The Schelling Point for being on the Discord server (see sidebar) is Wednesdays at 10 PM EST

SSC Survey 2017 Results

[None of these calculations were really double-checked and some of them might be wrong. If you’re really interested in accuracy, download the raw data at the bottom and see for yourself.]


Back in January I asked you to take the SSC survey. Thanks to the 5,500 (!) people who sent in responses. Below are some summaries of answers, alongside paraphrases of the relevant questions to jog your memory. If you want to see the actual questions (some of which are long) you can read them on the survey here. Please don’t try to take the survey; your answers will be ignored.

“What country are you from?”

“What is your biological sex?”
“What is your gender?”
“What is your sexual orientation?”

“Do you identify as asexual?”
“Do you prefer monogamous or polyamorous relationships?”
“Are you currently in a relationship?”

“As what race do you most identify?”

“What is the highest level of education you completed?”

“What is your current employment status?”
“In what field do you work or study?”

“How would you describe your religious views?”
“If you believe in a religion, which religion is it?”

“Whether or not you believe it, what is the religious background of your family?”
“What are your ethical views?”

“How long have you been reading Slate Star Codex?”
“How many of the 750 or so Slate Star Codex posts have you read?”

“How often do you comment on Slate Star Codex?”
“How were you referred here?”

“Do you read the SSC subreddit?”
“Do you read Unsong?”
“Have you ever clicked on the sidebar ads?”

“How much do you like SSC?”
“How often do you agree with the object-level points SSC makes?”

“Do you want SSC to focus more or less on the following topics?”

“Do you identify as a member of the following communities?”

“What is your opinion of the rationalist commmunity?”

“Where do you fall on a classic political spectrum?”
“How interested are you in politics?”

“Which of these political philosophies do you most identify with?”
“If you are an American, what party are you registered with?”

“What is your position on the following issues?”

“Have you taken the Giving What We Can pledge?”

“How often do you read the SSC comments?”
“What is your opinion of the SSC comment section?”

“Do you find an ideological bias among SSC commenters?”
“Are you bothered by a bias in SSC comments?”

“What is your opinion on the level of comment moderation?”
“What is your opinion on the level of identity politics discussion in the comments?”

“What is your opinion on the recent policy of requiring all commenters to register accounts?”
“What would you think of requiring new commenters to answer a knowledge question before being allowed to register?”

“Are you bothered by scratchy tags on your clothing?”
“How do you interpret the sentence ‘I have read this book and much like it’?”
“What direction do you see in the spinning dancer illusion?”

“Which do you think is more important when trying to learn a new skill: hard work or talent?”
“Do you find it hard to follow conversations in noisy areas?”
“How many of the three duplications of the word ‘the’ (ie ‘the the’) did you notice in this survey?”

“How do you interpret the Einstein mask illusion?
“Do you think of yourself as more detail-oriented or more big-picture?”

“How happy do you generally feel?”

“Do you have any of the following psychiatric diagnoses?”

“Have you been on SSRIs?”
“At what age did you start SSRIs?”
“Do you currently use SSRIs?”

“Do you think of yourself as more ‘growth mindset’ or ‘fixed mindset’?”
“Do you lift weights?”
“Regarding sleep, are you an ‘early bird’ or a ‘night owl’?”

“How often do you remember your dreams?”
“If you are transgender, when did you realize this?”
“Does your internal thought process feel verbal or nonverbal?”

“How concerned are you about bioterrorism as a threat to humanity?”
“How concerned are you about superintelligent AI as a threat to humanity?”
“Are you able to clearly visualize images in your mind through imagination?”

“Overall, how satisfied are you with life?”
“How much do you enjoy puns?”
“Do you think other people are basically trustworthy?”
“Do you think you are ‘a typical SSC reader’?”


Some means:

Age: 30.6

IQ: 138.5
SAT out of 1600: 1471.9
SAT out of 2400: 2218.68

(Congratulations to the person who got a 1650 on their SAT/1600, and the person who got a 2450 on their SAT/2400. You clearly have bright futures ahead of you.)

Income (mean): $96,443.5
Income (median): $57,000

The highest observed incomes were in the $10,000,000 range; I know some big venture capitalists read this so I didn’t delete them as obvious trolls. Removing everyone who makes over $1 million, mean income goes down to $79,000. But there were also people who put down incomes of 0 because they were students, unemployed, or homemakers. When these people are also taken out, the mean of the remaining 2,700 people goes back to $98,000, and the median to $75,000.

Below are some Likert scales. Note that the midpoint is not what you think. On a 1-10 scale, the midpoint is 5.5, not 5. On a 1-5 scale, the midpoint is 3, not 2.5.

Political spectrum: 4.55 / 10 (higher = further right)
Political interest: 3.75 (out of 5)

Global Warming: 2.0 / 5 (higher = more skeptical)
Immigration: 3.5 / 5 (higher = fewer restrictions)
Minimum Wage: 2.9 / 5 (higher = higher minimum wage)
Gay Marriage: 4.5 / 5 (higher = should be legally recognized)
Feminism: 3.3 / 5 (higher = more favorable)
Human Biodiversity: 2.7 / 5 (higher = more favorable)
Donald Trump: 1.7 / 5 (higher = more favorable)

SSC Science Articles: 3.5 / 5 (higher = want increased focus)
SSC Politics Articles: 3.5 / 5 (higher = want increased focus)
SSC Book Reviews: 3.2 / 5 (higher = want increased focus)
SSC Rationality Articles: 3.5 / 5 (higher = want increased focus)
SSC Silly Articles: 3.2 /5 (higher = want increased focus)

Happiness: 6.0 / 10 (higher = happier)
Life satisfaction: 6.3/10 (higher = more satisfied)

Charity: $3271.6 / year
Percent given to charity: 2.8%

Among people who were employed (not students, unemployed, or homemakers), the numbers were surprisingly similar. Median charitable donations were a very disappointing $300/person, and median percent charity was 0.7%. But 358 people (out of 3500 for whom I had good data) gave 10% or more to charity, and 25 people gave 25% or more to charity. 13 people gave more than $100,000 to charity per year.


Some more complicated things I was looking for. Everything in italics was “pre-registered”, ie guessed before looking at data and describing a future data analysis plan. I use { and } in place of the normal less-than and greater-than signs because I can’t be bothered to figure out how to not make them confuse the HTML.

Hypothesis 1: There will be a general ‘ability to tolerate ambiguity’ which links being able to see the spinning dancer go either direction, being able to see the face mask either direction, and being simultaneously aware of both meanings of the sentence about reading books. In other words, all three of these areas will correlate with each other. They might also correlate with liking puns.

Results: I ran correlations between SpinningDancer, FaceMask, ReadThisBook, and Puns. Since there were four variables, it came out to six (n * n-1 / 2) different correlations. Of these, three were positive and significant:

FaceMask x Puns: r = 0.03, p = 0.03
FaceMask x ReadThisBook: r = 0.08, p { 0.001
Puns x SpinningDancer: r = 0.05, p { 0.001

And three weren’t:

FaceMask x SpinningDancer: r = 0.01, p = 0.55
Puns x ReadThisBook: r = 0.01, p = 0.56
ReadThisBook x SpinningDancer: r = 0.02, p = 0.10

I don’t see any patterns in which ones worked or didn’t; in particular, I would have expected ReadThisBook to correlate with liking puns, since it was the same kind of verbal/linguistic ambiguity. The significance here was so good that I’m reluctant to just throw out this whole idea, but the effect size was pretty small and I’m honestly not sure what to do with this.

Hypothesis 2: There will be a general ‘tendency towards bottom-up processing’ which links being detail-oriented, noticing the duplicated “thes”, getting annoyed with tags on clothing, and not being able to tolerate noisy conversations. In other words, all four of these areas will correlate with each other.

Results: Again, the four variables made six different correlations. All six were positive and significant.

ClothingTags x NoisyConversation: r = -0.1, p { 0.001
ClothingTags x DuplicateThes: r = 0.08, p { 0.001
ClothingTags x DetailOriented: r = -0.03, p = 0.02
NoisyConversations x DuplicateThes: r = 0.05, p = 0.001
NoisyConversations x DetailOriented: r = 0.1, p { 0.001
DuplicateThes x DetailOriented: -0.03, p = 0.02

Once again there are some very impressive p-values but all the correlations are very weak. At this point I started wondering whether maybe my methodology was broken. I tried correlating these against a panel of a dozen political topics that I wouldn’t expect them to correlate with. ClothingTags correlated with two political topics, usually at around p = 0.01. NoisyConversation correlated with six political topics, again at similar levels, and so on for the rest. I can sort of make up stories about why this might happen (people who didn’t like noisy conversations didn’t like Donald Trump, and he is a loud kind of guy) but I’m not going to go that direction. The results weren’t just a general factor of people who like putting large numbers into surveys, because the correlations were just as frequently negative as positive. So I’m not sure.

Overall the correlations between real interesting psychological factors that seemed like they should be correlated were larger and more frequent than the correlations with unrelated political topics, but they were all so small, and everything is so noisy, that I’m not going to count this as a meaningful victory. The only ones that approached being interesting were the correlations between clothing tags, noisy conversations, and detailed-orientedness, which everyone already knows are all kind of autistic-y traits.

Hypothesis 3: People who used SSRIs during childhood (or maybe during puberty? or both?) are more likely to be asexual. In other words, asexuality rates for these groups will be higher than those of people who used SSRIs during adulthood and people who never used SSRIs. If sample size permits, I will try to exclude current users of SSRIs from all groups to rule out them being “asexual” because current SSRI use is ruining their libido.


(by age at which they started SSRIs, and asexuality rate)

Younger than 10: 7%
10 – 15: 9%
15 – 20: 9%
20 – 25: 7%
Older than 25: 4%
Never on SSRIs: 6%

Okay, I didn’t realize how many different categories I had, so my fancy preregistration is going to have to go. I am going to, in an ignoble unpreregistered way, combine everyone who started an SSRI at 20 or younger, with everyone who started it older than 20 or not at all. An independent samples t-test comparing mean asexuality between those two groups finds…not much.

More specifically: there were 4490 people who hadn’t taken SSRIs while young, and 435 people who had. The respective asexuality rates in the two groups were 6% and 9%. The difference was about p = 0.1. Adjusting out people currently on SSRIs did nothing whatsoever.

I conclude that my hypothesis was wrong, and taking SSRIs during puberty is not a risk factor for asexuality. Note that taking SSRIs not during puberty isn’t a risk factor for this either, and there was minimal difference in asexuality rate between people who had ever taken SSRIs and those who had not. Either permanent loss of sexuality from SSRIs is so vanishingly rare that a survey of 5500 people cannot pick up on it, or it is impossible to confuse with “asexuality” as an orientation and I should have asked the question some other way.


Okay, so much for fancy responsible hypothesis preregistration. Everything following is whatever interesting came out of a giant fishing expedition. Because of the previously noted tendency for things to be super-highly-significant in this dataset even when they’re sketchy, I’m including only things with a decent effect size (r } 0.1). Everything in this category automatically has p { 0.001. I’m not including results I think are obvious.

The more trustworthy you think other people are, the higher your life satisfaction (r = 0.19)
The more visual your imagination, the more likely you are to remember dreams: (r = 0.16)
The more trustworthy you think other people are, the less likely you are to like Donald Trump (r = 0.14)
The more trustworthy you think other people are, the more likely you are to support more open immigration (r = 0.19)

The more you describe yourself as having a growth mindset, the higher your life satisfaction (r = 0.18). Bizarrely, describing yourself as thinking hard work matters more than talent doesn’t predict life satisfaction at all. It’s just the words “growth mindset”

If you don’t like noise or noisy conversations, you are less likely to be a generally happy person (r = -0.13, -0.17)
Night owls are less happy than early birds (r = -0.13)

Weightlifting was positively linked to life satisfaction (r = 0.1), negatively to asexuality (r = 0.1) and to various right-wing beliefs at around r = 0.1. It was negatively linked to being a night-owl (r = -0.1)

The more liberal you were, the more likely you were to think SSC comments had a conservative bias, and vice versa.


You might have noticed some very positive feelings about the comment section. The average person rated the comments 3.5/5 (median: 4/5). On a scale from 1 to 5, where 1 was “biased conservative” and 5 was “biased liberal”, the average score (both mean and median) was 3, ie exactly in the middle. From a wisdom of crowds perspective, you rate the SSC comment section as literally the least biased it is logically possible to be.

79.9% of commenters said they were “not bothered” by any bias in the comment section. The 20.1% of people who were bothered by comment section bias (n = 908) were very slightly more liberal than SSC as a whole (4.05/10 compared to 4.55/10, where higher is more conservative). This group rated the comment section as having a very slight conservative bias (2.6/5, where lower is more conservative) but there was a high standard deviation. In other words, this group contained both people annoyed that the comments were too conservative, and people annoyed that the comments were too liberal, with a very slight preponderance of the former.

So this is people’s perception. Can we measure reality? We know that SSC as a whole is very slightly liberal, but what about frequent commenters? Here are the numbers, again on a political spectrum where 1 is maximally liberal and 10 maximally conservative:

1. Lurkers who never comment: 4.5
2. People who comment less than once a month: 4.7
3. People who comment at least once a month: 5.1
4. People who comment at least once a week: 5.2
5. People who comment many times a week: 6.3

So there is a really interesting tendency for conservatives to comment more often than liberals (maybe because they have more to disagree with?). But numbers in the last three groups were very small: out of the 5335 people for whom I had data, only 54 commented once a week, and only 45 commented many times a week. So they may not be able to bring the average up very much. Since tiers 1 through 4 were liberal (REMEMBER THE MIDPOINT IS 5.5) and only tier 5 was conservative, there’s probably an extremely slight preponderance of liberal comments on the whole.

I checked opinions of the 1100 people who comment once a month or more, and they were broadly similar to those of the general population.

I had a question in which I asked people to guess what percent of survey-takers would be right-wing (ie greater than 5.5 on the political spectrum question). The true answer is that of 5335 respondents, 1703 (31.9%) were 6 or above. 485 (9.1%) were at exactly 5 (technically to the left of center but not obviously so from the scale). 3,144 (58.9%) were unambiguously left of center.

So the correct answer to the estimation question was 31.9%. The average person guessed 34.9% (mean) or 35% (median), so you were pretty on-the-mark.

(In case you’re wondering, I was expecting to find that most people were lefties but thought everyone else was on the right. The first part was true, the second part not so much).

There was a very slight effect where, the longer someone has been reading the blog, the more conservative they are likely to be. Someone reading SSC more than two years is 4.7 on the scale; someone less than a year, about 4.3. My guess is that I got a few extra conservatives back when I wrote about the far right more.


Just for fun, I wanted to see how this community differed from the rest of the population, so I got a hundred Mechanical Turkers from the US and UK to fill out a slightly-edited version of the same survey. The sample size isn’t big enough to say anything for sure, and I’m not going to bother figuring out how to do t-tests across two different datasets, but here are some things I noticed.

MTurkers were 72% white, 6% black, 6% Hispanic, 10% Asian, and 3% other. They were 68% male and 32% female. They were much closer to representative than SSCers.

2% were transgender, 11% bisexual, 5% gay, and 8% asexual, about the same percent as SSC readers. This surprised me. We may be unusually LBGTQetc for the US population, but not necessarily for the Internet population.

Their average IQ was 111, compared to our 138. Their average SAT score out of 1600 was 1272, compared to our 1472.

Turks were more likely to have schizophrenia than SSCers, though no more likely to have family members. Of note, 1% of Turks (1 person) had a formal schizophrenia diagnosis, and 7% (6 people) thought they might have schizophrenia. An 8% schizophrenia rate in a population is unheard of. I don’t know if MTurkers are disproprotionately likely to have schizophrenia, or if I just got a weird sample.

1% of Turks were formally bipolar and another 6% suspected bipolar. Compare to a more normal 2% and 2% among SSCers. About half of the Turks with suspected bipolar were the same ones with suspected schizophrenia.

0% Turks were formally autistic and another 4% were suspected autistic, compared to 4% and 12% of SSCers. We are apparently like 5x as autistic as normal. Who ever would have guessed?

10% of Turks were formally depressed and another 20% suspected, compared to 18% and 16% for SSCers. Kind of ambiguous between us being more depressed vs. better at getting diagnosed.

17% of Turks were formally anxious and another 17% suspected, compared to 12% and 16% of SSCers. There is a group we are less anxious than!

4% of Turks were formally OCD and another 7% suspected, compared to 2% and 6% of SSCers.

2% of Turks were formally eating disordered and another 3% suspected, compared to 1% and 3% of SSCers.

7% of Turks were formally alcoholic, and 3% suspected, compared to 1% and 4% of SSCers. The Turk number, which seems very high, actually fits better with epidemiological estimates of prevalence (though I don’t know about formal diagnosis). We are fantastically non-alcoholic. Oddly, this does not seem genetic – we have the same number of alcoholic family members as the Turks do.

5% of Turks are formally drug addicted and 2% suspected, compared to 0.2% (!) and 2% of SSCers. This one might be genetic: 25% of Turks have addicts in their families compared to 14% of SSCers.

28% of Turks are atheist+nonspiritual, 11% atheist+spiritual, 25% agnostic, and only 30% theist (committed or lukewarm). These numbers are way more atheist than the general population, but way less atheist than SSC.

Turks were 4.5 on the political spectrum, indistinguishable from SSCers. On specific issues, they were a little more restrictionist on immigration, a little more pro-Trump, a little more feminist, and a little less pro-gay. The only place there was a large difference ( } 1 point) was on the minimum wage, which they almost universally supported.

Contra my predictions, SSCers were actually less annoyed by clothing tags than Turkers (2.4 vs. 2.9) and no more annoyed by noise (3.0 vs. 3.0). We were about equally detail-oriented, and worse at following noisy conversations (2.4 vs. 3.1). We were a little more likely to believe in talent instead of growth mindset (2.4 vs. 2.7) and equally likely to be night owls.

We were less likely to remember dreams (2.7 vs. 3.0), less likely to have strong visual imaginations (2.4 vs. 2.6) and more likely to think verbally (2.2 vs. 2.6). We believed people were a little more trustworthy (2.5 vs. 2.9).

We were slightly less satisfied with our lives (6.3 vs. 6.4) and vastly less happy (6.0 vs. 6.9), even though we were earning an average of $97,000 and they were working on Mechanical Turk. This was probably the most striking result.


I will be doing something with the meetup information shortly. Otherwise, this is all the data I have the energy to extract out of this right now. But there is a lot of stuff here. 5048 people kindly allowed me to share their data publicly, so I encourage anybody interested to play around with this and report what you find.

(Survey data as .XLS file)

(Survey data as .CSV file)

(MTurk data as .XLSX file)

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

Leave a Reply

684 Responses to SSC Survey 2017 Results

  1. wfro says:

    Not quite what I expected, on several fronts. I’m particularly surprised by the middling results on AI risk and detail-orientedness. Although a lot of these sorts of things are inherently answered contextually and our contexts are likely not representative of society.

  2. Null Hypothesis says:

    I don’t know what conclusions if any can really be drawn from it, but the ‘comment bias’ section being more or less perfectly symmetrical and with hardly any 1’s or 5’s makes me smile.

    The constant presence of bio-modal distributions gives me a similar feeling. And bi-modal in the true two-hump fashion, rather than just railing to either side and running away from the center.

    Also, I’m skeptical about the ‘138’ IQ being the average. That supposedly puts the average SSC reader above the 99th percentile in intelligence (~99.4). From reading the comments alone, I could almost believe that. But the statistics part of me says there simply aren’t enough geniuses of 160+ reading this blog to counterbalance how many 115 to 130 ‘neanderthals’ are surely present.

    However, The average SAT score percentiles are:
    SAT out of 1600: 1471.9 – 98.5 percentile
    SAT out of 2400: 2218.68 – 98.7 percentile

    Since people directly remember those, I’m more willing to believe those numbers aren’t inflated. Which are pretty damn close. That missing 1 percentile is significant, but not very. IQ with N(100,15) puts 98.5th percentile at IQ = 133. So maybe people are being honest after all. Especially since once your scores get that high on the SAT the last 100-200 points are more indicative of not making stupid-mistakes rather than being more intelligent.

    • Scott Alexander says:

      We have this argument every year. Points in favor include:

      1. Survey IQs mostly match survey SATs from IQ/SAT conversion tables.
      2. One year we asked ACT and that matched too.
      3. One time we made everybody describe which IQ test they took and in what circumstance, and the subset who took provably legit IQ tests given by provably legit psychologists weren’t any different from the rest.

      I don’t doubt that a lot of the overly high numbers are people who took a test as kids which wasn’t properly normed for kids their age or something.

      • tmk says:

        I suspect people with low or middling scores just don’t answer those questions, because it would make them feel bad momentarily.

        • Deiseach says:

          I really think the IQ question is biased due to the US representation on here. Outside the US, I don’t think people are tested as frequently or regularly or that it’s particularly a thing (unless you want to join Mensa). I will defer to my fellow non-Americans on this, but I honestly think even Really Smart Big Brain Trinity College Professors in Ireland wouldn’t have a clue about their IQ score.

          We are fantastically non-alcoholic. Oddly, this does not seem genetic – we have the same number of alcoholic family members as the Turks do.

          I wonder are those of us who use alcohol poorly more likely to be dipsomaniacs rather than alcoholics as such – that is, we indulge in bouts of drinking, often or mainly to control depression or whatever, but come off it when the effects get too bad and don’t touch alcohol afterwards, or don’t abuse it in the same manner (i.e. we can socially drink, we can drink infrequently or not at all, our danger times are drinking alone for self-medication but we can stop those)?

          I do find the results really interesting and actually broadly in line with what I was expecting, though I found the political orientation question very hard to answer – I would identify as conservative, but I’ll die roaring before I say “Yes, just like the UK Tories!” If I had to vote for a UK party it would be Labour but I’m not as on-board the social permissiveness train as they would be in parts.

          I will take the flattery of “Their average IQ was 111, compared to our 138” but I think I’m more on the Mechanical Turk side of the fence with this one.

          I also am immensely tickled by the “The highest observed incomes were in the $10,000,000 range; I know some big venture capitalists read this so I didn’t delete them as obvious trolls” (and if any of you were pulling his leg in this survey, tsk-tsk!) So we range from “poor as church mice” to “Scrooge McDuck” 😀

          I really hope a lot more of the lurkers are encouraged by this to come out and comment.

          This survey was enjoyable and I hope somebody does crunch the data to find out more One Weird Trick about SSC readers. Beannachtaí Lá Fhéile Pádraig daoibh go léir!

          • Doom Squid says:

            Lurker Coming in to comment. I discovered this after the survey was initially posted, but the results are really quite interesting. As someone who loves data of any type, whether useful or superfluous, this was a lot of fun to read through.

          • wintermute92 says:

            I wonder are those of us who use alcohol poorly more likely to be dipsomaniacs rather than alcoholics as such – that is, we indulge in bouts of drinking, often or mainly to control depression or whatever, but come off it when the effects get too bad and don’t touch alcohol afterwards, or don’t abuse it in the same manner (i.e. we can socially drink, we can drink infrequently or not at all, our danger times are drinking alone for self-medication but we can stop those)?

            This is definitely something I wonder about. The mental illness stats on depression and anxiety are pretty high, and SSC readers apparently report much lower happiness than MTurk users. All of which suggests the possibility of serious alcohol abuse without much correlation to addiction.

            More broadly, I’m frustrated that society talks solely about alcoholism (which ought to specifically be about chronic overuse with addiction) when it seems to mean a tripartite issue of overuse (binge drinking), unhealthy use (drinking to cope), and addiction (drinking compulsively). There’s even some data implying that this is why it’s so hard to assess the results of alcohol treatment programs – effective treatment of compulsive drinking is totally different than effective treatment of drinking to handle depression.

          • thedufer says:

            This assertion that Americans are significantly more likely to have taken an IQ test than the rest of the (developed?) world surprised me. My perspective is that the people I know have taken IQ tests are just people who were in for batteries of psychological tests at a young age for various reasons (I was generally a pain in the ass in early elementary school, for example). An IQ test seems pretty reasonable in that situation. Is this not, in fact, the norm in the US? Are people less likely to put their children through psychological tests in other countries? I’m very curious where my (largely unfounded) assumptions have gone wrong.

        • Evan says:

          I didn’t report my IQ because I don’t know my IQ. The only friends I knew who had their IQs tested were as children, and they were gifted, learning disabled, had ADHD, etc. I’ve always assumed there’s some standard battery of psychometric tests students with learning disabilities or whatnot take. Anyway, it was the vast minority of people I know who’ve had their IQ taken. I live in Canada.

      • Jules says:

        There is also the obvious bias that people who have higher IQs are more likely to get tested, and answer the question.

        • random832 says:

          If IQ tests are supposed to be periodically normalized to the distribution of IQ test results, shouldn’t that result in test results being deflated?

          Or do you think they’re normalized based on the mode rather than the median (the mode and median are the same in real normal distributions), to reflect a presumed normal distribution whose left tail is depressed to this effect?

          • Brad says:

            I don’t think IQ tests re-norm like that, though the SATs do IIRC. For IQ tests there’s a norm study done with a population recruited for the study when the test form is first published and that’s basically it.

        • GCBill says:

          This is what I suspected too. However, people have tested for this on the Less Wrong survey and found…nothing.

        • wintermute92 says:

          How true is this? In the state where I grew up, the entire public school population took at least one IQ-proxy exam (the results actually were IQ scores, they just didn’t explicitly state that fact). Obviously there are biasing populations like “prospective Mensa candidates” but I’m curious whether that’s swamped by bulk testing.

          • peterispaikens says:

            It’s highly regional – the concept of “official” IQ tests seems to be prevalent in USA (as indicated by “the state where I grew up”) but worldwide in many places virtually no people would take an IQ test, it would be limited to special situations e.g niche organizations like Mensa, diagnosis of mental or development problems, etc.

            For survey takers outside USA, I’d bet that the “most recent PROFESSIONAL, SCIENTIFIC IQ test” (assuming that people answer honestly) disqualifies *all* “normal” people (including e.g. the median university professor, who is also likely to not have ever been tested) and leaves only the outliers. Also, correlation with SAT wouldn’t reveal any issues because survey takers outside USA won’t also have a SAT score.

          • Deiseach says:

            I have to agree with peterspaikens; mostly over here any IQ testing (professionally administered) is only in cases of suspected or diagnosed developmental difficulties. For everyone else, as long as you’re not a complete dunce in school, it’s assumed you’re of “average” or “normal” intelligence, with the very bright being recognised as smart, but not that much more likely to be tested (unless they’re baby geniuses and so outstanding that they need to be tested to be properly accommodated).

            I have no idea how you’d correlate e.g “points for university as converted from Leaving Certificate grades” to SAT; what does “I got 625 points which is the maximum for the Leaving Cert” correspond to in the SAT?

      • Rachael says:

        Whereas I’m surprised it’s not higher.
        Mine is 177, according to an official invigilated Mensa test at age 15 (which I think is old enough not to have large errors in overcompensating for age), but I think my intelligence is only about average among SSC commenters.

        • Null Hypothesis says:

          IQ of 177 would place you at 99.9999846th percentile.

          Or, in a group of 1,000,000,000 you’re likely smarter than all but 16 of them.

          Internet-browsing world topping out at 3 Billion, and total SSC commenting population being greater than 80, suggests that’s unlikely.

          Assuming IQ tests are even valid at 5stds and beyond, you should be hard-pressed to find yourself of average intelligence in any group.

          • Rachael says:

            “Assuming IQ tests are even valid at 5stds and beyond”
            I think this is the relevant bit. I don’t think they correlate very well with actual intelligence above about 140 or so.

          • kriztw says:

            Also, historically the standard deviation wasn’t always 15 points. A standard deviation of 20-30 would make a score of 177 totally reasonable. From a quick search it looks like most IQ tests stop at around 160 because the uncertainty gets too large, so I’m curious which test was taken in this case.

          • wintermute92 says:

            I would toss out the percentile numbers altogether. General population IQ tests start to break down around >150 – the scores lose test/retest reliability. Specialized high-IQ tests do somewhat better, but still struggle hugely because there just isn’t enough data to normalize properly. I don’t doubt the score or the intelligence, but north of 150 the correlation collapses too hard for distributions to work.

            There are some Mensa++ societies that require these stats, but they’re usually mocked for requiring test scores that are known to be useless.

          • Tekhno says:

            “Assuming IQ tests are even valid at 5stds and beyond”
            I think this is the relevant bit. I don’t think they correlate very well with actual intelligence above about 140 or so.

            Isn’t this the same for the lower end (for more straightforward reasons)? I don’t know enough about IQ tests to say, but I’d intuitively expect really low IQ scores to be meaningless. For example, if you find more and more impaired people, eventually you get people who can’t do the test at all, so there shouldn’t be a smooth regression all the way to score 0.

          • Gazeboist says:

            Yes, but what does that have to do with anything?

          • Eponymous says:

            Sigh. Higher IQs correlate with outcomes as high as we can measure. The idea that they don’t is a myth. See Benbow/Lubinski and SMPY, among others.

            We can easily measure very low IQs. We just have to use different tests. For starters, we can use tests for kids, or developmental indicators. Just think of a child psychologist who has a checklist of what a normally-developing 8-year old should be able to do. Run an adult through the same checklist and see how they do.

          • Brad says:


            Sigh. Higher IQs correlate with outcomes as high as we can measure.

            And just how high can we measure, using standard statistics rather than ad hoc justifications for wild extrapolations?

        • rlms says:

          I think that it is easier to get a ridiculously high IQ the younger you are, because it is measured relative to age. So a 10 year old with the IQ-ability of a 140 IQ 25 year old will have much higher IQ than the 25 year old.

          • Rachael says:

            Indeed, that’s what Scott meant about kids taking tests improperly normed for their age, and what I meant about errors from overcompensating for age. I think 16 is counted as adult though, so 15 shouldn’t be adjusted very much.

          • Deiseach says:

            Giving IQ tests to children should always be counterbalanced by one when they’re older (mid-twenties at a minimum).

            A smart twelve year old may be measured as “we would expect this level of problem solving ability/factual knowledge/vocabulary from a thirty year old adult so their IQ is Huge Number” but what do you test that smart twelve year old against when they’re thirty? In some cases they may still hold up (Mozart was always going to be a musical prodigy whether at six or thirty) but I imagine a lot of Huge Number child scores would get reduced down to Reasonably Large Number adult scores.

          • Trofim_Lysenko says:

            Unfortunately, that doesn’t often happen. I was IQ-tested when I was 11-12 on two different scales and there was a 3-4 SD variation between the two scores (the lower was still 135+), but that was under the auspices of a child psychologist who specialized in precocious/troublesome youth.

            I don’t think most people care enough to shell out a couple hundred bucks to get formally tested as a 25-30 year old adult.

          • Deiseach says:

            Possibly the solution to this, so, is to put in on the next IQ survey question: What age were you when you took the test and what was your score? (If you have taken more than one test, give the results and age of the most recent).

            That at least would let us see for those who have taken IQ tests how many of them were “140 on test but that was at age eight”. How you decide to sort out that information is then up to interpretation.

          • MSwaffer says:

            Several things should be noted here.

            First, while general intelligence is a relatively stable construct in adulthood, it is a notoriously unstable measure in children (See Larson, Hartmann and Nyborg, 2008 but also Dreary, 2000) We know children develop at different rates for different reasons which means normed scores of standardized tests aren’t very reliable over time (See Samaroff Between the ages of 18 and 60 IQ remains relatively stable but prior to 18 and after 60 there are large variations.

            Thus, if many of the posters reporting IQ scores were reporting from tests taken as children, it’s very likely these are unreliable numbers and subject of course to self-reporting bias.

            Second, intelligence is not necessarily a single construct. There are numerous IQ tests that purport to measure g, however there is significant disagreement about whether or not general intelligence is meaningful. 100 years ago, this was all the rage but more recently we have begun to realize that intelligence is a complex construct. While Gardner’s multiple intelligences theory is lacking in supporting evidence, the idea of a single facet g has gone somewhat out of favor. A finer grained view of cognitive abilities is sometimes more useful. (See Schneider and Newman, 2015)

            Finally, there is a misconception that intelligence, particularly as measured in children, is important to their success. Psychologists discovered when analyzing Terman’s longitudinal study of gifted children that socio-economic status of parents does just as well predicting success as IQ of children. There is certainly some level of correlation between IQ and other societal measures of success but generally the correlation is not as strong as suggested by the popular press. So while it may be interesting that some children have very high IQ scores, it isn’t necessarily predictive. (See Strenze, 2007 but also Gottfredson, 1998.)

            For my money, I don’t think the IQ question on this survey has any bearing on reality. If the inference folks are hoping to take away is that they are reading the same things that really smart people on the internet read… that’s not really an inference that should be made from this data.

            (Full disclosure: I have never taken a standardized IQ test. )

            Deary, I. J., Whalley, L. J., Lemmon, H., Crawford, J. R., & Starr, J. M. (2000). The stability of individual differences in mental ability from childhood to old age: follow-up of the 1932 Scottish Mental Survey. Intelligence, 28(1), 49-55.

            Gottfredson, L. S. (1998). The general intelligence factor.

            Larsen, L., Hartmann, P., & Nyborg, H. (2008). The stability of general intelligence from early adulthood to middle-age. Intelligence, 36(1), 29-34.

            Sameroff, A. J., Seifer, R., Baldwin, A., & Baldwin, C. (1993). Stability of intelligence from preschool to adolescence: The influence of social and family risk factors. Child development, 64(1), 80-97.

            Schneider, W. J., & Newman, D. A. (2015). Intelligence is multidimensional: Theoretical review and implications of specific cognitive abilities. Human Resource Management Review, 25(1), 12-27.

            Strenze, T. (2007). Intelligence and socioeconomic success: A meta-analytic review of longitudinal research. Intelligence, 35(5), 401-426.

          • Cliff says:

            “First, while general intelligence is a relatively stable construct in adulthood, it is a notoriously unstable measure in children”

            Fortunately, we have SAT data from adulthood to verify the IQ scores.

            “Second, intelligence is not necessarily a single construct. There are numerous IQ tests that purport to measure g, however there is significant disagreement about whether or not general intelligence is meaningful. 100 years ago, this was all the rage but more recently we have begun to realize that intelligence is a complex construct. While Gardner’s multiple intelligences theory is lacking in supporting evidence, the idea of a single facet g has gone somewhat out of favor. A finer grained view of cognitive abilities is sometimes more useful. (See Schneider and Newman, 2015)”

            IQ certainly is the best measure of intelligence that we have. The evidence for g is fairly overwhelming, since all subtests correlate at very high levels with all other subtests.

            “Finally, there is a misconception that intelligence, particularly as measured in children, is important to their success. Psychologists discovered when analyzing Terman’s longitudinal study of gifted children that socio-economic status of parents does just as well predicting success as IQ of children. There is certainly some level of correlation between IQ and other societal measures of success but generally the correlation is not as strong as suggested by the popular press. So while it may be interesting that some children have very high IQ scores, it isn’t necessarily predictive.”

            IQ is the single most predictive piece of information that we have. It is the best predictor of on-the-job success, life outcomes, etc. This has been exhaustively studied, so I am not impressed with one study that believes socioeconomic status of parents is as important as IQ after controlling for IQ (I hope they did that at least).

          • wysinwygymmv says:

            MSwaffer’s claims seem pretty contrary to the conventional wisdom on IQ in these parts. Anyone care to challenge them or at least comment on what has changed?

          • Mark V Anderson says:

            MSwaffer’s claims seem pretty contrary to the conventional wisdom on IQ in these parts. Anyone care to challenge them or at least comment on what has changed?

            I note that MS is making claims about the validity of children’s IQ’s, not IQ’s in general. What she says (I wonder why I assume female?) is pretty consistent with what I’ve heard elsewhere (and anecdotally on this blog). There is certainly a lot of evidence that IQ is highly correlated with many signs of success in life, such as income and education, but she makes a good point that children’s IQ’s are not as consistent.

          • Eponymous says:

            Second, intelligence is not necessarily a single construct. There are numerous IQ tests that purport to measure g, however there is significant disagreement about whether or not general intelligence is meaningful. 100 years ago, this was all the rage but more recently we have begun to realize that intelligence is a complex construct. While Gardner’s multiple intelligences theory is lacking in supporting evidence, the idea of a single facet g has gone somewhat out of favor. A finer grained view of cognitive abilities is sometimes more useful. (See Schneider and Newman, 2015)

            Okay, so you cite one paper, published in HR management review, that cites the CHC hierarchical model.

            But what’s at the top of the CHC hierarchical model? That’s right, the g factor:

          • MSwaffer says:

            First, for those arguing that IQ is a good measure of general intelligence… I didn’t say otherwise. My argument was that g might not be a good predictor of other things. I did argue that IQ is an unstable measure in young adults which was countered with the fact that we have SAT’s to correlate. That might just mean that the same kids selected to take IQ tests are the kids who also select to take SAT’s… both of which are correlated with SES. In any case, this is a self-reported survey which means all of these measures are likely inflated. There’s probably a bit of a Lake Wobegon effect here.

            Second, several argued that IQ is the best predictive measure we have. That’s not supported by evidence. Angela Duckworth has been doing a tremendous amount of research in this field and she feels self-discipline is a better predictor. Her research is similar to others like Bandura and Zimmerman who argue that self-regulation and learning strategies are far better predictors of success than raw IQ. The literature simply doesn’t support the notion that IQ is the best predictor of success in all areas. So while I agree that this issue has been exhaustively studied, as someone who is in the field of educational psychology, I would disagree that the bulk of the evidence supports IQ as the best predictive measure. My perspective is admittedly biased by a focus on educational achievement but the entire field of education is predicated on influencing and predicting future behavior.

            To be clear, IQ correlates highly with a lot of areas and certainly is a decent predictor of many things and if your goal in this discussion is to feel good about the high IQ score you got as a teenager, by all means, you should feel good about that. But if your argument is that people with high IQ are successful, I would invite you to look up Chris Langan. Undoubtedly the most brilliant horse rancher in history. A single data point for sure, but a very clear example of how social capital plays a major role in success (and failure), regardless of your IQ.

            Lastly, the point here is not that IQ is useless, the point is that IQ alone is not helpful for most practical purposes. For example, at least one person completely missed the point of the Schneider and Newman article. This is an organizational behavior issue in which hiring someone just because they have a high IQ has proven to not always work well. What works better is understanding the specific abilities of the potential employee. These abilities may or may not be correlated with IQ but more importantly, IQ may mask high performance if it only exists in one or two areas.

            While I appreciate that taking a contrarian view of IQ in this particular crowd is probably frowned on, it seems that this would be the one community that would be able to set aside the self-congratulatory back-patting long enough to take a discriminating look at the construct.

            Duckworth, A. L., & Seligman, M. E. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological science, 16(12), 939-944.

            Zimmerman, B. J. (2008). Investigating self-regulation and motivation: Historical background, methodological developments, and future prospects. American educational research journal, 45(1), 166-183.

        • Brad says:

          I don’t think it is conceptually possible to norm an IQ test for 15 year olds to five standard deviations. There aren’t enough 15 year olds on the planet.

          Cognitive performance tests for children should really use some other name and scale to make it clear that they aren’t based on fitting to a normal curve and aren’t commensurable. (Or maybe the other way around since the quotient part fits their use case better.)

          • onyomi says:

            It also seems a problem that children develop at different rates. You could have two people destined to become IQ 130 adults, but one of them is already at that ability level at age 16, while another doesn’t reach it until age 20 (just as two children destined to be exactly the same height as adults will nevertheless likely not achieve that height at exactly the same rate).

        • MSwaffer says:

          If you took the test at age 15, you were likely given the Cattell Culture Fair Intelligence Test (CFIT). The Research Norm Table based on high school students for this test has a standard deviation of 24 while many IQ tests have a standard deviation of 15. (Incidentally, according to a review in the MMY, Cattell felt the correct standard deviation should be 25 “arguing that the reduced scatter in traditional intelligence tests is due to contamination of intelligence with achievement”.)

          For this reason, even though the survey indicated the parameters of the question, very likely many people reported their IQ from tests that do not have the same scale which means the data is at the very least dirty data.

          Cattell Culture Fair Test

          Mensa Test Score Conversion Table

        • Evan says:

          I’m curious. I don’t think I’ve ever run into someone whose self-reported IQ is as high as 177. What about SSC commenters would lead you to think or feel you’re only in the middle of the pack?

      • alexsloat says:

        Yeah, the only formal IQ test I ever took gave me a ludicrously high result(like, “top few people in the world” high) when I was 7. I put the number on the survey, because it’s the best I have, but I don’t actually believe it to be accurate.

      • Gazeboist says:

        Important point about the as-kids numbers being less trustworthy. I have a file from a therapist I saw when I was 7 or so that says I have a 138 IQ, and I think I entered that on the survey, but I don’t really believe it.

        Also relevant: I’d bet IQ tests billed as such are given to people with high IQs more often, and/or they’re more likely to keep the results. “All the children (that we talked to) are above average.”

      • JASSCC says:

        Those of us who reported SAT scores from before the 1995 re-centering would usually have IQ scores higher than one would predict from the current SAT results percentiles for the same scores.

        It would be interesting to see if the reported SAT scores jump up for people younger than about 39. I bet you can find the SAT re-centering in there.

        • vaniver says:

          I looked into this for a previous LW survey. I don’t remember seeing a sharp jump, but there definitely was a visible effect for the >~42 vs. <~35. (We also had one old person who got a 1600 back when it meant something, which impressed me.)

        • Steve Sailer says:

          “Those of us who reported SAT scores from before the 1995 re-centering would usually have IQ scores higher than one would predict from the current SAT results percentiles for the same scores.”

          No, the opposite happened. The SAT-Verbal test was made considerably easier in mid-1995, with scores going up, IIRC, maybe 70 points, while scores on the SAT-Math went up about 10 points.

          The SAT had been normed on mid-Century Northeastern prep schools. By contemporary standards, these Holden Caulfield-types weren’t particularly good at math, but they were very good on verbal tests.

          As the SAT democratized, verbal scores dropped sharply while math scores merely drifted slightly downward.

          In 1995, scoring was made a lot easier on the verbal half and a little easier on the math half to return both scores to a mean of 500.

          Please note that I don’t know what has happened in the last year or two since David Coleman revamped the SAT.

          • switchnode says:

            The SATs were made easier, yes. So any given score on a pre-’95 SAT would indicate a higher IQ score than the same SAT score on a post-’95 SAT.

      • Virbie says:

        > 1. Survey IQs mostly match survey SATs from IQ/SAT conversion tables.

        I thought these conversions were considered valid for tests taken after ’94?

        • Nornagest says:

          It’s possible to find both, sometimes in the same table. The differences are not all that dramatic (going from pre-’94 to post-’94 rules for my own SAT scores works out to a shift of something like four IQ points, if I remember right).

          Our demographics being what they are, post-’94 is the logical table to be using.

      • Steve Sailer says:

        Scott has done an outstanding job of creating content for people in the 99th percentile. I can’t imagine people much below about the, say, 97th percentile of intelligence finding SSC of much appeal.

        It would be interesting if Scott were given, say, a column in Wall Street Journal or the New York Times if he could recalibrate to appeal to, say, the 90th percentile.

        • Tarpitz says:

          You really think Scott’s work is inherently pitched so high? It’s always struck me as insightful and beautifully written, but not hard to follow. It might demand some time (due to length) and a certain propensity for… let’s say dispassion, and perhaps both those things correlate with IQ, but it doesn’t seem to me at all a given that the requirement for enjoyment should be so extreme.

          • Aapje says:

            Writing can be challenging even when it is not hard to follow…

          • Tarpitz says:

            Challenging in what sense?

          • Aapje says:

            A large inferential distance to the cultural orthodoxy, for example.

          • Tarpitz says:

            Ok, I can certainly buy that as a legitimate natural language interpretation of “challenging”, and I imagine enthusiasm for writing of that sort correlates with IQ, but I’m really dubious that that correlation is strong enough to do the work your interpretation seems to rely on from it.

    • sophiegrouchy says:

      I feel like when I hang out with rationalists they treat me as if I’m particularly unintelligent (perhaps because I’m in a caring profession rather than a STEM one), but I scored significantly higher on the SATs than the average SSC reader, so I’m not at all surprised at these results

    • yeah also skeptical of purported 138 average , but scott says it’s legit ..that’s even higher than gwern

      • Nornagest says:

        For producing stuff like gwern does, I think above a certain point drive, discipline, and personality would matter a lot more than IQ does. gwern’s work is clever and inventive but not world-beatingly so; its real strength lies in being ridiculously thorough.

      • doubleunplussed says:

        I’ve worked in a few physics departments, and basically every grad student I spoke to about it had been the top scoring student in their high school. That doesn’t mean they’re all Einsteins, but it does mean I’m completely unsurprised when I find out that the people I’m around are top 1% in any measure of intelligence.

        I think the correct response to finding out that some people’s IQs are 138 isn’t “wow, people are smarter than I thought” but rather “wow, I have seriously underestimated how many people there are who are not as smart as what I take for granted in my peers”

        • _bpl says:

          I’ve had this conversation in several engineering departments.

          I’ve never taken an IQ test, but did score in the 99th percentile on both SAT and GRE.

          The ~10 friends I know who also read SSC are of similar nature.
          Thus, my sampling bias results in a lack of surprise.

    • vaniver says:

      Historically, I’ve been one of the people that looks closely at this, and I think the conclusion should be that this crowd is ‘very smart’ but not ‘extremely smart.’ That is, I’m pretty sure the intellectual caliber of Less Wrong (and apparently SSC isn’t that different?) is something like a top technical undergraduate university, where it actually isn’t that surprising for a typical person to be at the 1 in 100 level, and there to be a handful of more exceptional people floating around.

      I do think the self-reported numbers are inflated. For example, in 2012 we had people both take (a free online Raven’s, which worryingly might be calibrated on the internet population rather than a general population) and self-report IQ, and the average self-report was 139 and the average test result was 126. (My backcalculated number from SAT scores was around 128; it’s not clear to me why Scott uses tables that aren’t accurate after 1994 given the age of respondents.)

      That aside, what I think is most ridiculous about this particular survey is that the reported IQ distribution is roughly a bell curve, centered on 138. Even if SSC has a filtering effect, where the smarter you are the more likely you are to read SSC, it can’t be strong enough to overcome the exponential decay of the bell curve. (And this effect is mostly silly because of underrepresentation in the sub-130 region, where IQ tests are well-calibrated and bell curvish, so we can’t just say “well, maybe the right tail is fatter than expected” which seems likely to be true.) To point at what I mean, 78 of the public respondants self-reported 120-125 and 71 self-reported 150-155, both roughly equidistant from the mean, with the latter over 100x as common in the general population.

      I suspect that we should mostly stick to SAT scores (since people are more likely to have taken it and more likely to remember it), and maybe IQ scores for tests taken after age 21 (but who does that, unless they’re trying to get into MENSA or something?).

      • Deiseach says:

        Took that one again, improved since the last time! I was 99, now I’m all the way up to 103!

        Proof that SSC makes you smarter, taller, handsomer/more beautiful, wealthier and successful! 🙂

        • Aapje says:

          Just keep taking it again and again until you are at 138, easy peasy.

        • Acedia says:

          That just makes me think the test sucks, since it’s obvious from the way you articulate your comments that you’re above average in at least verbal intelligence.

        • Tekhno says:

          I just took the same test and got 104. Completely baffled by the later questions with checkerboards and circles and triangles and squares etc. Can’t find the pattern in them at all.

          • Deiseach says:

            We’ve found our level, Tekhno!

            What dndnrsn says is right; it’s a lot of visual processing and as I may have mentioned once or twice, to this day I still occasionally get mixed up about which is right and which is left, have no sense of direction, and have little to no spatial processing ability.

            So this is the perfect test for me to end up as “Congratulations, you are officially a moron” 🙂

          • Tekhno says:

            When I was a kid I took a test and got a very high verbal score and a very low visual-spatial processing score (of course here, both our scores are still above average). The combination was expected in only 0.2% of the population.

            I’m pretty good at those 3D shape rotation tests though, even though I’m not great at matching up patterns, so there’s another split again.

            I think SSC selects for weird people as much as intelligent people.

          • Deiseach says:

            The combination was expected in only 0.2% of the population. …I think SSC selects for weird people as much as intelligent people.

            We’re the special 0.2%, Tekhno; what percentile does that put us into? 😀

          • wysinwygymmv says:

            I started trying to draw to strengthen my visual imagination. I hadn’t considered looking for tests of spatial reasoning to test whether it’s having an effect or not, but it’s a great idea! Anyone have any recommendations about how I can find or work up an assessment for this purpose?

            Also, I recommend grabbing some graphic pens or graphite pencils (the ones at the art store, not CVS necessarily) and an instructional book on drawing. It’s just a nice, relaxing hobby. Since it requires sustained attention I suspect it also provides at least some of the same benefits as mindfulness meditation.

        • dndnrsn says:

          The test is entirely visual processing, though, isn’t it?

          • Eponymous says:

            No. It’s similar to the Ravens Progressive Matrices, which is considered nearly a pure test of (fluid) g. Assuming it’s legit (which I don’t know).

          • dndnrsn says:

            How would it measure someone’s verbal comprehension, reading comprehension, etc? I was under the impression that was a component of intelligence.

          • Eponymous says:

            It doesn’t. The idea is that those things are all correlated with g.

            That said, a full IQ test would contain many more subtests. For instance the WAIS has four major components (I believe). But nevertheless, Ravens is considered a test of pure g, or as near as you can get.

          • dndnrsn says:

            I’d be interested in the reasoning for that, because the “what pattern comes next” type test seems like it would prioritize visual processing over verbal.

          • Eponymous says:

            It’s an empirical result.

            The short version is that people who score highly on Ravens (and other tests with high g loadings) tend to score highly on many other tests, very much including verbal tests.

            My guess is that Ravens is primarily tapping into generalized pattern recognition ability rather than visual processing per se.

      • phisheep says:

        I suspect that we should mostly stick to SAT scores (since people are more likely to have taken it and more likely to remember it)

        Only in the USA though. Wouldn’t work for the other 36% of us.

      • johnjohn says: seems really inaccurate.

        Annecdotally gave me a score of 126 the first time I took it, vs the 146 I got on an iq test given at a job interview.
        When I was drafted I also had to take an IQ test that put me in roughly the same percentile as a score of 146, though they don’t reveal the exact nr.

        I don’t really find the results from the survey that unbelievable, it’s just yet another form of filter bubble effect going on. High IQ people seem to gravitate toward high IQ people.

        I’m with doubleunplussed on this one

        I think the correct response to finding out that some people’s IQs are 138 isn’t “wow, people are smarter than I thought” but rather “wow, I have seriously underestimated how many people there are who are not as smart as what I take for granted in my peers”

        • vaniver says:

          I agree that I’d rather have everyone’s ASVAB or something, and that has issues. The question is whether it has more or fewer issues than the self-reports, and given that it lined up closer with the SAT scores, I’m happier going with ( estimate) than (self-report estimate).

      • JASSCC says:

        It is perfectly plausible that if there is a population with a characteristic that has a rough bell-shaped curve with mean m and std. dev. s, and a very small sample is drawn by some selection characteristic that pushes the sample mean to m+2.5s, that subset can exhibit a bell-shaped curve with a new mean n = m+2.5s. But the new sample std. dev might be different.

        Intuitively, imagine that a topic appeals to be people with a certain unusual skill averaging at the 2.5s level. There are a very large number (99.4% of the total population) with the skill at less than the 2.5s level. There is a very small number (about .6% of the total population) with that skill at > 2.5s level.

        But all of the right of the curve could potentially participate in discussion of the topic (though in practice only a small number do), and only a very small fraction very close to the 2.5s level can nearly do so and will attempt it.

        If both the the overall population and this new sample are very large, and the factors that go into having this skill at the 2.5s level are due to the sum of a number of independent random factors, the central limit theorem tells us that *both* the population and the small sample will take a bell-shaped curve. This is not about “overcoming the exponential decay of the bell-curve”. It is precisely due to the same factors that give rise to an overall bell-curve — addition of very many independent random variables.

        “Gauss is not mocked!” as the saying goes.

        • vaniver says:

          Suppose we have a general popuation pdf G(x), a filtering function f(x) bounded by 0 below and 1 above, and we multiply those together and renormalize with a constant to get H(x)=cf(x)G(x). We know H(x), we know G(x), we can solve for f(x) (with c being immaterial).

          I find the resulting f(x) implausible. Obviously it’s possible to construct a mathematically consistent one, but it doesn’t match my forward-looking expectations of how IQ impacts how likely one is to read SSC.

          • JASSCC says:

            Sorry, I was trying to reply and I accidentally clicked on the button to “report” this comment. I did not mean to do that. I don’t see any way to unreport your comment, so I’m leaving this note.

            But to the meat of what you wrote: without inventing a filtering function, biased samples happen all the time. It is a commonplace that biased samples often exhibit a bell-shaped curve. There is no need to construct the filter function.

            Here is an example, drawn from heights of WNBA and NBA players. This is a sample biased three or more standard deviations to the right of the general population mean:


            Note the women’s curve looks pretty close to a normal bell curve. The men’s curve shows some skew (in the *rightward direction*!), but both are at least 3 sd > than the general population average (the men’s numbers are about 3 std devs up, equivalent to an iq bias of 145, and the women’s numbers are slightly more extreme — I’m getting about 3.33 std. devs, which in IQ would be like a biased sample with a mean of 150) and look even more like a normal bell curve. And the total numbers of players in the bins are much smaller (a few hundred each) than the sample size of the SSC survey.

            So there’s no need to invent the filter function. A biased sample of a large number of people converging on an atypical mean will, by the central limit theorem, commonly exhibit a bell shape around that mean, for the precise same reasons the overall population exhibits a bell shape around a different mean.

            As another example, I’m having trouble finding one right now, but I believe the IQ distribution of college graduates is typically shown as a small bell curve centered around almost 1SD > the population mean.

        • Ilya Shpitser says:

          “It is perfectly plausible”

          It is logically possible, but extremely unlikely (e.g. not plausible at all), especially at these sample sizes. A much more likely explanation:

          (a) People are kind of full of themselves, and

          (b) Take IQ way too seriously, and

          (c) Use poor measures of IQ.

          • JASSCC says:

            Why is a sample size of thousands of people not more than enough to exhibit a bell-shaped curve by the central limit theorem? Isn’t this what we expect from a biased sample?

            None of this is to say that the self-reported numbers are accurate. They may be wildly inflated. But the bell shape of the curve does not seem to me to be evidence of that.

          • Ilya Shpitser says:

            Sorry, was busy with other stuff. See other comments here on why the CLT doesn’t apply here.

      • thedufer says:

        > That aside, what I think is most ridiculous about this particular survey is that the reported IQ distribution is roughly a bell curve, centered on 138.

        I think there are other responses that do a good job of showing why this isn’t actually that surprising. But at a higher level, I don’t think you’re approaching this from quite the right direction. The question isn’t whether the shape of the data is surprising – it’s whether you’d be more likely to get that shape via people making up numbers vs the filtering effect.

        It isn’t particularly obvious to me why the former would be more likely to give us a bell curve than the latter.

      • Eponymous says:

        Incidentally, is considered reliable? I took it a while ago and got something in the low 140s. It seemed fairly legit. Also, what’s it’s ceiling?

        I’m curious because I’ve never taken an IQ test.

      • MSwaffer says:

        That aside, what I think is most ridiculous about this particular survey is that the reported IQ distribution is roughly a bell curve, centered on 138.


        For those quoting central limit theorem etc. the issue here is sampling from a normal distribution. If you randomly sample from the entire distribution, you would get a normal curve centered at 100 with an SD of 15.

        If you sample and get a distribution centered at 138 (+2 SD) you would not expect that sample to be normally distributed, assuming your sampling was random.

        For example, run this simple code in R:
        d <- rnorm(1000, 100, 15)
        sorted <- sort(d)
        top16 <- sorted[840:1000]
        newsample <- sample(top16, 80)

        This centers the second sample on 130 for simplicity thus sampling from the top 16% of the entire population. This gives us a right skewed distribution which we would expect given we are sampling from the right side of the normal curve.

        The only way you would expect to get a normal curve with a sample mean 2 SD above the population mean is if you have a biased sample.

        This may be the case, but it seems extremely unlikely?

        Now if we had people randomly inputting IQ’s of what they think they had or wish they had.. I would fully expect a higher than population mean and a normal distribution for those responses.

        • Ilya Shpitser says:

          Not just a biased sample, but a sample biased in a very particular way!

          After all, any draw from a population that isn’t using a uniform distribution is a biased sample, and most of these draws will not lead to a normal if the general population is a normal. Polls and surveys are almost always biased samples, just not in the “right way.”

          To me, a much more likely explanation than a very special sampling mechanism that leads to a normal distribution of IQ with a mean of 138 is people are just status-thirsty, and high IQ is a high status marker around these parts.

          • JASSCC says:

            Fair enough, on both points. The biased sample will not be normal, and let us stipulate that IQ scores are probably inflated. But assuming you do have a biased sample, from a normal distribution, what distribution do you expect it to take? And would you expect it not to have a bell-shaped curve?

          • Ilya Shpitser says:

            There is no general answer to your question, it heavily depends on the sampling distribution used.

            You can get anything from a truncated normal to a uniform.

          • Matt Swaffer says:

            You would expect it to be a right skewed curve, not a bell curve.

            A bell curve indicates that values to the left of the mean are just as likely as values to the right of the mean. So if you have a group of people and the mean IQ is 100, that distribution would be expected to be a bell curve because an IQ of 105 is just as likely as an IQ of 95.

            If you have a group of people with a mean IQ of 138, you do NOT expect there to be just as many people with an IQ of 143 as you do 133. The reason is that it becomes increasingly rare to have an IQ higher than 138 and more common to have one less than 138.

            So the expectation would be that you would more people to the left of 138 and fewer, with larger values, to the right.

            In the R code I posted earlier, I simply took the top 160 values of the population of 1000 and sampled. IF this were a normal distribution, that would center around 130 but it doesn’t, the mean is around ~121 with a min value ~115 and max of ~150. To get a distribution that centers around 138, you actually need to sample from the top 1.25% of the population and that distribution is not normal.

            d <- rnorm(10000, 100, 15)
            sorted <- sort(d)
            topx <- sorted[9875:10000]
            newsample <- sample(topx, 100)

            There really isn’t any difference in doing the above math using randomly generated numbers and using IQ scores from the population because, by definition, IQ scores are normally distributed with a mean of 100, SD of 15.

        • John Schilling says:

          The only way you would expect to get a normal curve with a sample mean 2 SD above the population mean is if you have a biased sample.

          What would you consider an “unbiased” sample?

          I’m not going to try to decipher your code, because you don’t get to presume my familiarity with any particular programming language. But it seems to me that you are assuming a “high-IQ” subpopulation will consist of a random sample of the top X% of the distribution for the population as a whole, with any one person in the top X% being as likely as any other to be in the high-IQ subgroup and anyone outside the top X% being absolutely excluded.

          This is clearly not the case here. SSC isn’t Mensa, we don’t have a minimum IQ requirement, and we clearly do have people whose self-reported IQ is barely in the top 50% of the population. A more plausible model is that e.g. Scott’s writing targets an IQ of 150 or so, and the probability of any random person reading the blog approximates a normal distribution centered on 150. People who are very much smarter or very much less smart than Scott(‘s writing) are less likely to stick around, but there will be a few even at the extremes.

          The readership here is necessarily a convolution of two distributions, p(IQ|human) and p(SSCreader|IQ). The first is defined as a normal distribution, the second very plausibly could be. The convolution of two normal distributions is IIRC a normal distribution. So I’m not seeing a problem here.

          • MSwaffer says:

            R is a pretty common statistical language but you could easily do this in SPSS, SAS, Stata… pick a language. All I did was create a normal distribution then grab the top 16% of the distribution and randomly sampled from it.

            As Ilya mentioned above, in order to get a normal distribution centered around 138, you would need a very specific sample that would exactly inversely match the bias of the rarified population you are sampling from. So Ilya was correct that the sample is biased already… to be more precise in my wording, you would need to sample in a specifically biased manner.

            In other words, there are so few people with IQ’s above 138 that you would need a very large percentage of people at 148 to balance out relatively few at a mere 128. In other words, you are saying that only the smartest of the smart read SCC and even just regular old smart people are only marginally interested.

            Maybe. Statistically that seems extremely unlikely.

          • MSwaffer says:

            As for the assertion that Scott’s writing targets an IQ of 150 or more, his writing style certainly does not.

            For example, I grabbed the text from one of his top posts, “Beware the Man of One Study” and ran it through some readability indexes. The Gunning-Fox index is 14, Flesch-Kinkaid is 11. That means that post targets the reading function somewhere between junior in high school and sophomore in college. That’s roughly the same reading level as a typical journal article in a peer-reviewed journal.

            Granted that is much higher than the target for most writing on the web… but not exactly IQ + 150 level reading. That’s fine though… I don’t think that’s the author’s intent.

          • Steve Sailer says:

            Scott is clearly smarter than I am, but I similarly aim for an 11th to 12th grade reading level in my writing.

            Scott is a highly lucid writer. But it’s asking a lot of people below, say, the 95th percentile to stay focused on his concerns.

        • JASSCC says:

          But is this a fair comparison? By applying a top 16 cut off, your are modeling a case where, by analogy a person under IQ 115 cannot participate.

          Incidentally, the second sample is not centered on IQ 130. It has a lower bound of IQ 115, which is approximately the 84th percentile.

          The top 16 of the normal curve has 50% of density below 1.4, and 50% above. So this is centered at approximately 121.

          The histogram, by the way, is of course right-skewed (incidentally, in my example using NBA player heights, I mistakenly described it as right-skewed but got it backwards, I should have stuck with positive and negative) because you impose a sharp cut-off at 115 — you are using a *truncated* sample, not a biased sample. It is not a fair analogy to the results a biased sample would produce.

          Regarding Ilya Shpitser’s comment below (to which I can’t reply), I am more than happy to accept that the IQ results may be “garbage”. What I am asking is what would be expected from a non-truncated sample.

          Obviously there is no such sharp cut-off as was used in the example, so that’s a completely unrealistic model of the type of bias that may apply.

          I am curious about this topic, so I am unprepared to “let it go”, no matter how stridently you tell me to.

          • Ilya Shpitser says:

            Oh my god.

            A truncated sample is a type of biased sample. Almost every biased sampling process from a normal will not lead to a normal. Not just a truncated sample, almost any biased sample at all.

            The IQ reported on the survey is garbage. Just let it go.

          • Matt Swaffer says:

            I’d be interested in seeing what parameters would produce a sample with a normal distribution that is centered at 138 (+2.5 sd)

            If I were trying to develop an explanatory model between height and propensity to enter and succeed in the WNBA, that’s the kind of model I would want. Then of course I’d be interested in a theoretical explanation for why that model would fit SSC readership as it relates to IQ.

          • Matt Swaffer says:

            I hadn’t looked at the data until just now so I have been assuming the comment about a normal distribution was accurate.

            As it turns out, the IQ data from the survey as actually negatively (left) skewed. (Shapiro-Wilk test gives us w=0.864, p<0.001) This is even more unusual than finding a normal distribution.

            So either we really are all above average here in Lake Wobegon… or we think we are.

        • JASSCC says:

          OK, I’ll accept that the distribution may not be normal, but here’s my R code that produces a non-truncated biased sample by using a linearly increasing probability by the order statistics of the total population.

          Note that the shape of the resulting histogram is roughly a bell-shaped curve. (If I repeat the sampling, I do sometimes observe a small amount of right skew, but it is not always apparent.)

          Why is this shape so unlikely from a real-world biased sample?

          d <- rnorm(100000, 100, 15)
          sorted <- sort(d)
          biasedsample <- sample(sorted,5000,prob=seq(0.00001,1.0,0.00001))

        • JASSCC says:


          If you expect to find a right skewed distribution, what do you make of this biased sample of heights where the histogram exhibits a bell shape and a fairly obvious *left-skew* at height values approximately 3 sds above the general population mean?

          And what about the WNBA heights, which while even more extreme (about 3.3 sds) shows no obvious skew?

          • Matt Swaffer says:

            In your R code, if you look at the results of skew(biasedsample) you will see that even though the histogram looks mostly normal, you do have a slight skew.

            As for the WNBA heights, I think a good working explanation is that there is not a random sampling of people into the WNBA. In fact, the marginal propensity to enter and succeed in the WNBA is likely directly tied causally to height of the individual. In other words, taller people are more likely to enter the WNBA because there is a causal relationship to success. In fact, people who are taller than the rest of the population have a much better chance of succeeding than those who are only marginally taller.

            I am not sure a similar causal relationship exists with regard to reading SSC. I would certainly be skeptical of drawing that inference.

          • JASSCC says:

            MSwaffer, I can’t reply directly to your comment, so I’m doing the next best thing and replying up here:

            1. I agree that it is implausible that SSC readership has a similar selection effect going on (though it is more plausible that those who actually know an IQ score from a test were singled out for formal testing). I was merely using the WNBA curves I found as an example of a biased sample of height that in one case does not exhibit strong skew, and in the other case goes in the *opposite* direction from your prediction. Of course WNBA/NBA players are not randomly selected, but neither are SSC discussants / readers.

            2. As already stipulated, I accept that a quite likely explanation for the IQ mean at SSC is profound inflation / poor or selective reporting of anomalous results.

            3. BUT, you will note that throughout I have spoken about the bell-shape of the curve. If you look back at my comments, you will see I avoided speaking about normal curves until responding to comments specifically about those. (To be fair, I did apparently mistakenly invoke the CLT as the explanation for this bell shape, as a rough approximation of the normal when some of the strict constraints are relaxed, and I have been schooled against that.) The main reason why I spoke about bell-shaped curves is that I’m addressing the notion that we can dismiss the reported scores *on the basis of the shape of the curve*. Many, many non-normal distributions have a rough bell-shape. I was specifically addressing the point that we could reject the reported stats on the basis of the shape / skew of the curve.

            4. If you run my R snippet a few times, sometimes there will be more skew, sometimes less. My point is that I am unconvinced that it’s impossible or even highly unlikely for a biased sample to have a roughly symmetrical shape from essentially innocuous causes. Your approach of essentially censoring the first sample at the 1SD level loads the deck against symmetry.

          • Matt Swaffer says:

            You are correct that truncating the data as I did stacked the deck against symmetry. However, in order to get a sample that averages 138 from a distribution with mean 100 and sd of 15, you have to sample from an even smaller “section” of the data…. which simply accentuates the skew.

            You are also correct that it is possible to have a relatively normal and / or bell-curved distribution of a sample where the mean is significantly higher than the population. In those circumstances however there should be a plausible causal explanation for self-selection. My argument is that the most plausible explanation is not self-selection into readership but rather self-selection of reported IQ.

            To summarize, it’s one thing to explain a significantly higher than population mean in the sample. It’s another to explain the distribution. One is much easier to do than the other.

          • Ilya Shpitser says:

            I don’t have anything substantive to add anymore, but wanted to pop in quickly and spread warm fuzzies to everyone involved in this thread for (a) updating and (b) generally high quality discussion.

          • Steve Sailer says:

            What’s the average IQ of people debating in the comments, using sophisticated statistical models, what the average IQ of readers is?

            I don’t see these kind of arguments in the comments at too many other sites.

    • MoebiusStreet says:

      It looks to me like normalizing to SAT is fraught, because SAT scores have inflated over the years.

      When I took the test in 1983, I got 1390. That was the highest score that had ever been recorded in my high school. Nationally, 1600s did occur occasionally, but it was very very rare.

      Not long ago I was having a conversation about this with a much younger person, and he remarked, “what, was your school brand new?”. It seems today that 1600s are something to be proud of, but they do occur with some regularity, enough that you’d expect any given school to see them once in awhile.

      (for calibration – IQ measured at 139, but that was when I was in elementary school)

      • roystgnr says:

        The SAT scoring was recentered in 1996. And I think that’s the only score inflation there’s been; the whole point of the recentering was to raise the mean score back up around 500 because it had been slowly dropping over the preceding decades. Judging by this table a 1390 on the old scoring would have been between 1400 and 1470 on the post-1996 scale, depending on the exact breakdown. That’s still around 97.5th percentile, though, which ought to be “highest in your graduating class depending on how big the school is” level, not “highest in the history of the school” level.

      • LCL says:

        On the flip side: I took the SAT in two consecutive years in the late 90’s and scored 1600 and 1590. My tested IQ at age 14ish was ~127, which I believe to be accurate, roughly hit on the ravens the first time I took it, and reported on the survey.

        I did have the sense of being better at the discrete skill of taking standardized math or reasoning tests than people who were otherwise smarter. Mostly that reduced to:

        – figuring out what rule they’re testing in any given question
        – noticing which answer choices represent which potential rule variant
        – generating a simple example that I can test with a calculator to derive the correct rule (if I hadn’t memorized it already)

        And doing all of those things quickly enough to meet the time limit. Having a good mind-meld with the sort of people who write standardized test questions definitely helps, which was likely a cultural bias in my favor (I’m from an academic family). For vocab sections I didn’t have any such algorithm; I just happened to have read a freaking ton as a kid so I knew all the words.

        I also had the sense that I was scoring at the high end of my ability and wouldn’t have been able to hit the caps if the test were harder. In contrast to finding it easy but sometimes making silly or careless mistakes, as smarter people report.

      • Cypren says:

        I think you might be overestimating the rarity; 1600s happened with some regularity prior to the recentering, too. I got one in 1993, and it was considered unusual, but not the kind of thing where people talked about it enough that (even local) reporters were going to be hunting for an interview with a kid who aced the SAT. It was just accepted as something that a few unusually bright and extremely committed students in any given area would achieve. I’m pretty sure there were other people in my county who managed the same feat that year as well; it was hardly unheard-of, especially in a well-to-do, highly-educated population.

        • Brad says:

          Your reference to extremely committed gives away the difference. In 1983, at least in most communities, the SATs were still thought of as an aptitude test and very few people studied for them. By 1993 almost all the wealthy suburbs had gotten the test prep bug, and it was already spreading beyond those places.

          • Douglas Knight says:

            Here is an ETS spokesman, quoted in 1993, claiming only 4-12 perfect scores per year.

            But this gives numbers for several years. I’m not sure of the source.
            1986:9 1990:10 1994:25. So maybe the ETS spokesman was a couple years behind the times, but it’s still quite rare. Recentering increases the number of perfect score by a factor of maybe 35.

          • Cliff says:

            But the evidence is that inteisive test-prep raises scores a minimal amount, maybe 30 points if I recall

          • Brad says:

            As a mean? What does the distribution look like?

            If there’s was a more or less hard ceiling at 50 points, I think you’d see a lot less test prep. But if there’s a fat right tail of people getting 150 points (1250->1400 is huge in terms out outcomes) that’s going to generate a huge industry even if the modal benefit is 15 points.

          • Quixote says:

            By the late 90s – early 2000s, about 1000 people scored 1600s each year. So not a huge group of people but not only a handful. If you were in a large city with some population density you kinda had an ex ante expectation of who you thought would get a 1600 (people you know from your school or at other schools that you met through math team, or debate, robotics club, or other nerdy competitive activity). It was common enough that most people you expected to get it did, but a few people made boneheaded mistakes or had a bad day and got 1590s.

      • Steve Sailer says:

        Here’s a 1987 People Magazine article on five of the nine kids in the USA who got a perfect 1600 on the SAT (old-style):

        I located several of them online recently. One is a massively successful neurology researcher at Georgetown medical school, one is a distinguished high school teacher, and another rides a Harley and is popular with his blue collar extended family but I’m not sure exactly what he does.

    • Eric says:

      I think a major source of score inflation is that people preferentially report/remember their highest score. So if someone gets 1550/1600, 1420/1600 and 1440/1600 they will tell you their SAT score was 1550/1600.

      • Cypren says:

        It’s also a matter of the test itself being practicable through repetition and focused study. There is (or was, at least, when I took it 25 years ago) a very specific set of rules and patterns to the questions that are asked and the answers that are sought. I took the SAT twice and the PSAT in between, and my score improved about 130 points (on the 1600-point SAT scale) over the three attempts.

    • Deiseach says:

      “Well, that’s the survey results from SSC, where all the arguments are strong, all the graphics are good looking, and all the IQ results are far above average” 🙂

  3. idontwantawpsite says:

    Could you included ADHD in the next survey? I was disappointed you didn’t include it.

    • PedroS says:

      Maybe he intended to include it but got distracted 🙂

    • wintermute92 says:

      I was surprised by that absence, I’d love to see it next time.

      Partly because I’m guessing it’ll be higher-than-population, and partly because it seems to have all kinds of interesting correlations with high IQ populations, autism, depression and ‘plasticbrains’. I don’t know if the survey has enough power to get significance on those, but it’s probably a better candidate than almost any other population going.

  4. Tekhno says:

    I wonder how many of the libertarians supported basic income.

    • Luke Perrin says:

      The average support was 3.6 and libertarian support was 3.0, so a significant number. That’s only including people who selected the default “Libertarian, for example like the US Libertarian Party: socially permissive, minimal/no taxes, minimal/no distribution of wealth”. Several people gave a custom description of their political views as like “Libertarian + redistribution”.

      • Urstoff says:

        Need follow up question,

        “Which most accurately describes your libertarian views?

        1. Reason Magazine libertarian
        2. Lew Rockwell / Ron Paul libertarian
        3. Bleeding Heart / Niskanen libertarian

        • I don’t think any of those three fits me very closely, although Reason is the closest. But if you added anarcho-capitalist libertarian, you would then have to distinguish among subgroups of that category.

          • vaniver says:

            4. Caplanism
            5. I’m literally David Friedman

          • LHN says:

            Shades of the old geek code, where the highest rating for, e.g., Linux identification/knowledge would be “I am Linus Torvalds.”

          • Urstoff says:

            There will always be people left out, but those seem like the three major strands of current libertarianism to me. I guess maybe add Mises / ancap libertarianism as a fourth.

          • John Schilling says:

            There are enough people who seem to assume all libertarians are anarcho-capitalists(*) that I think any attempt to subdivide libertarians has to account for the category even especially if it turns out to be small.

            * e.g. people who make ancap-specific arguments and claim to have refuted libertarianism

          • I think the Lew Rockwell category would include Rothbardian anarcho-capitalists.

        • Cerebral Paul Z. says:

          I demand a “rrright-ving hippie” option.

  5. Jugemu says:

    Rather than using { and } you can probably use &lt; and &gt;, which should render as < and >.

  6. suntzuanime says:

    I find the results for perceived trustworthiness of others pretty shocking. What’s wrong with you people? Constant vigilance!

    • daniel says:

      We just want to be satisfied 😛

    • alexsloat says:

      “Trust but verify”, as the old saying goes.

    • Deiseach says:

      No, no, suntzuanime, all these trusting high-earning pigeons nice people will be amenable to loaning us big money to enable us funding a lifestyle of excess and debauchery starting up companies based on revolutionary tech that will re-write the laws of physics as currently understood! 🙂

    • wintermute92 says:

      That surprised me too. Also, I see “trusting is correlated with happiness” a lot but I almost never see deeper investigations. Two possibilities I’d love to see investigated:

      1. Trustingness correlates with happiness, but the distribution is far from normal. It produces lots of small benefits and then royally screws a handful of people. (e.g. Signing dubious employment contracts. Most people don’t get sued, but those who do are ruined.)

      2. Strong reverse causation – people become vigilant as a result of bad outcomes.

      • nhnifong says:

        What seems likely to me is them both being caused by a third factor of having a relatively pleasant childhood with parents who are nice to you. I don’t think our adult lives change that very much. Even if being overly trustworthy as an adult and being taken advantage of was a drag on one’s happiness, it wouldn’t put a dent in the happiness boost you got from having nice parents.

        • Tarpitz says:

          I have very nice parents, was very trusting as a child, became very suspicious in my teens as a result of bullying following a change of school, and as an adult am trusting, unhappy, somewhat dissatisfied and confident that my unhappiness is in no way the result of people taking advantage of me (which they sometimes do, but almost always only in mild ways I don’t really mind).

          • nhnifong says:

            Then maybe the threat level of one’s current environment is a better candidate to be the third factor causing both unhappiness and suspicion. I may have attributed to parents because I’m never in threatening environments.

            Sure is frustrating that no amount of bickering can establish causation! Yet, trust and happiness seem like things every human would have an innate understanding of. Normal play would have been enough to give you the data, so it’s unnerving that we can end up with different opinions on it. It makes me think we know nothing.

    • Evan says:

      “Trustworthy” is ambiguous here. It could mean people generally aren’t trying to cheat, lie or deceive you as opposed to the notion people generally provide reliable info. People often think their beliefs are fact when they’re actually fiction, often let slip things in a way not exactly representative of their opinions/beliefs, etc. There’s also Robin Hanson’s theory of signaling which makes out (what seems like) most behaviour to contain an element of deception. This is an abstract theory about how our brains are wired for climbing the social ladder, not about people making the conscious choice to constantly and deliberately spread falsehoods. I’m sure there are walks of life where the average person must always be wary of who they’re talking to, and I think it makes sense to be skeptical of tall tales even if you don’t challenge the person on them outright. But going through life constantly looking for reasons for why people around you are lying to you or each other seems like a recipe for disaster.

  7. cactus head says:

    That bimodal distribution of happiness and life satisfaction is pretty interesting. Here’s a scatterplot I made which functions as a visualisation of the joint distribution–you can see the two modes in the joint distribution that give rise to the two modes in both marginal distributions of life satisfaction taken alone, or that of happiness.

    I assume the correlation was left out as one of the obvious results. For the record, Pearson’s product-moment correlation (default from cor.test() in R) is 0.728 with a confidence interval of (0.715, 0.741).

    I think I might construct two subsets of the responses, a low-in-both subset and a high-in-both subset, and dig around for any significant differences.

    • Luke Perrin says:

      One person answered each of (1,10) and (10,1). They should get married or star in a sitcom or something.

    • Mediocrates says:

      I too was interested in the bimodality of the happiness distribution. Just for fun, I ran a very simple analysis, my favorite type of analysis, to see if it correlates with a few other responses. Basically, I binned the responses into Sads (1-5) and Happies (6-10) and did some TTESTs to see if there were any significant differences.

      Trump rating? No. (Sad!)
      IQ? No.
      SAT1600? No.
      SAT2400? Yes, oddly, but rather slight. Sad average = 2202, Happy average = 2227, p ~= 0.03.
      PoliticalSpectrum? Oh yeah. Sad average = 4.36, Happy average 4.64, p ~= 0.000004.

      So liberals are significantly sadder than conservatives, but not for the obvious reason (Trump). Maybe if I get time I’ll check a few other responses.

      • Mediocrates says:

        Okay, a few more before I have to get to work. I thought the effect might be driven by differing views on global warming, a depressing issue with a liberal/conservative split, but ehhhh, there’s only the slightest of differences (p = 0.045). Immigration actually does much better (p = 0.0045).

        Also, Happies are richer ($72K vs. $109K, p = 0.0055) but slightly older (30 vs. 30.9, p = 0.0005). Note that I didn’t try dropping the outliers of the income distribution a la Scott.

  8. ti4 says:

    Suggestion for next survey: Question about reading speed and if that correlates with nonverbal thought process.

    • Jules says:

      Another, which might be important for questions like the “book and much like it”: is English your native language?

      • JohnBuridan says:

        What bothered my about this question is that a “proper” Henry Fowler English would NEVER state, “I read this book and much like it.” The way we were taught in the old school was that if you want the sentence to mean that you read a book and similar ones, it should be rendered, “I read this book and many like it.”

        The reason for this is English teachers tried to categorize ‘much’ as an indefinite quality and ‘many’ as a definite quality. For example, I see many giraffes, but am still filled with much sadness.

        During the survey I both noticed both the ambiguity AND the “bad” English. The sentence could be ambiguous if you didn’t have an incredibly anal sense of the English language. I do have such a sense and thus felt deeply wronged.

        • Fahundo says:

          What if you read the book, and you read much like it, but the much like it that you read wasn’t in other books? At that point it becomes an indefinite quality, right?

        • Deiseach says:

          Yes, to me the more ‘natural’ way of using “much” in that context would be “This book, much like others of the celebrity biography genre, reduces its subject to the image on-screen without too much effort to lift the mask” or “Much like the preceding volume of the series, this book piles on the action but fails to develop the interior lives of its characters” or something similar. It is sort of a forced ambiguity to encompass both the “I have read many books like this” and “I like this very much/muchly” (and even there, using “muchly” might be seen as exaggerated or affected).

        • houseboatonstyxb says:

          I looked at how much red ink would be needed to turn it into good and unambiguous English.

          1. ‘and much similar material‘ — is a little squishy at the join, but unambiguous.

          2. ‘and like/d it very much’ — unambigous but seems hardly worth saying, and takes a lot of red ink to move those words around.

          3. ‘and much liked it’ — elegant minimal use of red ink, but easily missed if read without a monocle.

    • Montfort says:

      If we’re asking about comment enjoyability/balance/frequency again next time, I would definitely like to see another question about whether they read all open threads, just the non-hidden ones, or skip them (and/or whether they follow the OTs until the next one opens, or just skim the comments once and move on).

      I expect the subject matter and poster distribution are significantly different between the regular post comments and OTs.

      (Though maybe this whole thing will die down by then. We can always hope.)

    • Callum G says:

      I wonder if spotting the the repeated words inversely correlated with nonverbal thought process. I should learn how to use R.

      • Spookykou says:

        Odd, I didn’t notice any of the the duplication in the survey but yours jumped out at me instantly, before I even knew it was the subject of your comment. Maybe there is an art to hiding them in your writing that Scott has mastered that is confounding the data.

      • Tekhno says:

        I’d also expect slower readers to be better at spotting repeated words vs people who scan sentences and infer expected words. Possibly just a trade-off between two reading styles.

    • wysinwygymmv says:

      Anecdotally, I have an intensely verbal thought process and a fairly slow reading speed. I read every word, hear it spoken aloud in my head as I do, and anything I don’t understand I re-read until I do.

      I was astounded to find out that my partner does not “hear” the words in her head when she reads, and can read and converse at the same time.

      I still never spot the duplicated “the”s.

  9. Jules says:

    Would it be possible to include a key for these graphs? And maybe sort a lot of these results by value?
    I find the colour coding horrendous, but I guess that you tried to make it neutral.

    • Nancy Lebovitz says:

      I had some trouble with the formatting, too. I think a color block next to the result which corresponds with the color would help.

      Sorting results by value would help, but not necessarily for all questions.

  10. MarginalCost says:

    Interesting that about half of readers don’t think they’re typical. I’d be interested in seeing what that’s correlated with, but it doesn’t seem to be in the data release?

    • episcience says:

      Isn’t that about what you’d expect? i.e. there’s a clump of about 50% of the population around a median and then 25% on either side who feel further away from the median and therefore “atypical”?

    • neonwattagelimit says:

      I didn’t take the survey but I would have answered ‘no’ to that because I am not a STEM person and I get the sense that this community is very heavily tilted towards STEM people (which is borne out by the survey data).

      It would be interesting to see what ‘yes’ and ‘no’ are correlated with on that one. It’s not an unexpected result, though.

    • nelshoy says:

      I would naively expect individuals in most surveyed populations to focus on how they differ from the norm rather than on how they conform. I had expected SSC readers to use more Outside View “am I more or less typical than the average reader?” which would translate to a more accurate response.

    • mingyuan says:

      episcience makes a good point, but I think there are too many ways to define ‘typical’ here, and I wouldn’t necessarily expect them all to cluster. For example, someone who is a white male but not in STEM could consider themselves atypical, as could a programmer who is not a white male, as could a religious white male, as could a liberal who perceives conservative bias in the comments. Is the ‘typical’ SSC reader actually a straight white cis-male atheist programmer, or is there little correlation between those five metrics (other than maybe cis-male & programmer), so that 48% of readers could very reasonably believe themselves to be atypical if they’re different on at least one of those metrics?

      edit: I guess this is similar to what nelshoy was saying; from the outside view someone who meets four of the five criteria is still mostly typical. But I think that putting more weight on how one differs from the norm is not necessarily illogical. Even though I share intellectual interests with the other people who read SSC and am a pretty typical age with a pretty typical education level and IQ, what strikes me most is always being the only girl at meetups (and usually being the only non-white person as well).

      • Rachael says:

        I rated myself as typical despite being female, because SSC is one of the few places where I feel like I fit in and feel at home, where I feel like I’ve found “my tribe”. I’m typical among SSC readers in terms of thinking style and general approach to life. To me, demographics are much less important than that.
        I guess I was thinking something like “am I more typical here than I am in the average community?” as opposed to “am I more typical than the average reader?” Although I think I’d have answered yes to the latter as well.

        • simon says:

          I had some trouble thinking about what I meant by typical and one way I thought of it was in terms of likelihood ratio. As in, disregard the priors for someone to have the relevant characteristics, would someone with my relevant characteristics be relatively likely to be an SSC reader. I think that works out to somewhere close to your definition.

          Agreed on demographics not being the relevant characteristics, though perhaps that’s easy for me to say since my demographics are in fact not that far from average for SSC.

      • MarginalCost says:

        But this should be easy to figure out! Scott, can you update the dataset to include the “typical” question so we can find out what correlates with self-perceived typicalness?

    • Evan says:

      I’m a white guy in his 20s, so that’s the only reason I put myself down as typical. I didn’t need to think about it more than that.

    • Evan says:

      I’m a white guy in his 20s, so that’s the only reason I put myself down as typical. I didn’t need to think about it more than that. I wouldn’t be surprised if that’s how many people answered the question the way they did.

  11. tmk says:

    So there is a vast difference politically between the average reader and the average comment. (We don’t have a measure for average comment, I base it on my intuition). Surprisingly, it doesn’t bother the readers all that much.

    Has anyone studied the effects of variations in how much people comment online? I.e. someone who spends all day every day commenting will be as visible as a thousand people who comment a few times a month.

    Edit: I looked at the raw data.

    They are 86% white and 89% male, similar to the whole set. 88% hetero vs 82, 77% prefer monogamous vs. 67%. 51% single vs. 43% in the whole set.

    28% are libertarian, 13% conservative, 12% n*r*onary, compared to 24%, 6%, 5% in the whole set.

    Scott: Could you not have the system auto-delete comments that use a word that appears in the the survery and in this very blog post?

    • Cheese says:

      > Surprisingly, it doesn’t bother the readers all that much.

      For me personally it’s more that while my views don’t necessarily concord with that of the commentariat (or at least a significantly vocal amount of them), there are many regular commenters who are very intelligent and articulate in expressing their views (which diverge from mine). And it’s very useful to be exposed to those arguments and examine your own views in light of them.

      However I spend enough time reading this bloody blog. If I were to spend the extra time it’d be yet another impact on study time. So it doesn’t bother me at all.

      • Doom Squid says:

        This. Even though the comments don’t line up perfectly with my views, the way they express their differences causes me to think about their views and mine, not just randomly start hating them.

        • tmk says:

          I think I agree. What makes a forum particularly bad is when the dominant group are constantly taking cheap shots at anyone who disagrees. We get some of that here, but not too much. Scott does a decent job of moderating rude people. The Marginal Revolution comment section is just ridiculous.

          • Gazeboist says:

            (random tangent)

            The Marginal Revolution comment section is just ridiculous.

            Ever since Scott linked to that MR story about the apartment building in Mumbai, I have been confused about the Brazilian dude in the MR comment section who (a) everyone else knows and (b) absolutely hates India for some reason. Do you have any idea what’s up with that guy?

          • Brad says:

            I assume everyone still over there at this point is engaging in some kind of dadaist performance art.

          • tmk says:

            I don’t get the Brazil guy either. No matter what the topic is, he finds a way to talk up Brazil. But what’s worse, when he occasionally writes a sensible comment about something else, somebody else immediately replies just to shit on Brazil.

          • Cliff says:

            Yeah it’s terrible. I mean do people really want a 20-comment subthread on every post about some stupid Brazilian thing? But that’s what we get. Worse than spam.

  12. akarlin says:

    #SSCSoWhite! 😉

    Amazing degree of correspondence.

    LessWrong stats:

    White (non-Hispanic) 87.3%
    Middle Eastern 0.6%
    Hispanic 2.3%
    Black 0.4%
    Asian (Indian subcontinent) 1.9%
    Asian (East Asian) 3.5%
    Other 4.1%

    SSC stats:

    White (non-Hispanic) 88.5%
    Middle Eastern 0.8%
    Hispanic 1.7%
    Black 0.5%
    Asian (Indian subcontinent) 2.2%
    Asian (East Asian) 3.4%
    Other 2.9%

    The national figures are slightly tilted more towards the US relative to LW as a whole, presumably because it is in English and pertains more to US issues.

    • Deiseach says:

      I’m delighted for the majority on here who appear to be in relationships – awww, so heartwarming! I hope you all are very happy and good luck and long lasting to your marriages, relationships or significant significances 🙂

      (So we’re not all commenting whilst dressed in our nightwear and located in our parents’ basements or spare rooms!)

  13. Anon. says:

    You can use “& gt;” “& lt;” (without the spaces) to make greater than/less than signs.

  14. Siah Sargus says:

    First off, shout out to that one person who strongly dislikes Slate Star Codex.

    I’ll never understand you, but I respect your decision to come here anyway. (Hypothesis: It was probably just a misclick.)

    Second off, why aren’t you lifting weights? It’s pretty unambiguously beneficial, not just for the body, but for the mind.

    Maybe next time, more categories could be added, like Running, Yoga, Boxing, Dancing, and whatnot.

    • Mark says:

      I severely injured my back weightlifting and decided never to do it again.

      • Noumenon72 says:

        At 40, getting injured is one of the actual functions of weight lifting for me — I use light enough weights that the pains aren’t permanent, and I use the pain as a guide for how I am moving incorrectly and learn to do it pain free.

    • Callum G says:

      I started lifting weights and taking my health more seriously about a month ago. I realised how interconnected my physical health was to other tasks and also realised that attractiveness is one of only a handful of privileges that I can control. Beat that akrasia.

    • Tibor says:

      I don’t remember what my answer was (I guess I included myself in the “weight-lifters”), but instead of lifting weights, I do bodyweight exercises and bouldering. I find it more satisfying, cheaper and natural than weight lifting and it also saves me the time spent going to the gym and back home (except for bouldering, I can’t do that at home and I actually enjoy doing it with friends rather than alone). Check out YAGOG – You are your own gym, if you’re interested in bodyweight exercise, it is a great book and a great mobile phone app (and no, I don’t get any money from them 🙂 ).

      • Gobbobobble says:

        Thanks for the tip, will have to check that out!

        For me personally, weight lifting has negative connotations due to the sorts of high school/college students who would partake in it, and I dislike gym atmosphere so don’t care to learn now and have to spend a bunch of money on my own equipment. And I have strangely positive memories of my dad running my brothers and I through the bodyweight/calisthenics exercises he did in boot camp ¯\_(ツ)_/¯

        • Tibor says:

          Yeah, I think a lot of the YAGOG exercises are actually very similar to what’s done in the military. And really, most of those exercises are quite well known, however YAGOG has a good structure and a premade 10 week long workout programme (with four levels of dificulty). You end up working out about half an hour 4 times a week in the first 6 weeks and then 5 times a week in the last 4 weeks (but the last two days in each of those last 4 weeks are tabatas and stappers which only take 16 and 20 minutes each, respectively…not that you could go on for much longer with those 🙂 ).

          I think it might be a good idea to visit a gym instructor every now and then and have him check your technique. When I started, my technique was lousy (which is not such a big issue as with weights – the movements are natural, so you are not likely to hurt yourself, it will just not be as efficient). Then I went to the higher difficulty level and eventually actually came back to the second easiest level – and did the stuff properly. Even something as basic as push-ups requires quite a lot of attention. If you do it improplerly, it will seem really easy, but you’re not really working out the muscles you should be.

          There are videos in the YAGOG app which help with this too, but of course it will be even more efficient if you have someone see you and point out your mistakes. You can also take a video of yourself working out.

          • Skivverus says:

            Not to be too grammar-nazi, but wouldn’t the acronym be YAYOG?

          • Tibor says:

            @Skivverus: Hmm, interesting, I’ve never noticed that 🙂 Yes, it is YAYOG. I guess my brain turned it into YAGOG for some reason. I had to check that it is actually not spelled wrong in the app and it isn’t, but I’ve always read YAGOG until now 🙂

            And I had to re-read your comment because I first read your correction as “YAYGOG”. Well, Yagog is an easier “word” to to pronounce than Yayog, so that’s probably why I changed it subconsciously. It often happens to me with names that I learn that a name is actually spelled differently than I though. People usually just read the beginning and the end of each word, but with names that can be tricky.

        • Scott says:

          FWIW, I feel the negative “gym atmosphere” is very much more perceived than real. Gym-goers in my experience have 99% of the time stuck to themselves unless you ask for tips/a spot, and 100% of the time been friendly towards and non-critical of other gym-goers, regardless of their skill level.

          • Nancy Lebovitz says:

            In my case, a friend took me to a gym in the hope of getting me as a gym buddy.

            For me, that gym (I don’t know whether it was typical) was a sensory disaster. The combination of unremitting noise and visual drabness made it a miserable experience for me.

          • Gobbobobble says:

            I agree with Nancy. It’s not about people being unfriendly, it’s just the effect of being around a bunch of other people who are also exercising. I prefer to zone out (or retreat into a podcast) and the audio/visuals of a gym makes getting to that meditative state much more difficult.

            *Especially* when they do something egregiously obnoxious like having mirrored walls (at least in the cardio area, I can see how they’d be useful for form-checking) or multiple non-muted televisions. Layer on some garbage gym-wide music and it becomes a cacophonous hell. And since I’d rather not sweat all over nice sound-cancelling headphones, I’d have to crank up earbuds to dangerously high volume to drown out the noise.

            Maybe I’ve just been to crappy ones?

          • suntzuanime says:

            I’ve given up on trying to baby my noise cancelling headphones and just budgeted a couple hundred dollars a year for replacements. All the situations where you actually need noise cancelling headphones are situations that expose them to wear.

          • Cypren says:

            Pick up a good pair of noise-isolating earbuds that seal the ear canal. They tend to be more effective than noise-cancellation in any environment where the noise you’re trying to cancel isn’t a fixed-cycle low-frequency (such as airplane engines).

            Many earbuds in the $50+ price range can be fitted with Comply foam tips to produce very solid and comfortable noise isolation at a reasonable cost. I’m a big fan of Shure earbuds in general, but this is an area where audiophiles can discuss and debate at length. I’m happy to share more if anyone is interested, but I’ll leave it at that for now.

    • alexsloat says:

      I’m bad at going to the gym without a partner, and my usual partner is bowled over with school now. I should buy some dumbbells for home use or something, but I don’t really have the space.

      • Callum G says:

        What Tibor was saying with calisthenics and bodyweight exercise could help, but yeah, it is hard to beat the motivational power of having a partner.

        • Tibor says:

          I used to go to the gym with two friends, it worked for a while bit then it created this cascade effect – whenever one of us was too lazy to go, it gave the other two an excuse not to go either. I exhibit the same problem with the bouldering, but that is something “optional”, so I am fine with not going every week (and there is a whole group of people I know who go there, so usually someone is coming and then I come as well).

          For me, working out at home has a motivational advantage as well. The YAGOG exercises are fairly short, they never exceed 36 minutes (but you work out 4-5 times a week). And you can start at any time, literally five minutes after you decide to work out, you can be doing it already and an hour afterwards you’ve already had a shower and changed clothes back to normal. When I go to the gym, it takes me at least 15-20 minutes to get there and the same amount of time to get back. Those are significant transaction costs and they increase my laziness a lot more than the lack of a training partner.

      • Tibor says:

        My German apartment is 19 square meters (including the bathroom), I have an electronic drumset, two wardrobes, a bed, a spare collapsible bed (which however still takes up some space) for visits, a kitchen corner and of course a table. And still there is enough room there for me to do the bodyweight exercises. Without the drumset, I’d actually have quite a lot of space.

        • alexsloat says:

          Thing about giant American-sized apartments – you have an ugly tendency to fill them up with stuff. Yes, I have the physical space to exercise, but large heavy things in any of our current open spaces would probably be a tripping hazard. Actual storage space is always at a premium, because I like physical things and accumulate too damn many of them, and because mess expands to fill the home available.

          • Tibor says:

            Well, that’s the thing. With bodyweight exercises you don’t need anything that isn’t in your flat already anyway. The things I use while working out are – a towel, a table, a door and a couple of books for when I do “dips” – I only have a drum stool and a chair with wheels which are unsuitable for that, so I improvise with the table, the stool and a couple of nonslippery books. This is a bit easier in my Czech apartment which is much larger.

            I actually have a problem filling apartments. My room in my parents’ house was very small, so is my German flat. My Czech flat is quite big but I don’t know what to put there to make it feel more cozy (and I have superfluous stuff there like a bar table). It could be because I don’t spend as much time there, but I kind of enjoy using up a small apartment to the fullest rather than having big rooms full of nothing. When I see those 8m^2 Hong Kong apartments, well, it still feels a bit too tiny but I am not horrified like a lot of Europeans seem to be. That reminds me of a time when I was looking for a temporary place to stay in Prague and I visited a Vietnamese family who rented their, well, closet. I was almost going to take it since it was really cheap, the rest of the flat was nice and I felt like it could be fun living with them, but unfortunately the roof was at an angle and I could not stand up straight in that room 🙂 They could, so maybe they did not realize the problem.

    • JonathanD says:

      I have a wife, three young children, a full time job, and side work I do in the evenings after the kids are in bed.

    • thepenforests says:

      I mean, it may be unambiguously beneficial, but it’s definitely not unambiguously enjoyable.

      I’ve tried lifting weights and I really, really dislike it. So it ends up in the (very large) category of “things that would be good for me but I don’t do because they’re unpleasant”

      • LCL says:

        Yes, this. It’s simultaneously painful and boring, I dislike being sweaty, and I dislike the social context of the gym.

        I do it sometimes anyway because I know it’s good for me. But I don’t understand how anyone actually likes it.

        • Noumenon72 says:

          As a teen, I liked the continuous improvement and body changes. As an adult, I like the sense that I’m restoring normal function and learning how my body’s supposed to work. Minute by minute that means I like doing the movements exactly right and focusing on the mind-muscle connection that trains my nervous system.

        • pdbarnlsey says:

          You and an earlier comment both mention not liking the gym’s social context. I’ve gymed weekly in the UK and Australia, though not in the US, and have found that, unless you wish otherwise, it’s basically completely atomistic. No one talks to anyone, most people drown out the (terrible) house music with headphones, the only interaction is to occasionally ask a person if they have finished with a machine.

          Does this not match your experience, or is your complaint a broader one about the look/feel of the gym experience and the other users? Do you feel unspoken judgements, maybe?

          Just curious – you’re obviously welcome to not enjoy things and to not do the things you don’t enjoy – it’s a real source of centre in my life.

    • Fossegrimen says:

      I do ultra-endurance cycling and find that lifting weights tend to add muscle mass that I don’t want.

    • Dr Dealgood says:

      I’m currently lifting weights, and love doing it, but bodyweight exercises have historically had better results for me.

      When I lift weights, I put on a lot of muscle but don’t lose much (if any) fat unless I’m also on a keto diet. Even now where I’m doing an hour of cardio followed by an hour of lifting I’m still putting on weight and might have to go back into keto.

      When I do bodyweight exercises I don’t gain any muscle to speak of but drop fat reasonably quickly. Even with no cardio, doing a quick twenty to thirty minute interval workout before breakfast every day is really helpful.

      So it’s not quite as straightforward as you imply. In terms of health and attractiveness to women, bodyweight is probably a better choice for me. But I enjoy lifting a lot more and like having a muscular body even if it comes with some fat.

    • Brad says:

      Anyone want to run the correlations for weight lifting and ideology?

      • Anon. says:

        PoliticalSpectrum Average of Weightlifting
        1 2.07
        2 2.06
        3 2.05
        4 2.19
        5 2.34
        6 2.31
        7 2.44
        8 2.37
        9 2.42
        10 2.54

        Correlation = .1033

      • WashedOut says:

        I socialise within strength-training circles (as well as in non-gym-going circles) and these are my anecdotal findings:

        1) Weightlifters typically range from politically conservative to libertarian
        2) They care more than the average person about border control and national security
        3) They care about maintaining traditional masculine and feminine roles, and preserving very black/white definitions about gender and sexuality
        4) Very low in compassion; low in openness; high in industriousness, conscientiousness

        The general ideological narrative among avid weightlifters could be summed up in broad terms as:
        “It’s up to each individual to look out for and improve themselves. The world sucks by default and you will suffer if you don’t make yourself good enough to cope, and that’s just how it is. I won’t feel sorry for you but if you want guidance I can point you in the right direction.”

        • Tarpitz says:

          That is possibly the purest expression of Magic: the Gathering’s black colour philosophy that I have ever encountered, including from Mark Rosewater.

          Do they also pay life to draw cards, kill things really efficiently and practice necromancy?

    • Deiseach says:

      Second off, why aren’t you lifting weights? It’s pretty unambiguously beneficial, not just for the body, but for the mind.

      As God is my witness, as God is my witness they’re not going to lick me. I’m going to live through this and when it’s all over, I’ll never lift anything heavier than a shoebox again. No, nor any of my folk. If I have to lie, steal, cheat or kill. As God is my witness, I’ll never lift anything heavy again!

      Did my share of lifting, carrying, toting, digging, etc etc and bloody etc whilst a child to young adult, be damned if I’ll lift heavy weights recreationally or while there is no-one standing over me with a whip forcing me.

    • roystgnr says:

      Only *one* person chose Strongly Dislike? That’s almost a perfect bait for determining Lizardman’s Constant, and instead of 4% we got 0.02%? That’s more impressive than (and lends a bit more credibility to) the self-reported intelligence data, frankly.

    • Gazeboist says:

      I do not lift because I am completely unable to manage my time. I do intend to pick up HEMA over the summer though.

    • Acedia says:

      Cardio exercise feels good and natural, as if my body “wants” to do it. Lifting weights feels really bad and uncomfortable and I never got any sense of “flow” from it the way I do from cardio, even after doing it diligently for a year and becoming noticeably stronger. So I stopped.

      I know people who say the exact reverse of this. Not sure what causes the difference.

    • limitedimaginationrambling says:

      There are a few reasons I can come up with from extensive personal experience.

      One of the large over-arching reasons is that it actually is a net negative to my mental well-being. Weightlifting (and generally, all exercise for the sake of exercise) is stunningly boring. It fails to offer any mental or emotional engagement whatsoever, which makes it a matter of pure willpower to simply proceed through whatever routine I’m attempting. For me, there’s also no feeling of reward or accomplishment when the routine is complete, no matter how much of a struggle it might have been. To this day, I’ve never experienced the endorphin rush that causes some people to happily exercise with enthusiasm in anticipation of feeling good afterwards.

      Exercise also feels like an incredible time commitment due to its unpleasantness and the degree of preparation required to transition into doing anything else. I don’t know how some people exercise in the morning before work, because if I tried that, I’d be exhausted and grumpy for most of the rest of the morning.

      Gyms are also anxiety-inducing places, where I feel I don’t belong, and gym locker rooms/showers typically activate my disgust response. Incidentally, I feel it must be said that no man really needs to use the hand-dryer or their own hair dryer on their undercarriage at a locker room sink. This is not a place I want to visit on a regular basis.

      While more regular exercise would be good for me, exercise for the sake of exercise simply lacks a sense of benefit due to how unpleasant the task truly is. Most other unpleasant tasks or chores at home or in the office at least enable other things to be done, or directly improve the aesthetics of a place or thing that I have to interact with on a regular basis. Exercise would make my body look nicer, but only through an enormously greater commitment of time and effort that I could spend doing any number of other things.

      This isn’t to say I don’t enjoy physical activity otherwise. I love to hike, snowshoe, ski, etc. Those are engaging because of what I see, hear, and experience beyond the dull, repetitive tasks of ordinary exercise. It’s also fairly trivial to introduce other ways to get in a decent degree of daily activity; I use a printer on the first floor of my office instead of my floor, so I end up doing at least 20 flights of stairs on a typical workday. When I’m at home wasting time on the internet, (like now) I’m using a standing desk.

      Why don’t I lift weights/exercise? Mostly because I’ve not found a way to do so without feeling miserable.

      • nimim.k.m. says:

        This matches my experience quite well. Only that I’ve not visited gyms many times so my experience is pretty limited (I vaguely recall a couple of trips as a part of HS PE curriculum?). Also the prices listed seem very expensive, especially anything involving instruction, even the university affiliated chain with discount prices for students.

        So I’ve intermittently tried starting bodyweight exercise regimen [1] since teenager years, always failing to achieve anything. It’s absolutely boring, and after a couple of weeks of initial activity ran on sheer willpower I lose motivation (because I don’t have that much willpower). I’ve tried using music (fast up-tempo stuff I like) during the exercises and it works for a while, but in the end I usually sooner or later notice that I’ve stopped doing any movements and just listening to music.

        I’ve found controlling calorie input than calorie usage much much easier. Skinny nerd isn’t of course not very attractive on the scale of male attractiveness, but it’s marginally better than an overweight nerd.

        [1] Also, it’s surprisingly difficult to find a set of instructions how to do movements that are a) comprehensive enough for a person with no idea at all what they are supposed to be doing and b) also, preferably coming from trustworthy sources instead of sketchy-looking internet sites filled with ads of more or less questionable looking products. Looking at all the stuff I can find on the Google instantly activates the same mode in my mind I use to sort out through the spam email and other trash…

    • Subb4k says:

      >Second off, why aren’t you lifting weights? It’s pretty unambiguously beneficial, not just for the body, but for the mind.

      Because it’s way more fun to lift people? (at least IMO)

      In more general terms, there are plenty of sports out there, and a lot of people prefer different physical activities. Some people want to be all strength/endurance, while some want to exercise their balance/dexterity as well. Some people want calm and focus, other energy and high activity. Some people want competition, others cooperation, and yet others want to be by themselves. Some people have money to burn on equipment or gym memberships, others not so much.

  15. sty_silver says:

    I’m going to make a “you should do more of X” comment: you should talk more about AI risk. 10% at “very concerned” is disappointingly low. Consider the scope of the issue, the fact that many people have a favorable view of you (based on the survey), and that the AI persuasion experiment yielded measurable success.

    • Joe says:

      Something something Aumann. Doesn’t learning that a surprisingly high number of people disagree with you shake your confidence that you’re correct and they’re wrong even a little?

      • sty_silver says:

        I think you mean “wrong” as in “wrong about believing that the risk of superintelligence is real?”

        I might already not believe that the risk of superintelligence is real. But this does not imply thinking it is a minor issue. A 10% probability assigned to the risk being real is more than enough to make it one of the most important issues in the world right now.

        Note that you not need to believe in computer intelligence to care about this issue, as there are multiple paths to superintelligence. Assigning a 99% chance to all paths collectively failing is, I think, utterly unreasonable, and I don’t believe most people who claim to be skeptics would make a bet corresponding to their supposed confidence if their life were on the line.

        As an aside, it is a given that Scott thinks of it as a serious issue, so my confidence isn’t really relevant.

        • roystgnr says:

          Wait, what are the other likely paths? I’m sure we can raise the world’s IQ 50 points by fiddling with genes, but superintelligence is mostly an existential risk because of the possibility of a “hard takeoff”, which seems to only be a possibility in the artificial intelligence case.

          • sty_silver says:

            Whole Brain Emulation and technical enhancement of biological brains.

            I don’t agree that superintelligence is mostly an existential risk because of the possibility of a hard takeoff. I’d agree that the fast takeoff scenario is what we should worry about, but only because a fast takeoff is very likely. A slow takeoff might be more dangerous, because it still opens up doors for extremely dangerous technology, without necessarily providing means to handle it.

            Imagine if a bunch of competing powers gaining access to far more advanced weapons than we have now, with smarter-than-human but not superintelligent AI at their side but humans still in charge. That does not look like something we should hope for to me.

          • Joe says:


            I agree AI will be a seriously big deal, but it’s a much broader issue than just a ‘risk’. Given humans evolved to have morally valuable features like emotions, thoughts, hopes, dreams, consciousness – and animals seem to have evolved increasingly more morally valuable features quite tightly in line with their intelligence – it seems entirely plausible that AIs will be complex, messy, sentient creatures just like us, not a mindless blob of Smartonium. In that case a universe full of ‘unaligned’ AIs may not be a morally worthless dead wasteland but a vast civilisation, containing enormous amounts of utility.

            Honestly this basic possibility – AIs as conscious beings (at least some of them) – seems like it ought to be the default assumption, until we actually have any examples of entities with general intelligence but no morally valuable mind features. That the standard AI-risk stance on the AI consciousness question is to dismiss it as a separate and irrelevant issue is quite frustrating, since it seems to have such enormous, and enormously relevant, implications – depending on the answer, can shift potential futures from being terrible to wonderful, or vice versa.

          • sty_silver says:


            I’m one of those people who tries to drown all discussions about consciousness, because I think it’s a waste of time. But most people who do this may not do a good job of explaining why that is.

            Given humans evolved to have morally valuable features like emotions, thoughts, hopes, dreams, consciousness – and animals seem to have evolved increasingly more morally valuable features quite tightly in line with their intelligence – it seems entirely plausible that AIs will be complex, messy, sentient creatures just like us, not a mindless blob of Smartonium.

            I bolded the relevant parts. Animals and humans have evolved, AI has been programmed. There is simply no reason to expect that things like sleep, hunger, friendship, family, fear, dreams, a desire to live, common sense, hopes, morals, you name it, are going to be relevant concepts in/for an AI. In fact, it is safe to assume that those things won’t be there unless they are either programmed into it, or the AI is given an incentive to acquire some of them.

            The problem this goes back to is the usage of the term “intelligence.” People always tend to think that, because human intelligence comes with certain features, the AI must automatically also have those once it becomes similarly intelligent. But this is actually not at all the case. There is nothing fundamentally different about a fully general AGI and a pocket calculator. Both are fully deterministic systems that work in exactly the way you programmed. It is more accurate to think of “intelligence” as “problem solving capabilities” in the context of AI. The first AI undergoing the .-explosion will have really high problem solving capabilities. It might still be incredibly stupid in terms of human intelligence (depending on how you define it).

            This does not even mean that the AIs won’t be conscious (I expect they eventually will be, but only if we solve the control problem). It just means that, even if the first AI were conscious after undergoing the intelligence explosion, it would not change anything except giving it moral value. It would still be a fully deterministic system doing exactly what is programmed into it. That’s just how programming code works.

            I think the only way these kinds of concerns become relevant is if a) the AI arrives in a form other than fully artificial computer intelligence, or b) we choose to give it a goal like “modify yourself to be similar to humans without the bad parts,” or maybe c) if we end up getting to AI by simulating evolution.

          • Nancy Lebovitz says:

            “Both are fully deterministic systems that work in exactly the way you programmed.”

            Computers do not work in exactly the ways they are programmed.

            They are subject to natural law, but this doesn’t they’re predictable from knowledge of their programs. Or at least not predictable by humans. And probably not predictable by other computers.

            Computer programs live in a hostile environment. They’re vulnerable to physical damage in the computers they’re running on, incompatibility with other programs, and malware.

            Maybe a GAI could be so well-designed that it wouldn’t be affected by those factors, but I’m not counting on it.

            As for whether a GAI would be similar to humans, my guess is that it won’t be very similar. I *think* intelligence is possible without sleep but I’m willing to be surprised.

            I’m guessing that a sufficiently advanced computer program needs a self-model which is something like consciousness so that it can keep some track of whether it’s being damaged.

          • rlms says:

            I think (c) is a a much bigger possibility than you realise. Consider the rapid progress in the field of AI now. It isn’t being done by simulating evolution, but it also isn’t being done standard programming that results in transparent code. I don’t think current machine learning techniques will be straightforwardly extended to get GAI, but it seems unlikely that GAI will be made through a simpler process than that currently being used.

            I think you are also being overconfident if you believe non-determinism is necessary for consciousness. It clearly isn’t sufficient (bacteria are not conscious), so it isn’t equivalent, and it doesn’t seem obviously necessary. After all, classical physics is deterministic, and quantum effects largely seem irrelevant in the human brain. Some people (Roger Penrose) argue that those quantum effects do give rise to consciousness, but that’s a minority opinion.

          • sty_silver says:

            I think you are also being overconfident if you believe non-determinism is necessary for consciousness.

            You misunderstand me. I don’t believe that at all. I believe that a computer will necessarily be deterministic, whether or not it is conscious, hence why consciousness is not the problem. I’m fairly agnostic about whether or when it will be conscious.

          • Gazeboist says:

            fully general AGI

            You should consider defining this term here. It seems important.

          • rlms says:

            OK, but what does determinism have to do with anything? Humans are equally deterministic.

          • Joe says:


            People always tend to think that, because human intelligence comes with certain features, the AI must automatically also have those once it becomes similarly intelligent.

            I’m not saying these mind features would just appear on their own, for free, just by dint of the AI being of a certain level of intelligence. It’s the other way round: I’m saying they would be intentionally built into the AI, whether it is built by humans, or it, in some sense, evolves. They’d be put there for the same reason evolution put these features in our minds: because they are adaptive, they are functional components of an effective mind design. Not just random useless junk.

            We do in fact have a reasonable idea of the function of mind features such as emotions, consciousness, sleep, family, friendship, desires, morals. Yes there might well be different ways of solving the problems these features are designed to solve, but if we want to get a better idea of what future AIs will be like, that’s the kind of question we should be investigating, along with “How similar are other possible approaches?” and “What are the advantages of each kind of approach?” and “Where might this approach, or something like it, be a better solution than alternatives?”. The alternative is to assume there’s some very simple, fully general ‘intelligence algorithm’, which is possible, but doesn’t seem at all obvious.

            It seems like you’re equivocating between abstraction and implementation when you talk about ‘general problem solvers’: this might be a reasonable very-high-level description of what a successful AI would be able to do, but it says nothing about how that functionality would be achieved. Unless we have a particularly good reason to expect the aforementioned general intelligence algorithm to exist, surely the default assumption should be that AI minds will have to be built with many relatively-small components designed to do specific jobs, like animal minds (including ours) seem to be.

          • sty_silver says:


            Humans are equally deterministic.

            Excactly! The consciousness question doesn’t change anything about the goal orthogonality, and is therefore no argument against the ways of AI safety research. That was the whole point I was trying to make. I said that the AI was deterministic as part of this point, because I felt like that was misunderstood.

          • sty_silver says:

            I’m saying they would be intentionally built into the AI, whether it is built by humans, or it, in some sense, evolves. They’d be put there for the same reason evolution put these features in our minds: because they are adaptive, they are functional components of an effective mind design. Not just random useless junk.

            Note that this is a prediction about the actions of programmers.

            I think you need to show me evidence here. If we agree that those things don’t appear magically, then they are missing by default, so the burden of proof is on you to show that leading AI researchers are in fact planing to do this, or are going to do it despite not planing to now.

            To briefly return to my initial appeal, let me just say that I wouldn’t at all mind if this were discussed in a blog post. If current research is misguided, it’s incredibly important to shed light on that.

          • Joe says:


            I think current research is very open to the idea that AI designs will replicate many of the features of human minds. For example, see this paper coauthored by Nick Bostrom, which asks artificial intelligence researchers from different backgrounds and specialities various questions regarding future AI prospects. Relevant here is the question that asks “In your opinion, what are the research approaches that might contribute the most to the development of [human-level AI]?”, with respondents selecting all they see as relevant. The avenues most picked were ‘cognitive science’ (by 47.9% of respondents), ‘integrated cognitive architectures’ (by 42%), and ‘algorithms revealed by computational neuroscience’ (also by 42%).

            Also see this summary of an interview with an AI researcher (Joscha Bach), in which he describes the challenge of building human-level AI as discovering the mechanisms of the brain we don’t yet understand and working out how to build them into the AIs we write. This perspective, of comparing the functionality of AIs to that of human minds, and looking at what’s missing to see what we still need to achieve, seems quite common within AI research.

            And to repeat, I think the evidence from other animal species is very telling. If we saw no particular correlation between species we think of as more conscious, and species we think of as more intelligent, then the idea that you can have a human-level intelligence that’s not conscious would seem more plausible. As it is, there is a pretty strong correlation. (And consciousness seems to have evolved more than once.) It’s possible that this may not hold for AIs, but until we have any reason to expect that I don’t see why we shouldn’t expect adaptive future AIs to be conscious too – at least some of them. (Plausibly there will be many different levels of AI for different tasks, depending on what functionality is needed, and the lower of these might well be mindless automatons.)

    • alexsloat says:

      I’ve heard lots about AI risk, and I don’t buy it. AI risk proselytizers seem to think we’re going to go to sleep in the modern world and wake up with Skynet, and frankly I find that a hilariously unrealistic concept – even if AI gets to be self-modifying(which I doubt, as I expect that it is mathematically impossible for a system to be intelligent enough to understand itself fully), you can’t have anything like an infinitely fast intelligence explosion, because hardware limitations will prevent it.

      • FeepingCreature says:

        None of those are positions that represent the mainstream of AI safety.

        The worry is not any sort of “infinitely fast” explosion, the worry is an “unmanageably fast” explosion; or rather, what Bostrom calls the “treacherous turn”, where an AI behaves in a Friendly fashion up until it believes itself to be unassailable. (Probably once we’ve connected it to all our industrial nanoassemblers or w/e.)

        Regarding the self-modifyingness, MIRI have done a bunch of work on that; you should subscribe to their newsletter! see the “subscribe” link on the right. They have an RSS too! As it stands, that seems an awfully unreliable belief to hang a certainty of safety from. It should certainly seem at least plausible to you that AI can build stronger AI.

        • alexsloat says:

          To be fair, if humans can build something smarter than humans, a superhuman AI could presumably do the same thing. But doing so reliably, rigorously, and efficiently is unlikely.

          Also, a superhuman AI would view a super-superhuman AI as a threat just like how we would – it’s smart enough to know that it’s eclipsing itself. So it’s left with either experimenting on itself(deeply risky when you can’t know what you’re doing) or facing exactly the same issues that humans worrying about AI safety face.

          • LHN says:

            As with all technology issues, Silver Age superhero comics showed how this would play out: the people of the planet Colu built supercomputers, who took over the planet. They built a robot agent, Brainiac, who was upgraded by Lex Luthor (from 10th to 12th level intelligence!) and so surpassed them and became a galactic threat. Brainiac’s “descendant” Brainiac 5 in turn invented Computo, which outmatched him and threatened to take over the universe, and so on.

            (With Luthor and Brainiac 5 both experiencing failed attempts to box their respective improved AIs.)

            Has MIRI made any efforts in the direction of developing the only recorded successful countermeasure, superhumans in colorful skintight outifts?

          • Jiro says:

            I don’t remember the Luthor story and if it existed at all, was not present in most tellings.

            Brainiac 5 was, in most retellings, a desendant of Brainiac’s adopted son (who fought Brainiac). He was a standard comic book human-appearing alien, not an android or robot. (There was a brief period of a couple of years in modern times where Brainiac was not mechanical and during those years Brainiac 5 was really a descendant.)

          • LHN says:

            @Jiro: Well, I did specify “Silver Age”. Brainiac was first revealed to the readers to be a computer in 1964, when Luthor used a space thought scanner to uncover his origins.


            (Prior to that, the character had been a standard green-skinned humanoid alien whose schtick was having an impenetrable force field. The change was reportedly because they got into a trademark dispute with the manufacturer of a computer kit that used the name “Brainiac”. DC defused the issue by giving the manufacturer some free publicity.)

            That story introduced Colu and the Computer Tyrants, and retconned Brainiac 2 to an adopted son to keep the LSH’s Brainiac 5 unchanged
            (Previously, Brainiac 5 was presumed to be Brainiac’s natural descendant, which is why on his first meeting with Supergirl he was so worried she’d hold it against him.)

            In that story, Luthor broke Brainiac out of jail and performed the aforementioned intelligence upgrade. (And ever after, both Brainiac and Brainiac 5 would be described as having “12th level effector” brains, where Kryptonians are 10th and baseline humans I think 6th.) They teamed up and they basically won: beat Superman, then were briefly captured by the Kandorians but negotiated a release.

            (Brainiac erases Luthor’s knowledge that he’s a computer, with the idea that it’s supposed to be a deep secret. But everything I remember afterward has the fact well-known to Superman and his supporting cast.)

            That story pretty much defined Brainiac for the rest of the Silver Age, and to a large extent for the Bronze (during the later part of which he turned silver and started talking in BASIC), before being wiped from continuity in 1985. Since then, Brainiac’s been a lot of things– a circus mentalist possessed by an alien, Krypton’s Internet– but he generally tends to revert to the Coluan AI first established in this story.

            (You’re of course right that B5 was never an AI, and I didn’t say he was. He was an AI experimenter starting at the original Brainiac’s own amplified baseline intelligence, who had a number of fairly spectacular failures keeping his own more capable AI under control– one of those failures killed a Legionnaire– along with the occasional apparent success/ticking bomb.)

          • Also, a superhuman AI would view a super-superhuman AI as a threat just like how we would

            Why? Would it necessarily have our value system? Can you tell what an unspecified AI would do?

      • drethelin says:

        You don’t need to full understand yourself to be able to self-modify. Current advances in artificial intelligence tend to be highly modular: specifics like face recognition, navigating unfamiliar terrain, mapping, pressure sensitivity, etc. If you can be smart enough to understand submodules of your intelligence you can upgrade them without having to upgrade the whole thing.

        Hell, humans don’t even understand their submodules but can still train themselves to become WAY more effective at many subtasks with deliberate practice, which a digital entity would be able to do way more reliably.

        • Skivverus says:

          The “perceptual control systems” article discussion seems relevant here.
          Agreed, you don’t need to fully understand yourself to self-modify; more understanding does (not-quite-tautologically) tend to produce better results, though.

        • Cypren says:

          In that same vein, note also that humans do not fully understand most things that they work on. Our most complex modern achievements are not the creations of a single mind, and no single mind in existence can fully hold and understand all of the parts of, say, a Nimitz-class carrier, an F-35 or the Google search engine. Instead, we modularize, divide and pool intelligence through cooperative effort to create products which are larger than what a single brain can comprehend or manage.

          There’s absolutely no reason to believe that AI can’t do the same, and even more easily, since it can start from a certain baseline of knowledge and then clone itself infinitely to make as many new “workers” as necessary, limited only by the hardware it can find to run them on. It may be mathematically infeasible for any one intelligence to fully understand itself well enough to substantially alter its own design, but entirely within the capacity of many copied intelligences to learn and evolve in different directions to understand subcomponents of the original and then pool their acquired knowledge to substantially improve upon its design. Then they can copy the new design and begin the cycle anew.

      • James Miller says:

        “Skynet” could suddenly come about from an AI that was slowly getting smarter but hiding this fact from humans until it was smart enough to take over.

    • Gobbobobble says:

      Has anyone run the numbers yet on AI Risk Concern vs Profession? I’m curious how the (rather large) cohort that actually works with computers skews.

    • suntzuanime says:

      You could ruin SSC in an awful hurry by turning it into an outlet for propaganda about Important Issues.

      • sty_silver says:

        Is it propaganda if it’s true and without self-interest?

        But the concern is real. Still, I wasn’t suggesting doing it all the time. One post per season or something about just AI would already make me happy. Of course, blogging like this is art. I understand that you don’t just get to choose your topics.

        • Ilya Shpitser says:

          Worrying about people’s immortal souls is also valid and without self-interest, hence:

          “Excuse Me Sir, Do You Have a Moment to Talk About Jesus Christ?”

          • sty_silver says:

            If you do in fact believe in eternal torture, I think you should try to convert as many people as possible. In fact, it’s an even more extreme case: everything you do should be in service of converting as many people as possible.

            At the same time, you should not believe in eternal torture because the evidence is not there. It’s lucky for us that extremely religious people don’t tend to have coherent utility functions obeying laws of probability.

          • Skivverus says:

            Not necessarily; SSC is arguably full of people who will remain unconverted by the methods that work on “as many people as possible”.
            Quantity matters, sure, but that doesn’t mean your goals are best served by spending all your time optimizing for the center of the curve.

          • phisheep says:

            If you do in fact believe in eternal torture, I think you should try to convert as many people as possible.

            Ah, but it might not be an effective strategy to bang on about it *all* the time, as people may run away before you manage to save their souls.

            I wonder what the most effective strategy would be? Probably, for house-to-house proselytisation, very similar to the strategy a burglar would use before moving on to another area.

          • Ilya Shpitser says:

            “At the same time, you should not believe in eternal torture because the evidence is not there.”

            I agree, we shouldn’t believe in things if the evidence isn’t there.

          • Gazeboist says:

            You should not believe in eternal torture because, if you hear an argument for the possibility of eternal torture, someone is trying to mug you.

          • James Miller says:


            What if the argument comes from my own brain?

          • IrishDude says:

            What if the argument comes from my own brain?

            As Richard Feynman said “The first principle is that you must not fool yourself – and you are the easiest person to fool.”, so I’d only suggest carefully examining your brain’s arguments. The higher the stakes for being wrong, the more carefully I examine my beliefs.

          • Gazeboist says:

            What if the argument comes from my own brain?

            Well, the real point is to avoid getting mugged. If you prefer to believe the argument, you can; you just need to be aware that someone (perhaps an actual external person, perhaps simply your worse self) will probably try to use it to mug you.

            In the specific case of religion (more properly, proselytizing, even in a non-religious sense, but Shpitser brought up religious proselytizing), the “mugging” is usually someone causing believers to be obnoxious to other people. I’ve generally found that “don’t be a dick; treat people like adults; show some awareness” is an adequate rule for anyone to avoid that particular harm, pretty much no matter what they believe.

            Sty is rather badly wrong even if the conversion method is effective, is my point.

          • sty_silver says:

            Ah, but it might not be an effective strategy to bang on about it *all* the time, as people may run away before you manage to save their souls.

            Sure, I agree with that.

  16. arandur119 says:

    I didn’t know enough when taking this survey for it to bother me, but having learned in the interim I’m retrospectively bothered at the description of communism as “complete state control of many facets of life.” Sort of erases bottom leftists. I wouldn’t be surprised if a few anarchocommunists went with “social democratic” as the closest available option.

    • alexsloat says:

      A lot of communists object to it, but in practice it always works that way. I don’t regard anarchocommunism as serious for the same reason I disregard anarchocapitalism and anarcho-I-like-to-smash-things-ism. It’s fundamentally unstable – with no power structures, there’s no barrier to people forming new power structures for their own benefit. Anarchy has a shelf life of a holiday weekend, and thereafter becomes a dictatorship.

      • IrishDude says:

        I don’t regard anarchocommunism as serious for the same reason I disregard anarchocapitalism and anarcho-I-like-to-smash-things-ism. It’s fundamentally unstable – with no power structures, there’s no barrier to people forming new power structures for their own benefit.

        I don’t know exactly what you mean by power structures, but anarcho-capitalism has people willing to do violence on behalf customers, if necessary. They gain this authority to act on their clients behalf voluntarily, and in doing so gain power. There are barriers to people forming new power structures for their own benefit if it harms others, as others have designated security with power to stop the harms.

        To see this, note that at the global level there’s anarchism, with no supreme rulers, but national power structures are kept in check by other national powers so that we broadly have peace and stability across the globe.

        • nimim.k.m. says:

          What if the current global system of nation-states where you mostly have exit rights (but not necessarily enter rights) is more or less the nearest thing of the stable equilibrium of an anarcho-capitalist system when implemented in practice?

          • IrishDude says:

            The current nation-state situation could be a local optimum, but I’m doubtful it’s the only stable equilibrium. There’s been 10 new countries created in the last 20 years, and I suspect there’ll be more fracturing down to smaller polities in the future. 100 years from now I think it likely at least some U.S. states will have peacefully seceded. Once polities get small enough, it will be easier for a Free State-type movement of ancaps to establish a beachhead, dissolve government, and give the political system a chance to compete in the modernish era. I could be wrong, but in the long-run (next 100 years, say), it seems plausible to me.

            BTW, if you can’t leave your nation because you’re surrounded by political borders with men with guns keeping you out of their territory, you may have de jure exit rights but it’s not de facto. Enter and exit rights work together.

            EDIT: Here’s Marc Andreesen saying “The world is going to see an explosion of countries in the years ahead” and pointing to technology, particularly communication technology, as a causal factor. I buy this story, though the timeline is much fuzzier to predict.

          • nimim.k.m. says:

            Oh, the exact number of states and their borders are very local perturbations. I meant in the sense of the Westphalian / post- WW2 arrangement of the world structured into a state-like entities (bureaucracy and military and police and other bells and whistles — organized large security provider claiming a monopoly of a certain territory) that are very recognizable to us as states instead of, say, a libertarian myth-era Iceland or pre-historic nomadic tribes.

          • IrishDude says:

            I think having 10,000 nations instead of 196 would lead to more interesting political arrangements, with at least some of those nations looking pretty dissimilar to what nations look like today. Smaller polities allow clustering of rarer populations and subsequently those rarer populations can exert more political influence.

            It’s the Free State Project effect, where committed libertarians have moved to New Hampshire and moved the needle on a range of libertarian issues. I can see something similar happening with committed ancaps if they could ‘take over’ one of the 10,000 nations, and then work on dissolving the state. From there, if ancapism works well in practice, they’ll attract immigrants and also provide a model for the other smaller polities about how to promote human flourishing and economic growth, creating a copy-cat effect and increasing the number of ancap arrangements in other polities.

            Even if the current nation-state landscape remains, with more or less 200 countries, I think two factors are likely to lead to weakening states in the future:
            1. Many, many nations are likely to go bankrupt due to public debt and therefore have difficulty performing their state functions. This is likely to create pressures to devolve state functions down to more local levels and/or to privatize functions.
            2. Technology like Bitcoin, 3d printing, and Uber is likely to continue to create alternative structures to the state that allow people to opt-out of the state without moving. A more impotent state, particularly in the West where people are generally peaceful toward one another, can create space for privatized government functions and more ancap-like arrangements.

            In 500 years, do you think it certain nation-states will still be ruling the entire globe?

      • Tekhno says:

        A lot of communists object to it, but in practice it always works that way.

        Yes, but then you’ve left the realm of describing people’s self-identified ideological groupings and entered the world of debating the outcome of the ideologies. At the very least, there are differences in how state socialists and anarchist socialists behave. The ideology has an effect on what they do.

    • semiel says:

      I’m in that general quadrant (though I don’t consider myself a “communist”), and I just picked “other”.

  17. Elijah says:

    I’m a little taken aback by the HBD results- 27.7% favourable, 27.3% neutral- given that the survey responses are otherwise skewed towards the stereotypically leftist position on every issue(except the minimum wage).

    I wouldn’t be surprised if SSC commentators had the highest overlap between people sympathetic/neutral on immigration and people sympathetic/neutral on HBD of any community on the internet.

    • tmk says:

      I wonder how many just have a positive association to the word “diversity” and failed to notice the actual meaning of the term. Wasn’t the term deliberately created to appear inoffensive?

      • Deiseach says:

        It may be a case that people both believe that there is a difference in biological attributes between general populations and that this does not mean Group A, very much less that Individual B who is a member of Group A, is inferior, lesser, not entitled to full human rights or any of the rest of it.

        I had no idea that drinking milk was racist, for example. If someone is going to reveal the dread truth to me that milk-drinking mocks lactose intolerance in non-white populations, which is something that exists enough to be a real trait and an example of Human Bio Diversity and something that should be taken into account, they cannot then turn round and say that measured differences in group population intelligence do not exist and should never be considered.

        Mind you, I am not hugely convinced of the validity of the measurements as often quoted; I think there was an ideological bent in some tests and the only way we’ll get any useful data is proper testing. But since it’s racism to even think about such a thing, that’s not going to happen. So we’ll be stuck with the likes of Richard “Southern Irish are IQ 96 why no me being a Unionist has nothing to do with that” Lynn and his “work” as the type of material being used.

        • Nancy Lebovitz says:

          I think it’s hilarious that white supremicists used to say “We are mighty warriors!” and now they say “We can drink milk!”

          These are degenerate days.

          • Deiseach says:

            I would have thought the ostentatious milk-drinking was a joke, even if it is intended as “white supremacist” (listen, which would any of us prefer: a bunch of white supremacists taking the piss out of Shia LaBeouf’s camera by necking pints of milk, or Combat 18 beating up, assaulting and even killing people?)

            But seemingly some people are so earnestly endeavouring to make us aware of the institutional racism upon which the dairy industry Western civilisation was founded, that they end up making themselves look faintly ridiculous:

            What about racism? White nationalism? If you’re having trouble finding the connection between these institutions and milk, you’re not alone. You, along with the rest of the nation, have been so accustomed to hearing the benefits of milk that you probably didn’t even realize the subtle racism hidden in our health facts.

            It may not surprise you that the United States was founded on racism. That every institution we uphold has racist roots that are sometimes difficult to catch and even harder to fight against. This phenomenon affects our voter ID laws, state testing and, yes, even our federal dietary guidelines. But how can our health guidelines, a system meant to be built upon scientific fact alone, have racist messages? Where there is a deep-rooted tradition to suppress an entire race’s existence, there’s a way.

            The federal endorsement of milk in American diets contributes to the problem by uncritically pushing people to drink milk, despite the potential detriment it has on non-white people’s health.

            Yes, we were all encouraged to drink milk in order to suppress the African-American people. Even those of us not living in the USA because, after all, “Where there is a deep-rooted tradition to suppress an entire race’s existence, there’s a way”.

            As for that horror movie:

            This odd form of white supremacy also received cinematic attention through Jordan Peele’s horror movie “Get Out,” a movie that highlights racism in a post-racial America. Peele artistically addresses the new medium of hate with one of the film’s most eerie scenes, which shows a white woman meticulously sipping milk from a bendy straw. The scene would have gone unnoticed in the movie, but audiences were forced to notice the long, drawn-out frame of the woman taking a sip; Peele wanted people to notice.

            Well, now at least I know one more movie not to waste my time going to see. Honestly, if I were in the audience, I’d be going “Who the heck over the age of four drinks milk using a straw?” and not “Argh, what a chilling indictment of the hidden racism within our very dietary choices!”

          • Cerebral Paul Z. says:

            It might be worth bringing back the old milk marketing campaign we used to have in the US, featuring celebrities sporting milk mustaches. This time, they could make them all toothbrush mustaches.

          • Nancy Lebovitz says:

            That was a weird scene in a bunch of ways, but I don’t think it especially undercuts the movie which is an excellent example of psychological horror.

            Deiseach, I don’t know if you like horror, but if you do, I recommend the movie. It’s not just about race. It’s also about fear of people.

            Some spoilers follow.

            Peele says he didn’t know about the white racist use of milk as a symbol when he wrote the scene.

            On the other hand, he did think that drinking milk through a straw while eating dry Froot Loops was a teenage sort of thing to do, symbolic of Rose’s arrested emotional development. I never would have guessed any of that. I’m enough of a food snob that Rose eating Froot Loops just seemed weird.

            When I was watching the movie, my thought was that the standard scary woman in movies is the blonde driven by emotional possessiveness.

            No handy symbology for murderous brunette who’s part of an evil conspiracy, so go with android. Perhaps I should have thought Nazi, but I didn’t.

          • James Miller says:

            I can create poison gas by drinking milk, which makes me an even mightier warrior than those lactose tolerant white supremacists.

          • Hyzenthlay says:

            Yes, we were all encouraged to drink milk in order to suppress the African-American people.

            I remember hearing about a Domino’s pizza lawsuit once; their policy was to require all employees to have smooth-shaven faces, and apparently there’s this thing where a large number of African-American men have sensitive facial skin and experience irritation when they shave, so Domino’s was sued for racial discrimination and changed their policy.

            Which, okay. I mean, it seems kind of ridiculous to me to require all employees to have baby-smooth faces, and I can understand a group wanting to raise awareness about “Hey, this policy disproportionately impacts us in a negative way.” I have a lot of hypersensory issues and weird dietary stuff myself, so I can appreciate the exasperation that comes from navigating a society not designed for you.

            What strikes me as dishonest is this incessant framing of all these things as Institutional White Supremacy and part of the insidious legacy of colonialism and slavery.

            I mean, is nothing just coincidence anymore?

          • hlynkacg says:

            This is not a coincidence because nothing is ever a coincidence.

            …and that’s why we can’t have nice things.

    • Cheese says:

      I can perhaps answer as someone who marked themselves as favourable towards both positions. It might be a matter of semantics.

      By HDB do you mean ‘there are, on average, measurable differences between races in terms of various aspects of human physiology as a result of genetic differences’? Or do you mean ‘there are, on average, measurable differences between races in terms of various aspects of human physiology as a result of genetic differences AND we know enough about the level of heritability and interaction of all of these traits to drastically change social policy AND we’re ethically comfortable with all the aspects of doing so’?

      Because the former is really fairly indisputable IMO. The latter is where we start wading into muddy waters and i’d depart from a lot of the various internet communities that you refer to.

      • rlms says:

        Do you think there are average differences in intelligence due to genetic differences?

        • Deiseach says:

          Do you think there are average differences in intelligence due to genetic differences?

          Yes, because you can see that in a family. Looking at my own (a handy subject of inquiry near to hand), of my two nephews one is the arty/humanities type and one is the science/maths type (my sister says “Elder boy takes after you, younger boy takes after our youngest brother” in this). Both are excelling in their different fields of interest/ability but Elder is not good with the maths (like myself, as I first discovered when he was a youngling and I was drilling him in his maths homework).

          If this question is trying to smuggle in “do you think non-white people as individuals or indeed as a sub-population of a larger white population are dumber than white people?”, then the answer is going to be vastly more involved but it will not handily boil down to “You racist!”

          • rlms says:

            Wouldn’t differences in aptitude at different subjects between genetically similar people be evidence *against* the idea that they are caused by genetic differences? (yes, but weak) Regardless, I did miss the crucial words “between races” in my original question. I think that most people probably agree that genetics influences intelligence.

          • quanta413 says:

            Wouldn’t differences in aptitude at different subjects between genetically similar people be evidence *against* the idea that they are caused by genetic differences? (yes, but weak)

            No, it’s not even weak evidence for what you think. This question is ill defined unless you specify at least some sort of ordinal scale for measuring relatedness and the other trait of interest (and preferably not just an ordinal scale but something cardinal). The question is how different and how related.

            By a typical biological measure, siblings have relatedness 1/2 so it’s not surprising that they still would differ quite a bit. Relatedness between two people can vary from 0 to 1. Relatedness is usually between 0 to 1/2 in western societies with the exception of identical twins; interbreeding of course makes the true measure a little higher but western civilization has relatively low interrelatedness for certain historical reasons as any HBD enthusiast will promptly talk you into a stupor about.

            That’s why psychometrics people love twin studies so much. Twins are related=1 to each other so you can assume almost all differences between them are non-genetic. Epigenetic differences being relatively unimportant.

          • rlms says:

            Yes, I agree. Equally, two brothers being very similar isn’t evidence *for* the idea that genes are important (unless you specify a scale).

    • rlms says:

      Me too. I would suggest that maybe people interpreted it as “genetic differences between individuals are socially relevant”, but the question specifically mentions races. Maybe some people missed the word “genetically”.

    • akarlin says:

      I am not surprised by this at all.

      Higher IQ people tend towards (cultural) leftism – for example, see the overwhelming support for gay marriage – but the LW/SSC-sphere is ultimately about rationalism, and the facts are what the facts are when it comes to HBD. SA is himself reasonably open to it, e.g. his positive review of Garett Jones’ Hive Mind.

      I am not exactly a very representative example, but I identify as an EA’er but have used EA arguments to argue against open borders. I imagine that the LW/SSC-sphere attracts quite a few such oddballs. Another example: IIRC, in the last LW survey, more people identified as [“proponents of new reaction” to-get-past-the-word-censor] than as conservatives. Hard to imagine this being the case almost anywhere else.

      • Gobbobobble says:

        I identify as an EA’er but have used EA arguments to argue against open borders.

        I’d be interested in some expansion on this if you don’t mind, since it seems to come up fairly rarely.

        My personal opinion is that brain drain is sufficiently bad for poorer communities to outweigh the gains to richer ones. And can be argued to be a form of malign imperialism, in that it’s an (indirect) extraction of intellectual resources from poor countries to rich ones.

        Curious whether this meshes with your view.

        • akarlin says:

          I’d be interested in some expansion on this…

          It’s here: Immigration and Effective Altruism

          tl;dr – “One dollar of spending money goes about five times further in poor countries than it does in First World countries due to purchasing power differences… If conditions in Syria are so utterly unacceptable that young males have no choice but to emigrate, surely it would be more effectively altruistic to encourage them to settle elsewhere in the Third World – say, why not a relatively stable and Islamic but poor country, like Tanzania, Senegal, or Bangladesh? The $10,000 they pay the Italian or Greek mafias to smuggle them into Europe would probably be enough to buy a nice house there!

          It is also often a misuse of the refugees’ existing human capital. The Syrian refugee doctor or engineer will probably work as a taxi driver, or something, if he can find work at all (see Emil Kirkegaard’s welfare stats).

          My personal opinion is that brain drain is sufficiently bad for poorer communities to outweigh the gains to richer ones. And can be argued to be a form of malign imperialism, in that it’s an (indirect) extraction of intellectual resources from poor countries to rich ones.

          Yes! I’ve argued on very similar lines, though I called it “cognitive colonialism,” e.g. see here and here.

          • tmk says:

            This all assumes that you can convince Tanzania to take in the Syrian refugees. I could see a point that it would be more cost effective to help people where they are (or nearby is case of war) than move them to the west, but Trump just cut foreign aid.

    • Corey says:

      I think it’s in a feedback loop with the white/male skew, since the typical bailey-ed HBD position, from the point of view of a woman or a black man, is “it’s Totally Rational to not-hire/avoid people like you because Science Says you’re dumber/more violent”, which is a little unwelcoming.

      Like others, I think the motte position of “different people are different” is more widely held around here, though you do see the above from time to time.

      • Fossegrimen says:

        “it’s Totally Rational to not-hire/avoid people like you because Science Says you’re dumber/more violent”

        That would apply in the case where you hire entire populations. As long as you hire individuals it’s just nonsense. I would guess that this community can understand the difference.

        • Cliff says:

          And since you are insuring individuals, that is why men pay the same amount for car insurance as women, right??

        • Hyzenthlay says:

          That would apply in the case where you hire entire populations. As long as you hire individuals it’s just nonsense. I would guess that this community can understand the difference.

          Sure, if one candidate is obviously more qualified than the other then that candidate is probably going to get hired.

          But if you have two equally qualified candidates and one comes from a group that’s negatively perceived in some way, that could tip the scales.

        • Progressive Reformation says:

          It might also apply to cases where key information is deliberately hidden from the employer (say, creditworthiness, which correlates very strongly with many desirable traits), or (if deciding who to give responsibility to) when one knows that under-qualified minorities are hired as a matter of policy.

          Basically, the whole doctrine of “Disparate Impact” has been twisted so that any purely meritocratic hiring system (like the police used to have) is considered ‘racist’. The way they implement this doctrine is either by hiding key information from employers (so they resort to, um, prejudices), or by forcing the deliberate hiring of under-qualified minorities. Note that both of these make it rational to perceive e.g. black men as less qualified than their white peers even if on paper they are equal (or even better).

      • Cliff says:

        Where do women come into HBD? Just that there may be some genetic contribution to why there aren’t so many women at the highest levels of engineering and IT?

    • Anonymous Bosch says:

      I’ve been toying with the idea of writing a sort of “Racist’s Guide to Social Justice” in which I make a case for criminal justice reform, public accommodations laws, reparations, and other generally leftist positions while accepting accept nearly every major premise of the Sailer crowd regarding IQ disparity.

      • Nancy Lebovitz says:

        Please do.

        Would you also want to include representation in fiction?

      • akarlin says:

        Been done, ages ago, by Robert Lindsay (a leftist).

        I mean even Pinker in The Blank Slate (2002) furnishes a Rawlsian argument for why social justice (sanely understood), assuming human biological inequality, is very defensible.

        However, liberal unease and leftist aversion to the very concept of HBD means that in practice, discussions as to its political and social consequences have been largely monopolized by the (Alt) Right, and they are hardly very interested in social justice or EA at all.

      • phil says:

        I mean, I don’t think that’s a particularly good title,

        but I think the proponents of Social Justice movement would win a lot more arguments if they did just that

        I think most SJ proponents think Sailer’s premises are more far reaching than most of Sailer’s followers do (and as best I can tell, Sailer himself)

        I think a good chunk of Sailer’s popularity, comes from the perception its extremely rare for his opponents to argue in good faith

      • Dr Dealgood says:

        If you do, please consider genetic differences other than IQ. Two people with equal IQ can still have a lot of interesting and relevant differences caused by race.

        That’s the claim which gets the most press, and aside from height is the difference with the strongest evidence behind it, but isn’t the whole story.

        The most interesting IMO are monogenic traits, since they’re less ambiguous and you can get a better sense of possible mechanisms. MAO-A 3R and increased aggression, the GHRd3 mutation and a “fast” life strategy (low birth weight, accelerated childhood growth), ALDH2(2) and resistance to alcoholism, etc.

    • Aapje says:


      I don’t think that HBD validates white supremacy/Neo-Nazi’s very well, as it ranks Jews and Asians as most intelligent. It much more strongly invalidates a belief that outcomes have to mimic demographics. This latter belief is a collectivist notion often held by SJ people and this crowd is pretty individualist and SJ-critical.

  18. Iain says:

    So there is a really interesting tendency for conservatives to comment more often than liberals (maybe because they have more to disagree with?). But numbers in the last three groups were very small: out of the 5335 people for whom I had data, only 54 commented once a week, and only 45 commented many times a week. So they may not be able to bring the average up very much. Since tiers 1 through 4 were liberal (REMEMBER THE MIDPOINT IS 5.5) and only tier 5 was conservative, there’s probably an extremely slight preponderance of liberal comments on the whole.

    I don’t follow the logic here. For obvious reasons, people who post more frequently are going to have a disproportionate impact on the “average” comment. Tier 1 doesn’t post at all, and Tier 2 posts less than once a month, so they have basically no impact. The once-a-week crowd is barely larger than the many-times-a-week crowd, and the many-times-a-week crowd is more conservative (0.8 above median) than the once-a-week crowd is liberal (0.3 below median). That leaves the once-a-month crowd to make up the difference, which seems unlikely. You might just barely be able to make the numbers work out if you simultaneously make a) generous assumptions about low response rates from less frequent posters, b) generous assumptions about how many posts “more than once a month” works out to in practice, and c) conservative assumptions about how many posts “many times a week” works out to, but the simplest explanation of this data seems to be that SSC readers are more liberal, but the SSC comment section has a slight preponderance of conservatives.

    Somebody scraped the comments section a while back and tallied up the most frequent posters. I would be interested to see the total number of posts from the 45 most frequent posters, and what percentage of all posts they (or we, I suppose, since I was on that list) are responsible for.

    • rlms says:

      In the past 100 posts, the 45 most frequent posters were responsible for 49% of comments (which is quite a bit lower than I’d have expected I think).
      More data:
      Top n commenters: 5 10 15 20 25 30 35 40 45 50 60 70 80 90 100 200 300 400 500 1000
      Percentage of posts: 12 20 27 32 36 40 43 46 49 51 56 59 62 65 67 80 86 89 92 97

      There were 1933 unique commenters (you were 14th most prolific by comment count).
      (apologies for bad formatting)

      • Iain says:

        Great, thanks!

        I think those numbers support my point. If half of the posts are being made by the most prolific posters, and those posters are more conservative than the less prolific posters are liberal, then overall we would expect the comments to lean slightly conservative.

        Also, I post too much.

      • ChetC3 says:

        How long a time period does 100 posts cover? And what does the distribution look like per post? My recollection is that the last few times I checked, comments/poster for individual posts was much closer to a power law distribution. Of course, I think that was before Jill/Moon was banned (speaking of frequent posters).

      • rlms says:

        Bonus speculative data!
        Out of comments in the last 100 posts by the 5% (96) most prolific commenters who account for about 65% of comments, I estimate that 32% of comments were by left-wing authors, 40% right-wing, 15% libertarian, 4% other, 10% couldn’t be determined. Methodology: my program listed the names of authors by number of comments. Some I judged immediately, for others it displayed the first 50 words of 10 comments, ideally those containing political words (Trump, Clinton, liberal etc.). Some people who would probably identify as libertarians were probably classed as right-wing (and vice versa, but to a lesser extent). Some left-libertarians may have been classed as other/couldn’t be determined. I may have defined “left” more narrowly than usual; I considered Scott to be on the left/libertarian border and so put him as “other”. I may sometimes have been inconsistent in choosing to class people as other or unknown.

        • rlms says:

          Code now on GitHub here.

        • quanta413 says:

          Just wanted to quickly note that you are a hero for actually doing the boring work of scraping the comments.

          Seriously, thank you.

        • Quixote says:

          Posting to giver honor and internet points to rlms for contributing to the public good

        • IrishDude says:

          Nice work. It would be helpful to be more transparent about classification, in case others disagree with how posters were assigned. Any chance you can post who you assigned to which political affiliation?

        • rlms says:

          Updated speculative data!
          Running the program again, I found these stats: 32% left-wing, 36% right-wing, 18% libertarian, 6% other, 8% unknown. This time, I also stored the classifications of commenters (see them here).

          The number next to each name indicates how many comments I read before classifying them (i.e. how difficult they were to classify).
          Names in each section sorted by number of comments.
          Sandy was mistakenly classed as unknown, and should definitely be right-wing.
          Left-libertarians (Nancy Lebovitz and someone else I can’t remember) were classed as libertarian.

          • Whatever Happened To Anonymous says:

            I’m not sure what offends me more, that it took 4 full posts to label me as right wing, or that I’m labeled as right-wing rather than libertarian. Either way, I’m going to find a way to take offense, no matter what.

            In a more serious note, while there’s a bunch of classifications with which I’d disagree, I’d say the list is broadly correct as far as my perception goes. Would 0 mean that you already had a strong preconception of the posters’ political leanings?

          • dndnrsn says:

            Shouldn’t bean be listed under starboard, not right?

          • rlms says:

            @Whatever Happened To Anonymous
            Possibly I was trying to determine if you were libertarian or right-wing! But there is a lot of overlap there; I think a high proportion of the right-wing people could be reasonably classed as libertarian. A 0 usually means that I had a strong idea of their politics already, but in a few cases (mostly for the unknowns) it means they didn’t have any posts flagged as political (containing words such as “liberal”, “conservative” etc.).

          • Brad says:

            For reasons I’ve discussed elsewhere in this thread, I disagree that Aapje is a left wing poster. And moon/jill as well as Earthly Knight are now banned and won’t be contributing to the balance of posts from here forward. That’s #3-5 of the most prolific purportedly left wing posters.

          • Cerebral Paul Z. says:

            Bean has centerline turrets and can fire in whatever direction the tactical situation requires.

          • The Nybbler says:

            I don’t understand ‘0’. Does that mean classified without looking at comments at all? Anyway, I consider myself libertarian — right of center, basically a _Reason_ magazine libertarian. Not the _Reason_ of today, but the _Reason_ of the Virginia Postrel years.

          • keranih says:

            13 posts to figure out that I was right wing.


            I think that’s pretty good.

          • Spookykou says:

            I am happy that I am the least clearly left, left wing person, and I do consider myself mostly left wing if that counts for anything.

          • Randy M says:

            Oh, 30! I win! I’m the most centrist… or make the most content free posts.
            Yeah, probably the latter.
            And… this isn’t helping.

          • Nornagest says:

            What does “other” mean, as distinguished from “unknown”?

          • rlms says:

            “unknown” means insufficient data to classify, “other” means doesn’t fall into any of the three known categories. Most of the people in “other” expressed left-wing sentiments sometimes, but also sometimes not (considering the username Anonymous as one entity, it expressed strong political views but inconsistently, so it made sense to put in “other”).

            I think “other” probably skews left, in that some people would argue that some of its members should be classified as such. “unknown” probably skews right/libertarian, in that I have an impression that most of the people in it definitely have those views, but I couldn’t any political comments to back me up.

      • Scott Alexander says:

        How did you calculate this?

        • rlms says:

          Elaborating on Douglas Knight’s link (since I didn’t do it there), using a Python script (requests, BeautifulSoup for the scraping). If you look in the logs, it identifies itself with the User-Agent “Comments scraper”.

          • Matt Swaffer says:

            rlms, one interesting thing to try would be to adapt an authorship attribution algorithm to identify the political leanings of commenters. Typically this is used to figure out who wrote an anonymous comment / article etc., but it could be interesting to try to adapt it to identify other characteristics.

            If you did a supervised machine learning algo you could train it with the writings of known left / right / lib / …. authors and then see what comes out. Alternately, you might be able to straight up run a cluster analysis of the data and identify common characteristics of the clusters.

            In any case, it might be an interesting way to extend your current line of analysis.

      • Gobbobobble says:

        Subjectively, it seems like the “proper” posts’ draw a different mix of commenters than the open threads. Wonder if the stats are materially different for those.

    • Deiseach says:

      The once-a-week crowd is barely larger than the many-times-a-week crowd, and the many-times-a-week crowd is more conservative (0.8 above median) than the once-a-week crowd is liberal (0.3 below median). That leaves the once-a-month crowd to make up the difference, which seems unlikely.

      It does sound like we should try that “liberal commenters only” thread idea again, and this time let it be only for non-conservative identified readers, lurkers and commenters, just to get an idea of what are the views and if people would feel safer/more encouraged to comment if they weren’t wondering if conservatives were going to jump out of the wardrobe. If they got in the habit of leaving comments on the liberals-only threads, that might encourage them to comment more on the all-for-everyone threads?

      • Iain says:

        Oh, I guess I should probably have included a disclaimer: I don’t think that the slight conservative lean in the comments section is a problem that needs fixing. Speaking from the left, the entire reason I comment here is because it has a substantial contingent of reasonable, intelligent conservatives. Asking those conservatives to stop commenting is actively counterproductive, as far as I’m concerned.

        I would say something here about how conservatives should look at these numbers and bear in mind that, as the majority, they have a responsibility to self-police to keep the atmosphere friendly for non-conservatives, but honestly I feel like things have been pretty good lately. So, uh, keep up the good work, I suppose?

        • Gobbobobble says:

          as the majority

          Isn’t this conflating majority of posts with majority of posters? I thought the main takeaway was that while there are more left-leaning folk, the right-leaners post more frequently.

          • Iain says:

            If we’re talking about the atmosphere in the comments section, isn’t the majority of the most frequent posters what we care about? In any given community, it’s the “regulars” who tend to set the tone of the conversation.

            And to reiterate: I think the status quo is good, and I am not proposing any changes.

        • thenoblepie says:

          Oh, I guess I should probably have included a disclaimer: I don’t think that the slight conservative lean in the comments section is a problem that needs fixing. Speaking from the left, the entire reason I comment here is because it has a substantial contingent of reasonable, intelligent conservatives. Asking those conservatives to stop commenting is actively counterproductive, as far as I’m concerned.

          Same here. One of the reasons I love this comment section is that it confronts me with well-articulated arguments against my positions that I would otherwise never encounter. I wouldn’t want things to change.

    • suntzuanime says:

      I’d also like to point out that there are probably differences in the rates at which each poster makes political posts vs. neutral ones. If you post frequently but mostly about the mathematical minutiae of Bayesian statistics, whatever your radical internal political leanings are won’t have much impact on the political leanings of the comments.

  19. Tibor says:

    I wonder how come there are so many Finnish readers here. A higher percentage of Finns read the SSC than that of the British as well as any other non-English speaking nationality.

    I am also surprised than there are even fewer female SSC readers than there are female libertarians.

    • tmk says:

      All of they said the comment less than once a month or never.

    • lkbm says:

      It’s a very high number for the total population of Finns, but it’s also a low enough raw number that a couple Finns sharing a lot of posts on Facebook could probably cover it if they have the right friends. Or one professor at a university who has the students read a SSC article.

      The LW Diaspora 2016 survey also had a fairly high number of Finns (25):

      (If there is one person who thinks they’re responsible for a large portion of the Finnish readers, or has some idea of a major source, now’s a great time to start commenting.)

  20. hwold says:

    I see you removed NearestCity from the published answers (too easy to de-anonymize I suppose ?). Do you plan to do something with it later ? I would like to see a more fine-grained geographical distribution, especially in the subset of “MeetupInterest or MeetupInterest2”.

  21. Rock Lobster says:

    I’m surprised there aren’t more finance people here. I’ve noticed that people around here, at least commenters, don’t know much about finance stuff, and I don’t mean that critically. It would be good to talk finance sometimes.

    Of course I’m a bit out of place around here myself in that I don’t know the first thing about computers.

  22. Thecommexokid says:

    Of the 700+ total SSC blog posts, what percent coverage did you get with the “Favorite post” question? Any stand-out popular favorites?

  23. Anon. says:

    Some avg IQ stats:

    By sex:
    Female 138.7
    Male 138.5

    By gender:
    M (cisgender) 138.5
    F (cisgender) 138.7
    F (transgender m -> f) 139.3
    Other 140.1
    M (transgender f -> m) 141.6 (N=5)

    By race:
    Black 131.7
    Hispanic 133.2
    Middle Eastern 134.5
    White (non-Hispanic) 138.5
    Other 139.7
    Asian (East Asian) 140.1
    Asian (Indian subcontinent) 141.4

    By profession:
    Media, mathematics, hard science, physics, business were the highest, > 141.

    Psychology, social science, economics, art were the lowest, < 135.

    By education:
    2 year degree 135.56
    High school 136.05
    MD 137.13
    Bachelor’s 138.00
    Master’s 138.93
    None 139.03
    Other professional degree 139.80
    JD 141.17
    Ph D. 143.24

    By religious background (N>10 only):
    Hindu 142.9
    Mixed/Other 142.0
    Unitarian Universalism or similar 140.4
    Jewish 139.7
    Christian (Catholic) 139.4
    Christian (Other non-Protestant, eg Eastern Orthodox) 139.1
    Christian (Mormon) 138.7
    Christian (Protestant) 138.1
    Muslim 136.3

    There doesn’t seem to be a correlation with commenting frequency. No corr with EA/H+/SJ/LW.

    Political spectrum is sort of U-shaped, with 1s and 10s smartest and 3s/4s dumbest.

    Party affiliation: averages vary by less than a point, but the Rs are the smartest.

    Global warming: smart people care more.
    BioRisk/AIRisk: no corr.
    Immigration: smart people want more.
    Minimum wage: smart people don’t like it.
    HBD: no correlation with IQ at all!

    Detail-oriented: 140.3 for the least, 137.0 for the most!
    Weightlifting positively correlated.

    Also for some reason the smartest people want less book reviews.

    • Nevin says:

      Any correlations with theism/atheism?

      Political spectrum is sort of U-shaped, with 1s and 10s smartest and 3s/4s dumbest.

      This matches other data I’ve seen according to which moderates tend to be less informed.

      • Anon. says:

        By religious views:
        Atheist but spiritual 135.9
        Deist/Pantheist/etc. 136.6
        Lukewarm theist 136.7
        Agnostic 138.2
        Atheist and not spiritual 139.2
        Committed theist 139.9

        • Nevin says:

          Interesting — not huge differences, but they seem to follow the same kind of U-shaped pattern as political attitudes, with the extremes having higher IQs and the moderates having lower IQs.

    • nelshoy says:

      Interesting that race and IQ still follows the Black<Hispanic<White<Asian ranking you see in the general American population.

      -but religiosity does the opposite.

  24. Nevin says:

    I’ve tried to post a comment several times now with comparisons between this data and the PhilPapers survey data. It wouldn’t let me post, and then it wouldn’t let me edit a shorter post with that info. Anyone have an idea why the system is auto-rejecting this?

  25. Nancy Lebovitz says:

    I’m mildly interested how people handle reading comments. Use the drop-down list? (My favorite) Tilde-new? Just page down? Search on people/keywords? Something else?

    • Randy M says:

      Drop down lists if there’s one or two, ctrl-F, ~n if there are several so I see them in threading order rather than chronological.

    • Cerebral Paul Z. says:

      If the total number of posts hasn’t yet become unreasonable, I’ll just scroll through the whole thing. Otherwise I check the drop-down list: if there are fewer new posts than 20 or so, I use the list; else I scroll through the whole section after all, making heavy use of the Hide button. (I continue to long for a persistent Hide function.)

      Speaking of Hide, reading through the comments in the survey spreadsheet I was struck by the number of people asking for this already-existing feature. It might not be a bad idea for the top of each comment page to include a brief explanation of the commenting mechanics.

    • Corey says:

      I use the drop-down list, unless I’m away long enough for that to get really long, then I give up the thread.

    • Deiseach says:

      New post when there are very few comments: from the top-down.

      When a post has been gathering comments for a few hours/days: down to the very bottom of the page and scroll up. Keep my eye out for the green border around new comments and stop to read those.

      Yes, it would be easier to use a search method as outlined, but when did I ever do anything the sensible way?

    • IrishDude says:

      To start, I scan down through topic threads, reading the nested comments of the ones that are interesting and hiding/scanning past the ones that seem uninteresting. When I return to read comments at a later date, if many have been added I’ll usually use the drop down and click on a subset of posters that I find have interesting things to say.

    • John Schilling says:

      Until there’s more than ~200 comments, I page down and at least skim everything (or hide the entire subthread if it is going someplace I don’t want to follow). Past that, if there’s more than ~20 new comments I search on tilde-new to preserve thread ordering, drop-down list if it’s less than that.

      I also maintain an index of comment threads with discussions that I found particularly interesting at the time, so if I ever want to revisit e.g. your reviews of the old and new Ghostbusters movie and subsequent discussion, I know immediately where to head in the archives.

    • williamgr says:

      Usually ignore comments on ssc. If I’m interested:

      a) Search by keyword (for this post: “sorting” to see whether anyone else was mildly bothered by the seemingly random order of the keys (neither alphabetically nor numerically sorted) and “bi-modal”/”bimodal” for insights on the happiness bi-modality.)

      b) Search by people (David Friedman, Nancy Lebovitz, Deiseach, Scott himself).

      For unsong, read everything and use drop-down list upon returning.

  26. Tibor says:

    I am a bit confused about the transgender. “f -> m” means born as a woman and then had a sex operation or the other way around?

    • Loke says:

      f -> m means transitioning/transitioned from female to male.

    • John Schilling says:

      “f -> m” means born as a woman and then had a sex operation?

      Yes-ish. Had the operation, wants to have the operation, doesn’t want to have the operation but wants you to perceive them as a man anyway. It is now considered rude to insist on surgery before recognizing someone’s preferred gender choice.

      • Gazeboist says:

        Also there’s a lot more than a single operation. Transitioning is a damn long process compared to most other things-people-get-surgery-for.

      • Tibor says:

        My psychology would probably be a lot different in that case, so it is hard to say, but I think that if I were transsexual, I’d probably get breast implants but not the sex operation. It is called an operation but from what I gather it is mostly mutilation and it can severely reduce the sensitivity of the sexual organs (or what’s left of them). Maybe in a few decades they’ll be able to give you a sex operation worthy of its name. Since heterosexual men are unlikely to be attracted to you one way or another, there’s a little that the “sex operation” can offer, aside from aesthetically looking a little bit more like the other sex.

        I am not sure how well the penis can work when you gobble up so much estrogen though. But perhaps viagra can solve that.

        • Amy says:


          I’m a trans woman and I’m probably not going to be getting any operations, partly because I was lucky enough to start hormones early and I look female and attractive anyway, and partly because, yes, I agree that the current SRS/GCS operation is just not sophisticated enough for me to pursue, and I’d like to wait for science to advance. Also, there’s this misconception that trans bodies are somehow faked by plastic surgery and whatnot, which may have affected my preference to do it “all natural” – I can say that “actually, no, my body/breasts/curves/skin/etc are made of tissue bioidentical to cis women, produced by an identical process, triggered by an identical hormone”.

          Also, I think that in saying heterosexual men are unlikely to be attracted you’re generalizing from one example way too much. I know a lot of men who have no problems dating trans women, and personally guys try to hit on me so much it’s uncomfortable. It’s particularly strange that you specify heterosexual since I’ve never met a gay man who was into trans women.

          • Tibor says:

            There are also bisexual men. I said unlikely, that is not the same as never. If the man in question is only interested in women, it is definitely going to be a handicap if you were born as a man, since current medicine cannot really change your sex. It can at best make you more like the other sex and then it is the hormones that do most of that.

            Put in a different way – I think that the portion of men who find transsexual women attractive when they’ve undertaken the sex operation but not otherwise is going to be low. I think that they are either going to be interested anyway or not at all.

            You’re right that the hormones probably are often enough to give you breasts and alter other things. But as you write – you started with hormones quite early, I guess it is more difficult if you do that after puberty. Or am I mistaken?

          • Talk about starting hormones early got me imagining a different scenario. Suppose it becomes possible and acceptable for parents to choose the sex of their kid not by filtering sperm for X or Y chromosomes but by applying hormones, possibly starting pre-natal. You would presumably end up with kids who had genitals and chromosomes of one sex, physical appearance mostly of the other.

            From one point of view it is carrying the idea that gender is not sex to its limit, but it is doing it by choice of the parent not the person who has been modified.

            Horrifying or not?

          • Iain says:

            Probably horrifying. Trans people are a demonstration that sex is at least partly in your head, that the development of mental sex is not completely controlled by hormones, and that there are significant negative consequences on mental health if your mental sex and physical sex don’t match.

          • Amy says:


            I agree, most of what you said is right – being trans does decrease your appeal (that being said, for women the number of men interested in you is not exactly a bottleneck), and things do get harder for people who start hormones later because there are some things that testosterone does that are irreversible except by surgery. My only concern is that saying “hormones can’t change your sex” is a bit misleading because it implies that the only relevant characteristic is your genitals – remember if an MtF woman starts hormones before puberty, all other factors of her biology would be identical to a natal female. So one won’t be completely female, but definitely much closer to female than to male – just imagine rounding to the nearest gender.


            That’s a really thought-provoking idea! You could separate it into two cases:

            1. If you start hormones after birth. (Note: this is done for trans teenagers – you get prescribed hormone blockers to prevent damage from the wrong puberty until you can make a certain decision about transitioning, usually at 14 or 16. Of course, you have to have supportive parents, or be really good at planning for your future and buying hormones online (like me).) For most non-trans people however, this would be a really bad idea because the brain already undergoes sex differentiation before birth – meaning that you would be forcing them into a gender opposite that their brain was wired for – essentially having them experience the same dysphoria that trans people do (trust me, seeing a puberty you don’t want starting to wreck your body is horrifying). Although, there is an interesting theory that many people are cis-by-default – in the sense that there are people who are mentally male, and would suffer if they had to live as female and vice versa, but there is also a segment which really doesn’t care either way, and just live as their assigned gender because it’s the path of least resistance. If that’s true, perhaps cis-by-default people really could grow up as either gender (that would sound really nice assuming the child has a choice – imagine a society where at 12 years old, everyone tries out living as a boy and as a girl, and at 13 they decide which one they like best).

            2. If hormones are used before birth Then this could actually work. For instance, read about androgen insensitivity syndrome – this is when someone with normal XY chromosomes has a receptor mutation that prevents androgens like testosterone from having effects on their body. The person grows up completely female, and most never even know they have XY chromosomes. If you control hormone levels before birth, the brain, genitals, anything is fair game. The only exception is that in almost all cases the person is sterile, but there have been a few individual cases of XX males / XY females producing children.

          • Tibor says:

            Amy: Well, while the reproductive organs aren’t the only thing that there is to sex, I’d say it still is the most important thing. Maybe not for gender and maybe that’s what makes using two terms for a very similar thing useful, but while I am ok with calling someone who feels like a woman and more or less looks like one (sometimes it is quite obvious that that person was born as a man and sometimes you would not believe it) a she, it is still not quite the same thing. The best analogy (still pretty bad) I can think of is calling a wolf a dog. They’re similar (well, if it’s not a chiwawa or something 🙂 ) and the dog is not a cat, but it is still not quite a wolf.

            Iain: Well, the hormones affect your brain as well (which is what make it rather annoying when people claim that women and men only differ physically…the same hormones that change how you look also modify your brain and later in life, the decisions of how much of what hormone to produce is done by the brain, so they simply have to be different), so if you apply the hormones in the prenatal phase, you are probably going to change the brain as well. It might still give you a gender mismatch, but assuming that those hypothetical people know exactly how to apply the hormones, that is less likely. IIRC, gay men are statistically more likely to have been more (or less I am not entirely sure) exposed to certain hormones in the prenatal phase than other men, so this could also just affect sexual orientation.

            David: I would not call it outright horrifying, but if you apply the hormones simply on the whim of the parents, then you’re doing harm (unless you luck out and your child would have ended up being transsexual anyway, but that is extremely unlikely assuming the hypothetical science cannot foretell it). I would view it the same way as a mother heavily drinking or smoking during pregnancy, so definitely it would be something condemnable. I would think less of someone who did that as I do of mothers who smoke during pregnancy (not quite the same, because there is a scale from one cigarette in a weak moment to a pack or more a day and there is no corresponding scale with the hormones – I’d see that as more akin the pack a day).

          • @Amy:

            you get prescribed hormone blockers to prevent damage from the wrong puberty until you can make a certain decision about transitioning, usually at 14 or 16.

            Interesting–I didn’t know that.

            That suggests another of my maybe horrifying SF ideas, one I’ve actually discussed in print. Puberty can be the point at which serious conflicts arise between parents and children. I can imagine parents who, given the option, would like to keep their children pre-pubertal forever, or at least until they are in other respects adult and out of the house.

            Is there any reason why that wouldn’t work using hormone blockers? Do we know what a pre-pubertal adult would be like–are there any natural conditions that produce that result? I am pretty sure Maimonides discusses the legal issue of an adult woman who has not shown the signs of puberty.

            The only exception is that in almost all cases the person is sterile, but there have been a few individual cases of XX males / XY females producing children.

            If an XY female produced a child with an XY male, there should be a .25 chance of a YY child. Any idea what that would mean? As best I can tell by a quick google, it currently exists only in science fiction.

            I am now imagining a very patient mad scientist deliberately producing XY women in order to find out.

          • Jaskologist says:


            That’s basically what eunuchs are, assuming you get them at a young enough age. So there should be plenty of historical data to work from, at least on the male side.

          • The Nybbler says:

            If an XY female produced a child with an XY male, there should be a .25 chance of a YY child. Any idea what that would mean?

            Nothing much interesting. Either the gamete (ovum) fails to develop in the first place or the zygote perishes early on. There’s a lot of important genes on the X, you can’t do without it. The one report I find quickly suggests the Y-ovum is nonviable even in the very rare cases an XY-female is fertile; Y-chromosomes in offspring came from the father.


          • rmtodd says:

            DavidFriedman wrote:

            I can imagine parents who, given the option, would like to keep their children pre-pubertal forever, or at least until they are in other respects adult and out of the house.

            Is there any reason why that wouldn’t work using hormone blockers? Do we know what a pre-pubertal adult would be like–are there any natural conditions that produce that result?

            My first thought is that I’d worry about osteoporosis being an issue. IIRC there have been reports of young woman gymnasts, women who have spent their teens competing so vigorously that their hormone production shut down under the extreme stress, who are now showing up in their 30s with the kind of osteoporosis usually found in old post-menopausal women.

          • dndnrsn says:

            Don’t female gymnasts tend to have dangerously low bodyfat? Low bodyfat in men is no fun (it messes with hormone production – one of the reasons male bodybuilders use steroids is to make up for their body reacting to very low bodyfat % by reducing testosterone production) but in women it seems to cause more permanent damage. Damage also probably happens at a higher relative bodyfat % (relative since women carry a higher % of fat naturally – a man with 20% bodyfat is going to look a lot pudgier than a woman with 20%). In any case, is it the extreme stress, or the fact that they’re underweight? Or both.

  27. Nevin says:

    Some comparisons with the PhilPapers survey of professional philosophers:


    SSC Readers:
    64.3% Atheists
    16.7% Committed or lukewarm theists
    18.9% Other (incl. agnostics)

    72.8% Accept or lean toward: atheism
    14.6% Accept or lean toward: theism
    12.6% Other

    Political Philosophy

    SSC Readers:
    24.2% Libertarian
    5.7% Conservative
    25.6% Liberal
    29.1% Social Democratic
    12.4% Other (incl. neor., etc.)

    9.9% Libertarian
    14.3% Communitarian
    34.8% Egalitarian
    41.0% Other


    SSC Readers:
    46.9% Consequentialism
    6.2% Deontology
    11.6% Virtue Ethics
    6.9% Natural Law
    5.2% Contractualism
    23.3% Other

    23.6% Consequentialism
    25.9% Deontology
    18.2% Virtue Ethics
    32.3% Other

    Remarkably similar numbers on theism/atheism. Political philosophy is hard to compare because the options are pretty different, but SSCers are much more libertarian. In ethics SSCers are much more consequentialist and much less deontological. (You might count contractualism as a form of deontology, but that would still make SSCers less than half as deontological as philosophers.)

  28. Jacob says:

    >What would you think of requiring new commenters to answer a knowledge question before being allowed to register

    What would really be ideal is if you could have people answer a question specific to each post . This would do a good job of preventing trolling by regular users. Answering a question before registering would be a great idea too, though, obviously technical feasibility plays a role.

    • phisheep says:

      I remember that my answer to the knowledge question question was somewhat influenced by my inability to answer the question in the question.

      Which turned out to be really stupid of me as a copy of Kahnemann & Tversky’s paper was actually on my desk about three inches from my left hand at the time I was completing the survey. Though to be fair I didn’t realise I had it until a day later.

      (Have since tidied desk and discovered lots of other interesting stuff.)

  29. Waltus says:

    In future surveys, I’d be interested to see correlations among people who consider themselves “content creators”- people who write books or run a blog/podcast/youtube channel, or are otherwise involved in spreading their ideas.

  30. TK-421 says:

    Suggestion: maybe don’t use pie charts next time? They’re extremely hard to read, especially when the slices are unlabeled and not in any particularly obvious order. The bar charts, by contrast, are far more readable.

    • alexsloat says:

      At minimum, put a little block of each colour beside each answer so that we can easily see which is which. The results should be legend as much as numerical values.

      • houseboatonstyxb says:

        I much prefer pie charts, but this color legend would be helpful in other types of chart as well.

        Pie charts have advantages for us “right-brained”* people. I get more tags to remember by: more colors, proportion, shape, etc).

        * Yes, I know, I know.

    • taktoa says:

      Yeah, I’m surprised Scott doesn’t know that pie charts, due to the fact that humans judge pie slices by area rather than arclength, result in the actual ratio x / y being perceived as x^2 / y^2, so they are a really bad way to convey information.

      Note that I don’t like regular bar charts for percentages-that-sum-to-1 either, since they do not show that relationship very well; the optimal way (IMO) to prevent such information is a “stacked bar chart” like this.

      • simon says:

        For pie charts the ratio of areas is equal to the ratio of arclengths.

        • switchnode says:

          This is correct (I suspect taktoa was thinking of the occasional pitfall of pegging circular chart elements’ radii to data). However, humans are rather poor at comparing arc lengths, angles, and areas. We systematically underestimate acute and overestimate obtuse angles, the same sector can appear smaller or larger depending on its orientation, and it is extremely difficult to compare sectors across a chart or between charts (unless they are pegged to very obvious landmarks, like a 25% slice at 0 degrees).

          The stacked bar-chart does indicate that the relationship under consideration is part-to-whole, but is much less successful than the standard bar chart in permitting quantitative estimation and comparisons between near-equal fractions. (It is difficult to tell 10% from 11% in the stacked bar chart, trivial in the non-stacked.) I would use a standard bar chart for this type of data. Unit and value labeling, reader familiarity, and the discussion of methods in the text are, I think, sufficient to establish that the percentages sum to 1.

          (An auxiliary graphical element, such as a stacked-bar or pie chart, could visually confirm it, but would be redundant and in practice introduce the nuisance of an additional, different scale.)

  31. phil says:

    >>>The more trustworthy you think other people are, the less likely you are to like Donald Trump (r = 0.14)<<<

    that makes a lot of sense to me, from my vantage point, a lot of Trumps support, especially where he gets in in more IQ corners (and I include myself in that), seems to come from people who feel cynical or let down by why they perceive as the null hypothesis leader trustworthiness

    for example, they(I) assume some corruption or duplicity is a selecting factor for political success

    so, to pick one thing out, Trumps' lack of political experience isn't quite the disqualifying feature it would be for people who weren't oriented like that

    I think Trump's supporters tend to be pretty cynical about how trustworthy the average agent is


    whether than cynicism is warranted, is a separate question

  32. Deiseach says:

    Either permanent loss of sexuality from SSRIs is so vanishingly rare that a survey of 5500 people cannot pick up on it, or it is impossible to confuse with “asexuality” as an orientation and I should have asked the question some other way.

    Definitely think the question needs to be asked another way, because “ruining your libido” is not at all the same thing as being asexual. One is “I used to want/have/enjoy sex but since I started these tablets, nothing is doing even if I want it to be doing (and it’s wrecking my relationship/my life)”, the other is “never was interested/can be interested but the same way you might decide ‘I’m sort of hungry right now, I think I’d like a Big Mac’, not ‘I really want to/need to have sex'”.

    “Permanent loss of sexuality from SSRIs” is in the region of “I’m a wheelchair user now because a truck hit me and my legs had to be amputated”, as distinct from “I’m a wheelchair user because I was born without legs”.

    • Loke says:

      Well, I think Scott’s theory is that being on SSRIs while going through puberty could prevent one’s sex drive from developing normally in the first place. In which case the two conditions might be indistinguishable.

  33. Douglas Knight says:

    Note that the midpoint is not what you think. On a 1-10 scale, the midpoint is 5.5, not 5. On a 1-5 scale, the midpoint is 3, not 2.5.

    Is this is standard warning in the psychology of research?

    For the 1-5 scale, the survey-taker will subitize not make this mistake, while analyst might. But for the 1-10 scale it seems like they might make equal and canceling errors.

  34. nelshoy says:

    14% self-identifying EAs and only 2.8% average charitable donations? On a survey where you can lie?

    That’s a smaller percentage than the general population, despite possessing more income. Very disappointing, especially with how strongly causes like global warming were supported

    • Quixote says:

      I don’t know if the SSC commenters (excluding outliers) would do better by absolute $ donation amount than by %, but I would suggest that the comparison should be made on this basis.

    • mingyuan says:

      Well 25% of the readership are students and therefore (presumably) have no income, so maybe the self-identifying EAs are disproportionately college students and are unable to donate? I doubt that accounts for the entire discrepancy though.

    • Pablo says:

      14% self-identifying EAs and only 2.8% average charitable donations?

      Why not look at the percentage donated by the 14% who actually self-identify as EAs? These people, in fact, donate 6.00% of their income on average. Giving What We Can members donate 9.69%. Non-EAs donate 1.75%.


  35. anonanon says:

    >I don’t know if schizophrenics and self-diagnosed schizophrenics are disproportionately using MTurk because they don’t know what else to do with their lives, or if I just got a weird sample.

    MTurk pays like absolute crap, which means a significant proportion of turkers are people trying to earn a little extra on top of a fixed income, such as disability payments. It also trades off wages for flexibility, so I would not be surprised to see people who generally have trouble conforming over-represented.

    • behrangamini says:

      Either the MTurkers are more likely to have mental health issues, or the crap pay is encouraging random/creless clicking. In either case, the results of the ssc survey of MTurkers would be of interest to people using the service for surveys. Scott should consider publishing the results, if this hasn’t already been done. I suspect the ease of use of MTurk will win out over any doubts social and psych scientists may briefly have about the representativeness of the sample. This is reminiscent of the use of western college students for psych experiments and the drunk looking under the lamppost for his keys.

    • Cadie says:

      I noticed this as well. Even minimum wage jobs pay much more – working at McDonalds will get you at least twice as much money for the same hours worked and the pay is more consistent – so a US/UK sample is not going to be representative of the overall population or the general Internet-savvy population. Making fries or bagging groceries gets you significantly higher wages, and that’s saying something. Lyft/Uber also pays better and is almost as flexible as MTurk, though for that you have to have a decent car (cheap is okay, but it has to be clean and in good repair), insurance, and ability to drive, so it’s not suitable for some disabled and/or very poor people.

      Except where minimum wage and median wages are very low and technology access and English knowledge are relatively high or at least not terrible, such as India, MTurk pay is awful and selects for people who don’t have better options. Given that crappy US-minimum-wage jobs count as otherwise better options, this will bias the selection quite a bit.

  36. habu71 says:

    Ah, the SSC census results. I had been waiting for these.
    Here’s to hoping that Scott doesn’t end up like the census taker from the south that he wrote about in his previous post.

    • Randy M says:

      Or the state of Israel, for that matter.

      • Phigment says:

        It would really reduce the fun level of collecting user data if 70,000 people died as a result.

        On the bright side, it would give us a very strong plan for some effective-altruism-in-action.

        I bet we could prevent Scott from doing any more surveys really cheaply. If every survey kills 70,000 by plague, and he does one a year, then all we have to do is find someone in his general area and pay them a moderate salary to follow Scott around all the time and sabotage his survey-making ability. $40,000 a year/70,000 lives saved = $0.57 a head. That’s better than mosquito nets, isn’t it?

  37. Error says:

    I am curious whether bisexuality correlates with gender in this dataset.

    Context: Bisexual women are absurdly common in my social circles. I think I know more bisexual women than straight women. I’m wondering if that’s blind chance at work, or if it’s a nerd-girl thing along the lines of “the same brain chemistry that nudges nerdy women towards male-typical intellectual interests also nudges them towards male-typical sexual interests.”

    • Loke says:

      A quick calculation says that 32.5% of female respondents are bisexual, as opposed to 8.1% of male respondents. (Gay percentages are 3.4% for females and 3.9% for males, so pretty similar.)

      My impression is that female sexuality tends to be less visually-oriented in general, and therefore less dependent on gender. I don’t think it’s a nerd-girl thing in particular.

      • tmk says:

        Then again, only 85% of those with female sex are cis.

      • Virbie says:

        I read a pretty interesting paper to the effect that female sexuality’s higher fluidity may be a consequence of historical polygyny (1 husband, multiple-wives polygamy). I think the explanation was that it would help with intra-family conflict/tension or something like that.

        From Googling, I was able to find some crappy Daily Mail coverage of the paper.

        • Deiseach says:

          Did any of the authors of that study look at traditional accounts of polygamous cultures? There’s very little “sexual fluidity means the co-wives are happily bonking one another” and a whole heap more of “scheming against your rivals to have your son made the raja/caliph/emperor”. Look at the Ramayana, where the entire epic is kicked off by one wife being persuaded that the senior wife’s son being named the crown prince is going to mean a downfall in her own status and that of her son, and so she schemes to get that son exiled and her own son named heir. Or the Empress Regnant Wu Zetian, who showed little sign of “fluid sexuality makes harem relations more harmonious”.

          That’s rather more the Orientalism in art view of “life inside the seraglio” as depicted by Ingres than the reality, I would imagine. But “lesbianism for straight guys” has always been a seller, even if niche, so no changes there!

      • Error says:

        Google suggests the general female public has a bisexuality rate around 5-17% depending on how the question is asked, with personal identification (i.e. what Scott asked) coming in at the bottom of the scale. So it’s not just my circle. Bisexual women are hugely overrepresented in this space too, to the tune of two to six times the general population.

        I’m not sure whether personal identification is the appropriate direct comparison. I could easily see nerds answering differently for reasons having nothing to do with sexual interests, e.g. by failing to do doublethink like “well, I’m attracted to members of my gender but I’m not *really* gay/bisexual/whatever”.

        It looks like LGBT overall is also overrepresented here, though not to the same extreme. I’m not sure if I should draw any conclusions from that or not.

  38. Deiseach says:

    Interested to see that the age range on the survey is from 13 to 77, so plainly Scott offers something for everyone!

  39. R Flaum says:

    We were slightly less satisfied with our lives (6.3 vs. 6.4) and vastly less happy (6.0 vs. 6.9), even though we were earning an average of $97,000 and they were working on Mechanical Turk.

    This mirrors my experience. I used to be a programmer, and was miserable. I currently retrieve shopping carts from the lot for a supermarket, and am much happier. I think this is partly because I’m doing a less stressful job now (The fact that I’m getting a lot more exercise might also have something to do with it). I have unusually low expenses, though, so income difference will have a smaller effect on me than on most people.

    • hydro says:

      Interesting! Do you know what it was, exactly, that made being a programmer more stressful for you? Competitiveness? Long hours? General atmosphere?

      Having done a menial-labor-based job, I found my happiness response was exactly the opposite – though maybe I just haven’t achieved burnout yet.

  40. JASSCC says:

    It would be interesting to see some statistics on migration status, probably US-centered, given the residence country numbers. (Are you an immigrant, emigrant, permanent resident of a foreign country, etc.?)

  41. Whatever Happened To Anonymous says:

    SSC Science Articles: 3.5 / 5 (higher = want increased focus)
    SSC Politics Articles: 3.5 / 5 (higher = want increased focus)
    SSC Book Reviews: 3.2 / 5 (higher = want increased focus)
    SSC Rationality Articles: 3.5 / 5 (higher = want increased focus)
    SSC Silly Articles: 3.2 /5 (higher = want increased focus)

    The takeaway from this is that people want you to write a larger volume of posts.

    • JASSCC says:

      Cute. The real take-away is SSC readers make up five warring tribes, membership in each defined by greatly enjoying four of the five categories but intensely disliking the fifth.

      To make the tribes better tolerate their disfavored articles, you need to add a category everyone hates.

      Occasionally post about sports.

      • Randy M says:

        But not a silly post, nor a book review about sports, nor analyzing political effects of or on sports, nor the science of sports, and certainly not about the rational way to enjoy sports.

        Of course, mixing all those together is a pretty good way to kill Steve Sailer’s productivity for the day.

      • Deiseach says:

        You think you jest, but FiveThirtyEight is doing some sports predicting; I’m following them with great interest for Liverpool (because I’d love to see if Mathematical Scientific Analysis can make better sense of our results than the “play against the best of the best – we’re magnificent; play against middling to bottom sides – they thrash us” rule of thumb the long-suffering fans use) and now for the Champions League (given that Leicester City have made it to the quarter-finals!) I’m finding their forecasts for the latter very intriguing and am wondering if the odds are considering home advantage in the first leg or what.

        There you go, sports related commenting! 🙂

        • Gazeboist says:

          Unfamiliar matchups.

          (The linked page summarizes a character in a fighting game who is relatively weak overall, but does unusually well against certain extremely common top characters, in part because she plays very differently from other top level characters and most players don’t really know how to deal with her. This sort of analysis is, of course, much easier for fighting games than for almost any other kind of competition.)

          (Amusingly, as noted in the article, the character in question jumped all the way from 8th-best in the previous game to 10th in this one, a massive improvement. The secret, of course, is that there are 14 extra characters to be ranked against.)

          • nelshoy says:

            Woah, SSC has other smash fans? I can confirm that Samus is extremely disorientating for my sheik, but Samus mains will stereotypically never acknowledge this and insist that they won because they’re better/that their character is underrated.

            (As a single straight dude, I apparently really need to find some less male-dominated interests…)

          • Quixote says:

            I bet there are a bunch. Smash is way more spatial than most fighting games, and spatial skills are a hallmark of the nerdy math type skillset.
            For myself I used to play a ton of melee back in the day, I went to a decent number of tournaments (top 8ing a few Amtrak corridor regional events back before people got really good like they are now). I eventually stopped playing when brawl came out thoughts since I didn’t enjoy playing as metaknight. I would be interested in seeing how many SSC reader smash though and potentially seeing if an informal lower key SSC smashfest could be organized.

          • Error says:

            I play Smash. I’m good enough to beat everyone I know handily, but not good enough to compete respectably in tournaments. Partly that’s because I main Zelda, one of the weakest fighters in every game she’s appeared in.

          • Cypren says:

            I’ve been known to play a few rounds of Smash as well, though I’ve never spent the time required to get really good. I have a few friends who play in tournaments and are really into it, though.

            Marth all the way for me. I’m a good tipper. 🙂

        • nelshoy says:

          Not going to lie, fivethirtyeight kinda got me into sports. Somehow, checking how the ratings were changing over time led to me getting invested into which team wins, personally embarrassing as that is. 😛

          Honestly, I don’t see why so many rationalist types should viscerally dislike sports. I’d expect a lot of people just default to that from bad personal experiences. They’re all formally defined competitive systems with rational agents that you can do analytics on. I think much like how Nerds Can Be Bees but don’t naturally gravitate there, nerds can be sports fans too.

          I think a lot of it comes down to personal enjoyment of the games themselves. Basketball’s more fun to watch if you’re deeply familiar with the game. I know quite a few nerds who’ve always hated football, but will tune in religiously to watch CLG vs. TSM in League of Legends.

          • JASSCC says:

            I myself was a sports fan in my teenage years, and I still sometimes enjoy tennis matches. I wouldn’t disparage sports generally — I was mainly joking — but I’ve never enjoyed lengthy sports analysis. Occasionally I enjoy interviews with particularly sports stars when they talk about their training and motivation, in much the way I find it interesting to listen to anyone who has extremely rare skill talk about it.

            The best explanation I’ve ever heard of the appeal of sports chat is that so many significant and highly visible things happen in a typical professional game that even an intelligent child who knows the game can often make interesting observations. I don’t dispute that, but that requires a level of commitment to understanding the game inside and out that I’ve never shared.

          • nelshoy says:

            I think sports chat is just craving novelty in your Domain of Interest when not much else is going on. Probably the same impulse that makes people religiously follow political news and pundits.

            “How will X player signing with Y team affect their odds next year?”


            “How will Y minor newspaper endorsement affect Z candidate’s chance in the election in twelve months?”

        • switchnode says:

          the “play against the best of the best – we’re magnificent; play against middling to bottom sides – they thrash us” rule of thumb the long-suffering fans use

          Ah, must be students of History. The English, very memorably, have always preferred to fight against heavy odds!

      • IrishDude says:

        I’m not a fan of watching most sports, but I love the NFL. There’s managing salary caps, fielding a large team that needs to work together while requiring talented individuals, strategizing game plans and match-ups, executing on the field physically and mentally, and adjusting to events on the ground. It’s really entertaining to watch. The combination of football being both rawly gladiatorial and sophisticated strategy hits the primal parts of my brain and my prefrontal cortex.

      • registrationisdumb says:

        That will just add a sixth tribe.

    • JulieK says:

      I suggest next time asking people to pick one type of article they would like to see more of, and one they would like to see less of.

  42. Alexander Scot says:

    Very interesting! I’m amazed at how many white men there are, but even more surprised at the lack of straight, monogamous women. I drilled into the genders a bit, and while straight monogamous non-asexual men made up about 70% of respondents, the equivalent for women was less than 4%.

    I know this blog contains the occasional strongly-worded criticism of feminists, but I’m curious as to why so few straight women read this – there is little that I would regard as exclusively masculine. Anyone have any theories?

    • JASSCC says:

      Wild hypothesis: the prevalence of noxious sexist comments drove women out of participating in many web-based fora to the point that many women avoid any forum that has commenting as a salient feature, unless drawn to it by an overriding special interest in its central topics. The “straight” part is harder to account for, but perhaps gay commenters develop a thicker skin earlier?
      Or perhaps in the absence of anything approaching equal numbers of women, topics or discussion style drifts in a direction that gay women more easily engage with?

      • quanta413 says:

        Additional Wild Hypothesis: Not only are web-based fora notoriously noxious on average, but women are also more likely than men to get positive feedback from friends and family on their facebook posts, instagrams, pinterests, etc. and so invest more of their web online time in those places. Twitter is of course a horrifying cesspool even by internet standards where no it seems like no sane man or woman would spend much time (and yet so many do!).

    • Aapje says:

      Men are more likely to be systemizers and women empathizers. This forum seems to have a lot of systemizing debate and fairly little empathizing.

      But most likely, it is a confluence of factors.

      • Alexander Scot says:

        Could be – I think one of the reasons I like this blog and forum is the focus on rationality and empiricism over emotionally engaging anecdotes, but others may dislike it for the same reasons. I don’t see a reason why there’d be such a gay/ straight split (gay in this case also includes asexual, poly and trans people), but given that there are quite a lot of people of that community (I think it worked out to about 24% of all responders including Scott as far as I am aware), they may be encouraged in a way that straight women – seeing relatively few commenters like them – might not be.

        As a side-note, I wonder if including a question on Myers-Briggs types might add some depth on how people on this forum think. Or even Howart’s houses… bet there’s a ton of Ravenclaws here

        • hydro says:

          Maybe, given the prevalence of straightness/non-straightness in the general population, straight women are more weirded out by finding very few people like them than gay women are.

          (Anecdotally – before coming here, I had read HPMOR, went, “huh”, and moved on, later found a few things on LW, went “huh”, and moved on, then was linked here by some article on 538, went “huh”, and actually stayed. My experience is probably not the typical one, but I’m not really sure how it would be influenced by gender. Unless readers of 538 are also predominantly male? It may just be an effect of the routes to getting here also being heavily disproportionate, gender-wise.)

      • mingyuan says:

        I think there’s an element of socialization as well – it seems plausible that women are less likely to be attracted to the community because it’s less socially acceptable for a girl to be into something that emphasizes logic over emotion.

        Also, a 90% male community seems kind of self-reinforcing. Both because white guys will get their white guy friends into it, and because if you show up and everyone else there is a white male (and you’re not) it’s a little uncomfortable and you might not want to come back. Once it comes to be perceived as a white male thing maybe others will be turned off by that? That doesn’t explain how it got that way in the first place though – maybe because all the original big names in the community were white males? But that’s true of a lot of other internet communities too, and not all of them have the same demographic imbalance we do.

        I don’t know, I’ve had this conversation a lot since I’m often the only girl at meetups, and no one has a conclusive answer. Maybe people who have been involved in the community since the very beginning have a better perspective.

        • Nancy Lebovitz says:

          Scott said something to me about my being demografically unusual– I’m over 60 as well as being a woman– but I have no idea why other women my age aren’t showing up. I’m not them.

          • Gazeboist says:

            But why did you show up? There must be other women your age who share other, relevant characteristics with you. It’s a small world, but it ain’t that small.

          • Nancy Lebovitz says:

            It’s hard to say– I like the conversation. I like the courtesy rules– in particular, the rules against mocking people for ignorance (not completely followed, but compliance is fairly good). My exposure to Social Justice made the relative courtesy at LW and now SSC a great relief.

            My interests have enough overlap with the folks at SSC.

            I may be weirder than I realize.

        • Hyzenthlay says:

          Also, a 90% male community seems kind of self-reinforcing. Both because white guys will get their white guy friends into it, and because if you show up and everyone else there is a white male (and you’re not) it’s a little uncomfortable and you might not want to come back.

          Most people here don’t have user-pics of themselves, though, so gender and race isn’t evidence at a glance. I’ve been reading this blog for a while and I’m still sometimes not sure of commenters’ gender unless they have a male or female sounding username.

          Personally, I was surprised that only 10% of the commenters are female. I would have guessed that at least a third to 40% were.

          • Jaskologist says:

            I generally assume that everybody here is male, but I choose to picture all of you as attractive females anyway.

            Some of you make this difficult. Still, I persist.

          • dndnrsn says:

            According to the list (slightly out of date) of top commenters, ignoring Anonymous commenters and collapsing David Friedman and DavidFriedman into one, 1 of the most frequent commenters is Deiseach, known to be female; I do not know any of the others on that list to be female. However, on the most prolific commenters list, using the same measurement, 2/10 are confirmed female (Deiseach, Jill – who is currently permanently banned).

            So, the frequent commenters are not less female than the commenters as a whole, and may be more female than the commenters as a whole, but 9/10 are either “confirmed male” or “don’t know.”

      • JASSCC says:

        I think this may be an ingredient, but learned aversion to chats / comments / open discussions in general by women on-line as a rational response to abuse is the primary explanation.

        There was a book a few years ago pointing out that women are far more likely than men to receive nasty messages threatening them with violence (especially sexual violence) in response to things they posted online. Also, I believe I recall they found women were far more likely than men to report having used opposite-sex pseudonyms to attempt to get their points of view accepted.

        The way I see it, most American adults have lived in households with access to the internet for about 20 years or less. For maybe the first 10 years — up to about 10 years ago — it was a relatively new social phenomenon. In that time, I think women used to participate more in all discussion fora, and often were subjected to horrible sexist abuse. Some even adopted male persona to participate without risking such abuse. (This is not explainable by women’s preferences, as far as I can see, but only as a response to sexism.) But more and more they started abandoning discussion groups unless they had either an overriding interest or a direct, very personal interest.

        After another decade, I think that has become the norm. Women have simply been driven out / learned to stay away form most open discussion groups. Even a discussion group as friendly and well-policed as SSC (pretty much the upper end of the spectrum, in my experience), is unlikely to appeal to women to give it a try after that history. As a result, to the extent that there are male and female discussions styles, discussion styles probably end up gendered very male. The experience I imagine is that if a woman is interested in a particular article here, she is disproportionately likely to read it and then never bother with the comments (from a policy of staying away from comments), or any form of participation (such as filling in the survey).

        A possible counterpoint to my point would be that lesbians reported even more abuse than straight women online.

        • Aapje says:

          The male gender role teaches stoicism, which probably is the cause of higher physical pain insensitivity in men. It stands to reason that men are even more tolerant of psychological pain.

          • JASSCC says:

            That strikes me as plausible but also a leap, but even if so, I think a simpler explanation is that the large majority of the pain in question (e.g. abuse of a commenter in response to a comment) is directed at women rather than men.

        • Barely matters says:

          So, the study is found here:

          It’s results were that that women reported experiencing more online abuse than men in the subcategories of sexual harassment and stalking, whereas men reported experiencing more abuse in the categories of name calling, shaming, and physical threats. Overall the rates of online abuse are comparable between genders, and the only real quibble is whether sexual harassment is worse than physical threats.

          While 78% of respondents thought message boards were roughly equally friendly, there were 13% who considered them more friendly to men as opposed to the 6% who thought them more friendly to women.

          If these reports are to be believed, this lends credence to Aapje’s reading that it’s more a matter of perception and response to a given amount of abuse rather than a difference in the actual amount.

          • Nancy Lebovitz says:

            I wonder whether they count rape threats as physical threats or sexual harassment.

          • Barely matters says:

            Being that it’s self reported, I’d imagine it varies from respondent to respondent. As far as I know, the survey asked questions like “Have you been the target of online sexual harassment?” and tallied the total “Yes” responses.

          • JASSCC says:

            Thanks for this info. This makes me strongly reconsider my point, to the point that I would rescind the gist of my comments until I understand how to reconcile this with other things I have read (both selective presentation of facts from this study, and other experiments), for example one experiment claiming to find that feminine user names in chat groups attract far more attacks than masculine names (many times as many per hour). I now consider myself officially puzzled on this topic.

            The only way I can think to reconcile these is if the Pew study’s findings of nearly equivalent abuse received by gender (slightly more male) represents much more *time* spent in open, pseudonymous chat or discussion fora for the male vs. the female internet users included in the study. Then both claims could be true. For example, imagine that some women included in the study as “internet users” never or very seldom participate in open, pseudonymous chat and discussion fora where abuse is most common, and those who do perhaps experience a faster rate of abuse, and therefore could be more likely to quit such fora quickly or spend much less time actively participating. Over time, the larger percentage of men participating at any given time might then eventually receive an equal amount of abuse, but at a much lower rate.

            But now I’m speculating.

          • Aapje says:

            @Barely matters

            As far as I know, the survey asked questions like “Have you been the target of online sexual harassment?” and tallied the total “Yes” responses.

            If that is the method, it is very poor, because we know from rape victim surveys that people answer differently if asked about whether they experienced specific behavior vs more generic categories. This is especially true for men, as sexual victimization is heavily gendered in most people’s minds, so men are less likely to recognize themselves as being sexually victimized.


            I one played a computer game where it was very popular to make rape threats and such (although this was surely used as a metaphor for in game violence).

            It is perfectly consistent with the data if women get more harassment in fora and yet men spend way more time playing violent games where the in game chat is…colorful.

          • Barely matters says:


            Your point about the possibility of different timescales is a really good one. I’m sure followup studies could ask more detailed questions in that vein and get more thorough answers. It might even be found in the data of this study itself, but didn’t seem relevant enough to the news reporters to be included. Either way, it’s a good question.

            I’m inclined to agree. Self reporting in general has shown very a poor track record in studies that can later be independently verified. I think the reason for its use is that “Due to experimental limitations, we haven’t found a way of generating solid evidence either way” is a profoundly unsatisfying answer. And given that the choice here is currently between vague anecdotes and systematically counted anecdotes, I could charitably consider them to be trying to do the best they can with the limited resources at hand.

            So I’d agree that it says very little aside from “In our sample, we did not find evidence of significant differences in the rates of online abuse”

          • Aapje says:

            @Barely matters

            For diets, a common method is to let people write down their meals as they eat them, to reduce the problem of people leaving out meals (usually snacks). It might be interesting to try this method with online harassment, where people white down the harassment as it happens. Harassment is relatively common for young adults, so if you focus on this group, you should be able to get decent data when collecting during a limited period.

            I would expect that the pre-experiment instructions still could make a huge difference though and there is a risk that this is influencing the outcome.

            And I think that gamers would be less inclined to write down harassment as it happens, if doing so harms their gaming outcomes.

        • Also, I believe I recall they found women were far more likely than men to report having used opposite-sex pseudonyms to attempt to get their points of view accepted.

          The opposite pattern may exist in World of Warcraft. It’s a common belief that most of the female characters are actually male players–my wife has been asked online if she is “really a girl.” I haven’t seen any actual data, but raiding with voice chat on in a group with a fair number of female characters only one of the voices was female, which is a small sample but at least consistent with the belief.

          Googling, it looks as though about a third of the characters are female. There seems to have been an old Nielsen report showing a pretty large minority of players female, but it’s limited to the age group 25-54, which I suspect eliminates a lot of younger players as well as a few older ones.

          • LCL says:

            When I played years ago: in cases where I found out the player’s gender, a majority of female characters were played by males. But it’s a biased sample because I only found out the gender of the players I closely associated with, i.e. serious try-hard gamers.

            I also, over time, came to make my own characters exclusively female (I am male). If you asked me why at the time, I might have said “they’re nicer to look at.” In retrospect, I think I was lonely and creating a series of cool virtual girlfriends.

          • Brad says:

            Do you guys play zoomed all the way in or something? I can tell if a toon is a boomkin or a tauren but much beyond that it is just various colorful blobs.

          • Leit says:

            You’re that guy that stuns all the casters on the trash before Botanist because he didn’t realise the blue pulse was coming from him, aren’t you?

          • Sivaas says:

            If you’re even fighting the Tel’arn trash, hopefully you’ve long ago installed DBM and are thus announcing to the raid group that you have Arcanic Release through /say. You’ll probably stun the group once or twice while you learn the mechanic exists but if you’re doing it more than that the problem isn’t how far zoomed out you are, DBM announces and debuffs should be giving you the information you need. A blue glow around your character is not giving you anything more.

            I agree with Brad: at a raiding level you should be at max zoom out, possibly using script commands to extend that max zoom (I can’t remember if you still have to do that in Legion, I should probably double-check my settings. I know there was some controversy over that and the max zoom available is lower than it was in WoD). You want to see as much of the fight area as possible, because things you have to deal with will be coming from every direction and considerable distance.

            But still, you’re looking at your character in the login screen, your kill screenshots, and occasionally zooming in to see transmog. That’s probably enough to want your character to look good.

          • Brad says:

            I never got into the whole xmog thing. My guildies say I look like a hobo. Don’t care.

          • Loquat says:

            I haven’t gotten to know many other players in World of Warcraft, but when I was doing more serious gaming in Warhammer Online, Aion, and Rift I knew plenty of men (and women) who played female characters but only ever met one woman besides myself who played a male character. And for me it’s more about aesthetics than gender preference per se – in WoW my main is a female troll just because that’s the race/gender combo that best meets my preferences for looking reasonably tough but not like an outright monster or a steroid-abusing blimp.

    • rlms says:

      The lack of women is presumably due to most readers being STEM types (and hence largely male). The lack of straight monogamous women is very interesting. Are LW rationalist women much more likely to be polyamorous than LW rationalist men?

      • The Nybbler says:

        STEM in general is MUCH less lopsided in terms of gender than the SSC survey. Even engineering and software development don’t go much below 20% female.

        • rlms says:

          Are you saying engineering and software development are each around 20% female? That doesn’t match what I’ve heard for programmers, where a brief google suggests it’s closer to 10%. I don’t know about engineering.

          • Gazeboist says:

            IIRC high-single-digits in pure math is as low as it gets. Your 10% figure for programming-ish disciplines sounds about right, with a caveat that subdisciplines are unbalanced. Systems work is very male dominated; application-type stuff less so.

          • The Nybbler says:

            Figures for the US


            Software developers, systems + applications: 20% women. This is the largest computer-related occupation and is what a tech company software engineer wold be considered; the second largest is computer systems analysts at 35.7% women. “Computer Programmers”, a term I now see mostly used by “body shops” doing business programming, comes in at 22.6% women

            Architecture and engineering occupations: 14.2% women, varying from 6.4% for mechanical to 26% for architecture. I thought this was around 18% but that might have been undergrads rather than employed. Still well over 10%.

            No figures available for mathemeticians.

    • Daisy says:

      Well, women in weird nerdy subcultures are disproportionately bisexual, I’m assuming as a function of Openness to Experience. So I think there’s not much of a puzzle on why so few of the women who did respond are straight.

      Why so few women overall responded is a different question. But I didn’t respond, and I guess I might have if I’d felt enough warm fuzzies about this community to overcome my ADD.

    • Deiseach says:

      I’m curious as to why so few straight women read this

      Just for myself, I didn’t come here via the usual route (Less Wrong), it was actually via a series of religious blogs: Leah Libresco was mentioned as a reasonable atheist who would engage with religious types instead of blanket “You’re all stupid! Read Dawkins!” so I followed her, then she mentioned here (I think this was in conjunction with CFAR where she worked for a while) and I ended up here and I’ve been hanging round the joint since 🙂

      Less Wrong, the very, very little I’ve seen of it, would definitely not have brought me here. Granted, I’ve only seen it now in its (relative) decline, not when it was getting off the ground, but the bits I read make me go “Nooooooo…. some of this is okay and interesting and truthful, but in general the tone is not for me nor me for here”. I am assuredly not a STEM person so that’s probably a big handicap there for me being able to engage with the mindset and topics and level of converse. And I kinda took against Eliezer Yudkowsky due to something he wrote which seemed to disparage the role of and need for beauty; the particular instance in the essay that got my goat, paraphrasing very heavily and going on memory, was how walking down streets and seeing the stained glass windows of churches really irritated and annoyed and offended him because this was a terrible irrational waste of valuable wall space and in a proper future rational world these would be solid blank concrete walls where you could, if necessary, project on the inside talks and videos and Useful Informative Non-Fictional Material. I also picked up that he didn’t much approve of the notion of churches being permitted to be blatantly existing out in the open like that in the first place, but eh, kneejerk atheist was something I’d expect in the context and wasn’t particularly concerned about. It was “What do we need all this useless beauty for, when we could be making sensible use of that wasted space?” that gave me the Public Commissioner from “Hard Times” heebie-jeebies:

      ‘This is a new principle, a discovery, a great discovery,’ said the gentleman. ‘Now, I’ll try you again. Suppose you were going to carpet a room. Would you use a carpet having a representation of flowers upon it?’

      There being a general conviction by this time that ‘No, sir!’ was always the right answer to this gentleman, the chorus of No was very strong. Only a few feeble stragglers said Yes: among them Sissy Jupe.

      ‘Girl number twenty,’ said the gentleman, smiling in the calm strength of knowledge.

      Sissy blushed, and stood up.

      ‘So you would carpet your room — or your husband’s room, if you were a grown woman, and had a husband — with representations of flowers, would you?’ said the gentleman. ‘Why would you?’

      ‘If you please, sir, I am very fond of flowers,’ returned the girl.

      ‘And is that why you would put tables and chairs upon them, and have people walking over them with heavy boots?’

      ‘It wouldn’t hurt them, sir. They wouldn’t crush and wither, if you please, sir. They would be the pictures of what was very pretty and pleasant, and I would fancy —’

      ‘Ay, ay, ay! But you mustn’t fancy,’ cried the gentleman, quite elated by coming so happily to his point. ‘That’s it! You are never to fancy.’

      …‘You are to be in all things regulated and governed,’ said the gentleman, ‘by fact. We hope to have, before long, a board of fact, composed of commissioners of fact, who will force the people to be a people of fact, and of nothing but fact. You must discard the word Fancy altogether. You have nothing to do with it. You are not to have, in any object of use or ornament, what would be a contradiction in fact. You don’t walk upon flowers in fact; you cannot be allowed to walk upon flowers in carpets. You don’t find that foreign birds and butterflies come and perch upon your crockery; you cannot be permitted to paint foreign birds and butterflies upon your crockery. You never meet with quadrupeds going up and down walls; you must not have quadrupeds represented upon walls. You must use,’ said the gentleman, ‘for all these purposes, combinations and modifications (in primary colours) of mathematical figures which are susceptible of proof and demonstration. This is the new discovery. This is fact. This is taste.’

      I have since been informed that this isn’t really a fair representation of his views and that he will leave room for frivolity and fancy, but it’s hard to erase first impressions.

      So why not more women readers? I honestly don’t know, unless it is that Rational Space is heavily majority male and that’s why, if the usual route to finding here is “be involved someway, even as a lurker, in Rational Space”.

    • Barely matters says:

      In this case, we might be looking at a subset of the broader question regarding the lack of women in subcultures generally. I became interested in this when I realized that every subculture I participate in skews heavily, heavily male and tried to branch out to communities with less lopsided ratios. If anyone sees the commonality between the skydiving, scuba, weight training/martial arts, SlateStarCodex, gaming, programming, and metal communities, aside from the fact that they’re comprised of almost entirely men, let me know. Medium distance running is the only physical subculture I’ve found that reaches about a 60-40 female-male ratio.

      As it turns out, there are relatively very few niche communities that don’t skew heavily male, and I would be very, *Very* grateful if anyone here has input into some that break this pattern (Bonus points for other high G/IQ groups like the commmentariat here that skew female).

      • Libertarianism skews heavily male, with one noticeable exception. Students for Liberty, a large and active libertarian group, appears, from casual observation, to be about thirty to forty percent female.

        I’ve asked SFL people how they do it. One answer I got was that they get the men to dress decently!

        It matters, and not just because appealing to women as well as men doubles the potential audience. Movements to improve the world face a serious public good problem, since if I improve the world I improve it for everyone. One solution is tie-ins, indirect benefits to participating. One of the biggest, not just for political movements but for many other voluntary groups, is the opportunity to socialize with people rather like you while doing things together. It doesn’t have to be heterosexual romantic socializing, but that’s a very big part of it.

        • Barely matters says:

          You might be onto something here.

          Skytrash still wear beat up jumpsuits from the 80’s, martial arts groups jump around in wacky pyjamas, and I’m not even going to try to defend the clothing and grooming habits of a typical gamer, programmer, or metalhead. Maybe niche aligned Lululemon equivalents are the key.

      • dndnrsn says:

        Is there a reason to believe women participate less in subcultures in general than men do? If they don’t, for every man doing something niche, there’s gotta be a woman doing something else niche. Perhaps they’re just in different niches, and the niches women tend to prefer are less visible, or come to mind less as “subcultures”?

        • Barely matters says:

          As far as I can tell, yes. finding conglomerated data is hard, so it ends up being a bit of a patchwork, but it appears that women are somewhat overrepresented in mainstream activities like Netflix watching, childcare, churchgoing, post secondary education, fashion and celebrity gossip, recreational shopping, and reading fiction. Since these groups are absolutely gigantic compared to most niche subcultures, a relatively small female skew in a mainstream activity counterbalances male skews in a lot of smaller groups.

          For example: The United States Parachute Association currently licenses about 35000 members at a rate of approx 90% male. So if all 28000 of those surplus male skydivers (80% of the total population) immediately quit flying and enrolled in post secondary education, it would balance out almost 1% of the 2.9 million man gap.

          So you’re right, it could just be an issue of defining what we’re calling a subculture and what is part of the mainstream. Or that their female dominated subcultures exist in similar numbers, but are less visible. I’m open to being convinced, but if this is the case, I haven’t seen any evidence for it yet.

      • psmith says:

        If anyone sees the commonality between the skydiving, scuba, weight training/martial arts, SlateStarCodex, gaming, programming, and metal communities,

        Women less likely?

        there are relatively very few niche communities that don’t skew heavily male, and I would be very, *Very* grateful if anyone here has input into some that break this pattern

        Theater? Dance? Fanfic?

        • switchnode says:

          The affluence-, education-, and social-class-filtering mechanisms discussed in “Black People Less Likely” are unlikely to apply to women, although the social-acceptance mechanism might. Parallel majority-female communities are theoretically possible, but less likely on the internet (where gender presentation is not always apparent) and don’t in fact seem to exist except possibly with regard to gaming (ime).


          Yep, 90% female.

          Virtual pet sites are about 60–70% female. (Roleplaying sites are about 50/50.)

          Knitting culture? It’s a thing but I’m not familiar with it.

          I can’t think of any other female-dominated “niche” interests off the top of my head. Does following e.g. mommyblogging, fashion, or celebrity gossip, to the point of being a regular commenter on various fora with established communities, put a woman in a “subculture”? I suspect (but don’t have evidence) that by raw numbers it does, but since those interests are coded feminine and/or a large number of women have a passing interest in them, they appear ‘mainstream’, or at least feminine-mainstream. Are there similar phenomena for men?

          • psmith says:

            The affluence-, education-, and social-class-filtering mechanisms discussed in “Black People Less Likely” are unlikely to apply to women

            Is joke.

            Are there similar phenomena for men?


          • The Nybbler says:

            Are there similar phenomena for men?

            Sports (watching, not playing), cars, and their intersection at auto racing would be traditional ones, though they are less male-dominated than they once were.

          • Nancy Lebovitz says:

            Ravelry is a major knitting and crocheting site, and I think it’s mostly women.

            Making Light is a fannish discussion site and I think it’s about 50/50.

        • Hyzenthlay says:


          I used to be involved in various fanfic communities, and they definitely skew female. Same for online art communities.

        • Barely matters says:

          I think women less likely might follow a similar dynamic.

          I really don’t think Scott wants to open that can of worms though, given that the big Loving More surveys pretty consistently find that polyamory communities have almost twice as many men than women. Yikes.

          Dance is a good call though. Does theater really skew female? I thought it was relatively evenly split.

          • Tarpitz says:

            Complex question. Young, aspiring actors skew heavily female, but filter effects relating to the number of parts available mean that the older and more successful an actor, the more likely they are to be male. By way of illustration, when I posted audition notices for my first production on the London Fringe (think off-off-Broadway) – unknown director, tiny theatre, play no-one had ever heard of, short run, lots of rehearsals, little chance of exposure, no pay – I received applications for parts in the following ratios (applicants:parts):

            Male over 40: 0:1
            Female over 50: 3:3
            Male 20s: 4:1
            Female 40-50: c.5:1
            Female late 20s: c.80:1
            Female early 20s: c.200:1

            Lighting and sound skew heavily male, stage management, production management and direction skew somewhat male, set design skews somewhat female, costume design skews heavily female. Critics I think skew female, as probably do audiences. Men in every area of theatre theatre skew gay. Overall I endorse “theatre skews female” mostly because young actors who haven’t yet packed it in are such a large chunk of the theatre people population.

          • Barely matters says:

            I appreciate this comment so hard.

            And have a new level of respect for how rough it must be for young aspiring actresses.

            Thank you for taking the time to go into so much depth and detail.

          • Tarpitz says:

            Acting is a generally poor career choice. I’m 34. For the past four years, my income from acting has ranged from c. £5000 to c. £10,000 per annum, for between 10 and 20 weeks of work each year. That may make it sound like the pay’s pretty consistent, but it actually includes weeks where I was paid £1500 and weeks where I was paid nothing. I would guess that put me above the 95th percentile for acting income of men my age who attempted a professional career at some point, and considerably above the median for those still making a go of it. For women, the odds genuinely might be ten times worse. Given changes in the specific nature of the work I’ve been doing (I’ve recently broken into feature films, basically) I’m optimistic that my situation will improve considerably in the coming years, but that’s far from guaranteed.

            In short, there’s a reason why the stock response to, “What advice would you give a young person considering a career in acting?” is “Don’t.”

          • Aapje says:

            There is a reason why the ‘Porn Valley’ is close to Hollywood.

          • Tarpitz says:

            I’m sure there is, but – and this may be a difference between the UK and the US – the people I’m talking about as “young aspiring actors” are not for the most part clueless kids who have gone to the big city thinking they’re going to be movie stars. London is New York, not Hollywood; I was advertising on Casting Call Pro, not StarNow. These people are in the majority of cases either graduates of 3 year undergrad drama school courses or 1 year postgrad drama school courses which followed a 3 year undergrad academic degree elsewhere. If they chose to pursue another career it would be accountancy or teaching, not hairdressing.

            The people you’re talking about exist, of course – the director of a fringe show I performed in some years ago showed me a StarNow application she’d had for one of the female parts from a girl who announced that she had no qualifications or experience but was studying for her drama GCSE (ie was at most 16 and probably younger) and accompanied her CV with a series of selfies in increasing stages of undress; I only hope her parents cottoned on before she ran across one of the many less scrupulous “employers” on that site. But she and people like her are not “theatre people” – they’re another tribe entirely, even if an even tinier percentage of them end up working with the already tiny percentage of us that get into movies. And yes, I’m sure that tribe skews female too.

          • Barely matters says:

            Congrats on making it to feature films!

            You’ve definitely gone further than I managed early in my career. Hilariously, given Aapje’s comment, I went from background, to dancing the red lit stage, to doing skin flicks, before transitioning into skydiving because it seemed safer and more stable.

            So I know what you mean. It’s good work when work is plentiful, pays a lot by the hour, but gross earnings are always dismal.

            I’m always happy to hear when someone is moving on up.

          • dndnrsn says:

            before transitioning into skydiving because it seemed safer and more stable.

            Beyond this being an excellent sequence of words, what’s the story here? Like, training people to skydive? How does one make money as a skydiver?

          • hlynkacg says:

            My curiosity has been piqued as well.

          • Aapje says:

            I think that the only real money there is as an instructor/tandem-diver. My completely fact-free/observational belief is that a lot of people have sky-diving on their bucket list and do it just the once, so there are relatively many instructors compared to the number of dives.

          • Barely matters says:

            Tandem is definitely my favourite as far as work jumping goes, though I also teach accelerated freefall and static line. I have a few medals from competition, but it’s a pretty niche sport, so there’s not too much in the way of cash or sponsorship deals aside from the Redbull airforce and the like.

            Tandem is either love it or hate it for jumpers. It’s ease scales with burliness and sociability, because 90% of the game is in quickly and efficiently instructing your students about all the things they need to do in order to avoid killing themselves and you, then overpowering and stabilizing the inputs they don’t realize they’re performing when they inevitably freeze and panic. If you’re small to average, it gets to be a lot of effort at high stakes. I’m a pretty big guy (for you), with an almost Phelpsian wingspan, so I got lucky in terms of having the right shape to make it easy.

            AFF is usually the favourite among professionals, because it’s longer term training through 7 student levels, ending with your student being licensed to self supervise. You’re not physically connected to your student, so whereas in Tandem the dynamic is ‘We need to get this right, because we’re in this together’, in AFF it’s ‘you need to get this right, and no matter what happens, I’m going to be fine’. AFF can be fun, but I find it to be a huge pain in the ass between having to speedpack your parachute to make the next flight, the suspense of having to watch your student land on his own, and really repetitive, longform classes. The melee of intercepting a spin or correcting a rollover can be ridiculously satisfying though.

            I’ve have to jet off to work for now, but I can give more detail if anyone’s interested.

          • Tarpitz says:

            Thanks, and I’m in turn delighted to hear of someone finding a way of moving on to something else actually enjoyable, rather than the Hell of real jobs!

          • Barely matters says:

            Oh, I didn’t quite make it out unscathed, but I’m almost there.

            I’m taking time out from flying to make some bucks in the short term and investing the proceeds until I start making enough capital gains to cover my expenses. Jumping is great when you’re in a good location, but it really doesn’t pay the bills on its own. (Old guys on the dropzone and aging strippers in the club have way more in common than either would like to admit.)

            You know the line, ‘Do what you love and you’ll never work a day in your life: Because you’ll have to fight through six million other people who are competing for that sweet gig.’

      • Callum G says:

        Anecdotally I’ve been involved with some subcultures that are majority women: Yoga, book clubs, feminist groups and dancing come to mind. Book clubs could be the group that skews towards high IQ or at least towards intelligent discussion.

        • At a slight tangent, one would think there would be a tendency towards even sex ratios in groups. Men, after all, are interested in socializing with women and women with men for purposes of courtship, which should make a largely male group attractive to women and a largely female group attractive to men.

          But it doesn’t seem to work that way, presumably because other factors outweigh that one.

        • Nancy Lebovitz says:

          I think there’s a lot of evidence that a great many men at least want time off from associating with women. That’s why men form all-male clubs.

          I’m not sure whether the desire to spend time away from men is comparably strong for women.

          • dndnrsn says:

            I don’t know if “time off from associating with women” is the right way to put it. Men behave differently when there aren’t women around, both for better and for worse, as I would assume women behave differently when men are not around.

            In my (outsider) experience, all-female spaces tend to have an edge of politicization all-male spaces don’t, although there are exceptions (beyond obvious ones like changing rooms) – an ex of mine would have “girls’ nights out” but only women being present didn’t seem to be the entire point – she expressed annoyance at a friend’s girlfriend showing up.

          • Aapje says:

            Men behave differently when there aren’t women around

            Or do men behave differently when there are women around?

            My perception is that (some) men feel a need to posture to impress women, which can be tiresome to other men.

          • hlynkacg says:

            6 or 1 half dozen either way, my own experience is that the social dynamics in an all male environments tend to be a lot more “open” than it is in mixed company.

          • Barely matters says:

            My perception is that (some) men feel a need to posture to impress women, which can be tiresome to other men.

            This is actually exactly my issue. The key point is that it’s not linear in scale, and the amount of posturing is much worse when women are present but locally scarce.

            Gaming is the most salient example for me. I’ve been in excellent groups that have roughly even gender splits, and excellent groups that are all male. But the second it’s all guys except for a single woman, without fail it becomes “DM’s girlfriend special treatment bullshit forever.”

        • hls2003 says:

          Regular church attendance skews towards women.

        • Barely matters says:

          Those are all really good picks.

          I was going to say that most of the suggestions so far tend to be pretty sedentary (Like knitting, fanfic, book clubs, feminist groups), but realized that I took the ‘active’ tag on the request out. Dancing and yoga are good counterexamples. Guess I need to go buy some $200 stretch pants!

          I’m already active in the local hoopdance community, and it has a slight female skew (Which is still a lot more male leaning than I would have expected).

    • apprenticebard says:

      I wouldn’t have shown up show up as part of the 4%, since I put “yes” for asexual, but I’m female and wouldn’t consider dating anyone non-male or not-OK-with-monogamy, which is… most of the things? I guess?

      I read (almost) every post and a lot of the comments, mostly because they’re consistently thoughtful and insightful (and charitable, this place is so much more charitable than so many other places), and because Scott is pretty good at explaining complex ideas in ways that I can understand (the comments sometimes less so, but they offer a lot of diverse viewpoints and lots of cool follow-up information). SSC is also way more accessible than the limited LW material I’ve seen, since Scott doesn’t have as much of a mathematics background as certain other people. Amateur interest in economics and psychology probably makes this place more enjoyable, too.

      I don’t comment because my perception is that the vast majority of commenters are smarter or more knowledgeable than I am, and I have a hard time believing that I’d have anything to contribute. I mean, I’m not technically stupid—I don’t remember my exact SAT score and I’ve never taken an IQ test, but I know on other standardized tests I’ve always been 95th-99th percentile across the board. On the other hand, people here are very intelligent, and are clearly highly educated in lots of areas that I know nothing about. Plus I’m easily intimidated even by people who are really nice. (Possibly this is somewhat more common among women? Hard to say.)

      As far as gender splits in general, my impression is that there are plenty of women involved in nerdy activities, but they hang out in different spaces than men. I spend a ton of time with people who write fanfic, and as others have said, the vast majority of fanfic writers seem to be women. Plus, sometimes women are involved in an activity, but not the subculture surrounding that activity. Like, I played D&D/Pathfinder every week in high school, which I understand is generally held to be more popular with men. On the other hand, I never went to a large organized meetup or tried to be part of the larger tabletop gaming community, and my group was mostly female (we had one male classmate playing with us regularly, and the usual DM was the father of three of the other players). It was incredibly fun, but when you hear a couple stories about girls playing with strangers and the strangers deciding to roleplay raping the girl’s character or something, you don’t have a lot of incentive to venture beyond the games you play with your friends (since, after all, your goal is to have fun).

      Obviously that’s not happening with SSC commenters, since even lurkers are way less likely to be women. My guess would be that it has more to do with the content of what Scott posts than with the larger community surrounding him, since a lot of people said they don’t even read the comments. But I couldn’t tell you what it is about the content, exactly. I come here in part because I don’t think Scott is overly analytical or that the content is particularly inaccessible, compared to similar places, and a lot of the material seems like it should be equally interesting regardless of gender. I don’t know why there would be so few women. I was pretty surprised by those numbers.

  43. Lasagna says:

    OK, first glance surprises to me:

    How many readers are single (43.2%!!!) and how young (median age of 30) they are. They go together, sure, but that still seems really high. Guys – you’re mostly guys – take it from someone who got married late: get married early, start having kids, stop with all the other bullshit. It’s so, so much better. You’re all mostly pro-monogamy (good), so get moving. Most everything else is noise.

    Biggest disappointment: how few people were against gay marriage. It would be nice to occasionally interact with SOMEONE who’s against it. What happened to all the anti-gay marriage folks? Did absolutely everyone change their mind, are they afraid to speak up, or do I live in one of those absurdly impossible bubbles Scott has talked about before? I don’t get it.

    • Randy M says:

      There were conversations about gay marriage in some of the recent open threads. The open threads go by fast, but they were in the last 3-5 threads ago.

      edit: This thread, for one. And towards the end of this one.

    • Deiseach says:

      Arguing about gay marriage is probably fruitless; in my country we’ve legalised it, and in the USA it’s probably only a matter of time before most states legalise it. All we can do is argue about is it or is it not a bad idea for social mores and culture in general, and that is going to take twenty-fifty years before we have enough data to judge (as with divorce).

      As Randy M says, we’ve discussed this before and I’d imagine going over it again would only generate more heat than light. Not worth clawing each others’ eyes out over it.

      • JASSCC says:

        Did you miss the 2015 news? You can get gay married virtually everywhere in the U.S., except for a few places fighting the ruling:

        Except from wikipedia article:
        Marriage summary
        On June 26, 2015, the U.S. Supreme Court held, in Obergefell v. Hodges, that the Fourteenth Amendment requires states to offer same-sex marriage and recognize same-sex marriages performed elsewhere.[1]
        The federal government has recognized all legally performed same-sex marriages since June 26, 2013 and the decision in United States v. Windsor.
        All 50 U.S. states, 4 of 5 U.S. territories, and the District of Columbia are fully compliant with the Obergefell decision, but some counties in Alabama (10), Kentucky (2), and Texas (1) have either stopped issuing marriage licenses (10) or continue to refuse to issue marriage licenses to same-sex couples (3).

    • suntzuanime says:

      Are you offering to set me up with somebody?

    • Fahundo says:

      take it from someone who got married late: get married early, start having kids, stop with all the other bullshit. It’s so, so much better.

      I feel like you’re trying to trick people so you don’t have to be alone in your misery.

    • JulieK says:

      I had been expecting it to be even younger and more single, actually.

    • nimim.k.m. says:

      get married early, start having kids, stop with all the other bullshit

      Where in the world you live where it is straightforward act of “getting it done”?

    • daniel says:

      Take it from someone who married early: no rush.
      But according to my anecdotal second-hand evidence, having kids early-ish does sound like a much better deal than one would naively assume.

    • Jaskologist says:

      From the Talmud:

      Rava said – and the same was taught in the school of Rabbi Ishmael – Up to the age of 20, the Holy One, blessed be He, sits and waits [to see] when [a man] will marry; once he reaches twenty and has not married, [God] says, May his soul perish!

      Rav Hisda said, The reason I excel my colleagues is that I married at sixteen. Had I married at fourteen, I would have said to Satan, An arrow in your eye!

  44. JASSCC says:

    On “Hard work” vs. “Growth Mindset”: it’s absolutely true the wording makes a difference to me.

    On considering it, I realized that to me “hard work” means striving at a task so intensely that it’s at least unpleasant. What makes it “hard” is that unpleasantness. “Hard work” conjures images of things like laborers digging a ditch under the sweltering sun. Practicing the piano is not hard work. Learning a language is not hard work. Struggling to shift a 300 pound weight across a room is hard work.

    By contrast “diligence” or “persistence” captures something closer to the meaning (to me) of “growth mindset”.

    • Daisy says:

      Yes, I was surprised at the equivalence drawn between “having growth mindset” and “believing hard work is more important than talent.”

      I thought growth mindset was about not labelling yourself and forming all-encompassing irrational beliefs about yourself based on a necessarily limited sample of experience: i.e. it’s a mindset, a sort of algorithm for forming more useful self-concepts, not a single belief about the world. I also remember thinking about one of Scott’s previous discussions of growth mindset that he seemed to have missed the point in a way I couldn’t quite articulate, and also that given the way he interpreted the concept his extreme scepticism about it was more than warranted.

      In my understanding of it, Carol Dweck tests for people who use the “growth mindset” algorithm by testing for beliefs that are correlated with a tendency to use the algorithm, because you can’t test for the algorithm directly. But Scott asking people on this survey about “growth mindset” probably is testing directly for people who use the algorithm, given the audience and that only people who find “growth mindset” useful will say yes.

      Also, I predict that as more and more people start learning that “growth mindset” type test responses are socially desirable, Dweck’s tests will become increasingly useless, because lots of people will agree with growth-mindset type beliefs on tests without actually having internalised the algorithm.

      • houseboatonstyxb says:

        By contrast “diligence” or “persistence”

        I agree that ‘hard work’ is a negative expression and something positive or neutral would be better.

        On considering it, I realized that to me “hard work” means striving at a task so intensely that it’s at least unpleasant. What makes it “hard” is that unpleasantness. “Hard work” conjures images of things like laborers digging a ditch under the sweltering sun. Practicing the piano is not hard work. Learning a language is not hard work. Struggling to shift a 300 pound weight across a room is hard work.

        I disagree with your division that puts all non-physical work as not hard. Piano, language, any sort of learning can be ‘hard work’ — depending on the attitude, hours, etc.

        • JASSCC says:

          On consideration, I agree with you, rather than the perspective I stated. I was sort of interrogating my own thought process and realized I had internalized the notion that hard work entails something like physical suffering.

  45. deluks917 says:

    I wonder if its possible to semi-randomly sample the rationalist community and have them take IQ tests. Maybe someone could convince a random rationalist meetup to all take a 30 minute IQ test?

    • Deiseach says:

      Maybe someone could convince a random rationalist meetup to all take a 30 minute IQ test?

      IQ test: 30 minutes

      Arguing over methodology, sampling, test questions, etc.: 6 months

      Arguing over results: 3 years and possibly some internecine wars

  46. tcheasdfjkl says:

    The thing where the longer you’ve been reading SSC the more conservative you are actually seems totally intuitive to me in the most straightforward way possible. You (Scott) are more conservative than I am, and so reading this blog actually has sort of made me more conservative. I would be unsurprised if this were true for many others who are like me.

  47. nadbor says:

    I have high life satisfaction, growth mindset *and* I believe in talent over hard work. Made me smile when I read that this is actually a trend.

    For me at least, the explanation is pretty simple. I don’t believe in growth mindset. I have growth mindset. I know innate ability is more important. I also know that I am blessed with more innate ability than most – so all it takes is a little effort to acquire any skill I want. Easy to see how this combination would lead to life satisfaction.

    I wonder if others had similar reasons for giving those answers.

  48. Pablo says:

    I don’t comment as often as I used to, which never was very much in the first place, but I did take this survey and hung around to see if the results would show up. I found my response and this enabled me to discover that I must have reported myself as female on the survey when I am actually male. Survey falsification for the win!

  49. The more liberal you were, the more likely you were to think SSC comments had a conservative bias, and vice versa.

    And in both cases you were right, which illustrates the problem with “bias.” Bias only makes sense relative to truth. If you believe the truth is left wing, you will correctly view opinions at the population as biased relative to that truth, and similarly the other way around if you believe the truth is right wing.

    • JASSCC says:

      Bias *can* also mean identity-based or personal animus. For example, it’s possible a person who identifies as left-wing could be accused of anti-right wing bias for disagreeing with right-wingers even when they actually agree on a point of policy out of bias against anything they have to say. I guess you could formulate this as ignoring the “truth” of actual agreement to gin up disagreement, but in this case the more important aspect of bias, to me, is an absence of fair treatment.

      But I think the whole thing is a muddle. I have an extremely hard time locating myself on a left-right scale or understanding the shifting definitions of left and right in American politics. (Apparently strong opposition to free international trade is now “right”?) So it’s hard for me to even sort out the players.

      • Cypren says:

        The problem with the left-right scale is that it’s too often confused with the Democrat-Republican scale. As the political interests and coalitions of the parties change, people are more likely to change their definitions of “left” and “right” to go along with the latest party platform than they are to change where they place the party itself along a consistent left-right axis.

        Hence global trade is now a “left” position by dint of Trump vehemently opposing free trade and Democrats vehemently opposing anything Trump supports. Meanwhile religious freedom (a traditionally “left” position up until the 1990s or so) has become a “right” position in the last few decades as Republicans have started using it to try to defend Christian traditionalism from the secular victories in the culture wars. Absolute freedom of speech (also traditionally a “left” position) is also likewise becoming a “right” position as the authorities cracking down on speech have shifted from religious traditionalists policing “immoral” and “unpatriotic” speech to secular multiculturalists policing “hate speech” and “climate deniers”.

        If anything, what it really shows is how weakly tied to meta-principles most political movements are, and that categorizing them on an axis is never more than a temporary assignment of convenience, one with a short shelf life.

  50. pipsterate says:

    Interesting that I seem to have a lower IQ than the average here, and also a much lower appreciation of puns. I would expect more intelligent people to prefer more complex types of humor.

    • JASSCC says:

      I think the nerd-appeal to puns is that they can be extremely complex / abstruse.

      In well-crafted pun stories (stories where the pay-off or punchline is a pun), the pun fits like a key into a lock. The pleasure in such a story is realizing the whole thing works as a puzzle prompting you up to fill that particular pun in as an answer. If it’s done really well, you even might be able to realize exactly what the pun *must* be when you get to the penultimate line of the story, and then the payoff is confirmation on the next, final line.

      • Joe says:

        This comes to mind as a good example.

        • One Name May Hide Another says:

          Ha. Enjoyed it. Thanks for sharing.

          • Mark says:

            I honestly thought that the point of puns was to be annoying and that that was where the humour came from.

            Are there people who genuinely enjoy the punchline of that story, in and of itself?

            I kind of enjoy that kind of thing in the same way I might enjoy getting rick-rolled – ah, well done, you got me.

          • Nancy Lebovitz says:

            Mark, I enjoy some puns.

            In the case of that story, it was more an appreciation of having been cleverly had because the story was so long and had major themes about the purpose of life–and it had quite a harrowing beginning– and at the end, it was a pretty forced pun.

            Edted to add: Mostly I was annoyed. It was a story which deserved a better ending.

          • Nancy Lebovitz says:

            Here’s a pun that I actually enjoyed:

            Which home appliance will be the next to spy on people?

            Your Roomba is getting the dirt onyou.

          • One Name May Hide Another says:

            I kind of enjoy that kind of thing in the same way I might enjoy getting rick-rolled – ah, well done, you got me.

            Yes. I think that describes my own reaction well. I felt trolled, but didn’t mind it very much. The whole thing was absurd, but highly entertaining in many ways.

            I also do agree with Nancy: I was a tad disappointed. A good story like this deserves an actual conclusion.

          • Rachael says:

            Mark: I didn’t enjoy the pun at the end of that story, because it was anticlimactic and because it doesn’t actually work in my dialect. But I do enjoy puns in general. Scott’s excellent lists of Swifties make me grin with sheer delight.

          • JASSCC says:

            That story was both a pun story and a shaggy dog story. I’m not so much a fan of the shaggy dog genre, because I don’t like they pay off well enough to pay the price. A shorter pun story is more fun.

            Another and nearly opposite form of pleasing puns are those that come cleverly up in conversation, especially when the punster manages to produce one after another, without reusing a root idea. That’s not so easy a trick, and it’s amusing and impressive in the way that juggling several balls is.

    • onyomi says:

      Just a theory: I remember the first time I saw a cat macro (lolcat). I found it to be REALLY funny. At the time, I remember thinking to myself: “am I genuinely amused by this or am I just amused by how stupid it is?” I honestly wasn’t sure, which made it funnier.

      In other words, the combination of being undeniably funny yet also so stupid it shouldn’t be funny may itself be funny to people prone to overanalyzing and/or taking themselves too seriously.

      I am not smart enough to enjoy the Swifties.

      • Aapje says:

        My theory is that intelligent people like surrealism.

        I like surrealism.

        • nimim.k.m. says:

          Case in point. In intellectual circles, Monty Python is often considered the finest example of the comedy created by the Western civilization.

          Regarding the internet-memespace, my favorites include monorail cat GIFs and doge.

  51. P. George Stewart says:

    *raises Spockean eyebrow* Fascinating.

    Heartwarming to see the political balance – and overall tolerance of the other “side”.

  52. Dave says:

    For those who would like to dig further into the raw data, I’ve posted it as a repository here ( is a site for collaborative exploration and publishing of datasets, and includes functionality for querying, enhancing, and visualizing data. If Scott would prefer his dataset not to be hosted on, I will of course take it down immediately.

    (Disclosure: I am an employee of, as well as a fan of Slate Star Codex.)

  53. Fahundo says:

    35% of you think upvotes would improve the comments section. 77% of you didn’t immediately see both meanings in the “I have read this book” question. Explain yourselves.

    Also: I first saw the spinning dancer about 9 years ago, and back then I was able to get it to reverse direction, but I was not able to when I took this survey.

    • suntzuanime says:

      It would be “many like it” if you were talking about other books that you’d read that were similar to the book in question. There’s a “much [reading material] like it” reading, but “reading material” is enough separated from the sentence context that it’s a bit of a stretch of an elision.

      I’m assuming the upvotelikers just haven’t thought deeply about the dynamics of internet communities.

      • Fahundo says:

        It would be “many like it” if you were talking about other books that you’d read that were similar to the book in question

        Yes, but wouldn’t you still at least stop and wonder if the person meant to say “many” instead?

        • suntzuanime says:

          Given that there’s a reasonable interpretation that doesn’t require rescuing, I would need some sort of context cue to make me assume that someone meant to say something other than what they actually said.

          Without context, I think it’s actually much more likely for someone to tell you about a book they’ve read and liked, rather than about a book they’ve read that’s similar to other books they’ve read.

          EDIT: Maybe the conflict stems from you taking a less strict view of the term “immediate”? I did notice the the second reading quickly, but it wasn’t my immediate interpretation of the sentence, so I didn’t answer “both meanings were immediately apparent to me”. If you have to “stop and wonder”, from my perspective that precludes immediacy.

          • Fahundo says:

            I’m pretty sure I took “immediate” to mean before I read far enough to know what question was about to be asked.

            What excuse do the 10% who saw the “I have read much like it” but not the “and I liked this book” interpretation have?

          • suntzuanime says:

            Dialectical variation such that “I much like [x]” isn’t possible?

          • Mark says:

            If the things you read weren’t books then you would say “much like it”.

          • switchnode says:

            What excuse do the 10% who saw the “I have read much like it” but not the “and I liked this book” interpretation have?

            The question asks “how do you interpret” (emphasis added), not ‘how might it be interpreted’; perhaps they considered the (subtextually) intended meaning more important than the surface ambiguity.

            Without context, I think it’s actually much more likely for someone to tell you about a book they’ve read and liked, rather than about a book they’ve read that’s similar to other books they’ve read.

            The usual form of the quote (although possibly not the original, as it seems to be apocryphal) is “I have read your book and much like it”, which would alter the likelihood ratio. I wonder whether Scott chose this version on purpose?

      • Kaj Sotala says:

        I’m assuming the upvotelikers just haven’t thought deeply about the dynamics of internet communities.

        Or have found that having upvotes empirically seem to provide better results than not having them, even taking the negative dynamics into account.

        On each site that I’ve seen that uses them (many, but most of my experience is from Less Wrong and Reddit), they do reliably highlight higher-quality comments and make the comment sections actually readable. Whereas on major sites that don’t use them (SSC included), finding the quality comments often becomes too much of a chore to be worth the effort.

        • nimim.k.m. says:

          I think the major problem with reddit style upvote system seems to be that it’s easy to find threads that start with highly upvoted comments. But what if there’s highly voted, highly valuable comment in a downthread of poor comment as a response to less well-received comment? And then there’s the tendency to form echochambers and posting content just to get more upvotes (especially if the upvotes are summed up as a permanent karma score).

          And sometimes it’s the controversial comments in the upvote / downvote system that are the most interesting (much of this of course depends on the crowd who is doing the voting and how much you agree with them).

          I think a much improved upvote-system (over the popular presently used ones) would present first the whole comment threads ranked by the upvote and controversiality tallies of all comments part of that thread with some randomization, and also maybe hide the actual vote counts. You’d get the benefit of highly voted content being easily found, but you’d avoid the karma-seeking behavior and ill feelings caused by making a comment you thought was fine but alas, wasn’t, and the ensuing “this is going to be downvoted to hell” preludes. Ranking “controversial” (lots of upvotes and downvotes) favorably and some randomization would be beneficial for weakening the echochamber effects.

    • Mark says:

      I think that ordering (or hiding) based on votes would be bad, but that votes would be good.

      (1) It would make it easier when scanning through comments to see things that are perhaps good but that you might have missed.

      (2) It would make it easier to see where the community has strong bias, but lacks good arguments.

      As for the reading thing – I’m not exactly sure how reading works, but I would guess that “I have read this book and much like it (many things like it)” is sufficiently unusual that there is no automatic connection in my brain to that meaning.
      I would guess that we normally read by parsing text at a level beyond that of the individual words – you’re looking at the phrase, and you’re associating that with something that you’ve seen before.

    • Kaj Sotala says:

      I tend to find the comments section basically unreadable without upvotes. Takes too much time to find the quality/interesting comments, whereas with an upvoting mechanism there would be no need as the quality comments would be immediately obvious.

      (Yes, I did end up reading this comment section so it’s not quite a totally universal rule, but in general I’ve ended up reading the comment section less and less as the number of comment(er)s has grown.)

      • Deiseach says:

        Yeah, but that depends heavily on “comment with most upvotes/most popular comment” = “highest quality/most interesting comment”. That might not be a bad assumption round these parts, but my fear is that upvote systems can devolve into popularity contests.

        And there’s always the chance that you could miss out on a comment that would be interesting to you just because it hasn’t attracted a lot of votes from others.

        • Hyzenthlay says:

          That might not be a bad assumption round these parts, but my fear is that upvote systems can devolve into popularity contests.

          I’m very much against upvoting systems for this reason. That and the echo-chamber effect. My experience with this feature on Reddit is that the comments which get the most upvotes are the ones which reaffirm the majority views of the subreddit. Dissenting opinions get downvoted, regardless of quality.

          Also, I feel like the system influences how I read comments. If all comments are presented neutrally, I have to process them myself and decide how I feel about them, which makes my brain work a little harder. If they have a +1,000,000 or -1,000,000 points next to them, that is going to influence my perception right off the bat.

          And, too, it influences how I comment. If there’s no upvote system I’m more likely to just say what I’m really thinking. If there’s an upvote/downvote system I’m more likely to craft a response that I think will get a positive reaction. It’s hard to escape this effect even when I try to work against it. Also, if I comment on something and it starts getting downvoted I’m prone to editing and censoring myself because my lizard brain wants to be popular.

          tldr: Upvote/downvote systems have a hivemind effect on a community, a lack of this system encourages individual thought.

          • Skivverus says:

            So, agree/disagree/respect instead of upvote/downvote, perhaps?

          • Amy says:


            Or something like Strongly agree/ Strongly disagree / Charitable/Uncharitable / Informative / Changed my mind / Rude / Spam

            I hypothesize that adding more specific options would trigger more System 2 thinking and reduce political hive-mind u/d voting. People would think about what label applies best instead of just mashing up and down buttons. Anyone interested in testing this?

          • Nornagest says:

            I guarantee agree/disagree/respect will turn into upvote/downvote in practice, just with two upvote butons.

          • Skivverus says:

            Maybe so, but where has this been tested? Discord comes to mind, but they have an arbitrarily-large-number-of-emotes method of reactions, which isn’t quite the same.

            I think there’s definitely a tradeoff curve between information gathered and usability; the “like” button is simpler and more used than upvotes/downvotes, which in turn are simpler than what we’re proposing here. I’d advise caution in skipping straight to eight options.

          • Nornagest says:

            Slashdot’s the seminal example of an upvote/downvote system with multiple upvote types. There are lots of upvote-only systems with multiple types (Facebook is probably the biggest). I don’t know of anywhere that does agree/disagree/respect specifically, but I very much doubt this specific phrasing is magic.

        • daniel says:

          I don’t think a voting system automatically means you can’t sort by chronology and whenever you don’t read all the comments there’s always a chance that you’ll miss one because of its position.

          I’m more worried attempts to get laughs than popularity contents per se and I think the vote systems that give a separate vote for funny and interesting/insightful mitigate this.

        • John Schilling says:

          From today’s Atlantic:

          “people tend to be extremely similar in their vulgar and prurient and dumb interests and wildly different in their refined and aesthetic and noble interests”

          QED: upvote systems drive you heavily towards stupid jokes and other lowbrow silliness, and political shibboleths once tribal dominance has been established.

          • Iain says:

            (It is only with the benefit of hindsight that I realize the obvious absurdity of responding to a post about why upvoting is a bad idea with a simulated upvote. To be clear: my +1 is not intended as some sort of weird joke, I agree completely with John Schilling’s post, and I should not be trusted with a keyboard.)

          • Deiseach says:

            QED: upvote systems drive you heavily towards stupid jokes and other lowbrow silliness

            Oh dear God, you strike me aghast with horror at the vision of an SSC filled with nothing other than pun posting forever and ever and ever and ever!

          • hlynkacg says:


            Between that and a boot stamping on a human face, I’ll take the boot!


          • John Schilling says:

            …an SSC filled with nothing other than pun posting forever and ever and ever and ever!

            There will also be Tom Swifties. And links to cute cat videos, even though we don’t get those now.

          • dndnrsn says:


            Nothing else but a boot? Solely?

          • hlynkacg says:

            @dndnrsn, I see what you did there ಠ__ಠ

    • Urstoff says:

      Upvotes only make sense to me for sites with enormous numbers of commenters. SSC doesn’t seem to be anywhere near that yet.

    • daniel says:

      upvotes: I kinda like reading comments, some of them are significant to put the posts in context but there are a lot of them, upvotes seem like a better way to zero in on the important comments than skimming and reading them randomly when I don’t have time to go through all of them. Also, when writing a comment, how do I know if anyone found it useful? how do signal to someone I found their comment insightful without leaving lots of trash “yeah, cool” comments on the site?
      I am worried about their possible negative impact though so I wouldn’t say I’m 100% pro upvotes.

      Meanings: I read the sentence very quickly, I might not have even parsed it before reading the rest of the questions and was left with the vague feeling that I only noticed one meaning, I generally read with very little attention to detail usually context saves me from mistakes.

      Spinning dancer: Same here, it drove me crazy, I must have spent several minutes trying to regain the ability to have it change directions at will.

      • Fahundo says:

        Spinning dancer: Same here, it drove me crazy, I must have spent several minutes trying to regain the ability to have it change directions at will.

        This is actually what I was interested in hearing more about. I still have no idea how I was able to lose the ability.

    • LCL says:

      Upvoting would be a service to people who don’t have time to read all the comments. I am often such a person, and would appreciate the service.

  54. Mark V Anderson says:

    I was very disappointed at how few found the mandatory registration burdensome, and how many think it somehow improved the discussion. I looked over the data set to try to determine other similarities between me and the other 296 people who think this is overly burdensome, but I don’t see anything. We don’t seem to have similar political views or demographics.

    I can see why there would be little downside to readers of comments, since there are are already so many comments that exclusion of some for whom it is too burdensome might be a feature rather than a bug. But I notice that I am well in the minority on this even amongst frequent commenters.

    I find that annoying comment sites are that way because there is so much ad hominem instead of reasoned discussion, and people seem to comment just to bait other commenters, instead of making actual arguments. The worst sites have regulars that do this all the time. I don’t see how registration would change such activity at all, because the nasty cat-calling is done by regulars that would register anyway. It isn’t one off trolling that destroys comment boards, it is the constant back and forth of ad hominems. Not that I’ve ever noticed that in SSC anyway.

    There was discussion of Marginal Revolution upthread. That is a site where the commenters are probably at a similar level of intelligence as SSC, but are much less civil. It is the regulars fighting back and forth that is the most annoying.

    I realize that this is a losing battle, based on the survey numbers, but sometimes I need to yell into the wind.

    • suntzuanime says:

      Did you not used to read the comments before the switch? There were a lot of Anonymouses who ran around saying nasty things. I don’t know why you think posting “just to bait other commenters, instead of making actual arguments” is limited to named posters. I guess if you’re worried about the constant back and forth of ad hominems, it solves the problem to have named posters just receive them with no way to reciprocate? As an occasional target of anonymous harassment in the pre-registration era, I gotta say I do not find that solution compelling.

      I’m on the balance not a fan of the registration requirement, but if you think that there wasn’t any positive aspect to the tradeoff you’re blind.

      • Mark V Anderson says:

        I guess I’m blind. I did read the comments before the switch. I remember it was a lot easier to make comments then. I have not noticed less trolling. I do think the comments have improved a bit in there being less back-biting and fewer pile-ons attacking folks. But I think that had nothing to do with the registration and had more to do with the conscious efforts of posters to try to avoid that. And maybe the end of the US elections helped a bit.

        I guess if you’re worried about the constant back and forth of ad hominems, it solves the problem to have named posters just receive them with no way to reciprocate?

        Sorry, I have no idea what this means.

        • suntzuanime says:

          You mentioned “the constant back and forth of ad hominems”. In the pre-registration era it seems to me that some of the worst users of ad-hominems were anonymous. However, this did not lead to a back and forth, as ad-hominems were hard to deploy against anonymous users, because there was no specific hominem to be ad. My point is, this still isn’t good, the reciprocity is not the problem.

          EDIT: I actually do remember getting fed up and making some nasty remarks about anonymouses as a class, so it’s not even clear that anonymous posting works that way. Unless everybody is anonymous. That’s the solution they use on 4chan, and it’s by all accounts a highly civil discussion forum with little to no trolling.

          • Fahundo says:

            I remember there were a couple of anonymi(sp?) to whom people gave nicknames and were able to track based on posting style.

          • Nornagest says:

            “Anonymice” is traditional, I think.

          • Mark V Anderson says:

            However, this did not lead to a back and forth, as ad-hominems were hard to deploy against anonymous users, because there was no specific hominem to be ad. My point is, this still isn’t good, the reciprocity is not the problem.

            Ah. Actually that is my point, that one off ad hominems were no big deal and easily ignored. In fact in my world doing them anonymously are a feature not a bug, but it does prevent the reciprocity, which to me is the biggest problem.

    • Cypren says:

      I would offer that the quality of discussion in an online community is based primarily on two things: an aggressively enforced, clearly-defined and fairly-applied moderation policy, and a barrier to entry. The necessity of the former should be obvious. The latter serves two purposes: first, it establishes a cost in time and effort to participate, dissuading drive-by “me too” and “u suck” responses from people with no real investment in the community who merely want to stroke their own tribal loyalties or eject some venom on an outgroup. Second, that same cost puts some teeth into the moderation policy by requiring a bit of effort to circumvent bans.

      It’s fascinating how low the costs can be and still be effective. One of the very few other communities out there that I’ve encountered that has a high-quality comments section would be MetaFilter. One of the primary reasons, in my view, is that there’s a one-time $5 cost to comment. This is a trivial amount of money for most people, but enough to trigger loss aversion and make people think twice about behavior that would result in a ban.

      I suggest that the registration process here serves the same purpose, to a lesser degree. Forcing someone to come up with a new throwaway email address and go through the signup process again every time they want to spew some “outgroup sucks” venom takes some of the fun out of the process. It makes it require just a little effort, and that effort is more than a lot of trolls are willing to incur for a drive-by spitting.

      • Mark V Anderson says:

        I would offer that the quality of discussion in an online community is based primarily on two things: an aggressively enforced, clearly-defined and fairly-applied moderation policy, and a barrier to entry.

        This makes logical sense, but I haven’t seen it to be true. Of course I am apparently blind 🙂 and didn’t notice things being worse before (Sun, I am not taking offense here, but I am honestly confused). In my experience, it is the ongoing contributors on most blogs that cause the problems. But I am clearly in the minority here.

        I am also not in favor of Scott’s reign of terror — I feel like he bans folks too easily (although it’s not too big a deal if it’s only for a few weeks). But it is true that SSC is much more civil than most other blogs, so maybe that has had a good effect.

    • JohnBuridan says:

      I voted with you.
      And it wasn’t until yesterday that I capitulated and began commenting again. I also hate the the obstructing, obsidian, insidious wordpress bar taking up my precious screen space.

      Also, I haven’t noticed some big difference in the comment section, but apparently others thought they did.

      Maybe everyone who voted against the login system has authority issues?…

      • Iain says:

        You can get rid of the WordPress bar. If you click on the “Logged in as JohnBuridan” link underneath “Leave a Reply”, it will take you to your WordPress profile, where you can uncheck “Show Toolbar when viewing site”.

  55. Amy says:

    I would suggest next time using “sex assigned at birth” rather than “biological sex” – not only does the latter play into some harmful stereotypes (e.g. “biological sex” vs “gender identity” is just a few conceptual steps away from “objective sex” vs “subjective gender” or even “physical sex” vs “imaginary gender”), but it’s also vague and can be misleading. Brain differences, and secondary sex characteristics that develop from hormones would definitely qualify as biological factors – an MtF person who started hormones before or early on in puberty would be much closer to female than male when you look at her biology as a whole, since hormones mediate almost all sex differentiation. Also you have trans intersex people who were assigned the wrong gender, with or without knowing they were intersex – e.g. one of my MtF friends is actually 47,XXY, but only learned that after she transitioned (for all I know, I might be the same – I never tested my DNA). Using “sex assigned at birth” is unambiguous and avoids both these issues.

    • Loke says:

      The actual survey did say “sex assigned at birth”; Scott just used a simpler phrasing when reporting the results.

    • Hyzenthlay says:

      I would suggest next time using “sex assigned at birth” rather than “biological sex” – not only does the latter play into some harmful stereotypes (e.g. “biological sex” vs “gender identity” is just a few conceptual steps away from “objective sex” vs “subjective gender” or even “physical sex” vs “imaginary gender”), but it’s also vague and can be misleading.

      I agree that “biological sex” as a phrase is vague and “sex assigned at birth” is better. But I do think of gender as being inherently more subjective than physical characteristics. Sex-related characteristics (whether you’re talking about chromosomes or secondary sex characteristics or hormones or brain structure) can be measured and observed. Gender can’t.

      I mean, I’m in my thirties and I still haven’t completely figured out what my gender is or what gender itself is. Some people do have a very clear and stable sense of their own gender identity that manifests early in life, but lots of people don’t.

      I wouldn’t say gender is “imaginary,” at least no more so than a concept like capitalism or spirituality is imaginary. But I would say that gender is inherently subjective, and trying to define it in a non-subjective way would inevitably involve stereotyping.

  56. TomA says:

    In your book review of “Seeing like a state”, I gather that you found the information contained therein to be fascinating, but ultimately did not know quite what to make of it all (e.g. how it might become concretely applicable to your life goals or important interests).

    It seems to me that your survey and the accompanying results may fall into the same category. You now have some detailed insight into a self-selected cohort of SSC readers. How does that create advantage or benefit (as opposed to being just an interesting exercise in data-gathering)?