the brakes may be less rational and more innate. There may be a tendency for women not to relax the filters but instead to keep them high and then complain when no one passes them, resulting in the common complaint of “All the good ones are married or gay”.

          • Aapje says:

            Or there may just be a (huge) lag. Reducing expectations is not very pleasant.

  28. toastengineer says:

    Seems like we need to hash out what our profile of a “harasser” is. Everyone seems to imply they’re thinking of some kind of serial predator who gets off on making people uncomfortable, which definitely is a kind of person who exists. Somehow I suspect, however, that the majority of “sexual harassment” is less that guy and more “I’m incredibly desperate for affection and emotional attachment due to my abusive/neglectful/incompetent childhood and for similar reasons have no idea what the current rules are for politely asking for it.”

    If we’re talking about the latter, there are definitely women who are more skilled at dealing with that kind of situation than others. There was this really good looking girl in high school who was in Calculus Club and such and if you made a pass at her she’d give you a very clear “hell no and please don’t show this dialogue box again” and then you’d be good friends. Seems like most women will give you some variation of “ask again later” which you’re supposed to interpret as “fuck no,” but some people don’t work that out for quite a while indeed, which seems like it would lead to a lot of unintentional harassment-like behavior.

    So the average woman who is a programmer has had years of experience dealing with basement dwellers and knows how to shut down an unwanted advance by a well-meaning incompetent immediately. Someone who has less experience with that kind of person is more likely to end up in a shitty situation when they do run in to them. The sort of woman who becomes a coal miner is probably similarly skilled at dealing with whatever the rural equivalent of such is.

    • Darwin says:

      As a socially awkward and highly neurotic person, I certainly empathize with the fear that I might accidentally harass someone by doing something I think is friendly/normal but is actually awful and traumatic for them.

      But thinking rationally and looking at the evidence available to me, I just don’t think it’s very likely that this actually accounts for the majority of sexual harassment in the world, or even somewhere like the tech industry.

      Socially awkward men may make a lot of women uncomfortable unintentionally, but hopefully most are not reporting this as harassment. I think there are actually very few people who are clueless enough to accidentally cross the line into harassment, without being driven by some type of malice or callousness or aggression that would make me happy to classify them as a perpetrator.

      • Simon_Jester says:


        I’m sure that a certain fraction of all sexual harassment comes from desperate clueless people… But after you’ve harassed enough people that way, you would have to be truly, willfully obtuse to not realize there’s a problem. Especially once you get past the high school level, which is completely inapplicable to the workplace.

      • Steve Sailer says:

        “Socially awkward men may make a lot of women uncomfortable unintentionally, but hopefully most are not reporting this as harassment.”

        Hopefully, but shouldn’t America’s vast legal and social science fields have investigated this highly relevant question by now? I asked about this a lot in the wake of the 1991 Anita Hill brouhaha, but nobody seemed to have an answer back then.

        I then wrote an essay in late 1992 predicting that, from what I’d heard from Arkansawyers, President-Elect Bill Clinton risked impeachment over some unwanted sexual advance he likely had made to some state employee even though no doubt most of his sexual advances had turned out welcome. But nobody would publish it back then, so it was a big surprise to the world when it turned out Clinton had sexually harassed Paula Jones, for which he had to pay $850,000.

    • cryptoshill says:

      Theres a good bit of evidence (some linked ) that rapists at least (as a subset of general sexual harassers) are very frequently repeat offenders. The same model could be applied to sexual harassment, as rape and sexual harassment are considered linked.

      I do agree that creating a precise definition of what harassment *is* or *isn’t* is extremely important and is a question that the wider discourse is avoiding, seemingly on purpose.

      I had some loose pontifications on attributing the perceived high rate in “STEM” fields to recent popularity and culture shock. Many of our models for politically correct sexual signaling are written by self-described feminists who I will cluster with the nascent “thinkpiece industry”. I find it plausible that the popularity of STEM has diverted people who would normally be prolific members of the thinkpiece-industry into technology fields, and are expecting a level and ability of communication that introverted “nerdy/math” people would consider at a minimum unnecessary, or even potentially exhausting.

    • mdet says:

      In my personal experience as a man, the vast majority of sexual harassment I’ve witnessed has been men who deliberately take joy in groping or sending lewd messages to women, people who view this behavior as fun pranks. I have occasionally witnessed the poorly made but sincere attempt at flirting/friendship that crossed into potentially uncomfortable territory, but it’s not a majority in my experience.

      • toastengineer says:

        But again, you hear “sexual harassment” and think “groping” and “Richard Stallman” and I hear the same phrase and hear “tried to start a conversation in an elevator” or “asked her out again after being rejected.” As near as I can tell that really is what the phrase usually means; see orangecat’s comment below.

        • mdet says:

          I agree that we should be extra careful not to conflate the two categories, and that they are sometimes conflated. And now that I think about it, I do remember seeing (and unfortunately, making) some more “awkward pass” type incidents. But in my memory the responses to those really were a proportionate “Umm, no thanks” and very distinct from incidents like where I witnessed a guy text a dick pic to the girl standing next to him just for the lolz

    • orangecat says:

      Seems like most women will give you some variation of “ask again later” which you’re supposed to interpret as “fuck no,” but some people don’t work that out for quite a while indeed, which seems like it would lead to a lot of unintentional harassment-like behavior.

      Right. I’ve heard that Google has instituted a policy where if you ask a coworker out and the response is anything other than “yes”, you’re required to treat it as a “no” and not ask again. That seems like a reasonable approach.

      • Evan Þ says:

        It sounds like a defensible approach for coworkers; it sounds like a really bad blanket policy for the wider world.

        • The Nybbler says:

          The problem in both cases is there’s no downside to a woman who uses an initial non-yes as a filter to see which guys are really interested. Backing that up with official sanction just makes it work better as a filter.

          Further, the only story of harassment by asking for a date I heard (on internal social media) at Google was about someone who asked exactly once, rather diffidently. This resulted in the woman in question “crying in the stairwell” (this is basically a trope), and of course got her tons of support with and a lot of hate aimed at Google men (fortunately she did not name the particular guy)

          • Aapje says:

            Backing that up with official sanction just makes it work better as a filter.

            More strict filtering is not always better, especially when nothing passes through. I use a search engine to reduce billions of possibilities to a few dozen, but having zero results is usually not the desired result.

          • orangecat says:

            Yeah, it’s not perfect. Hopefully it makes that filter less viable as women realize that they can’t expect guys to put their career at substantial risk.

          • John Schilling says:

            Hopefully it makes that filter less viable as women realize that they can’t expect guys to put their career at substantial risk.

            Women can expect guys to put their career at substantial risk. And they won’t be wrong. Willingness to make sacrifices is a big part of what they are usually trying to filter for.

          • David Speyer says:

            The downside is all the men she wants to date but doesn’t, because they move on to someone else when she says no. This response seems to think a woman would never be interested in a particular man and want to say yes to him.

  29. Leah Velleman says:

    I wonder whether something gets lost here when we conflate the startup world (stereotypically an unsupervised bunch of pushy, risk-taking young single people) with — and I use this term lovingly — dinosaur tech companies (stereotypically a bunch of mild-mannered old married people, with a strong HR department looking over things).

    The cultural gap between startups and dinosaur tech seems much bigger than, say, the gap between independent and chain restaurants, or small and large universities. If that’s true, then that’s an argument for treating startups and dinosaur tech as separate sectors. (And while some companies might be hard to sort into one category or the other, that’s true for any set of categories. “Is this job in econ or finance?” and “Is this job in business or law?” will turn up ambiguous cases too, but that’s not a reason to ditch those distinctions.)

    I’m pushing this question because I have a hypothesis, which is that startups really do have a harassment problem — but that the people speaking up about it, and the news outlets quoting them, have misrepresented this as a problem across all of tech, ignoring the fact that most tech workers are at boring old companies with low harassment rates. I suspect this misrepresentation has happened partly because startup folks like to represent themselves as The Real Tech Industry, The One That Counts, and partly because companies tend to transition from “startup” to “dinosaur” over time and news outlets can’t be bothered to keep track.

    • SamChevre says:

      I have no idea about startups, but dinosaur tech is a really useful category.

      I have a friend who works remotely for a big-name tech company as an engineer. His description of visiting headquarters–which he does reluctantly every few months–is “it’s like a fraternity house.”

      But I work with programmers; several of the programmers I’ve worked with are now retired, and when I worked with them had pictures of their grandchildren on their desks. They were superb programmers, but they’d been programming since APL was a hot language.

    • gbdub says:

      Yeah, I was going to say the same thing. The media talking about “techbros” tend to conflate tech/STEM with “Silicon Valley startup culture” when the latter is actually a very small fraction of the former. And a fraction that draws a disproportionate share of the alpha male type that probably does a larger share of harassing.

      • knockknock says:

        True Alphas can more likely get what they want without harassing anyone — their advances are much more likely to be welcome or at least socially skillful even if very forward. It’s the Betas that likely get rebuffed and reported for harassment, however the people involved might define it.

        • dndnrsn says:

          What if a “True Alpha”, whatever that is, enjoys pushing other people’s buttons and boundaries to show their dominance, their superior place in the status hierarchy, whatever? You’re modelling sexual harassment as passes that fail, and it isn’t necessarily that, or only that.

          • Simon_Jester says:

            I think this model of alpha/beta male dynamics has at least four categories, not two. The relevant categories are at least:

            1) Alpha males, who possess natural charisma and authority
            2) Comfortable beta males, who lack natural charisma and authority, but who nonetheless have social skills on a minimal level
            3) Insecure-aggressive beta males, who try to ape alpha male status but do so with toxic, destructive behaviors, and call them…
            4) Zeta males, let’s call them, who just utterly lack the ability to function within a social framework, due to limitations like mental illness or just plain behavior problems.

            Now, one perspective is to claim that there’s no such thing as ‘true alphas,’ that it’s betas and zetas all the way down, but some of the betas are bullies and impersonate charismatic/authoritative behavior well enough to be called ‘alphas.’

            Another is to conflate the betas and zetas into a single group and lump the socially successful alphas into another, without recognizing the differences that cause typical betas to behave far differently from the edge cases.

          • dndnrsn says:

            What would this predict regarding sexual harassment, assault, etc?

          • Simon_Jester says:

            The first perspective I describe, which denies the existence of what I will call for lack of a better term ‘positive’ alpha males and says it’s betas and zetas all the way to the bottom and all the way to the top…

            Well, it would predict that sexual harassment is nigh-universal among powerful men, because powerful men will tend to be insecure betas trying to reassure themselves of their worth and strength by transgressing against others. It would largely deny any value to be had in male role models, because just about any prominent male is as likely to be an asshole as any other.

            The second model would, by contrast, tend to attribute most sexual harassment to the ‘creepers, freaks and geeks’ among the “beta males,” while largely ignoring sexual harassment committed by the ‘alphas’ who are successful and charismatic and not sexually desperate enough to behave this way.

            Neither model really encompasses reality in my opinion.

            I believe that there are definitely ‘true alphas’ who possess charisma but also the kind of self-security and dignity that prevents them from abusing others. At the same time there are also insecure bets who’ve connived their way into powerful positions but sustain those positions by constantly “kicking down and kissing up,” as it were.

            At the same time, there are ample numbers of ‘betas’ and ‘zetas’ who do not cause problem behavior on the subject of sexual harassment, so strategies that look only at the top OR at the bottom are missing things.

          • dndnrsn says:

            I think you are being a bit… I don’t know what the right term here is. Idealistic? To say:

            I believe that there are definitely ‘true alphas’ who possess charisma but also the kind of self-security and dignity that prevents them from abusing others. At the same time there are also insecure bets who’ve connived their way into powerful positions but sustain those positions by constantly “kicking down and kissing up,” as it were.

            How do you tell who is who? By their behaviour? Is someone who seems like a “secure alpha” who turns out to have been abusing people all along – do they switch retroactively? Do people abuse others because they lack security and dignity themselves, or are there other reasons why they abuse people? The split you seem to be describing between noble alphas and abusive conniving upstart betas – sounds a bit “the kings in yore were great men, before the usurpers came”.

            How would you falsify the theory? The problem with the ranking theories is that they’re retrospective – can you predict a given person’s behaviour? Or does his behaviour show that instead of being a 7th level True Alpha he was in fact a 6th level Beta/2nd level Conniver (prestige class)? Is a dominant male in the ape species closest to us kind and benevolent to those below him? I’m no ape expert, but I think they’re usually not. What if most everyone kisses up and kicks down, just some people have fewer above them?

    • Mary says:

      dinosaur tech companies

      Remember how long the dinosaurs lasted?

  30. b_jonas says:

    > The takeaway that most real researchers take from the EEOC claims is that the lowest-paying and most mundane occupations – retail, restaurant work, hotel work, etc – have much higher sexual harassment rates than the prestigious occupations people generally talk about.

    Could this be because of a difference in the age of the workers? The lowest paying occupations are ones that people start to work in without spending five years of their lives in college, so there’s more young people in there. I suspect if the workers are younger, there’d be more sexual harassment at work.

    • Darwin says:

      I think this is a very likely causal factor. Another one is probably bargaining power – employees who complain or stand up for themselves in those industries can be replaced very quickly and with very little fallout.

      • Aapje says:

        Churn is probably pretty high anyway, and the cost of getting a new worker up to speed pretty low, so if someone is offended enough to leave it is less of a burden on the employer than in other occupations.

      • Mary says:

        This is also a factor for aspiring actress, which is notorious for the casting couch.

    • ChrisA says:

      Most likely it is because of IQ that lower paid professions have greater rates of sexual harassment. If you have ever worked in a low paid profession you know the quality of the “jokes” are very low and people in them are constantly doing things that are the result of bad judgement, like drugs, drinking, bad relationships and so on. All these claims about tech people being introverted nerds is the obverse of this, higher IQ people are simply not as stupid a low IQ people (duh), which paradoxically makes them less fun and more inhibited, and less willing to engage in reckless behaviour.

  31. brmic says:

    Thank you for posting this and the data file. FWIW, I tried to reproduce the results and couldn’t reproduce the correlations between female victimization, male victimization and male perpetration. fem vic vs. male vic is 0.65, same as yours. fem vic vs. male perp is 0.01 for me, and male vic vs. male perp is 0.21 for me. Everything else more or less checks out.

    As a reviewer, I’d say the combination score is not convincing, especially since it ignores all considerations of different male to female ratios in the various industries.
    Also, if you have two measures with r = 0.8, Fig 6 is not a good idea IMHO. It’s probably just noise. (Also, it should be a dotplot centered around 1, because the relevant info is distance from 1:1 ratio.)
    Instead, I’d focus on the correlation between female victimization at work and female victimization outside work of 0.65 (for me) and the same for males at 0.59, which also leads to the conclusion that there’s a strong ‘people in fields’ effect, without having to go through the combination score. If you’re so inclined, you might then do the at-work by outside-work ratios and end up a kind of cross-validation set, where you can see whether the bad fields for women are bad for men as well. Of course, once you then consider sex ratios per field. it’s story time all over again. Still, e.g. men report similar levels of out of work victimization in computers (20%) and Health Care (24%), but at work victimization of 4% and 12% respectively, which strongly suggests that Health Care is worse.

    • brmic says:

      For those wanting to play along at home, here’s my R code. Please point out any errors you find.

      # Libraries ===================================

      # Functions ==================================
      tableNA <- function(...) table(..., useNA="always")

      # Data aquisition and cleanup =====================
      k <- read_delim(".../ssc2018public.csv", ";", escape_double = FALSE, trim_ws = TRUE)

      # options(tibble.print_max = Inf)
      # Serparate people responding in any way to the harassment items -----
      # no idea what to do about the '...ends in' answers
      k[which(!k$Harassment1 %in% c('No', 'Prefer not to answer', 'Yes')), 'Harassment1'] <- NA
      k[which(!k$Harassment2 %in% c('No', 'Prefer not to answer', 'Yes')), 'Harassment2'] <- NA
      k[which(!k$Harassment3 %in% c('No', 'Prefer not to answer', 'Yes')), 'Harassment3'] <- NA
      k[which(!k$Harassment4 %in% c('No', 'Prefer not to answer', 'Yes')), 'Harassment4'] <- NA

      k_nohas <- k[which($Harassment1) &$Harassment2) &$Harassment3) &$Harassment4)), ]
      k <- k[which(!$Harassment1) | !$Harassment2) | !$Harassment3) | !$Harassment4)), ]

      # is non-response systematic? ----------------------------------
      t.test(k_nohas$Age, k$Age) # no problem

      tab1 <- cbind(table(k_nohas$Sex), table(k$Sex))
      fisher.test(tab1) # potential problem
      round(prop.table(tab1, margin = 2),2) # ignore it

      # Merge 'Profession' levels -----------------
      k$prof <- k$Profession
      k[which(k$prof %in% c('Computers (AI)', 'Computers (other academic, computer science)', 'Computers (practical: IT, programming, etc.)')), 'prof'] <- 'Computers'

      k[which(k$prof %in% c('Health Care (mental health)', 'Health Care (other)')), 'prof'] <- 'Health Care'
      k[which(k$prof %in% c('Statistics')), 'prof'] <- 'Mathematics'
      k[which(k$prof %in% c('"Other ""hard science"""', 'Physics')), 'prof'] <- 'Hard Science'
      k[which(k$prof %in% c('"Other ""social science"""', 'Psychology')), 'prof'] <- 'Social Science'

      # neuroscience ?

      # Subset for legibility of following code ================
      kw <- k[which(k$Sex == 'Female'),]
      km <- k[which(k$Sex == 'Male'),]
      ko <- k[which(k$Sex == 'Other' |$Sex)),]
      stopifnot(nrow(kw)+ nrow(km) + nrow(ko) == nrow(k))

      # Main analysis ================================
      # Percentage Women harassed at work, should be 21%
      round(100* prop.table(tableNA(kw$Harassment1))[3], 2)
      # Percentage Men harassed at work, should be 6%
      round(100* prop.table(tableNA(km$Harassment1))[3], 2)

      # Figure 1 ---------------------------------------------------------------
      tab1 <- round(100*prop.table(table(kw$Harassment1, kw$prof), margin = 2)[3,],2)
      names(tab1) 30]

      barplot(tab1[colSums(table(kw$Harassment1, kw$prof))> 30], col = 'blue',
      main = 'Figure 1: Percent women reporting sexual harassment victimization at work, by occupation type',
      cex.names = 0.5)

      # Figure 2 ---------------------------------------------------------------
      tab2 <- round(100*prop.table(table(km$Harassment1, km$prof), margin = 2)[3,],2)
      names(tab2) 30]

      barplot(tab2[colSums(table(km$Harassment1, km$prof))> 30], col = 'blue',
      main = 'Figure 2: Percent men reporting sexual harassment victimization at work, by occupation type',
      ylim = c(0,14), cex.names= 0.6, las = 2)

      # Figure 3 ---------------------------------------------------------------
      tab3 <- round(100*prop.table(table(km$Harassment3, km$prof), margin = 2)[3,],2)
      names(tab3) 30]

      barplot(tab3[colSums(table(km$Harassment3, km$prof))> 30], col = 'blue',
      main = 'Figure 3: Percent men admitting to sexual harassment perpetration at work, by occupation type',
      cex.names= 0.6, las = 2,
      ylim = c(0,5))

      # Correlations -----------------------------------------------------------
      # cbind(tab1, names(tab2), tab2) # check labels match
      # cbind(tab1, names(tab3), tab3) # check labels match
      # cbind(tab2, names(tab3), tab3) # check labels match
      (c1 <- round(cor(as.numeric(tab1), as.numeric(tab2)),2))

      plot(as.numeric(tab1), as.numeric(tab2), xlab = '% female victimization', ylab = ' % male victimization',
      main = paste0('Correlation = ', c1))
      abline(lm(as.numeric(tab2)~as.numeric(tab1)), col="red", lwd = 2)
      lines(lowess(as.numeric(tab1), as.numeric(tab2)), col="blue", lwd = 2)

      (c2 <- round(cor(as.numeric(tab1), as.numeric(tab3)),2))

      plot(as.numeric(tab1), as.numeric(tab3), xlab = '% female victimization', ylab = ' % male perpetration',
      main = paste0('Correlation = ', c2))
      abline(lm(as.numeric(tab3)~as.numeric(tab1)), col="red", lwd = 2)
      lines(lowess(as.numeric(tab1), as.numeric(tab3)), col="blue", lwd = 2)

      (c3 <- round(cor(as.numeric(tab2), as.numeric(tab3)),2))

      plot(as.numeric(tab2), as.numeric(tab3), xlab = '% male victimization', ylab = ' % male perpetration',
      main = paste0('Correlation = ', c3))
      abline(lm(as.numeric(tab3)~as.numeric(tab2)), col="red", lwd = 2)
      lines(lowess(as.numeric(tab2), as.numeric(tab3)), col="blue", lwd = 2)

      # Figure 4 --------------------------------------------------------------
      tab4 <- data.frame(cbind(tab1, tab2, tab3))
      tab4$tab1s <- tab4$tab1/max(tab4$tab1, na.rm = TRUE)
      tab4$tab2s <- tab4$tab2/max(tab4$tab2, na.rm = TRUE)
      tab4$tab3s <- tab4$tab3/max(tab4$tab3, na.rm = TRUE)
      tab4$combiscore <- (tab4$tab1s + tab4$tab2s + tab4$tab3s)/3

      tab4$combiscore2 <- rowSums(scale(tab4[,1:3]))/3

      rownames(tab4) <- names(table(k$prof))

      barplot(tab4$combiscore, col = 'blue',
      main = 'Figure 4: Combination score of sexual harassment by industry',
      ylim = c(0,1), names = rownames(tab4),
      cex.names= 0.6, las = 2)

      # Figure 5 -----------------------------------------------------------
      tab51 <- round(100*prop.table(table(kw$Harassment2, kw$prof), margin = 2)[3,],2)
      tab52 <- round(100*prop.table(table(km$Harassment2, km$prof), margin = 2)[3,],2)
      tab53 <- round(100*prop.table(table(km$Harassment4, km$prof), margin = 2)[3,],2)

      tab5 <- data.frame(cbind(tab51, tab52, tab53))

      tab5$tab1s <- tab5$tab51/max(tab5$tab51, na.rm = TRUE)
      tab5$tab2s <- tab5$tab52/max(tab5$tab52, na.rm = TRUE)
      tab5$tab3s <- tab5$tab53/max(tab5$tab53, na.rm = TRUE)
      tab5$combiscore <- (tab5$tab1s + tab5$tab2s + tab5$tab3s)/3

      barplot(tab5$combiscore, col = 'blue',
      main = 'Figure 5: Combination score for out-of-work sexual harassment by industry',
      ylim = c(0,1), names = rownames(tab5),
      cex.names= 0.6, las = 2)

      # Correlation ====================================
      (c1 <- round(cor(tab4$combiscore, tab5$combiscore),2))

      plot(tab4$combiscore, tab5$combiscore, xlab = 'combination score at work', ylab = 'combination score out-of-work',
      main = paste0('Correlation = ', c1))
      abline(lm(tab5$combiscore~tab4$combiscore), col="red", lwd = 2)
      lines(lowess(tab4$combiscore, tab5$combiscore), col="blue", lwd = 2)

      # Figure 6 ==================================
      tmp <- tab4$combiscore/tab5$combiscore

      barplot(tmp, col = 'blue',
      main = 'Figure 6: Ratio of at-work harassment to out-of-work harassment by industry',
      ylim = c(0,1.2), names = rownames(tab5),
      cex.names= 0.6, las = 2)

      # Figure 7 ==================================
      cols <- c('blue', 'grey', 'grey', 'red', 'grey', rep('red', 3), rep('blue', 2), 'red', 'blue', rep('grey',4))
      barplot(tab4$combiscore, col = cols,
      main = 'Figure 7: Ratio of at-work harassment to out-of-work harassment by industry',
      ylim = c(0,1), names = rownames(tab4),
      cex.names= 0.6, las = 2)

      # Figure 8 ==================================
      # female victimization at work and outside of work
      tab8 <- round(100*prop.table(table(kw$Harassment2, kw$prof), margin = 2)[3,],2)

      (c1 <- round(cor(as.numeric(tab1), as.numeric(tab8)),2))

      plot(as.numeric(tab1), as.numeric(tab8), xlab = 'female victimization at work', ylab = 'female victimization outside work',
      main = paste0('Correlation = ', c1))
      abline(lm(as.numeric(tab8)~as.numeric(tab1)), col="red", lwd = 2)
      lines(lowess(as.numeric(tab1), as.numeric(tab8)), col="blue", lwd = 2)

      # Figure 9 ============================
      # male victimization at work and outside of work
      tab9 <- round(100*prop.table(table(km$Harassment2, km$prof), margin = 2)[3,],2)

      (c1 <- round(cor(as.numeric(tab2), as.numeric(tab9)),2))

      plot(as.numeric(tab2), as.numeric(tab9), xlab = 'male victimization at work', ylab = 'male victimization outside work',
      main = paste0('Correlation = ', c1))
      abline(lm(as.numeric(tab9)~as.numeric(tab2)), col="red", lwd = 2)
      lines(lowess(as.numeric(tab2), as.numeric(tab9)), col="blue", lwd = 2)

      • Calion says:

        Pastebin exists for a reason.

        • brmic says:

          Apologies! I was wondering whether I could put that somewhere instead of posting a wall of text, but couldn’t think of anything suitable. Never used pastebin before, though now that you mention it I of course remember it being used by others.

  32. Jiro says:

    Compare all the anecdotes and popular lore about how immigrants are criminals. It’s totally false – immigrants have crime rates well below native-born citizens. But we only know that because there have been really good studies. If the studies hadn’t been done, and all we had to go on was the daily lurid stories about a Mexican guy knifing someone, who would believe it?

    This is a piece of muddled thinking that, unfortunately, has been here before.

    If the daily lurid stories are about a Mexican guy knifing someone, and you’re comparing that to crime rate statistics, you’d better make sure that your crime rate statistics are also about Mexicans (and about illegal Mexicans, if appropriate).

    Using statistics that are about “immigrants” lumps together different groups of immigrants, both by ethnic origin and legal status. It also ignores any possible problems with second generations not assimilating, if the “Mexican guy knifng someone” is a citizen whose parents are illegal Mexican immigrants.

  33. MartMart says:

    1. Harassment is a means to an end: the goal might be sex, or exercising power, or limiting women from competition for jobs, or raising one’s image among their peers. Whatever the goal is, most harassers do so for some reason
    2. Harassing is a skill. Doing it poorly does not achieve whatever goal the harasser had in mind, while doing it well (it’s a terrible word, maybe someone has a better one) might. If it is a skill, it’s most likely a social skill.
    3. Thus, people with poor social skills make poor harassers and are likely to fail to achieve whatever goals they have. They are therefor likely to realize their failure rate (sooner or later) and move on to other strategies (or give up).
    4. People with good social skills are more likely to achieve their goals thru harassment, and get away with it, and so more likely to do so again in the future.
    5. So fields that are dominated by people with poor social skills are likely to have less harassment.

    Not sure how to test that..

    • Simon_Jester says:

      Perhaps instead of saying “doing it well,” we should say “doing it proficiently?” It’s pretty easy to imagine a ‘proficient’ sexual harasser- one who is good at coercing people into not reporting their actions, or who is good at tricking victims into thinking they’ve invited the behavior, or who is good at pushing past people’s boundaries but not so far that they snap and react harshly.

    • Watchman says:

      I’d replace means to an end with power here. Power literally (if using a French derivation is literal) means the ability to do something so is to a degree synonymous with means to an end, but also allows for the unintentional creation of hierarchies that would result from successful harassment even if such a hierarchy was not the object. Being able to assert your sexual identity over others is clearly a statement of power, and the use to which you put that power is less relevant.

      This then replaces skill with power to harrass which involves not only skill but status and the known outcomes of earlier incidents of harassment (so if there is a 1970s culture in an office all moustachioed men will be empowered to harass). Social skills are a key component of this as they are situational and knowing how to navigate the social rules of your workplace would presumably make harassment that much easier.

      These changes have two benefits for your hypothesis. Firstly any follower of Foucault will automatically agree with you. Probably more usefully your hypothesis becomes less rooted in motivation and ability and is more a recognition that social context is important to harassment whilst retaining its thrust that a preponderance of those with lower social skill (=less power to harrass) will make an industry less prone to harassment. I’m not sure it’s going to be the whole story but it is a sensible case.

  34. Garrett says:

    How much of the intentional harassment (I’m discounting dementia patients, etc.) is predicted by the social skills of the harasser? That is, someone who is socially adept might either think that they are more likely to have their advances be acceptable or that they will be able to more easily get away with it.

    In contrast, most of the STEM fields listed involve people who are sub-average as it pertains to charisma/social skills and are generally aware of that fact.

  35. oppressedminority says:

    The industries that rank lowest in EEOC’s data tend to be small industries with very few women – for example, taken seriously the WaPo’s graph shows that mining has the least problem with sexual harassment of any industry in the world. Is this thanks to their uniquely progressive culture – or because there are practically no female miners? I’m going to say the second one.

    Are you trying to find the industries with the lowest sexual harassment, or the most progressive industries?

    From my totally non-scientific, casual observation of the world, it would appear that the more progressive an industry, the more sexual harassment will occur. This makes sense as progressivism (despite all the well-meaning lies it tells itself) sets the stage perfectly for sexual harassment to occur:
    -large number of women in the workplace
    -delayed marriage
    -sexual freedom as an ethos
    -large number of feminist men

    It’s basically a powder keg drenched in rocket fuel.

    • Ozy Frantz says:

      I’m not sure why a large number of feminist men would cause more harassment to occur, even from an anti-feminist perspective: the anti-feminists I know spend quite a lot of time complaining about feminists making men too scared to flirt.

      FWIW, my totally non-scientific, casual observation of the world says that the most important factor is age. Millennials, regardless of political affiliation, basically all agree that the prototypical examples of sexual harassment (boss demands to have sex with employee or the employee will be fired, displaying porn at work, saying a fellow employee has great tits, repeatedly asking out another employee who has clearly rejected you) are wrong. The disagreements are mostly about whether anti-sexual-harassment law primarily targets that kind of unacceptable behavior or generally acceptable behavior (such as asking people out once).

      Conversely, in the sixties and seventies, there was not a consensus that this behavior was wrong, instead of just being boys being boys and something women have to put up with to get ahead. (To be clear, the rationalizations differed, but both liberal and conservative men did a lot of sexual harassment.) A fair number of the most iconic MeToo cases, such as Weinstein and O’Reilly, are men getting fucked over by the fact that behavior that was acceptable for 90% of their careers is unacceptable in the last 10%.

      • Simon_Jester says:

        With, I infer, the caveat that by “behavior that was acceptable,” we mean something close to “behavior they could get away with without punishment?”

        I’d hesitate to call “the long-delayed reckoning for all the horrible shit Bill’s been doing to people he knew couldn’t stop him” a case of “Bill getting fucked over.” 😛

        • Ozy Frantz says:

          The latter, yes, as a feminist and a Millennial I agree with the general consensus that Weinstein’s and O’Reilly’s behavior was morally wrong.

          • oppressedminority says:

            I dont even think Weinstein and O’Reilly believe their behavior was in any way acceptable.

          • Simon_Jester says:

            I suspect that they believe it was right for them to treat women that way in the same way that embezzlers often rationalize a justification for why the money they stole was really theirs all along.

            It’s very easy for human beings to come up with excuses for ignoring a moral principle.

        • SamChevre says:

          I would say it a bit more narrowly: “behavior that was acceptable” means somethinng close to “behavior that the majority of people with knowledge of the field think should not be punished.”

          So behavior that is gotten away with because of disparities in power, blackmail, and so on is not “acceptable”, even if it’s not punished. On the other hand, behavior that is unpleasant, but everyone expects it and sees it as “how this work is”, is “acceptable”. (So, as a father, being vomited on is acceptable: most people would expect it goes with the role.)

          • Douglas Knight says:

            How do you apply that to the examples at hand? Aren’t disparities in power, both in Hollywood, and in media that might report on it, seen as “how this work is”?

          • Simon_Jester says:


            Rewind to the Roaring Twenties or the ’30s. Suppose Howard Hughes (to pick a guy who did a lot of the same shit Harvey Weinstein did towards actresses in his movies) tried to coerce an unmarried woman into having sex with him. Suppose the woman’s father came looking for him with an axe… Well, it’d be a big scandal, but not many people would be asking “wait, why did Mr. Exampleman do that?”

            Everybody would get that what Howard Hughes did in that situation was wrong, in that it was a thing that in a perfect world would not happen, and that might reasonably be viewed as “wronging” some or all of the other people involved.

            But it likely wouldn’t be viewed as a matter of public interest or a reason to boycott Howard Hughes’ movies. Or if it was, it’d be viewed thusly by the same kind of generically prudish people who’d boycott Hughes’ movies for portraying a woman in a bathing suit.

            So is Hughes coercing starlets into bed with him in the Roaring Twenties behavior that at the time was ‘acceptable?’ As in ‘yeah, that’s how work is, and it’s not a problem, everyone should just get used to it, geez/’

            Or was it ‘unacceptable, but something you could get away with by being as rich and powerful as Howard Hughes?’ Because Hughes could get away with coercing women into having sex, on account of being very wealthy and powerful. Someone without his money and power would be treading on rather more dangerous ground.

          • SamChevre says:

            The Howard Hughes example is helpful. I’m intending to say “”it [was] ‘unacceptable, but something you could get away with by being as rich and powerful as Howard Hughes?’ ”

            People know that the powerful get away with unaccepatble behavior; they might expect it. I’m defining “acceptable” as “file off the serial numbers, and few people would think the person acted wrongly” not “people would think if it was Mr Big he’d probably get away with it.”

          • Aapje says:


            I think that a lot of people would also have responded with:

            “Everyone knows that Hollywood is Sodom and Gomorrah, so any decent woman would stay away from them in the first place”

          • Simon_Jester says:

            @Aapje , you’re right, and I mentioned that:

            But it likely wouldn’t be viewed as a matter of public interest or a reason to boycott Howard Hughes’ movies. Or if it was, it’d be viewed thusly by the same kind of generically prudish people who’d boycott Hughes’ movies for portraying a woman in a bathing suit.

            The ‘generically prudish’ people I’m talking about are the same people who’d say ‘no decent woman should go to Hollywood.” Instead of specifically addressing the object-level problem of “this behavior towards women is terrible,” they try to take it a step back by building fences on the ground and telling ‘decent women’ not to cross the fence.

            This does have the benefit of letting most women be relatively safe from certain abuses, IF they never leave the fence.

            The catch is, societies like this tend to (in effect) tolerate the fence being crossed without official social sanction, by designated indecent women (gasp, the painted jezebels!). These women are subsequently related terribly. But who don’t really matter very much, not from the point of view of prude culture, because by being painted jezebels they’ve lost all claim to society’s protection from terrible behavior by males.

            In other words, some variation on the Madonna-whore complex. Madonnas don’t go to such places, so a woman in such a place must be a whore, which means they can be treated however the strongest man around likes.

            This is, suffice to say, not a viable social norm to use in a society with anything resembling legal equality of the sexes. Or in which women in general are considered to have agency rather than being a sort of freestanding extension of the personage of a male head of household.

          • Aapje says:


            I think that you paint a one-sided picture that makes it seem that men get all these protections that women only get if they conform strongly.

            My belief is that men are partially in a similar situation, where they only get protection if they stick to their gender role, but also that certain protections are simply never or much more reluctantly granted to them.

            When I was being beaten up, the adults had no willingness to intervene and they at most gave me the advice to fight back. It was a good lesson that men have to ensure their own safety. People will often not take the effort to help us, while they will take the effort to help women in similar situations.

            The same pattern is visible in many other situations.

            So given this, the situation is more complex: women do get more restrictions, but they also get more protection in return.

            So how do we then get to equality? Do we actually start treating women like men, letting them be beat up without intervening? Or do we start treating men like women, putting restrictions on them and giving them protections in return? Or do we try to treat everyone better than they are now? Is that viable or are there strong negative consequences to this? Do we need to find alternative solutions to limit or negate those negative consequences?

            Or does society just continue as we do now, by denying the truths in favor of a very simplified narrative that is secretly rejected and resisted by a great many people who know that it is nonsense?

      • albatross11 says:

        As one datapoint, my mother in law has some very unpleasant stories of interactions with men in her PhD program in the 70s, both other students and professors. Lots of there were what would now be unambiguously seen as sexual harassment, probably without it being 100% clear to the men that what they were doing was actually wrong rather than just a little socially out-there and forward.

      • oppressedminority says:

        Again, I must emphasize that this is just a general impression, with no formal study or empirical data to back this up. But there is a good working theory behind it and anecdotal evidence that supports it.

        Feminist men are generally those who could not be successful by being manly men, and try to appeal to the opposite sex using virtue signaling. It doesn’t generally work, they get frustrated that women go for the “bad guy” instead of the “good feminist guy”. Harassment or assault may then follow with greater than average frequency.

        That’s the theory. The anecdotal evidence I will provide in this context is nicely captured here. Granted, all of these are not for sexual harassment, but Matt Hickey, Devin Faraci, Robert Marmolejo, Glen Fleishman, Joss Whedon, Andy Signroe, Michael Hafford, Sam Kriss, Rupert Myers, etc… are a nice representative sample.

        • rlms says:

          Golly, some people have a lot of time on their hands.

        • Calion says:

          Okay, your mention of Joss Whedon intrigued me, so I followed it up. His crime seems to have been cheating on his wife with some of the many gorgeous women who looked up to him that surrounded him all the time, and lying about it for years. That makes him human, not a sexual abuser.

          • oppressedminority says:

            Agreed. I edited the post to reflect that.

          • Jaskologist says:

            I argue. He slept with the actresses working under him, didn’t he? That’s your classic power imbalance, and pretty much what Weinstein did. 9 times out of 10, I’d bet Weinstein was also dealing with women who knew exactly the bargain they were making.

          • John Schilling says:

            I argue. He slept with the actresses working under him, didn’t he?

            I don’t believe any of the women Whedon slept with have been identified, or have come forward even in the post-Weinstein era. Google suggests that his ex-wife once accused him of sleeping with “his actresses” among others, but that’s pretty weak in isolation.

            “Unfaithful womanizer” is probably the greatest accusation that can be made with confidence and thus ought to be the limit for accusations made in public. Depending on the context, it may be sufficient.

        • arlie says:

          *rolls eyes*

          So if a feminist is someone who believes that women should be treated as well as men, and the opposite of a feminist man is called “manly” then “manly” means “treats women badly” or at least “believes women should be treated badly, even if he doesn’t have the power and/or gumption to do so himself.”

          Or to put it in soundbite form, “manly” is a synonymn for “contemptible bully”.

          That’s not what the word means in the English I grew up speaking. “Bully” clusters with “coward” and other words that cluster with “unmanly”… even though they can also be applied to females … where they do NOT cluster with “feminine”, just in case you were unclear.

          • Conrad Honcho says:

            Are you suggesting that feminism is the radical idea that women are people too?

          • oppressedminority says:

            By feminist men, I mean self-declared feminists, not people who believe in the equality of the sexes. The former group is quite small, the latter is the overwhelming majority.

            By manly men, I mean men who possess stereotypical masculine traits. These are the men that most women are attracted to, just like women who possess stereotypical feminine traits are those that most men are attracted to.

          • vV_Vv says:

            So if a feminist is someone who believes that women should be treated as well as men


          • Aapje says:


            So if a feminist is someone who believes that women should be treated as well as men

            What do you call a person who believes that men should be treated as well as women?

        • mdet says:

          Counter-examples: Nick Offerman, Common, Patrick Stewart, John Hamm, Daniel Craig, Will Smith, Chris Hemsworth, John Legend (may or may not be “manly”)

          I don’t think either of our lists are anything more than anecdotes though.

      • Conrad Honcho says:

        I’m not sure why a large number of feminist men would cause more harassment to occur, even from an anti-feminist perspective

        It’s the right-wing version of “that preacher who rails against the gays must be overcompensating because he’s secretly gay.”

        e.g., James Franco wears #TimesUp pin decrying bad behavior against women in general to mask his own awful behavior against specific women.

      • vV_Vv says:

        Millennials, regardless of political affiliation, basically all agree that the prototypical examples of sexual harassment (boss demands to have sex with employee or the employee will be fired, displaying porn at work, saying a fellow employee has great tits, repeatedly asking out another employee who has clearly rejected you) are wrong.

        Are these examples actually the most common or are they just stereotypical?

        • Ozy Frantz says:

          The disagreements are mostly about whether anti-sexual-harassment law primarily targets that kind of unacceptable behavior or generally acceptable behavior (such as asking people

          out once).

      • Aapje says:

        @Ozy Frantz

        My perception is that a subset of feminist men is drawn to the ideology because it lets them blame their bad traits on others: “I’m not a bad person/man since all men are bad, so my bad behavior doesn’t reflect very badly on me, especially since I’m a feminist who tries to be better.” Hugo Schwyzer, who was your co-contributor to The Good Men Project, seems to have been such a person. More examples.

        Of course, this is not sufficient to conclude that feminist men are more likely to abuse, but the possibility is there.

        • oppressedminority says:

          How could I forget Hugo Schwyzer? Thanks for reminding me.

        • Protagoras says:

          I knew Hugo’s father. He taught me a lot about Kant, and also had amusing stories about being a grad student at Berkeley in the 60s. So it always made me feel a little weird seeing Hugo’s troubles discussed in the various places that they were highlighted; I couldn’t help but feel a little sorry for Hubert (who is dead now, but not soon enough to avoid having known about some of his son’s activities).

  36. Conrad Honcho says:

    Compare all the anecdotes and popular lore about how immigrants are criminals. It’s totally false – immigrants have crime rates well below native-born citizens. But we only know that because there have been really good studies.

    This is attacking a strawman. Nobody is concerned about immigrant crime. They’re concerned about illegal immigrant crime. Nobody thinks the Indian guy on an H1-B visa coding at Microsoft is out stabbing people. They are concerned that the Mexican drug mule who hopped across the border illegally is stabbing people.

    The rates at which illegal immigrants commit crimes are hard to pin down because:

    1) We don’t know exactly how many illegal immigrants there are here.

    2) Crime is a segregated enterprise. Criminals tend to victimize people like them / in close proximity to them, so many victims of illegal immigrant crime are illegal immigrants themselves, so their crime reporting rates will be lower as illegal immigrants tend to avoid interacting with law enforcement officials.

    3) For some reason public officials are loathe to release clear statistics on illegal immigrant crime. Maybe it makes them look bad. Dunno.

    However, here is an article using SCAAP (State Criminal Alien Assistance Program that partially reimburses states and localities for the cost of incarcerating certain criminal illegal aliens) data to show that no, rates of crime by illegal immigrants in states with large illegal alien populations for which we have data are in fact higher than people who aren’t illegal aliens.

    It seems unfathomable that a population selected for poverty (which correlates with crime) and disregard for the rules of society is somehow less criminal than people who are not specially selected for poverty and disregard for the rules of society, but that’s what the media would have us believe. Particularly when there are whole classes of crimes that illegal aliens are very likely to engage in like identity theft and fraud just so they can get documentation for employment, utilities, benefits, etc.

    We’ve also got a Simpson’s Paradox thing going on here, because when we’re talking about illegal immigrant crime rates versus “native born” crime rates we’re not comparing similar demographic cohorts. Why don’t these studies ever break out the crime rates for “illegal Latinos” versus “native born Latinos” and compare those rates to those of “native born blacks” and “native born whites?” I think it’s because it would then no longer serve as effective eulering political rhetoric. The stereotypical voter complaining about “illegal immigrant crime” is the white suburban Republican, who then gets the eye-roll “ugh, don’t you know immigrants (tee hee I didn’t say illegal no one will notice!) commit crimes at lower rates than native born Americans?” response, implying “native born Americans like you, whitey.” When in reality the statistics, if compiled honestly, might show that, “ugh, don’t you know illegal immigrants are slightly less criminal than your current criminal racial underclass?!” This would not be effective rhetoric.

    And another thing, even if our saintly illegals were less criminal than our native born lily white suburbanite Republicans, it still wouldn’t matter, because every crime an illegal commits is one that wouldn’t have happened if our government were doing its job of guarding the borders.

    So it’s all a strawman and lying with statistics. People care about “illegal immigrant crime,” not “immigrant crime,” the studies do not show illegal immigrant crime rates are lower than native born crime rates, the studies do not break out extreme confounders like race, and even if they did and showed what the political rhetoric would have you believe, it still wouldn’t matter because those who care about this care about absolute numbers and not rates.

    Also, I think sexual harassment is bad and people shouldn’t do it.

    • albatross11 says:

      Can I suggest that we move this object-level question somewhere else? We’re already in a CW-heavy discussion; adding a second unrelated CW-heavy discussion seems unlikely to improve the heat/light ratio.

      • Jiro says:

        Ultimately, this discussion is here because Scott made the bad argument conflating “immigrants” and “illegal Mexican immigrants” here. I mean, Scott would be perfectly within his rights to have people not be able to rebut the things he says where he says them, since it’s his blog, but that would lead to less rational discourse.

      • Conrad Honcho says:

        I’m happy to discuss it elsewhere, but perhaps Scott should choose a different example illustrating statistics not supporting media fears. Maybe something not as CW-heavy, like “the summer of the shark.” tl;dr shark attacks were down over averages in summer 2001, but the media reported every single shark attack for some odd reason, making people think sharks were going nuts.

    • Calion says:

      This line of reasoning would be more convincing if the people who were upset about illegal immigration were not to a very large degree the exact same people calling for limiting legal immigration still further.

      • albatross11 says:

        Compare with gun control advocates and people who are the most upset about the dangers of widespread gun ownership.

      • Conrad Honcho says:

        I don’t understand your reasoning. Someone wants to eliminate illegal immigration because it’s illegal and because some illegal immigrants commit crimes, but also wants to limit legal immigration because it causes downward wage pressure and therefore they fraudulently believe legal immigrants commit crimes?

        “Some people think illegal drugs bought from street dealers are bad for you, but this is silly because Tylenol from Walgreens is perfectly safe.”

        • arlie says:

          *sigh* At an individual level, it’s not a problem.

          But when there’s a strong correlation between people-wanting-fewer-legal-immigrants and people-stressing-how-many-crimes-other-than-immigration-law-violations-are-commited-by-illegal immigrants, it’s reasonable to ask if the whole things is caused by something else.

          We see the same thing with the strong correlation between people who loudly oppose legal abortion, and people who loudly oppose teaching post-puberty young people about contraception. In their case, the common factor seems to be a desire that extra-marital sex have the worst possible consequences. (They’d probably say it was religion and/or opposition to pre-marital sex.)

          • Evan Þ says:

            But see Scott on the principle of charity, using exactly this example. He analyzes your specific argument at the end:

            The pro-lifers seem to be doing much the same thing. Sometimes sex has unwanted consequences. There’s a convenient way to avoid those consequences – abortion – which would be very helpful. But if that convenient way to avoid the consequences hurts another person – the fetus – then we’re back into “take the consequences of your own action” mode.

            So while I agree there is a certain element of “you must bear the consequences of your own action” going on here, I don’t think it is opposed to claiming the main issue is the rights of the fetus, nor that the pro-lifers are particularly denying it, nor that it’s anything especially evil.

          • Aqua says:

            Does it even matter how many crimes illegal immigrants do really? If it’s greater than zero, seems like a win to stop it at the border. Even if it is lower on average, you are reducing crime overall

          • Aapje says:


            A lot of conservatives seem to believe that sex education causes promiscuous behavior and thus more unwanted pregnancies. So from their perspective, they are reducing the need for abortions by objecting to sex education.

          • albatross11 says:


            Your argument proves too much. Exterminating the human race would decrease crime even further, so it’s a win, right?

          • Conrad Honcho says:

            *sigh* You got me. There’s nothing I can do to argue with a mind reader. I don’t really have a problem with people sneaking into my country illegally, flaunting the rules of the society, murdering, raping and robbing my fellow citizens while the government I pay for refuses to enforce the laws the citizens voted for, the same sorts of laws against illegal entry that every nation on the planet has and enforces without significant issue, and would absolutely enforce against me if I were to try sneaking into another country without permission. It’s all a ruse. I actually love all the violence, lawbreaking and murder. I just secretly hate Mexicans because a taco truck killed my whole family and so I make up the whole big front about not liking murderous illegal invaders as a clever ruse to hide my frothing racial hatred.

          • Aqua says:

            I don’t think it proves too much. I’m saying you might as well prevent illegal immigration, since it’s illegal by itself.

            It’s not that important for illegal immigrants to commit crime at a higher rate than avg. You could view any crimes commit by illegals as sort of “extra crimes” that we didn’t expect to have occurred since we expected there to be no illegal immigrants (ideal case I guess)

            I’m saying I find it weird how the crime rate of illegal immigrants really enters the argument at all. Imagine illegal immigration was actually undeniably good, like 100% of them became doctors or something and it solved healthcare magically. This would then be an argument for increasing/restructuring legal immigration and funneling these people into that system. It would be silly to keep calling it illegal.

            In the real world, illegal immigration seems to sometimes be argued as it would be better if it was “illegal but not really that illegal” (sanctuary stuff) which I always found strange. If current illegal immigration is actually net good, why isn’t it becoming legal? If we are deciding it is in fact net-bad, maybe it should actually be enforced?

          • Simon_Jester says:


            In the real world, illegal immigration seems to sometimes be argued as it would be better if it was “illegal but not really that illegal” (sanctuary stuff) which I always found strange. If current illegal immigration is actually net good, why isn’t it becoming legal? If we are deciding it is in fact net-bad, maybe it should actually be enforced?

            This is very easily explained.

            The US used to admit very large amounts of unskilled or semi-skilled immigrant labor. Extensional definition of this mindset would look like: Ellis Island, that peasant great-grandpa of yours who came to America knowing six words of English with nothing but the shirt on his back and the will to work, Little Italies and Little Germanies and for that matter Chinatowns in cities dating back to the American 19th century or earlier, and of course Emma Lazarus’s The New Colossus, which got built onto the Statue of Liberty for a very good reason.

            “Give me your tired, your poor, the wretched refuse of your teeming shore.” That was legal immigration. Lots of it.

            Fast forward to the 1920s, which was very not coincidentally the high water mark of the Ku Klux Klan and the isolationist movement in post-Reconstruction American politics. A lot of American nativist whites (one may note irony there) felt that all this “wretched refuse” contained too many undesirable people, like Italians and Slavs and Jews and Chinamen. And that this threatened America’s ‘national character’ in some way.

            This resulted in some very sharp quotas being clamped down for the number of immigrants the US would admit legally, and to a very large extent those quotas are still in force.

            The thing is, demand for motivated immigrant unskilled and semi-skilled labor didn’t go away. It just became politically impossible to acquire such labor legally. Thus, a vast, prodigious industry arose in the field of smuggling such labor into the country and permitting it to remain illegally. There was no such system in place back in 1900, because there really didn’t need to be.

            So in short, a perverse situation has arisen. There is a great demand for motivated immigrants, including those with desirable skills or for that matter no skill at all besides a strong back and willingness to roll up one’s sleeves and get to work. But there is a legal regime designed by nativists, and which nativists are very hostile to any change of, that “caps” the lawful supply of such immigrants.

            Things like the H1B visas, student visas, and family reunification that lets a citizen bring their relatives into the country legally are all elaborate workarounds devised to deal with the reality of restrictive quotas imposed by nativists, by creating extra channels through which legal immigrant labor can flow into the economy and satisfy demand.

            But at the same time, even with all this in play, it’s like trying to put out a forest fire with a garden hose. Thus, illegal smuggling of immigrants into the countries is also still a thing. And it’s strongly demanded by the economy, until and unless we remove the nativist quota systems

            Which leads to things like sanctuary cities and amnesties. The idea being that yes, illegal immigration is in fact against the law, but the laws in question are perverse and leading to bad outcomes, so we will civilly disobey or decline to enforce those laws, depending on whether or we’re local government (which offers sanctuary) or federal government (which offers amnesty).

            And YES, the obvious solution is to simply create a pathway that permits legal immigration on a scale large enough that illegal immigration would decline. But the existing system does not permit that; it takes a decade or more to acquire citizenship in many cases, and there are pretty sharp limits. And changing or streamlining this system would probably require major Congressional supermajorities, on an issue where the public is divided.

      • gbdub says:

        It’s only a problem if you think the only valid solution to illegal immigration is to reduce demand by increasing legal immigration.

        But generally such people are proposing we solve illegal immigration on the “supply” side.

    • mdet says:

      I agree with albatross and don’t want to inflame a tangential debate, but do you have a response to this American Conservative article that you can quickly direct me to? The article concludes that Hispanics in the US do not have a substantially higher incarceration rate than White people at the state level (where most violent offenses are) when you control for age (since criminality is very related to age).

      I think your point about crime against illegal immigrants (and therefor crime *by* illegal immigrants) being under-reported is good, as is the one about illegal immigrants being “selected for” crime. And the article deliberately looks at crime by Hispanics instead of by (illegal) immigrants, which weakens their argument but I think is defensible.

      • Conrad Honcho says:

        That article is an example of the exact strawman / missing the point error I was upset with Scott for making. It starts with a string of quotes about specifically illegal immigrant crime, and then spews eight pages of statistics about crime rates of Hispanics versus whites, regardless of immigration status and never provides any figures for the crime rates of illegal immigrants, which is the matter at hand.

        No one is particularly concerned about legal immigrant crime (with a carve-out for Islamic terrorism, which is a separate issue with entirely different motivating factors than generic “crime” and only related to one small subset of immigrants). The concern about crime is strictly over those of who enter the country illegally.

        Why is it so difficult to understand the difference between legal immigration and illegal immigration?

        • mdet says:

          I wouldn’t call this a strawman, because they explicitly acknowledge what they’re doing. The article basically says “Look, there’s no explicit numbers on crime by illegal immigrants, so I’m going to look at the amount of crime by Hispanics as well as the crime rate in states / cities near the border with large Hispanic populations. This is the best proxy for ‘illegal immigrants’ I can think of”.

          You, me, and the author all agree this isn’t as strong as a database of “crime by illegal immigrants” would be, but I think the difference is that some people consider this data to be “close enough”, and you don’t (which is fair).

    • Steve Sailer says:

      Illegal immigrants typically arrive in America when they are older than the age at which joining a street gang seems like a really good idea.

      Also, since Mexico took over the cocaine business from Colombia in the 1990s, there has often been more money to be made in gangs in Mexico than in America, which tends to attract bad guys to Mexico rather than to the US. Before the death of Pablo Escobar in 1993, crime didn’t pay all that well in Mexico, and the brief reign of liberalism over the criminal justice system in the US in the 1960s and 1970s, with its short prison sentences, had made America an attractive destination for bad guys. But lately Mexico is more attractive to Spanish-speaking criminals than is the US.

      On the other hand, more than a few of the sons of illegal immigrants think joining a street gang sounds cool.

      On the other other hand, the average crime rate for African Americans is so high compared to every other group that in many cities illegal immigrants pushing out blacks would lower the overall crime rate, as has happened in Los Angeles.

  37. Ozy Frantz says:

    Your preregistration includes biology and psychology as female fields, but the chart has them grayed-out (assuming that we consider “social science” to be basically the same thing as “psychology”).

    Calling law a female profession seems dubious to me: only a third of American lawyers are female. It might be better to rename the categories “thing-oriented” and “people-oriented,” and point out that at least one people-oriented category is majority male.

    • dick says:

      > only a third of American lawyers are female

      More specifically, that wiki article says that the legal profession generally (lawyers, judges, law school profs, etc) is 34% female, but it also says that 60% of US attorneys are female. The impression I get is that law is at or close to gender parity at the low end of salary/status and not at all close at the high end. I would think that’s true of healthcare and the media as well?

      In any event, one of the sources says that new law school grads are almost exactly 50-50, which would suggest that the field is, if not actually at gender parity, approaching it in a way that programming/math/etc are not.

      • INH5 says:

        According to BLS statistics (see Table 11 on this page), only 36% of lawyers are women, but 87% of paralegals and legal assistants and 77% of “miscellaneous legal support workers” are women, with legal occupations as a whole being about 50/50.

        Lawyers being majority-male might be just because law school students only recently reached gender parity, but one thing to check is how many law school grads became paralegals or assistants instead of lawyers, and whether female law grads are more likely to take that career path.

        • albatross11 says:

          I wonder how assortative mating affects this. Lots of people meet their eventual husband or wife either in school or at work, and people increasingly tend to marry someone with comparable educational and professional accomplishments to their own. But then, when they decide to have kids, it’s almost always the woman who pulls back from work a bit (maybe dropping out entirely, maybe going part-time for awhile). This is especially true in really high-powered, high-workload fields–it’s *hard* to have both parents in 60-hour-a-week jobs with constant travel when there’s a 3 year old and an 8 year old at home.

          The reason I think assortative mating matters here is that the woman with a law degree and a decent law career who is married to a guy with a stable but not spectacular job will probably keep her own job–two parents with normal office jobs can manage kids with a little daycare, and they may need the extra money. The same woman married to a guy who’s a partner in a big law firm and is working crazy hours and pulling down crazy money is a lot more likely to decide to stay home with the kids–her husband’s crazy hours mean that it’s hard for her to keep a job plus manage kids even with daycare, if they don’t want to hire a nanny or something, and they probably don’t need the money anyway.

          It seems like might be driving some of the tendency of high-end educated women to drop out of the workforce.

  38. Nancy Lebovitz says:

    Is there any information from people who’ve moved from one field to another about whether they’ve been subject to more or less sexual abuse in various fields?

    • Steve Sailer says:

      A woman who had switched from the marketing research industry to “gaming” (i.e., casino gambling) industry told me that men in the gaming industry were awful (of course, so were the women) and she was glad to be back in marketing research.

  39. Deiseach says:

    The industries that rank lowest in EEOC’s data tend to be small industries with very few women – for example, taken seriously the WaPo’s graph shows that mining has the least problem with sexual harassment of any industry in the world.

    Now it doesn’t, but that’s the fault of bleeding heart liberals like the Earl of Shaftesbury and his 1842 Parliamentary Report on The Employment and Conditions of Children in Mines and Manufactories. Because of him and others, you could no longer have five year old children sitting in darkness (candles are expensive!) for twelve hours a day six days a week down the mines – if you wanted child labour after the 1842 Act they had to be at least ten – and it was down to Victorian middle-class prudery that having naked adult men working alongside half-naked pubescent girls was deemed unacceptable:

    These two girls were, at the time spoken of, aged 14 and 13 respectively; they had been at work in the pit eight years, and the flippant manner in which the younger of the two gave her evidence, shows what effect her employment and associations have had upon her moral feelings.

    She said:-

    “There are other girls that hurry in the same way, with belt and chain. Our breeches are often torn between the legs with the chain. The other girls’ breeches are torn as often as ours; they are torn many a time, and when they are going along we can see them all between the legs, naked. I have often; and that girl, Mary Holmes, was so today; she denies it, but it’s true, for all that”

    …”In the Flockton and Thornhill pits the system is even more indecent; for though the girls are clothed, at least three-fourths of the men for whom they hurry, work stark naked, or with a flannel waistcoat only and in this state they assist one another to fill the corves 18 or 20 times a day. I have seen this done myself, not once or twice, but frequently.

    Neither do the girls or the men attempt to gainsay the fact.

    When it is remembered that these girls hurry chiefly for men who are not their parents, that they go from 15 to 20 times a day into a dark chamber, which is often 30 yards apart from any one, to a man working naked, or next to naked, it is not to be supposed but that where opportunity thus prevails sexual vices are of common occurrence.”

  40. MNadolsky says:

    Toward the end of the article, one possible interpretation of the data was not entertained: Could it be that nerdy types are simply less likely, for some reason, to sexually harass others, while artsy/media types are more likely?

    Claims of sexual harassment are claims of motivations. Behavior is or is not harassment dependent on the motivation of the person behaving (or, according to some, according to the interpretation of the behavior by the person being behaved toward – which is based on that person’s interpretation of the motivations of the person behaving). This is the fundamental difference between a well-meaning flirtatious advance and sexual harassment in most likely a significant number of these cases.

    When I see claims about motivations of other people, especially people the claimant is not very close to or generalized groups of people, my go-to is to consider that it is projection. As far as I can tell, predicting the motivations of others based on their behaviors is extremely inaccurate and predicting them based on what people report about their own motivations is more accurate but still not great – but predicting peoples’ motivations based on what they most readily believe are the motivations of others is *highly* accurate.

    Take someone who has racial animus toward some group and someone who does not have racial animus toward any group, present them with a complete stranger stating an opinion in favor of immigration control that does not refer to race, and ask them why they think that person holds that belief, in my experience the former will readily believe that opinion is based in racism while the latter will not.

    All that being said, could it simply be true that people in the media, male and female, are simply the type of people that are willing to sexually harass others, and project those motivations onto people they don’t like – the nerds?

    • albatross11 says:

      One confounder that would be interesting to check (assuming you could get data on it) is prevalence of dating/romance in the workplace. If you assume that some fraction of sexual harassment is a failed attempt at dating/romance in the workplace, then you would expect to see the same industries with a lot of sexual harassment have a lot of workplace romance/sex going on. And those obviously correlate with age and marital status, as well.

  41. herculesorion says:

    Something I’d like to see is a correlation between “percentage of women” and “prevalence of harassment”. Like, does a potential harasser become encouraged by a target-rich environment? Or maybe he figures that if there are few women in a workplace, each of them receives a disproportionate amount of the office’s social capital and would be more able to respond to unwanted behavior?

    It also does suggest an explanation for why there are so many reports of how harassment is such a problem; as with liberals and wealth inequality, the people who complain of the problem are intentionally putting themselves in the place where the problem is worst. (And–as with wealth inequality–they’re assuming that they’re Smart People Who Don’t Make Mistakes and therefore the problem is the fault of someone else, probably their political enemies who they don’t like anyway so it’s a win-win.)

  42. albatross11 says:

    To my mind, the most important step in thinking about this is realizing that the prestige media are an incredibly bad source of information about this issue. They’re so wrapped around the axle w.r.t. being on the right side in the culture war and getting clicks/outrage farming and confirming their readers’ and writers’ biases that their coverage of the whole issue is about one step up from the Weekly World News’ coverage of the latest Bigfoot sighting.

    On a lot of issues, the prestige media impart negative information–after watching/reading, you think you know something about the issue, but a lot of what you know ain’t so.

    • gbdub says:

      Frankly it feels like at least some of it is projection. Media (not just Hollywood) seems like it might be uniquely bad for women.

      As others have noted, this may have something to do with the relatively low demand for jobs relative to supply. It’s a more social field in general, plus your career advancement is much more tied to “who you know”.

  43. greghb says:

    Some people who reviewed these data suggested that the media focuses more on a narrative where STEM workers are dismissive of women, rather than one where they harass them. I don’t find this accurate – a quick glance at any media source will show they focus on both narratives.

    God help me, but I’m going to defend “the media” on this point. There are several possible narratives:

    1. Women are under-represented in tech.
    2. Tech is unique among industries for its under-representation of women.
    3. Women are sexually harassed in tech.
    4. Tech is unique among industries for its sexual harassment of women.

    From my media-consumption vantage point, there is plenty written about 1, 2, and 3. (In a sense 2 subsumes 1, so maybe there is less on 1 per se.) But I really don’t see much on 4. I tried “a quick glance at any media source” (googling “nytimes women in tech harassment”, skimming the articles), and most of it is 2 alone, 3 alone, or, at worst, a juxtaposition of 2 and 3. I really don’t see 4.

    Since #MeToo got its big push from the Weinstein story, I think the media can’t get away with saying sexual harassment is uniquely bad in tech. It’s just cemented in the popular mindset that, if you’re going anecdote-for-anecdote, Hollywood is pretty bad. When journalists do go looking for hard numbers, they usually find the under-representation numbers and report on them — which is fine. But I am actually impressed, with a somewhat low bar, that the media doesn’t do more to say, “and sexual harassment of women must also be disproportionally high in tech, because representation of women is so low.”

    There are a million media-consumption vantage points, so I’m not sure there’s much value in saying, “From my vantage, I see X!” I guess if nothing else, maybe I’m curious to see examples of the media pushing 4. I’m sure there are some. It’ll open my vantage point a little to learn where to find them. But I do disagree with the sweepingness of the original claim.

    • gbdub says:

      Post-Weinstein maybe, but pre-Weinstein tech was definitely portrayed as uniquely awful for women, specifically because the (overrepresented) men there were particularly bad to them. Maybe it was portrayed as more general sexism/discrimination, but it definitely felt like harassment was being pushed as part of that. E.g. Dongle-gate.

    • rlms says:

      That’s also my impression. The articles I found from searching “Silicon Valley sexual harassment” (e.g. this and this) are about VCs harassing female founders etc., not male engineers harassing their female coworkers.

    • Conrad Honcho says:

      The most common reason cited for women being underrepresented in tech is because of men in tech being awful. Harassment in general, stereotyping, hostility, etc, but not necessarily sexual harassment.

    • Garrett says:

      The trick is that by pointing out (2) and (3) frequently, in close proximity, and with much eyebrow wiggling, the major media organizations can imply (4) while still having the deniability of not having said it.

    • greghb says:

      I guess to be fair, “sexual harassment” is vague. I like the taxonomy described in this article (scroll down to the infographic). Under this framework “sexist hostility” is a category that includes behaviors like “condescension because of someone’s gender” and “sexist remarks”. It differs from “sexual hostility”, “unwanted sexual attention”, and “sexual coercion”. In short, sexist vs. sexual. I could believe that the media attributes under-representation of women in tech to an exceptional excess of sexist hostility as defined in the above. But that doesn’t really sound like what Scott is talking about here.

  44. apollocarmb says:

    One major problem is the vagueness of the term “sexual harassment”. One person’s sexual harassment could be another person’s funny joke.

    I am really surprised you just glossed over this fact.

  45. vV_Vv says:

    Compare all the anecdotes and popular lore about how immigrants are criminals. It’s totally false – immigrants have crime rates well below native-born citizens.

    This certainly isn’t the case in Europe.

  46. gbdub says:

    I wonder if the big difference in reported rates depending on whether you ask “were you harassed” vs. “did you experience X behavior ever” are primarily due to issues of frequency/persistence?

    E.g. an occasional inappropriate comment might cause you to answer “Yes” if asked the latter, but not the former – and really it probably isn’t “harassment”. I personally wouldn’t feel harassed unless such comments were pervasive and/or persisted for a while. At the risk of typical minding, I do feel like most people would react in a similar way.

    I don’t know, I’ve long been thinking similar thoughts w.r.t. “underreporting” of sex crimes in general – while there are probably some percentage of unambiguously awful things that go unreported out of fear etc., the “right” reporting rate is probably quite a bit less than 100% because we really don’t want to make a federal case out of everything that violates the strict letter of the law. (On the other hand we probably don’t want to take the “minor” stuff completely out of the law, because something that is annoying but harmless if rare or in certain contexts could be traumatizing if persistent or in other contexts)

  47. Nate the Albatross says:

    I’ve been a manager in retail, coffee and finance and I have some observations on the physical and professional environments which could explain a lot.

    In both retail and restaurant turnover is very high and training is minimal. it is very likely to be a first job for many, many people. Even at what would be considered a medium sized business – say a coffee chain with 200 stores – the manager may be the sole interaction the employees have with anyone regarding human resource issues. Not to mention mom-and-pop retail/restaurant where your manager is the owner and also handles all HR duties. In media – high ranking employees often get a say over the operation of the business and significant control. The Screen Actors Guild recently issues a policy against private auditions at homes or hotels. Retail, restaurant and media are jobs that feature many hours of dead time where there are no customers, and often have areas out of sight like stockrooms, kitchens, and dressing rooms. Underdeveloped policies, conflicts of interest and opportunities abound. Health care is a little different as they have policies, but they are also an industry where there are private rooms with beds. So decent policies in health care but far more opportunity.

    By contrast mining and tech are very different environments. Mining is regulated for safety and involves driving heavy machinery. Federal regulators see mining operations policies – including their HR policies. Employees in tech and mining usually receive some education (truck driving, coding) and are taught following rules and procedures is very important. In mining because of safety concerns and in tech because the tech won’t work if someone goes off script. A company that gets REALLY particular about writing procedures for safety or technology will also be good at writing other kinds of procedures. Compared to the subjective enforcement of a retail or restaurant manager and the difference becomes very clear.

    And the physical environments are very different too. Neither mining or tech happen in a food truck. Miners are on a mountaintop, and they can’t take a break for two hours because there aren’t any customers. Tech workers are in cubes, factories or labs – often with open or glass viewing allowing everyone to see everything that is going on.

    When we look at the state of HR policies and the physical environments, some industries are far more susceptible to harassment than others. Tech and mining are particularly unappealing to potential harassers in they have heightened scrutiny, well developed HR policies and hardly any opportunities. This isn’t to say it never occurs, but that it is lower on average makes sense.

  48. Doug says:

    > More surprising were the lower harassment rates in STEM and male-dominated fields

    Maybe men in these groups are just less sexually aggressive and more introverted. It’d be interesting to aggregate # of lifetime sexual partners by industry and correlate with levels of sexual harassment.

    Not that I’ve ever been involved in sexual harassment on either side of the table… But my guess is that most cases don’t look like Alice walking over to Bob’s desk and grabbing his crotch out of the blue. Most situations probably escalate from some fun, flirty, or social interaction that one party takes to far. Programmers don’t know how to flirt and avoid socializing, so they never get into these situations.

    Existing literature on sexual assault shows that it disproportionately occurs in environments of “revelry” and “carousing”. I don’t see any reason workplace harassment would deviate from this pattern. Pharma reps get drunk with their coworkers a hell of a lot more than mathematicians.

  49. terran says:

    Does professional have explanatory power for experiencing harassment, after we control for personaliaty traits? Yes

    > glm(Harassment1==’Yes’ ~ Gender + Income + AutismSpectrumTest + BigFiveE + BigFiveN + BigFiveA + BigFiveC + BigFiveO + Profession,data=ssc,family=’binomial’) %>% Anova
    Analysis of Deviance Table (Type II tests)

    Response: Harassment1 == “Yes”
    LR Chisq Df Pr(>Chisq)
    Gender 115.94 5 < 2e-16 ***
    Income 0.07 1 0.792
    AutismSpectrumTest 0.23 1 0.634
    BigFiveE 1.31 1 0.253
    BigFiveN 1.19 1 0.276
    BigFiveA 0.28 1 0.597
    BigFiveC 0.18 1 0.670
    BigFiveO 1.83 1 0.176
    Profession 97.71 22 1.62e-11 ***
    Results erroneous – see followup comment

    For those not familiar with analysis of variation, this is essentially saying
    that after we control for all other things listed, Profession is still highly
    significant in explaining whether Harassment 1 was answered “Yes” – that if we
    provided a random feature with 22 categories, it would have a probability of
    1.6 * 10^-11 of having the same level of explanatory power as the actual
    Profession data.

    If Profession were relevant only because it was serving as a proxy for personality
    type, we would instead have found that after we added the personality information
    directly into the model, Profession had a low explanatory power. We instead see
    that the explanatory power of the Big 5 and Autism personality measures is very
    poor here, while Profession remains strong. The conclusion is that Profession
    is not just serving as a proxy for personality type of the reporter.

  50. terran says:

    Correction – I found an issue with how the strings were converted to numbers. The qualitative conclusion about profession doesn’t change, but some of the personality traits of the reporter are significant:

    > glm(Harassment1=='Yes' ~ Gender + Income + AutismSpectrumTest + BigFiveE + BigFiveN + BigFiveA + BigFiveC + BigFiveO + Profession ,data=ssc,family='binomial') %>% Anova
    Analysis of Deviance Table (Type II tests)>% Anova

    Response: Harassment1 == "Yes"
                       LR Chisq Df Pr(>Chisq)    
    Gender                56.31  5   7.01e-11 ***
    Income                 0.03  1    0.86908    
    AutismSpectrumTest     5.07  1    0.02429 *  
    BigFiveE               7.30  1    0.00688 ** 
    BigFiveN               0.16  1    0.69081    
    BigFiveA               5.47  1    0.01937 *  
    BigFiveC               3.14  1    0.07659 .  
    BigFiveO               7.87  1    0.00503 ** 
    Profession            76.00 22   7.29e-08 ***
    Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

    The ANOVA does not show the sign of the coefficient; O, E, and A are all positively correlated with reporting of having received harassment.

    However, when we look at harassment outside of work, the results are qualitatively similar, which suggests an argument that profession is still serving as a proxy for personality or lifestyle factors which are not adequately captured by the Big Five questions, and thus they do not serve as an adequate control.

    > glm(Harassment2=='Yes' ~ Gender + Income + AutismSpectrumTest + BigFiveE + BigFiveN + BigFiveA + BigFiveC + BigFiveO + Profession ,data=ssc,family='binomial') %>% Anova
    Analysis of Deviance Table (Type II tests)

    Response: Harassment2 == "Yes"
                       LR Chisq Df Pr(>Chisq)    
    Gender               195.87  5    < 2e-16 ***
    Income                 0.88  1   0.348216    
    AutismSpectrumTest    15.01  1   0.000107 ***
    BigFiveE              25.06  1   5.57e-07 ***
    BigFiveN               2.49  1   0.114481    
    BigFiveA               8.12  1   0.004366 ** 
    BigFiveC               3.30  1   0.069320 .  
    BigFiveO               7.04  1   0.007953 ** 
    Profession            76.46 22   6.14e-08 ***
    Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

  51. MB says:

    A sufficient explanation is that people who go into the legal field, the arts, the humanities, social sciences, healthcare, writing, etc. want power over other people (admit or deny it, it’s obviously true), while those who go into programming, STEM, mining, mountain climbing, sailing, or what have you are looking for power over things/nature — which may lead to power over people, but less directly (uncharitably, this could be viewed as passive aggression).
    Thus, harassers and abusers, who presumably derive some pleasure from their personal power over their victims, would naturally be more drawn to occupations in the first group.
    If this is a gross oversimplification, so be it, but it does explain the unusual groupings detected in this study.

  52. Deej says:

    My guess is it’s

    1. Nerdy men less likely to harass other people than more confident people, for reasons of niceness and shyness.

    2. Nerdy women less likely to be harassed. Not sure of reasons for this/less sure if true.

    3. Media stuff is a version of people being much more angry/disgusted/confident-in-speaking-out when nerdy men do harass.

    • SaiNushi says:

      “2. Nerdy women less likely to be harassed. Not sure of reasons for this/less sure if true.”

      caveat: nerdy women less likely to notice harassment until it gets really far (like, deliberately groping).

      source: self in middle/high school, until I found some friends who opened my eyes to just how much of what people were saying was meant sexually.

  53. Watchman says:

    I think there’s a case that the post and comment thread have been unduly generous to the media here. The media obsession with sexual harassment in certain industries cannot be understood as a rational analysis akin to Scott’s here. For a start most journalists are not that numerate (ever see actual probability discussed by a non-expert commentator?) and tend to repeat figures presented to them without anything more than perfunctory analysis. So even conscientious journalists will accurately report dubious claims, especially where they reinforce prior assumptions.

    Furthermore we cannot give the media the benefit of assumed good will here. The media is not a monolithic entity but its members are mostly in a echo chamber where they not only read and listen to each other’s output but also regularly physically interact, for example columnists appearing as guests on news channels. The members of the are also likely to occupy linked social circles due to the fact that national media tend to be focused in one place, and most people’s social lives are to some extent influenced by work. So the media as a collective is quite capable of producing a signal (whether right or wrong) and then reinforcing it internally through repetition, and then accept it as orthodoxy since this will be the message they consume and hear repeated (and remember journalists are meant to listen to what they are being told…) so they think it’s normal. Now media has its outliers who push different interpretations but the problem here is this is normalised on a sort of “Well Fox would say that wouldn’t they” way: there is an expectation certain outlets and columnists will or may offer a contrary opinion but that is the role within the media that outlet or columnist has. That is to say the putting forward of contrary arguments is regarded as a norm within the industry and there seems to be a view these are not serious challenges to an orthodoxy but rather a reinforcement of that orthodoxy by the predictable agents (in the sense of a person or organisation that can choose its own actions) within the media who are putting out the standard polarised opposite views that make it seem this is normal.

    How then does this relate to tech? Well in the US tech is not in the same circles as media. It is centred in northern California not New York or Washington. It is tied in with industry not politics. It promotes individualism rather than collectivism (media consumption is often tied to group identity). And therefore the media have a narrative that it is a threat, mainly to them but also to the jobs (they’re correct if they ignore history…) and safety of their consumers. Hence the fact self-driven car accidents and Amazon couriers’ conditions are in the media a lot; the incredible technological advances that allow cars to steer themselves and the happiness of Amazon couriers are mentioned a hell of a lot less. Note a similar pattern seems visible in India where the Delhi-centric media are negative about the Bangalore focused tech industry, picking up on stories of sexual misconduct therein whilst ignoring much worse elsewhere until an outcry forces them to acknowledge it. Although there is plentiful cooperation and cross-pollination between industries that does not stop a central media narrative being produced and reinforced about tech – this is a social construct not the work of individuals.

    As a proof of sorts of this I offer the UK where if either tech or media can be given a focus then it seems to be central and easy London. This means the sectors coexist with each other and the business, law and politics sectors also centred in London. Interestingly even the ever-vigilant for sexual harassment Guardian, which is extremely London based, does not have a narrative that the UK tech industry is a hotbed of inappropriate sexual behaviour. It does run this line about the US tech industry (centred a long way from their New York headquarters) though on occasion. Where the tech industry can be the media’s ‘other’ then it will be focused on and perhaps relatively uncommon behaviours will be portrayed as norms. Where it is part of the same immediate society then the media is less likely to be in a position to create narratives about at least their local industry as it is not the ‘other’.

    So it is likely that the media has created a narrative about tech, which is mainly applied where tech is not focused in a location proximate to that of the media. And this narrative probably has one further effect, presumably not deliberate, that in the US at least public opinion is going to be that tech has a much higher rate of sexual harassment than media (outside Hollywood). Every indicator provided here and every possible rationale I have read in this thread suggests that media has a profile (female dominated, customer focused, supply of workers generally outstripping demand, prevalence of narcissists, more socially adept people) that suggests sexual harassment should be much more common than in tech. The prevalent assumption amongst feminist columnists that sexual harassment is a common (female) experience might reflect personal experience rather than a political filter on reality due to their industry. Yet this is not reported, for one of two broad reasons: either my hypothesis of high rates of sexual misconduct in the media is way off the mark or the media is focusing on other sectors whilst ignoring their own problems. The second is at least consistent with the representation of tech as a hotbed of abusers, since it is about perception, othering and protection of the in group. And the failure to address these problems, which should be pretty well instantly recognisable to any humanities or social sciences graduate, is why I feel the media is getting off lightly here.

  54. synaxarion says:

    Building off what some other people have said, if people in certain fields (hypothetically the STEM/nerdy/male ones) are more likely to be introverted, it’s also possible that taking a ratio is biased because people in the interpersonal/creative/female fields are more likely to participate in more social events and therefore are more likely to experience out-of-work harassment? Of course, that would explain the trend opposite of what we’re seeing, but just wanted to explain why I found the use of a ratio to eliminate differences in standards for harassment questionable.

  55. AnthonyPearce says:

    please tell me, how can i apply for become a author account

  56. eelcohoogendoorn says:

    A second reason the low harassment rates in STEM were surprising is conventional wisdom claiming the opposite. In media coverage, STEM fields are portrayed as full of “techbros”, and uniquely unwelcoming to women. How do we square this with the data?

    I am aware this narrative goes around; but it has about as much to do with my perceived reality as say, flat earth creationism. And I say that as a person who is quite well grounded in both nerdy and not-nerdy spaces.

    Compared to any other setting that humans inhabit, other than perhaps funerals or somesuch, nerd-dominated places are overwhelmingly sexless. Yes, I have seen awkward guys pounce on the single female wandering into their territory; aiming to impress her with a mansplaining contest about boardgames or math; and I have cringed.

    But if you want to get your ass grabbed or a tongue shoved down your throat or not be treated with reverence for your mere existence; well then literally any place else would be a better bet.

    Which isn’t to say bad things do not happen in nerdy places. But there is a reason ‘mad men’ was not set in a software company. Unlike Scott Aaronson I do not self-identify with the rejected nerd stereotype, but I think he explained it rather well. Nerds are just easy to bash, thats all. The fact that no women joined their clubs is a nice stick to bash them with. And the people who said it was their own fault are not even wrong; but excessive sexual aggression was not their mistake.