2014 Survey Results

Scott Alexander

The number of Asians (both East and South) among American readers is pretty surprisingly low - 43/855 ~= 5%. This despite Asians being, e.g., ~15% of the Ivy League student body (it'd be much higher without affirmative action), and close to 50% of Silicon Valley workers.

[-]someonewrongonthenet11y320

Being south asian myself - I suspect that the high achieving immigrant-and-immigrant-descended populations gravitate towards technical fields and Ivy leagues for different reasons than American whites do. Coming from hardship and generally being less WEIRD, they psychologically share more in common with the middle class and even blue collar workers than the Ivy League upper class - they see it as a path to success rather than some sort of grand purposeful undertaking. (One of the Asian Professional community I participated in articulated this and other differences in attitude as a reason that Asians often find themselves getting passed over for higher level management positions, as something to be overcome).

Lesswrong tends to appeal to abstract, starry-eyed types. I hate to use the word "privilege", but there is some hard to quantify things, like degree of time talking about lesswrong-y key words like "free will" or "utilitarianism", which are going to influence the numbers here. (Not that asians don't like chatting about philosophy, but they certainly have less time for it and also they tend to focus on somewhat different topics during philosophical... (read more)

2laofmoonster11y

East Asian - mostly agreed. I think WEIRDness is the biggest factor. WEIRD thought emphasizes precision and context-independent formalization. I am pretty deracinated myself, but my thinking style is low-precision, tolerant of apparent paradoxes and context-sensitive. The difference is much like the analytic-continental divide in Western philosophy. I recommend Richard Nisbett's book The Geography of Thought, which contrasts WEIRD thought with East Asian thought. 37 Ways Words Can Be Wrong (and LW as a whole) is important because of how brittle WEIRD concepts can be. (I have some crackpot ideas about maps and territories inspired by Jean Baudrillard. He's French, of course...)

3skeptical_lurker11y

Is affirmative action being used against Asian even though they are a minority?

[-]Nornagest11y140

"Used against", to me, implies active planning that may or may not exist here; but the pragmatic effects of the policy as implemented in American universities do seem to negatively affect Asians.

4skeptical_lurker11y

Ahh, the old 'malicious or incompetent' dichotomy.

3Nornagest11y

I'm a big believer in Hanlon's razor, especially as it applies to policy.

[-]epursimuove11y130

There's pretty unambiguous statistical evidence that it happens. The Asian Ivy League percentage has remained basically fixed for 20 years despite the college-age Asian population doubling (and Asian SAT scores increasing slightly).

-5Robin11y

2Vaniver11y

I've noticed this for a while. Might be interesting to look at this by referral source?

[-]Vaniver11y230

Calibration Score

Using a log scoring rule, I calculated a total accuracy+calibration score for the ten questions together. There's an issue that this assumes the questions are binary when they're not- someone who is 0% sure that Thor is the right answer to the mythology question gets the same score (0) as the person who is 100% sure that Odin is the right answer to the mythology question. I ignored infinitely low scores for the correlation part.

I replicated the MWI correlation, but I noticed something weird- all of the really low scorers gave really low probabilities to MWI. The worst scorer had a score of -18, which corresponds to giving about 1.6% probability to the right answer. What appears to have happened is they misunderstood the survey, and answered in fractions instead of percents- they got 9 out of 10 questions right, but lost 2 points every time they assigned 1% or slightly less than 1% to the right answer (i.e. they mean to express almost certainty by saying 1 or 0.99) and only lost 0.0013 points when they assigned 0.3% probability to a wrong answer.

When I drop the 30 lowest scorers, the direction of the relationship flips- now, people with better log scores (i.e. close... (read more)

2Luke_A_Somers11y

I've always wanted to visit 100. Can you show the distribution of overall calibration scores? You only talked about the extreme cases and the differences across P(MWI), but you clearly have it.

3Vaniver11y

Picture included, tragic mistakes excluded*. The percentage at the bottom is a mapping from the score to probabilities using the inverse of "if you had answered every question right with probability p, what score would you have?", and so is not anything like the mean probability given. Don't take either of the two perfect scores seriously; as mentioned in the grandparent, this scoring rule isn't quite right because it counts answering incorrectly with 0% probability as the same as answering correctly with 100% probability. (One answered 'asdf' to everything with 0% probability, the other left 9 blank with 0% probability and answered Odin with 100% probability.) Bins have equal width in log-space. * I could have had a spike at 0, but that seems not quite fair since it was specified that '100' and '0' would be treated as '100-epsilon' and 'epsilon' respectively, and it's only a Tragic Mistake if you actually answer 0 instead of epsilon.

0Luke_A_Somers11y

Yeah, that's not a particularly strong scoring method, due to its abusability. I wonder what a better one would be? Of course, it wouldn't help unless people knew that it was going to be used, and care. Fraction correct times this calibration score? Number correct times the product rather than the average of what you did there? Bayes score, with naming the 'wrong' thing yielding a penalty to account for the multiplicity of wrong answers (say, each wrong answer has a 50% hit so even being 100% sure you're wrong is only as good as 50% sure you're right, when you are right)?

2Vaniver11y

The primary property you want to maintain with a scoring rule is that the best probability to provide is your true probability. I know that the Bayes score generalizes to multiple choice questions, which implies to me that it most likely works with a multiplicity for wrong answers, so long as the multiplicity is close to the actual multiplicity.

2Luke_A_Somers11y

I think the primary property you want to maintain is that it's best to provide the answer you consider most likely, otherwise it's best to say 'sdfkhasflk' - 0% to all of them you aren't certain of. Multiple choice would making the scoring clearer, but that constraint could well make the calibration easier.

0MTGandP11y

Sort-of related question: How do you compute calibration scores?

0Vaniver11y

I was using a logarithmic scoring rule, with a base of 10. (What base you use doesn't really matter.) The Excel formula for the first question (I'm pretty sure I didn't delete any columns, so it should line up) was: =IF(EJ2,LOG(EU2/100),LOG(1-EU2/100))

[-]redlizard11y230

MIRI Mission/MIRI Effectiveness .395 (1331)

This result sets off my halo effect alarm.

[-]William_Quixote11y220

Once again pandemic is the leading cat risk. It was the leading cat risk last year. http://lesswrong.com/lw/jj0/2013_survey_results/aekk It was the leading cat risk the year before that. http://lesswrong.com/lw/fp5/2012_survey_results/7xz0

Pandemics are the risk LWers are most afraid of and to my knowledge we as a community have expended almost no effort on preventing them.

So this year I resolve that my effort towards pandemic prevention will be greater than simply posting a remark about how it's the leading risk.

[-]blacktrance11y290

Clearly, we haven't been doing enough to increase other risks. We can't let pandemic stay in the lead.

1Ander11y

Get to work on making more AIs everyone!

[-]Fluttershy11y120

Givewell has looked into global catastrophic risks in general, plus pandemic preparedness in particular. My impression is that quite a bit more is spent per year on biosecurity (around 6 billion in the US) than is on other catastrophic risks such as AI.

[-][anonymous]11y110

Pandemics may be the largest risk, but the marginal contribution a typical LWer can make is probably very low, and not their comparative advantage. Let the WHO do its work, and turn your attention to underconsidered risks.

427chaos11y

Money can be donated.

4someonewrongonthenet11y

I'm not so sure about that. Isn't the effective altruist focus on global poverty/disease reducing the risk of pandemic? I know very little about epidemiology, but if seems as if a lot of scary diseases (AIDs, ebola...) would never have spread to the human population if certain regions of the third world had better medical infrastructure.

2William_Quixote11y

That's fair. It's certianly true that poverty reduction also reduces pandemic risk. But it does so inditectly and slowly. There are probably faster ways to reduce pandemic risk than working on poverty.

[-]iarwain111y210

WHY ISN’T THERE AN OPTION FOR NONE SO I CAN SIGNAL MY OBVIOUS OBJECTIVITY WITH MINIMAL EFFORT

This is why I didn't vote on the politics question.

This is a sign that most Less Wrongers continue to neglect the very basics of rationality and are incapable of judging how much evidence they have on a given issue. Veterans of the site do no better than newbies on this measure.

Theory: People use this site as a geek / intellectual social outlet and/or insight porn and/or self-help site more than they seriously try to get progressively better at rationality. At least, I know that applies to me :).

[-]FeepingCreature11y330

This definitely belongs on the next survey!

Why do you read LessWrong? [ ] Rationality improvement [ ] Insight Porn [ ] Geek Social Fuzzies [ ] Self-Help Fuzzies [ ] Self-Help Utilons [ ] I enjoy reading the posts

3MTGandP11y

And then check if the "rationality improvement" people do better on calibration. (I'm guessing they don't.)

1homunq10y

[ ] Wow, these people are smart. [ ] Wow, these people are dumb. [ ] Wow, these people are freaky. [ ] That's a good way of putting it, I'll remember that. (For me, it's all of the above. "Insight porn" is probably the biggest, but it doesn't dominate.)

0Sabiola11y

[x] All of the above

7[anonymous]11y

Is that actually nonobvious? It's sure as hell what I'm here for. I mean, I do actually generally want to be more rational about stuff, but I can't get that by reading a website. Inculcating better habits and reflexes requires long hours spent on practicing better habits and reflexes so I can move the stuff System 2 damn well knows already down into System 1.

[-]Unnamed11y190

I decided to take a look at overconfidence (rather than calibration) on the 10 calibration questions.

For each person, I added up the probabilities that they assigned to getting each of the 10 questions correct, and then subtracted the number of correct answers. Positive numbers indicate overconfidence (fewer correct answers than they predicted they'd get), negative numbers indicate underconfidence (more correct answers than they predicted). Note that this is somewhat different from calibration: you could get a good score on this if you put 40% on each question and get 40% of them right (showing no ability to distinguish between what you know and what you don't), or if you put 99% on the ones you get wrong and 1% on the ones you get right. But this overconfidence score is easy to calculate, has a nice distribution, and is informative about the general tendency to be overconfident.

After cleaning up the data set in a few ways (which I'll describe in a reply to this comment), the average overconfidence score was 0.39. On average, people expected to get 4.79 of the 10 questions correct, but only got 4.40 correct. My impression is that this gap (4 percentage points) is smallish compared ... (read more)

[-]Unnamed11y160

Details on data cleanup:

In the publicly available data set, I restricted my analysis to people who:

entered a number on each of the 10 calibration probability estimates
did not enter any estimates larger than 100
entered at least one estimate larger than 1
entered something on each of the 10 calibration guesses
did not enter a number for any of the 10 calibration guesses

Failure to meet any of these criteria generally indicated either a failure to understand the format of the calibration questions, or a decision to skip one or more of the questions. Each of these criteria eliminated at least 1 person, leaving a sample of 1141 people.

I counted as "correct":

any answer which Scott/Ozy counted as correct
any answer to question 1 (largest bone) which began with "fem" (e.g., "femer")
any answer to question 2 (Obama's state) which began with "haw" (e.g., "Hawii")
any answer to question 4 (Norse god) which began with "od" or "wo" (e.g., "Wotan")
any answer to question 8 (cell) which began with "mito" (e.g., "Mitochondira")

These seem to cover the most common ... (read more)

[-]Unnamed11y130

And here's an analysis of calibration.

If a person was perfectly calibrated, then each 10% increase in their probability estimate would translate into a 10% higher likelihood of getting the answer correct. If you plot probability estimates on the x axis and whether or not the event happened on the y axis, then you should get a slope of 1 (the line y=x). But people tend to be miscalibrated - out of the questions where they say "90%", they might only get 70% correct. This results in a shallower slope (in this example, the line would go through the point (90,70) instead of (90,90)) - a slope less than 1.

I took the 1141 people's answers to the 10 calibration questions as 11410 data points, plotted them on an x-y graph (with the probability estimate as the x value and a y value of 100 if it's correct and 0 if it's incorrect), and ran an ordinary linear regression to find the slope of the line fit to all 11410 data points.

That line had a slope of 0.91. In other words, if a LWer gave a probability estimate that was 10 percentage points higher, then on average the claim was 9.1 percentage points more likely to be true. Not perfect calibration, but not bad.

If we look at various s... (read more)

[-]Jacob Falkovich11y160

Myth: Americans think they know a lot about other countries but really are clueless.

Verdict: Self-cancelling prophesy.

Method: Semi-humorous generalization from a single data series, hopefully inspiring replication instead of harsh judgment :)

I decided to do some analysis about what makes people overconfident about certain subjects, and decided to start with an old stereotype. I compared how people did on the population calibration question (#9) based on their country.

Full disclosure: I'm Israeli (currently living in the US) and would've guessed Japan with 50% confidence, but I joined LW (unlurked) two days after the end of the survey.

I normalized every probability by rounding extreme confidence values to 1% and 99% and scored each answer that seemed close enough to a misspelling of Indonesia according to the log rule.

Results: Americans didn't have a strong showing with an average score of -0.0071, but the rest of the world really sucked with an average of -0.0296. The reason? While the correct answer rate was almost identical (28.3% v 28.8%) Americans were much less confident in their answers: 42.4% confidence v 46.3% (p<0.01).

Dear Americans, you don't know (significantly) less about the world than everyone else, but at least you internalized the fact that you don't know much*!

Next up: how people who grew up in a religious household do on the Biblical calibration question.

*Unlike cocky Israelis like me.

[-]Baisius11y150

I'm losing a lot of confidence in the digit ratio/masculinity femininity stuff. I'm not seeing a number of things I'd expect to see.

First, my numbers for correlations don't match up with yours. With filters on for female gendered, and answering all of BemSexRoleF, BemSexRoleM, RightHand, and LeftHand, I get a correlation of only -0.34 for RightHand and BemSexRoleM, not -0.433 as you say. I get various other differences as well, all weaker correlations than you describe. Perhaps differences in filtering explain this? -.34 vs -.433 seems to be high for this to be true though.

Second, Bem masculinity and femininity actually seem to have a positive correlation, albeit tiny. So more masculine people are... more feminine? This makes no sense and makes me more likely to throw out the entire data set.

Thirdly, I don't see any huge differences between Cisgender Men, Transgender Men, Cisgender Women, or Transgender Women on digit ratios. I would expect to see this as well. I get 95% confidence intervals (mean +/- 3*sigma/sqrt(n), formatted [Lower Right - Upper Right / Lower Left - Upper Left]) for the categories as follows:

F (Cis): [0.949 - 0.996 / 0.956 - 1.004]
M (Cis): [0.962 - 0.978 / 0.

... (read more)

[-]simon11y120

There isn't necessarily any problem with a small positive correlation between masculinity and femininity. The abstract of what I think is the original paper (I couldn't find an ungated version) says that "The dimensions of masculinity and femininity are empirically and logically independent."

[-]gwern11y110

https://pdf.yt/d/zguK1Egu2kFycsA- / https://www.dropbox.com/s/4ky5snzyteftmz2/1974-bem.pdf

2Baisius11y

It's not clear that this maps to colloquial use of the terms "feminine" and "masculine" then. I think most would consider them opposite ends of the same spectrum.

7Nornagest11y

There are aspects of the Western gender roles that are opposed to each other at least to some extent: emotionality vs. stoicism, active vs. passive romantic performance. But there are also aspects that aren't. Blue is not anti-pink. Skill at sewing doesn't forbid skill at fixing cars. These might resolve in people's perceptions to positions on some kind of spectrum of male vs. female presentation, but they won't show up that way on surveys measuring conformity with stereotype. Indeed, that suggests a possible mechanism for these results. Assume for a moment that people prefer to occupy some particular point on the perception spectrum. But people often like stuff for reasons other than gender coding, so it'll sometimes happen that people will be into stuff with gender coding inconsistent with how they'd prefer to be seen. That creates pressure to take up other stuff with countervailing coding. If people respond to that pressure, the net result is a weak positive correlation between stuff with masculine and feminine coding.

[-]gothgirl42066611y150

I would be really interested in hearing from one of the fourteen schizophrenic rationalists. Given that one of the most prominent symptoms of schizophrenia is delusional thinking, a.k.a. irrationality... I wonder how this plays out in someone who has read the Sequences. Do these people have less severe symptoms as a result? When your brain decides to turn against you, is there a way to win?

I also find it fascinating that bisexuality is vastly overrepresented here (14.4% in LW vs. 1-2% in US), while homosexuality is not. My natural immediate interpretation of this is that bisexuality is a choice. I think Eliezer said once that he would rather be bisexual than straight, because it would allow for more opportunities to have fun. This seems like an attitude many LW members might share, given that polyamory a.k.a. pursuing a weird dating strategy because it's more fun is very popular in this community. (I personally also share Eliezer's attitude, but unfortunately I'm pretty sure I'm straight.) So to me it seems logical that the large number of bisexuals may come from a large number of people-who-want-to-be-bisexual actually becoming so. This seems more likely to me than some aspect or... (read more)

[-]Gunnar_Zarncke11y140

I also find it fascinating that bisexuality is vastly overrepresented here.

I don't. Compare it with the OkCupid data analysis. Bisexuality could be more of a signal. Admittedly at least in the (quite large) OkCupid data.

5gothgirl42066611y

Oh, wow, that's incredibly strange/interesting, I had never seen that before. Thanks for sharing. The fact that young bi men are almost always closeted gay men, while old bi men are almost always closeted straight men, is particularly baffling.

7Izeinwinter11y

The first part does not actually follow from the data with any rigor - "Go online to meet people of the same sex, find opposite-sex partners in real life" is a perfectly reasonable strategy, simply because online dating avoids the whole "I'm straight" shot down in flames thing, which must get really old really quickly. The older guys listing bisexuality and only messaging women, tough? Ehhr.. what?

7SeekingEternity11y

Best guesses at an explanation for that one: 1) A lot of older men had some homosexual experimentation in their past, decided that they therefore count as bi, but are now only interested in heterosexual relationships. 2) A lot of older men choose to signal what they believe to be the desirable characteristic of "sexual adventurousness" to their actual target sexual partner, which is younger women.

1I_fail_at_brevity11y

Plus, there may be many bisexual men specifically looking for a partner they can breed with. Based roughly on barely remembered male fertility age statistics, I'd guess men would be most interested in fathering children in the 25-45 age rage, and there does seem to be a bit of a hump in the data in that range.

7Nornagest11y

Hypothesis: a large fraction of young men in those results are coming to terms with their sexuality, while a large fraction of old men are trying to signal sexual adventurousness?

4gothgirl42066611y

Yeah, that's what I thought too. I'm just surprised that bisexuality would be something so many men imagine (perhaps correctly?) women are attracted to.

6Vaniver11y

I don't find the first part baffling; there's a trope that many gay men go through bisexuality on their way to accepting their homosexuality. (I had a brief period where I identified as bi because I wasn't fully ready to identify as gay.)

5Toggle11y

Same. It's easier to tell people that you have a left hand than it is to tell people you're left-handed, so to speak.

5Username11y

Nah, selection bias. You don't go on OK Cupid as a bi man to find men - that's Grindr or other similar sites. Much easier and quicker and more straightforward. But if you're a bi man looking for women, OK Cupid is a good place to go.

3Error11y

I'd be interested to see the orientation numbers broken down by sex/gender. My personal experience is that geek/nerd women seem to be bisexual at surprisingly high rates. I'm wondering if having typically-male personal pursuits (e.g. LW) is correlated with typically-male sexual interests (i.e. liking women). I'm in that boat. Feels like I'm missing out on half the potential fun. :-(

9Vaniver11y

Using the "Sex" (not gender) and "Sexuality" columns, omitting blanks, asexuals, and others: Male Heterosexual: 999 Male Bisexual: 142 Male Homosexual: 40 Female Heterosexual: 79 Female Bisexual: 62 Female Homosexual: 6 So the male/female ratio by sexuality is: Heterosexual: 12.6 Bisexual: 2.3 Homosexual: 6.7 The sexuality percentage by sex is: Male: 84.6% / 12.0% / 3.4% Female: 53.7% / 42.2% / 4.1% So while female bisexuality is almost as common as female heterosexuality here, the total bisexual ratio resembles the male bisexual ratio closely, as you would expect from the male/female ratio being so high overall (8 men per woman in this restricted sample).

[-]Error11y130

almost as common as female heterosexuality here, as you would expect

I initially misparsed this as "the female bisexuality rate is as expected." I see that isn't what you meant, but had to re-read two or three times. Just FYI.

I feel like a 42.2% bisexuality rate among LW women is surprising enough to say something, but I'm not sure what.

7JohannesDahlstrom11y

It is interesting. IME in real life and in OkCupid, female self-identification as bisexual correlates quite strongly with the geek/liberal/poly/kinky meme complex (edit: mirroring your experiences, didn't read carefully enough). Out of my top matches in OkCupid, over 80% of women interested in men seem to self-report as bisexual. However, also IME, bisexual identification usually doesn't imply being biromantic! Many of those women have had, or would like to have, sexual experiences with other women, but still may prefer men in romantic relationships almost exclusively. FWIW, I support adding a question about romantic orientation in the next survey.

6buybuydandavis10y

Great line from OkCupid:

4SeekingEternity11y

Anecdotally, this matches my experience (both on OKC and the "bisexual but hereroromantic" thing with three of my four most recent sexual partners).

3Vaniver11y

Grammar modified to be clearer, thanks for pointing that out.

0roystgnr11y

All I've come up with is a half-formed joke about how human females really are intrinsically attractive after all.

1Error11y

While an appealing hypothesis, if that were the case I would expect roughly the same percentage for the general public. The wiki of a million lies suggests the actual rate for the general public is in the low single digits.

5Vulture11y

As clever as this phrase is, it is tragically ambiguous. I'm guessing 65% chance Wikipedia, 30% RationalWiki, 3% our local wiki, 2% other. How did I do?

6Error11y

I meant Wikipedia. I've actually never heard the phrase applied to any other wiki. It's certainly not original to me.

2Vulture11y

Thanks!

3Alsadius11y

None of the other wikis you list are big enough to have more than maybe 75,000 lies.

1Nornagest11y

Are you counting talk pages? I'd expect those to have a higher density of lies than the main namespace.

0Alsadius11y

Sure, but I'd expect that smaller wikis have exponentially less talk, because there's fewer people to do the talking.

-4Unknowns11y

It seems that women are borderline bisexual by nature. For example, heterosexual women are significantly more likely to want to dance with other women than heterosexual men are to dance with other men, and the same thing is true for all sorts of other activities that have some kind of borderline relationship with sexual activities. So perhaps there is a kind of implicit bisexuality there which is more often made explicit in the case of Less Wrong women than other women, perhaps on account of higher introspection or the like.

8KPier11y

I am suspicious of this as an explanation. Most straight-identified women I know who will dance with/jokingly flirt with other women are in fact straight and not 'implicitly bisexual'; plenty of them live in environments where there'd be no social cost to being bisexual, and they are introspective enough that 'they are actually just straight and don't interpret those behaviors as sexual/romantic' seems most likely. Men face higher social penalties for being gay or bisexual (and presumably for being thought to be gay or bisexual) which seems a more likely explanation for why they don't do things that could be perceived as showing romantic interest toward men (like dancing or 'joking' flirting) than that women are borderline bisexual by nature.

7NancyLebovitz11y

Men who aren't bisexual are missing considerably less than half the potential fun, since the proportion of men who are gay or bisexual is fairly low.

4gothgirl42066611y

Yeah, but gay men are also more promiscuous.

0Alsadius11y

Is your comparison "than straight men" or "than straight women" here?

-3cameroncowan11y

I think being bi is simply being open-minded to all the potentials of relationships. But I agree the number of people to whom you might be engaged in sex or romance does not significantly increase. But I think the dual sexuality thing is dumb because sexuality is fluid. If I had a nickle for every time I went to bed with a "straight" man we could have a nice dinner.

[-]randallsquared11y140

May is missing from Birth Month.

[-][anonymous]11y110

I think it's pretty astounding that nobody at Less Wrong was born in May. I'm not sure why Scott doesn't think that's a deviation from randomness.

2imuli11y

May is in the data, a copy-paste error is much less astounding than nobody being born in May. 119 respondents, nothing surprising here.

4gjm11y

You might consider the hypothesis that FrameBenignly appreciates this and was making a joke. This seems much more likely to me than that s/he actually thinks no one said they were born in May. (Of course, maybe I'm missing a meta-joke where you pretend to take FrameBenignly at face value just as s/he pretended to take the alleged survey data at face value. But then maybe you're now missing a meta-meta-joke where I pretend to take you at face value...)

5[anonymous]11y

I'd make a triple-meta joke, but there's a two-meta limit on all month of birth jokes.

2imuli11y

Oh no! I forgot to leave my evidence.

0TheOtherDave11y

I see what you did there.

[-]ilzolende11y140

Thanks for showing us that there are autistic cryonics patients in the world. I am more likely to sign up when I am old enough to legally do so without parental permission, because now I know I wouldn't be the only autistic person in the future, no matter what happens when people develop a prenatal autism test.

8NancyLebovitz11y

I believe that if cryonics works, people will tend to associate with those from their home era.

[-]Vaniver11y130

Thanks for doing this!

[This question was poorly worded and should have acknowledged that people can both be asexual and have a specific orientation; as a result it probably vastly undercounted our asexual readers]

I find the "vastly" part dubious, given that 3% asexual already seems disproportionately large (general population seems to be about 1%). I would expect for asexuals to be overrepresented, and I do think the question wording means the survey's estimate underestimates the true proportion, but I don't think that it's, say, actually 10% instead of actually 4%.

0Richard Korzekwa11y

Why do you expect this? It seems reasonable if I think in terms of stereotypes. Also, I guess LWers might be more likely to recognize that they are asexual.

1Vaniver11y

Mostly the negative relationship between intelligence and interest in sex / sexual activity, especially when nerds are involved.

1Lumifer11y

There might be a negative relationship between intelligence and success in having sex, which is a different issue not connected to asexuals.

1alienist11y

Well, it's possible the asexuals got that way from accepting that they were never going to have sex. Also smart people are more likely to take ideas seriously, including the idea prevalent in many social circles that having a sex drive is evil. See Scott Aaronson's recent comment about how he once begged to be chemically castrated.

3Lumifer11y

While many things are possible, I don't think this is quite the way it works with asexuals... On the contrary, I think smart people are more likely to recognize that certain "prevalent in many social circles" ideas are bullshit or outright malicious. Scott Aaronson's problems in this respect did not arise because he is very smart.

0lalaithion11y

I think that, while it is indeed possible for asexuality to arise that way, most evidence seems to point away from that conclusion....

0Wes_W11y

Keep in mind also that other non-heterosexual orientations are also overrepresented, and I don't think anyone is quite sure why, but the same effect maybe applies to asexuals.

4ssica300311y

I think it's less a case of over-representation and more a case of a group of people who believe strongly in giving honest answers to survey questions in order to get good data and who are reasonably sure their privacy will be protected. Most surveys on sexuality suffer from reluctance to self-report. This (and last year's) figure for bisexuality in particular is more in line with my anecdotal & lived experience than 'official' survey data on the topic (bisexual people <1% population). Bisexual people (and bisexual men in particular) do exist! Yay! (We knew that, lol).

4alienist11y

That's probably a function of your social circle.

0[anonymous]11y

What actually demonstrates this? I know plenty of nerds with healthy sex drives.

4Good_Burning_Plastic11y

See e.g. the post "Intelligence and Intercourse" on the blog Gene Expression (though it appears to only mention studies about people in the US).

[-]Fluttershy11y110

Good job on running the survey and analyzing the data! I do wish that one of the extra credit questions had asked whether or not readers were fans of My Little Pony: Friendship is Magic.

[-]Nate_Gabriel11y90

P Supernatural: 6.68 + 20.271 (0, 0, 1) [1386]

P God: 8.26 + 21.088 (0, 0.01, 3) [1376]

The question for P(Supernatural) explicitly said "including God." So either LW assigns a median probability of at least one in 10,000 that God created the universe and then did nothing, or there's a bad case of conjunction fallacy.

9epursimuove11y

4Scott Garrabrant11y

Conjunctions do not work with medians that way. From what you quoted, it is entirely possible that the median probability for that claim is 0. You can figure it out from the raw data.

3TheMajor11y

I don't understand. Since existence of God is explicitly included in the question about the existence of supernatural things, everybody should have put P(God) < P(Supernatural), and therefore the median also is lower (since for every entry P(God) there is a higher entry P(Supernatural) by that same person). So the result above should be weak evidence that a significant proportion of the LW'ers fell prey to the conjunction fallacy here, right?

2Scott Garrabrant11y

No, I think that a god that does not interfere with the physical universe at all counts as not supernatural by the wording of the question. My point was that the median of the difference of two data sets is not the difference of the median. (although it is still evidence of a problem)

0[anonymous]11y

Something else I noticed: Agnostic: 156, 10.4% Lukewarm theist: 44, 2.9% Deist/pantheist/etc.: 22,, 1.5% Committed theist: 60, 4.0% A true agnostic should be 50% on the probability of God, but we'll say 25-75% as reasonable. A lukewarm theist should be 50-100%. I don't like the deist wording, but we'll say 50-100% for them, and 75-100% for the committed theists. We then get: 10.4.25+2.9.5+1.5.5+4.75 = 7.8% P God as our lower bound Compared to the 8.26% actual That's assuming all the atheists assigned a 0% probability to God. So it seems everybody is very close to their minimum on this; even likely below the minimum for some of them. My guess is a lot of people have some major inconsistencies in their views on God's existence.

[-]Vaniver11y130

1319 people supplied a probability of God that was not blank or "idk" or the equivalent thereof as well as a non-blank religion. I was going to do results for both religious views and religious background, but religious background was a write-in so no thanks.

Literally every group had at least one member who supplied a P(God) of 0 and a P(God) of 100.

0Richard Korzekwa11y

Okay, I'll bite: What does someone mean when they say they are Atheist, and they think P(God) = 100% ?

6jbay11y

According to Descartes: for any X, P(X exists | X is taking the survey) = 100%, and also that 100% certainty of anything on the part of X is only allowed in this particular case. Therefore, if X says they are Atheist, and that P(God exists | X is taking the survey) = 100%, then X is God, God is taking the survey, and happens to be an Atheist.

3Nornagest11y

Either they're actually a misotheist, or they're using a nonstandard definition of "God" or of "atheist" (though I think at least the former was defined on the survey), or they misunderstood the question, or they're trolling.

1Alsadius11y

Presumably "Yeah, God exists, but why should I care?". Or trolling/misunderstanding the question.

0Val11y

Wouldn't that be the very definition of a deist or an agnostic, instead of an atheist?

0Alsadius11y

I didn't say that they were good at defining terms.

[-]whateverfor11y80

Do you have some links to calibration training? I'm curious how they handle model error (the error when your model is totally wrong).

For question 10 for example, I'm guessing that many more people would have gotten the correct answer if the question was something like "Name the best selling PC game, where best selling solely counts units not gross, number of box purchases and not subscriptions, and also does not count games packaged with other software?" instead of "What is the best-selling computer game of all time?". I'm guessing mos... (read more)

[-]Vaniver11y110

I'm curious how they handle model error (the error when your model is totally wrong).

They punish it. That is, your stated credence should include both your 'inside view' error of "How confident is my mythology module in this answer?" and your 'outside view' error of "How confident am I in my mythology module?"

One of the primary benefits of playing a Credence Game like this one is it gives you a sense of those outside view confidences. I am, for example, able to tell which of two American postmasters general came first at the 60% level, simply by using the heuristic of "which of these names sounds more old-timey?", but am at the 50% level (i.e. pure chance) in determining which sports team won a game by comparing their names.

But it seems hard to guess beforehand that the question you thought you were answering wasn't the question that you were being asked!

This is the sort of thing you learn by answering a bunch of questions from the same person, or by having a lawyer-sense of "how many qualifications would I need to add or remove to this sentence to be sure?".

0whateverfor11y

OK, so all that makes sense and seems basically correct, but I don't see how you get from there to being able to map confidence for persons across a question the same way you can for questions across a person. Adopting that terminology, I'm saying for a typical Less Wrong user, they likely have a similar understanding-the-question module. This module will be right most of the time and wrong some of the time, so they correctly apply the outside view error afterwards on each of their estimates. Since the understanding-the-question module is similar for each person, though, the actual errors aren't evenly distributed across questions, so they will underestimate on "easy" questions and overestimate on "hard" ones, if easy and hard are determined afterwards by percentage that get the answer correct.

0Vaniver11y

That seems reasonable to me, yes, as an easy way for a question to be 'hard' is if most answerers interpret it differently from the questioner.

[-]Vulture11y80

Yayy! I was having a shitty day, and seeing these results posted lifted my spirits. Thank you for that! Below are my assorted thoughts:

I'm a little disappointed that the correlation between height and P(supernatural)-and-similar didn't hold up this year, because it was really fun trying to come up with explanations for that that weren't prima facie moronic. Maybe that should have been a sign it wasn't a real thing.

The digit ratio thing is indeed delicious. I love that stuff. I'm surprised there wasn't a correlation to sexual orientation, though, since I se... (read more)

[-]Richard Korzekwa11y140

I remember answering the computer games question and at first feeling like I knew the answer. Then I realized the feeling I was having was that I had a better shot at the question than the average person that I knew, not that I knew the answer with high confidence. Once I mentally counted up all the games that I thought might be it, then considered all the games I probably hadn't even thought of (of which Minecraft was one), I realized I had no idea what the right answer was and put something like 5% confidence in The Sims 3 (which at least is a top ten game). But the point is that I think I almost didn't catch my mistake before it was too late, and this kind of error may be common.

[-]James_Miller11y100

I was confident in my incorrect computer game answer because I had recently read this Wikipedia page List of best-selling video games remembered the answer and unthinkingly assumed that "video games" was the same as "computer games".

8[anonymous]11y

The correct answer is Tetris. The question should have been what is the best selling personal computer game of all time? Mobile phones are technically computers too. I'm not sure how much difference that would have made.

0habeuscuppus11y

I interpreted the question to include mobile devices and answered Tetris with high confidence. It would be interesting to see the results of the question if we accepted either Tetris or Minecraft as the correct answer, since both are correct depending on whether or not "computer" was meant to mean "IBM PC Compatible" or "video game playing platform"

1emr11y

On the computer game question: Isn't there an implicit "X is true and X will be marked correct by the rater"? You'd hope these two are clearly aligned, but if you've taken many real-world quizzes, you'll recognize the difference.

0devas11y

I think the computer games question has to do with tribal identity-people who love a particularly well known game might be more inclined to list it as being the best seller ever and put down higher confidence because they love it so much. Kind of like owners of Playstations and Xboxs will debate the superiority of their technical specs regardless of whether they're superior or not.

9Jiro11y

I think the computer games result has to do with it being a bad question. There are many legitimate answers depending on how you interpret the question, including my answer that Minesweeper sells as a bundle with Windows and thus has probably sold more copies than anything else.

5Vulture11y

Is it really a "bad question"? Shouldn't a good calibrator be able to account for model error?

1devas11y

Depends on whether you consider "being able to comprehensively understand questions that may be misleading" to be a subset of calibration skills.

0devas11y

Good point, I hadn't thought of that.

[-]orthonormal11y70

It looks to me like everyone was horrendously underconfident on all the easy questions, and horrendously overconfident on all the hard questions.

I think that this is what correct calibration overall looks like, since you don't know in advance which questions are easy and which ones are tricky. I would be quite impressed if a group of super-calibrators had correct calibration curves on every question, rather than on average over a set of questions.

0[anonymous]10y

Can you explain your point more fully? I don't think I understand or can identify the antecedents. For instance, what you mean by calibration..

0Douglas_Knight10y

Right, Dunning-Kruger is just regression to the mean.

0orthonormal10y

No, that's false. It's possible (and common) for a person to be wildly overconfident on a pretty broad domain of questions.

[-]tog11y70

It's interesting to compare these results to those of the 2014 Survey of Effective Altruists. These will be released soon, but here are some initial ways in which effective altruists who took this survey compare to LessWrong census takers:

Somewhat less male (75% vs 87%)
More in the UK
Equally atheist/agnostic
More consequentialist (74% vs 60%)
Much more vegan/vegetarian (34% vs 10%)
Witty, attractive, and great company at parties

0Username11y

For someone who isn't an EA, and therefor shouldn't take this survey, is there a place where we can see the results?

0RyanCarey11y

AFAIK, the results aren't out yet, but they'll go on effective-altruism.com when they are.

0tog11y

Correct, Peter Hurford is working on them and will I believe finish them soon.

[-]Unnamed11y70

I think that there are better analyses of calibration which could be done than the ones that are posted here.

For example, I think it's better to combine all 10 questions into a single graph rather than looking at each one separately.

The pattern of overconfidence on hard questions and underconfidence on easy questions is actually what you'd expect to find, even if people are well-calibrated. One thing that makes a question easy is if the obvious guess is the correct answer (like a question about Confederate Civil War generals where the correct answer is R... (read more)

0Rob Bensinger11y

I would count minor typos (like 'mitochondira') and spelling errors (like 'mitocondria'), and trivial variants (like 'mitochondrium') as correct. I'd count major typos, where the pronunciation would be majorly different from the correct name, as neither correct nor incorrect -- ditch the data, since it may require too many judgment calls about whether someone's zeroing in on the correct name, and since the respondents themselves weren't told how much to make their calibration correct-spelling-sensitive. So my system says: Heisenberg Hawai'i Odin Jacob Indonesia

2Sabiola11y

Hawaii is both in the 'Also correct' and the 'Neither' list.

0Rob Bensinger11y

One uses a typewriter apostrophe ('), the other doesn't (`).

1arundelo11y

No, bbleeker is saying that "Hawaii" (no apostrophe) is in both lists.

5Rob Bensinger11y

Ah. "Hawaii or Arkansas" is its own entry, and was typed in as an answer.

[-]John_Maxwell11y70

This is not a human universal - people who put even a small amount of training into calibration can become very well calibrated very quickly. This is a sign that most Less Wrongers continue to neglect the very basics of rationality and are incapable of judging how much evidence they have on a given issue. Veterans of the site do no better than newbies on this measure.

Can someone who's done calibration training comment on whether it really seems to represent the ability to "judge how much evidence you have on a given issue", as opposed to accurately translate brain-based probability estimates in to numerical probability estimates?

8Vaniver11y

As I interpret it, the two are distinct but calibration training does both. That is, there's both a "subjective feeling of certainty"->"probability number" model that's being trained, and that model probably ought to be trained for every field independently (that is, determining how much subjective feeling of certainty you should have in different cases). There appears to be some transfer but I don't think it's as much as Yvain seems to be postulating.

0John_Maxwell11y

Have you done calibration training? Do you recommend it? I think I remember someone from CFAR saying that it was kind of a niche skill, and to my knowledge it hasn't been incorporated in to their curriculum (although Andrew Critch created a calibration android app?)

0Vaniver11y

I've done a moderate amount of training. I think that the credence game is fun enough to put an hour or two into, but I think the claim that it's worth putting serious effort into rests on the claim that it transfers (or that probabilistic judgments are common enough in your professional field that it makes sense to train those sorts of judgments).

2John_Maxwell11y

CFAR calibration games I tried the game for a while... many of the questions are pretty hard IMO (especially the "which of these top-10 ranked things was ranked higher" ones), which makes it a bit difficult to learn to differentiate easy & hard questions. Other calibration quizzes

3Vaniver11y

I think that's necessary? You want to have questions you can answer at every credence band between 99% and 50%, and that implies questions you're only 60% sure of the answer of.

2Unknowns11y

Are these two things significantly different?

4lmm11y

Imagine someone who acted appropriately towards particular risks (maybe not very artificial ones like betting, but someone who did things like saving an appropriate proportion of their income, spent an appropriate amount of their free time doing fun-but-dangerous things like skydiving), but couldn't translate their risk attitudes into numbers.

0rule_and_line11y

I'm having difficulty replacing your quotation with its referent. Could you describe an activity I could do that would demonstrate that I was judging how much evidence I have on a given issue?

0Username11y

Make money playing poker, maybe?

0rule_and_line11y

Ah! That sounds like a great one! So, folks like Chris Ferguson are presumably doing both activities (judging how much evidence as well as accurately translating brain estimates to numerical estimates). But if I go find a consistently successful poker player who does not translate brain estimates to numerical estimates, then I could see how that person does on calibration exercises. That sounds like a fun experiment. Now I just need to get the grant money ... Sidenote, but how would I narrow down to the successful poker players who don't translate brain estimates to numerical estimates? I mean, I could always ask them up front, but how would I interpret an answer like "I don't really use numbers all that much. I just go by feel." Is that a brain that's translating brain-based estimates to numerical estimates, then throwing away the numbers because of childhood mathematical scarring? Or is that a brain that's doing something totally outside translating brain-based estimates to numerical estimates?

[-]lirene11y70

I only identify with my birth gender by default: 681, 45.3%

I'm surprised at this. Is there a special term for "only identifying with one's gender by default" or keywords I can use to look for statistics for among the general population? (a brief googling didn't uncover anything). I would've guessed this number to be much lower, and now I'm wondering whether this is signaling or whether my model of other people in this particular instance is completely wrong.

7Kaj_Sotala11y

It's not an established term in the sense that people would have conducted research on it, as far as I know: unless I'm mistaken, it comes from here.

[-]arundelo11y100

And here's the Internet Archive's copy of Ozy's original "Cis By Default" post.

Edit: And now they've reposted it to their current blog.

1lirene11y

Thank you both for providing the links. I will wait and see whether the percentage stays the same in the 2015 survey...

4FiftyTwo11y

Personally I was surprised so amny cis people strongly identified with their gender [tpical mind etc...]

1[anonymous]11y

I think I actually put that down. Reasoning: gender is mostly, IMHO, social performance. Since I was born male, and was mostly trained to perform maleness, that's mostly what I perform. But it's not a conscious, reasoned-out choice made by weighing reasons for and against. It's just the coincidence of what happened. A counterfactual me who had been born female and raised to perform femaleness, but was otherwise the same, would not be so masculine-leaning as to transition. Weirdly enough, my normal behavior tends to code as more masculine than average. This is, as people would obviously reason if they could think at all clearly ;-), overcompensation for spending half my youth being told to stop being such an oversensitive pussy and man-up already (sorry for the language, but the experiences were rather severe, actually).

0alienist11y

I don't understand how this counterfactual is supposed to work. Does being born female include not having a Y chromosome and thus no SRY gene, thus no testosterone bath in the womb and the resulting cascade? If so it's unclear what you mean by "otherwise the same".

2[anonymous]11y

Well yes, of course.

[-]Ixiel11y60

Thanks for all the hard work!

[-]Kaj_Sotala11y60

This was the first time that I failed to take the survey, because I kept going "I don't have a scanner at home for the finger ratio thing, so I'll do it when I'm at the university" and then each time when I was at the university, I forgot. (I wasn't there often.)

[-]simon11y60

Any correlation between digit ratios and gender default?

For the warming question, much of the difference between responses is likely due to differing judgments on what is "significant" as opposed to on what will/has been actually happening. I think it would be better to try to split these components. A first question could ask for example, how much do you think human influence (e.g. CO2 emitted due to human activity) has raised the temperature at the current moment (give a value that you think has a 50% chance of being above or below the correct ... (read more)

0buybuydandavis10y

I remember being annoyed with this question, as I am annoyed with most "news" articles on "the debate", with gibberish such as "Climate denier". That's not even English. "Significant" and "soon" are both a problem here. I think your question overly complicates, as it's only one component of warming, "due to human activity". Just CO2? Greenhouse emissions? All the forests and herd animals ever killed? Transformation of species populations? Pollution? How about what the average global temp will be in the 2050s, according to some existing instrumental measurement regime?

[-]Gram_Stone11y50

The questions about sexual orientation and gender identity seem to be quite well thought-out, as I would expect, but I have some suggestions in that regard for next year's survey:

1) Include the option to report as a man who has sex with men or a woman who has sex with women.

This could either be incorporated into the question of sexual orientation, or, perhaps more appropriately, because it is not technically a sexual orientation, it could be included as a standalone question.

2) Include the option to report one's sexual orientation using the Klein Sexual Or... (read more)

4[anonymous]11y

In my opinion, sexual orientation should always be relative to one's gender, not one's biological sex. The MSM or WSW identifier violates this to some extent (particularly as it is used by, e.g., the CDC) and causes, again imo, more problems than it's worth. The KSOG also seems plagued by this conflation.

5Gram_Stone11y

I can see how it would be wrong to claim that someone else is a MSM or a WSW, but I don't see how it could be wrong to give someone the option to self-report that way since they won't use the label if they disagree with it, and especially since this data won't be used for anything. Can you elaborate on what sorts of problems you think it would cause? At any rate, I agree that sexual identity should be self-determined, and therefore relative to one's gender identity rather than one's biological sex. To that end, perhaps the question on sexual orientation should include the option to identify as androphilic, gynephilic, or ambiphilic, or perhaps there should be separate questions for identifications of orientation that incorporate gender identity and those that do not.

4[anonymous]11y

MSM is a confused category with multiple edge cases. Does "male" refer to sex or gender? It has somewhat troubling connotations with promiscuity that don't always make sense, as it is sometimes applied to otherwise asexual males in homosexual relationships. It's not at all clear that the demographic it was designed to apply to (otherwise straight men who have specifically anal sex with other men) feel more comfortable identifying as MSM than gay or bisexual. I can understand that you feel this is the correct label for your situation, but it is somewhat fraught with historical baggage. I don't think we actually know this.

4Gram_Stone11y

I wasn't aware of these controversies, but I don't see how the historical baggage is relevant at LW. I think that most everyone here is going to realize that a 'man who has sex with men' is a person who self-identifies as male and has sex with others whom they identify as male. Furthermore, people will only be applying the label to themselves. You just implicitly gave me the option to use it, so I guess I should ask: Is there anything that makes me different from anyone else, or any reason that you wouldn't give everyone the same option simultaneously? I concede that point. It's also true that those who don't want their data to be used have the option not to take the survey or to have their data removed from the public results. On the other hand, I'm not sure how many people that would discourage or how much it would skew the results. Also, I missed this: I agree. I see that this is the case in its use of the terms heterosexual, homosexual, etc. We could fix this simply by modifying it to use the the terms androphilic, gynephilic, etc., and including a term each for gender identity and biological sex. It would make no such conflation while preserving the information provided by the existing labels, and thereafter provide even more information. Furthermore, if people want to interpret the data using the previous labels of heterosexual, etc., one could derive that from the data produced from the scheme that I just proposed, or there could be a separate question with the original labels. (Separate questions or the use of both gender-loaded and gender-neutral terms in the same question would also implicitly provide data on the respective popularities of gender-loaded and gender-neutral terms for sexual orientation, though it would be skewed by people whose usage patterns are changed by the question itself. That could be rectified by a follow-up question asking what term one used to self-identify before reading the previous question, if you want to go nuts with it. LW

2[anonymous]11y

I don't normally bother to vote in threads that I participate in, but I've decided to cooperate and upvote your comments as well. It's not really surprising that it was downvoted. The minutiae of queerness is not exactly interesting to most people.

2Gram_Stone11y

As long as you're cooperating with my debiasing efforts, the karmabomber downvoted my original comment as well. Even if you disagree with it, upvoting it is the impartial way to go, as I have done. On the topic, do we still disagree, or have we reached a common conclusion? Like many aspiring rationalists, I like to actually resolve things. To be clear, my current conclusions are: 1) people should be allowed to self-report as MSM or WSW, and; 2) the KSOG should be added but slightly modified to include the '-philic terms' instead of the '-sexual terms,' and to include gender identity and biological sex. Using this modified KSOG would also consolidate some existing questions into the KSOG. (Not necessarily mutually exclusive) alternatives to #2 include: adding the -philic terms to the existing orientation question (possibly along with another question asking what one's self-identification was before the previous question to account for the influence of the question itself); adding a new separate question using the -philic terms, and; including a modified or unmodified version of the Kinsey scale. I can dig that, I don't like to have my time wasted either, but it was necessary to understand my suggestion (like I said, it comes with the MSM territory, everyone has a different reason for using it, and those unique reasons are the whole reason that including the term could potentially yield more detailed and interesting data; preferences about sexual preferences? That's one example that sounds to me like something LWers would be very interested in. Maybe you personally are just not interested in sexology. I'm not talking about sharing anecdotes or something like that). I put a 'long anecdote warning' so that people can skip it if they want. (Now that I've thought of it, next time I'll put a TL;DR at the end as well, and I've added one.) Whoever did it obviously didn't do it because they really care about curating the site, because they downvoted a bunch of other unrelat

[-]Petter11y50

This post had more statements of the type “p < 0.01” than I would expect at LW. I recently read “Frequentist Statistics are Frequently Subjective” here.

3Luke_A_Somers11y

It's easy to calculate, and as long as you keep in mind what it means, it's not bad to include.

[-]shullak711y50

It looks like the median age is 27.67, but I'm curious to see the age-range breakdown as I've frequently assumed I'm "old" for the group (over 40). Oh....never mind.....just saw the link to the Excel spreadsheet and will sort it myself.

[-]Vaniver11y160

1421 respondents supplied their age: 1292 (90.9%) of them were less than 40. The modal age is 25.

5shullak711y

Thanks Vaniver. I am going to take this to mean that I'm young at heart.

[-]lmm11y50

As always, thanks for doing this.

people can both be asexual and have a specific orientation

Huh? This is worded as a question about orientation rather than practice, so people who have an orientation have an orientation, no? Or is the issue something else?

Why are some results ordered and some listed alphabetically? That's a bit confusing.

What's this easy becoming calibrated through training? Bonus for a mainstream-ish source rather than LWsphere.

4falenas10811y

People can be asexual but, say, homoromantic.

4lmm11y

Sure. And such people would be asexual and answer the survey as asexual. What's the issue?

6philh11y

Romantic preferences were implicitly tied into that question, because there wasn't a separate question for them, so such a person could reasonably answer homosexual. Alternatively, one could have weak sexual preferences in a particular orientation, like "having sex with guys is kind of fun if there's nothing better to do, but I wouldn't have sex with girls".

5lmm11y

Did anyone want to know about romantic orientation? I mean, if that's something people are interested in then sure, let's add a question for it, but I don't think that's a problem with a question about sexual orientation. That's a problem with any kind of discrete categorization of sexual orientation, nothing to do with confusing it with romantic orientation. You could argue for e.g. Kinsey scale (adjusted to include asexuality at similar precision) rather than the four choices we have, but there will always be edge cases.

4Kaj_Sotala11y

I think most people don't realize that sexual and romantic orientation even can be distinct, so if there's a question about "sexual orientation", one never knows what's actually meant - sexual orientation, romantic orientation, or both - unless it's explicitly specified.

3therufs11y

Wits and Wagers?

[-]RicardoFonseca11y40

So I noticed a few people from Portugal in the public raw data. I am one of those people and from what I saw in the data, the remaining are all lurkers.

Viva pessoal! Want to join the community, even if just to say hi? =)

[-]Gondolinian11y40

SAT Scores out of 1600/SAT Scores out of 2400 .844 (59)

I'm surprised that this correlation wasn't higher. They're both pretty much the same test, right?

Three explanations I thought of:

I'm missing something/I have an inaccurate model of the difference between the two tests.
There's a lot of random difference between SAT scores from different testings. If this is true, I would expect there to be a correlation of around .844 between one test score and a later test score under the same grading system.
SAT scores are correlated with age (no idea whethe

... (read more)

[-]gwern11y140

They're both pretty much the same test, right?

I thought they were partially not the same because they added the writing subtest.

If this is true, I would expect there to be a correlation of around .844 between one test score and a later test score under the same grading system.

The reliability of recent SAT tests seems to generally be ~0.9 according to one random PDF I found (and has long been high). If I am understanding the formulas in this page correctly, then in this application, reliability simplifies to the Pearson's r of the 2 scores*, and that reliability of 0.9 is pretty similar to the LW old/new correlation r of 0.84.

So this may be simply what one would expect from people taking the SAT twice, without having to invoke the lowered correlation caused by the additional sections and any other tweaks they've made.

* Specifically, I'm looking at Artifactual Influences, #3: reliability, where I think we can reuse the example: for test-retest, assume the LWer doesn't get dumber or smarter and the true correlation would be 1; the reliability of the old SAT should be 0.9, the reliability of the new one should be 0.9 too, so you get '1 sqrt(0.9 0.9)' or 'sqrt(0.9 * 0.9)' or 's... (read more)

[-]Vaniver11y110

Any ideas?

The psychometric term for #2 is test-retest reliability, and the numbers I've seen for the SAT range between .8 and .95, so I think that is a complete explanation for this phenomenon.

If the 2400 scores (which came later) are higher than the 1600 scores, that's evidence for #3, but comparing them is difficult because they do test different things and are normed differently.

[-]Tenoke11y40

What's up with a whole 10% being 'Atheist and spiritual'? It doesn't seem to be a family thing, as you get only 4.9% with that belief in the family section, and the numbers don't match up with the P(Supernatural) question.

I was worried about this last year when it was 8.1%, and the number seems to be increasing. Is this Will Newsome's post-rationality faction or what?

[-]Kaj_Sotala11y130

"Spiritual" doesn't necessarily imply a belief in anything supernatural: as Wikipedia puts it,

Since the 19th century spirituality is often separated from religion, and has become more oriented on subjective experience and psychological growth. It may refer to almost any kind of meaningful activity or blissful experience, but without a single, widely-agreed definition.

I've sometimes marked myself in the "atheist and spiritual" category in the surveys, though not always since I've been a little unsure of what exactly is meant by it. When I have, I've taken "spiritual" to refer to practices like engaging in meditation (possibly with the intention of seeing through the illusion of the self), seeking to perceive a higher meaning in everything that one does, enjoyment of ritual, cultivation of empathy towards other people, looking to connect with nature, etc.

[-]fubarobfusco11y110

It seems to me that religious belief and religious practice should be distinguished. The current questions are about religious belief and family background. Perhaps a question like this:

How many times in the past year have you done the following practices? Estimates are OK. The question here is whether you did the thing, not whether you believed in it, made it a habit, or did it voluntarily.

Attended a religious service?
Attended a "regular" religious service (e.g. a weekly church or synagogue service, Muslim daily prayers, etc.; not a wedding, funeral, or holiday service)?
Prayed to God, gods, saints, or other religious figures outside of a religious service?
Done yoga, qigong, or another movement practice derived from religious or spiritual beliefs?
Used or participated in psychic or fortune-telling practices (palmistry, ouija, I Ching, tarot)?
Meditated?
Meditated for thirty minutes or more in a single session?

5Tenoke11y

I'm sceptical that this interpretation makes sense in a question about religious views, but I guess it may explain it.

[-]Leonhart11y190

Data point: I picked this option, because of a grab-bag of vaguely related positions in my head that make me feel dissatisfied with the flat "atheist" option, including:

I enjoy and endorse rituals such as the Solstice celebration, as opposed to the set here who are triggered by them (ETA: not in any way claiming they are wrong to be so triggered, or don't have reasons)
I find the Virtues, and other parts of the Sequences with similar styling, to be deeply moving and uplifting, and consider this element of our house style to be a strength rather than a liability
We worry too damn much about the c-word, in a pointless attempt to appease the humourless, and we've compromised too much of our aesthetic identity doing it
Scott's Moloch isn't actually the Devil, but maybe acting as if is a good strategy for recruiting all parts of our minds to the fight. Ditto for Elua
After some experimentation, I think I understand better what the mindstate associated with "worshiping" actually feels like (really damn good) and suspect that the emotional benefits are totally available even if you know the targeted god doesn't exist

(I actually wish it was reversed to "religio... (read more)

4covaithe11y

I quite like this formulation, and if I had thought of it at survey time I might well have answered 'atheist(spiritual)' instead of 'atheist(nonspiritual)'. Regarding emotional benefits: I sing in moderately serious classical choirs, where inevitably much of the music is set to religious texts. I get some but not all of the emotional benefits from this that I used to get from religious worship, back when I was a committed theist. I think I would get more benefits if the texts were not religious, and still more if the texts were humanist / rationalist / expressed beliefs that I actively profess.

2Kaj_Sotala11y

There's Postrationality, Table of Contents, though the author hasn't written any follow-up posts yet.

7Richard_Kennaway11y

Postrationality appears to stand in the same relation to rationality as Romanticism did to the Enlightenment. That is, a falling away from the Way, not a progression past it; the easy, broad path and not the strait and narrow path that must be walked to hit the small target of truth.

6fubarobfusco11y

I can't tell if this is an Ideological Turing Test failure, or just a lie.

1Leonhart11y

Upvoted for informing me that "straight and narrow" was a malformation. Also, yes.

0Leonhart11y

Ah, yes. I read that page and scrunchyfaced, back when Scott posted the map. (Although I seem to remember reading other things on the same blog that were better thought out, so maybe the author was having an off day.) I hope that something more rigorous and interesting comes along. The defensible heart of the position, it seems to me, could be something along the lines of "Yes, we must be ready to relinquish our beliefs with the slightest breath of the winds of evidence. But exactly so long as we do believe A, let's really believe it. Let's not deny ourselves the legitimate Fun that can reside in savouring a belief, including any combination of robes and chanting that seems appropriate."

7Kaj_Sotala11y

A question about religious views seems like the perfect place to signal that you're not religious but still like some of the things that are commonly associated with religion.

[-]Gunnar_Zarncke11y40

I'd like to see the correlations between Bem Sex Roles and actual Gender answers.

[-]Stuart_Armstrong11y30

Thanks Alex and Ozy!

[-]dspeyer11y30

Do I understand correctly that a more masculine finger ratio correlated strongly to support for feminism in both men and women?

I am also amused to note that, despite our extreme sex ratio, our BEM gender masculinity and femininity are almost exactly equal -- way below error.

2simon11y

No, as I interpret what Yvain wrote, a more feminine digit ratio correlated to support for feminism in men and the correlation was not quite statistically significant in women.

[-]Gunnar_Zarncke11y30

Among the correlations the first I found surprising are

Minimum Wage/Feminism .378 (1286)
Immigration/Feminism .365 (1296)#

I wouldn't have guessed these from the corresponding memeplexes and can't see a plausible relationship. Anybody volunteer?

[-]blacktrance11y310

Support for a higher minimum wage, increased immigration, and feminism are all typically left-wing positions, so it's not surprising that they're found together.

5Gunnar_Zarncke11y

Thanks. Maybe it's obvious for you but it does surprise me. Maybe it's more clear over ther in the U.S. I'm nonetheless surprised by the magnitude if they are only connected via such a unspecific bucket as 'left-wing'.

[-]falenas10811y160

I'm pretty sure it's more than just that, a lot of feminist ideas are about helping typically underprivileged communities. I've seen a lot of stuff on feminist areas about helping the poor and the undocumented as an extension of that.

2alienist11y

That's simply the inside view of blacktrance's point.

9gjm11y

How do you know? I mean, we have (at least) two hypotheses. 1: lefties classify people on a scale from "oppressor" to "oppressed" and favour the oppressed. 2: lefties are interested in helping people in bad situations, and there are some categories of people who are systematically much more likely to be in bad situations. It looks as if lefties themselves tend to say #2, whereas various non-lefties say #1. One explanation, indeed, is that #1 is correct and #2 is what the poor self-deluded lefties think they're doing. Another is that #2 is correct and that #1 is what the poor deluded non-lefties mistakenly see it as. Why should we favour the second of these over the first? (I'm mostly a leftie. It looks to me as if there's some truth in both #1 and #2, but it seems to me that #1 is a consequence of #2, a special case, more than it's a fundamental motivation that gets rationalized as #2. I expect there are people for whom #1 is primary but don't see any reason to think that's the usual case. It seems like the obvious way for someone to feel #1 but need to deceive themselves that #2 is their real motivation would be if they are themselves in an "oppressed" group, so that #1 is a matter of self-interest. I think a majority of the lefties I know -- and for that matter of the non-lefties -- are fairly "privileged" and for those the "#1 primary, #2 is rationalization" account seems awfully implausible.) [EDITED to add: 1. I would be interested to hear from either of the people who downvoted me why they thought this comment merited a downvote. It looks OK to me. 2. In the 8 hours since it was posted, I am down 31 karma points. I wonder what account name The Artist Formerly Known As Eugine_Nier is using now.]

3blacktrance11y

I'm not a progressive, but I don't see 1 and 2 as mutually exclusive. 1 is just a different way of stating 2 - leftists classify people on an oppressor-oppressed axis, where the oppressed are people perceived to be in bad situations.

3gjm11y

I think "oppressed" is more specific than "in a bad situation", and "oppressor" is much more specific than "in a comfortable situation". Saying that lefties classify people on an oppressor/oppressed axis suggests that they're addicted to what's sometimes called "politics of envy" -- it's not enough to help the poor, the rich must be made to suffer becaucse they are evil oppressors, etc. I'm sure there are people who think (and feel) that way, but I think it's a straw man if presented as an analysis of how lefties generally see the world. I think most lefties would agree with me that when people are in bad situations it doesn't have to be because anyone's oppressing them. They might just have been unlucky, or they might in some sense have done it to themselves (one place where Left and Right commonly disagree: on the left, this is not usually taken to mean that they shouldn't be helped). And I think most lefties would agree with me that someone very comfortably off is not necessarily oppressing anyone. (There are some who would say that our society systematically favours some groups and screws others over, to the benefit of the former at the expense of the latter, and that that means that being a rich straight white educated able-bodied man does make you in some sense an oppressor. I, and I think many others, largely agree with the first bit of that but think the conclusion that the rich (etc.) are oppressors is misguided: to benefit from an oppressive system is not necessarily to be an oppressor. And I, and I think many others, think "oppressed" is too strong and too specific a word to describe the ways in which things are bad for most statistically-disadvantaged groups.)

0falenas10811y

I agree, but at the time I posted this blacktrance had yet to make their post.

[-]blacktrance11y100

As Arnold Kling suggests, progressives think of issues on an oppressor-oppressed axis. Women, poor people, and immigrants are all seen as oppressed, which is why feminism, raising the minimum wage, and support for more immigration are positions that are often found together.

7Prismattic11y

In my experience, libertarians tend to think highly of Arnold Kling's taxonomy, and liberals and conservatives do not. I regard it as a Turing test fail.

8roystgnr11y

Could you elaborate on your experience? The liberal philosophizing I've seen seems to go even further than Kling does. He suggests a possibly-subconscious implicit common thread, whereas they often talk explicitly about "punching up versus punching down", or redefine various subcategories of prejudice to only mean "prejudice plus power". I can think of cases where there's a clear position among the U.S. left wing but that position isn't unambiguously objectively described as "support the oppressed against the oppressor", but even in those cases the activism for that position is usually given that framing.

8Prismattic11y

(Prefacing this by noting that I am not going to get into a normative discussion here of whether liberal values are better or worse than libertarian values. I'm only addressing the question of whether Arnold Kling is accurately framing liberal values.) I'll leave speaking about what's wrong with the conservative frame for an actual conservative (from my also-outside perspective, it doesn't sound particularly accurate). But as far as liberalism goes, I think what Kling describes might be an accurate depiction of, say, "social justice" blogs, but those are a subset of liberalism, not the essence of it, and it doesn't describe the way the blue tribe people I grew up around (New England, middle class, disproportionately Jewish) reasoned, nor do I think it captures the way the more wonkish liberal bloggers reason. More specifically, libertarians think that only libertarians care about freedom, while liberals think that libertarians are privileging one particular, controversial, definition of freedom -- the negative liberty of freedom from government (and only government, and in some but not all cases, specifically Federal but not local government) coercion. The liberals I have always known also think that maximizing freedom is the goal, but we define it as something like the autonomy in practice to flourish. So for example, some (not all) libertarians think the Civil Rights Act reduced aggregate freedom, but pretty much all liberals think it increased it. There is a similar divergence in attitudes about net effect on freedom with regards to regulatory interference in freedom of contract between parties with unequal bargaining power. Etc.

-11alienist11y

3Luke_A_Somers11y

Turing test fail? Where was blacktrance trying to pass as having a particular political position?

4blacktrance11y

I think he meant that Kling, being a libertarian, failed the Turing Test when describing the framework behind the progressive and conservative viewpoints.

2Luke_A_Somers11y

I get that. But he wasn't even TAKING the Turing test. He described it fairly accurately, if in terms that people on the inside wouldn't have used. So?

1William_Quixote11y

The problem is that he didn't describe it accurately. He described it with a very simple bias that is different from the actual view and makes bad predictions oi you try to use it to think instead of using it to caricature political opponents.

1Luke_A_Somers11y

I'd suggest that the description was pretty good for being so short. In particular, it was adequate to explain this one feature that someone was not able to anticipate, and a lot of other major features, and it is in fact true of many - not all - components of the left.

2Prismattic11y

Correct.

2JoshuaZ11y

Can you expand on the last sentence?

5fubarobfusco11y

— Wikipedia

0[anonymous]11y

Prismattic is referring to an ideological Turing test. This is a way of suggesting that Arnold Kling is failing to understand liberals and conservatives in a similar way to how they understand themselves.

[-]gwern10y20

Looking at 2013/2014 effective altruist donation amounts: http://lesswrong.com/r/discussion/lw/m6o/lw_survey_effective_altruists_and_donations/

[-]BoilingLeadBath11y20

If anyone else is interested in them, I'm willing to score, count, and/or categorize the responses to the "Changes in Routine" and "What Different" questions.

However, I've started to try and develop a scheme for the former... and I've hit twenty different categories (counting subcategories) and will probably end up with 5-10 more if I don't prune them down.

What sort of things do you think might be interesting to look for?

(Though I haven't started to do work on paper, the latter seems like a much simpler problem. However, if you have tho... (read more)

[-]tog11y20

To compare, are there any public stats on LessWrong readership, such as how many new and returning visitors the site gets?

[-]Capla11y20

Hmmm...How strongly does this agrees with the sequences factor correlate with having read the sequences?

[-]goocy11y20

Charity: 1996.76 + 9492.71

For a statistician, this is insane. In this case, this would mean that a sizable chunk of responders actually receives money from charity.

You seem to assume that every dataset has an inherent mean and standard deviation. But means and standard deviations are the results of modeling a gaussian distribution, and if the model fit is too bad, these metrics simply don't apply for this dataset.

The Lilliefors test was created for exactly this purpose: it gives you the probability that a dataset is not normal distributed. Please use it... (read more)

[-]satt11y130

But means and standard deviations are the results of modeling a gaussian distribution, and if the model fit is too bad, these metrics simply don't apply for this dataset.

Means and standard deviations are general properties one can compute for any statistical distribution which doesn't have pathologically fat tails. (Granted, it would've been conceptually cleaner for Yvain to present the mean & SD of log donations, but there's nothing stopping us from using his mean & SD to estimate the parameters of e.g. a log-normal distribution instead of a normal distribution.)

3Meni_Rosenfeld10y

Is the link to "Logical disjunction" intentional?

0satt10y

It isn't! Thanks for catching that, I've fixed the link.

2gjm10y

You can indeed compute means and standard distributions for any distribution with small enough tails, but if the distribution is far from normal then they may not be very useful statistics. E.g., an important reason why the mean of a bunch of samples is an interesting statistic is that if the underlying distribution is normal then the sample mean is the maximum-likelihood estimator of the distribution's mean. But, e.g., if the underlying distribution is a double exponential then the max-likelihood estimator for its position is the median rather than the mean. Or if the distribution is Cauchy then the sample mean is just as noisy as a single sample.

0DTX10y

I'd expect a Pareto distribution for charitable donations, not log-normal, and that's exactly what the histogram looks like: Looks like alpha >> 2, so the variance is infinite.

5satt10y

Thanks for prompting me to take a closer look at this. The distribution is certainly very positively skewed, but for that reason that histogram is a blunt diagnostic. Almost all of the probability mass is lumped into the first bar, so it's impossible to see how the probability distribution looks for small donations. There could be a power law there, but it's not obvious that the distribution isn't just log-normal with enough dispersion to produce lots of small values. Looking at the actual numbers from the survey data file, I see it's impossible for the distribution to be strictly log-normal or a power law, because neither distribution includes zero in its support, while zero is actually the most common donation reported. I can of course still ask which distribution best fits the rest of the donation data. A quick & dirty way to eyeball this is to take logs of the non-zero donations and plot their distribution. If the non-zero donations are log-normal, I'll see a bell curve; if the non-zero donations are Pareto, I'll see a monotonically downward curve. I plot the kernel density estimate (instead of a histogram 'cause binning throws away information) and I see which is definitely closer to a bell curve. So the donations seem closer to a lognormal distribution than a Pareto distribution. Still, the log-donation distribution probably isn't exactly normal (looks a bit too much like a cone to me). Let's slap a normal distribution on top and see how that looks. Looks like the mean is about 6 and the standard deviation about 2? Wow, that's a far closer match than it has any right to be! Admittedly, if I ask R to run a Lilliefors test, the test rejects the hypothesis of normality (p = 0.0007), and it remains the case that the donations are neither log-normal nor power-law distributed because some of the values are zero. But the non-zero donations look impressively close to a log-normal distribution, and I really doubt a Pareto distribution would fit them better. (A

0gwern10y

As I understand it, tests of normality are not all that useful because: they are underpowered & won't reject normality at the small samples where you need to know about non-normality because it'll badly affect your conclusions; and at larger samples like the LW survey, because real-world data is rarely exactly normal, they will always reject normality even when it makes not the slightest difference to your results (because the sample is now large enough to benefit from the asymptotics and various robustnesses). When I was looking at donations vs EA status earlier this year, I just added +1 to remove the zero-inflation, and then logged donation amount. Seemed to work well. A zero-inflated log-normal might have worked even better. Also, you don't have to look at only one year's data; you can look at 3 or 4 by making sure to filter out responses based whether they report answering a previous survey.

0satt10y

I agree that normality tests are too insensitive for most small samples, and too sensitive for pretty much any big sample, but I'd presumed there was a sweet spot (when the sample size is a few hundred) where normality tests have decent sensitivity without giving everything a negligible p-value, and that the LW survey is near that sweet spot. If I'd been lazy and used R's out-of-the-box normality test (Shapiro-Wilk) instead of following goocy's recommendation (Lilliefors, which R hides in its nortest library) I'd have got an insignificant p of 0.11, so the sample [edit: of non-zero donations] evidently isn't large enough to guarantee rejection by normality tests in general. Certainly. It might be interesting to investigate whether the log-normal-with-zeroes distribution holds up in earlier years, and if so, whether the distribution's parameters drift over time. Still, goocy's complaint was about 2014's data, so I stuck with that.

[-]taryneast11y20

Is there any correlation between calibration and CFAR-attendance?

(you mention "previous training in calibration", but I thought I'd specifically ask just in case)

[-]RicardoFonseca11y20

Here is a complete list of Country counts (ordered by count) (ordered alphabetically).

Note that only countries from people who allowed their results to be public are listed. It is possible for someone from a non-listed country to have completed the survey but chose to remain private.

Inside square brackets are the full counts provided by Yvain (which include private results).

People who have not chosen a country are not listed. People with multiple countries are counted multiple times.

[-]hxka11y20

It looks to me like everyone was horrendously underconfident on all the easy questions, and horrendously overconfident on all the hard questions.

Have you ranked questions by their easiness before you looked at the results?

[-]Leonhart11y20

I want to be friends with the write-in worshiper of CelestAI mentioned :) PM if you like!

3Luke_A_Somers11y

I'd be on board with Celestia, but CelestAI? No.

0MathiasZaman11y

I think I know who that would be, but would like my suspicions confirmed. I should probably ask them.

[-]Richard Korzekwa11y20

Under the profession listings, it says 35 people and 4% for Business. 35 is 2.7% of 1500.

[-]seez11y20

Could someone explain or link to an explanation of the significance of the feminism-digit ratio connection? Why is it exciting?

2Viliam_Bur11y

I was imagining a ritual where a devoted feminist cuts off an end of their finger, as a symbolic castration of the internalized patriarchy. But the official version is that digit ratios are influenced by sex hormones. So if something correlates with the digit ratio, it suggests that it could correlate with the sex hormones. Essentially, the question seems to be: "Are our political opinions about feminism influenced biologically by our sex hormones?" Of course, even if we see a correlation, it wouldn't immediately prove that feminism is determined biologically... but it is a piece of data that is interesting and easy to collect, so we collect it.

2Baisius11y

http://en.wikipedia.org/wiki/Digit_ratio

[-]blacktrance11y20

Thank you for doing this survey.

I would be interested to see the correlations between political identification and moral views, and between moral views and meta-ethics.

(Also, looking at my responses to the survey, I think I unintentionally marked "Please do not use my data for formal research".)

[-]Teerth Aloke6y10

Who is here from a country that does not hold elections?

[-][anonymous]10y10

[This comment is no longer endorsed by its author]Reply

[-]EternalStargazer11y10

Birth Month Jan: 109, 7.3% Feb: 90, 6.0% Mar: 123, 8.2% Apr: 126, 8.4% Jun: 107, 7.1% .Jul: 109, 7.3% Aug: 120, 8.0% Sep: 94, 6.3% Oct: 111, 7.4% Nov: 102, 6.8% Dec: 106, 7.1%

[Despite my hope of something turning up here, these results don't deviate from chance]

It would appear 0% of lesswrongers were born in May. Which is strange, because I seem to remember being born then, and also taking the survey.

0Meni_Rosenfeld10y

I was born in May, and I approve this message.

[-]cameroncowan11y10

Thank you for this, very informative!

[-]AlexMennen11y10

P Global Catastrophic Risk/MIRI Mission -.174

I am so confused.

Also, I'd like to see an aggregate calibration graph that includes people's answers to all of the calibration questions.

[-]Gunnar_Zarncke11y10

I think one logical correlation following from the Simulation Argument is underappreciated in the correlations.

I spotted this in the uncorrelated data already:

P Supernatural: 6.68 + 20.271 (0, 0, 1) [1386]
P God: 8.26 + 21.088 (0, 0.01, 3) [1376]
P Simulation 24.31 + 28.2 (1, 10, 50) [1320]

Shouldn't evidence for simulations - and apparently the median belief is 10% for simulation - be evidence for Supernatural influences, for which there is 0% median belief (not even 0.01). After all a simulation implies a simulator and thus a more complex 'outer w... (read more)

[-]gwern11y100

Shouldn't evidence for simulations - and apparently the median belief is 10% for simulation - be evidence for Supernatural influences

A simulation is still a naturalistic non-supernatural thing, and it would just mean we see less of the universe than we thought we do. The question was, after all:

What is the probability that there is a god, defined as a supernatural intelligent entity who created the universe?

0Gunnar_Zarncke11y

See my answer to hairyfigment. Does that help?

3Luke_A_Somers11y

I think you're looking at it backward. You are trying to understand what the implications of a survey response are. This is the explanation. Your philosophical objection to the logic behind the explanation doesn't make it not the explanation.

2[anonymous]11y

It depends on your definition of supernatural, and most people on LessWrong seem to have a very narrow definition of supernatural. I think Eliezer once wrote a post about it, but I don't believe he cited any references. Some definitions of supernatural would require many people on here to revise their estimate significantly upward. I took the lack of a definition to mean we should use any and all possible definitions of supernatural when considering the question, which is why I picked 100 percent. There's actually been a discussion on whether simulations imply God, and most answered no. I thought the reasoning some used for that was rather peculiar. That discussion of course didn't include any citations either.

5Leonhart11y

You're thinking of this one, and he cited Carrier, and we have this argument after every survey. At this point it's a Tradition, and putting "ARGH LOOK JUST USE CARRIER'S DEFINITION" on the survey itself would just spoil it :)

0[anonymous]11y

Oh yeah, that one. I'd probably just get annoyed if they said to use Carrier since I hate that definition, so I guess the status quo works for me.

2Dallas11y

I actually calibrated my P(God) and P(Supernatural) based on P(Simulation), figuring that getting an exact figure for cases where (~Simulation & Supernatural) are basically noise. I forgot what I actually defined "God" as for my probability estimation, as well as the actual estimation.

0hairyfigment11y

I don't know what you mean by "complexity" here, but it doesn't seem based on MML like the arguments you refer to.

0Gunnar_Zarncke11y

Think of a simulation as an implementation of the rules of physics. The rules of physics have a certain complexity in the sense of necessary bits to encode them (basically whatever complexity measure you prefer, just may not assume prior knowledge of course). If the implementation is faithful, then an observer in the simulation can at best determine this complexity. By occams razor he would take that as the complexity of his universe. If the implementation has bugs (e.g. rounding imprecision or range limits or erroneous special cases) these create a kind of different physics which to the inhabitants could in pricinciple be discernible and and increase the complexity measure - possibly quite a lot because it exposes parts of the implementation logic with all its details. I seem to remember that physicists considered this and actually looked for these kind of rounding errors. Other kinds of complexity could hide in e.g. the random number generator used to make probabilistic choices. Now I wouldn't call the above kind of 'bug' supernatural nor the implementators gods. But another kind of the simulations divergence from the normal rules could count: If the implementor chose to a) select specific instances of probabilisitc choices or b) locally diverge from the 'physics' and establish e.g. an information channel between the simulation and the simulator. These could create effects compatible with theology. Strictly a physicist wouldn't call this supernatural as it has a perfectly satisfactory constructivistic mechanistic explanation (it just assumes one more level of indirection). But if we take 'our universe' to mean that what the implementor called the simulation, then it seems to fit.

1hairyfigment11y

And I don't see why you would choose to ignore the outer world like that. Modern physics includes all sorts of (potential) alternate realities! ETA: Also, the parent seems to me like the opposite of what you said before, namely, "a simulation implies a simulator and thus a more complex 'outer world' doing the simulation".

0Gunnar_Zarncke11y

I don't 'choose' to ignore the outer world. To make 'supernatural' or 'god' not trivially empty sets I have to come up with a sensible concept (in the sense of either appealing to intuition or theology or both) of these. One is to take 'universe' to mean some entity differentiatable (at least conceptually) from an outer/enclosing universe (which includes its 'creator' or creaing process). The concept of a creator god is of course quite common in theology and even if this leaves open the source of outer universe came from that doesn't necessarily preclude the nesting to begin with. Note: I didn't down-vote you. I think you may validly ask why I chose this distinction.

2hairyfigment11y

I know you didn't down-vote me, that was Eugene Nier. Why you didn't up-vote me is another question. And others have mentioned the "fundamentally mental" definition which seems at work here. The simulation hypothesis would technically give us reason to ask whether it really makes sense to assume the programmer(s) run on their own set of rules, which don't mention minds. And yet it still feels like we shouldn't believe the outer world will change its laws in the sense of all air turning to butterflies (or the outer-world equivalent) in the next second (our time). This feeling even seems somewhat reasonable, since a change to the outer world could easily affect ours. The grue-bleen problem means we would still need some more technical form of Occam's Razor to even explain what we/I believe there.

0Gunnar_Zarncke11y

Huh? How do you know? I know how to find out about whole posts. But comments? I tend to not up-vote questions which just ask e.g. for clarification except where this has clearly a general or novel aspect.

0FeepingCreature11y

If there's one simulation, there are many simulations. Any given "simulation God" can only interfere with their own simulation. Interfered-with simulations diverge, not-interfered-with simulations converge. Thus, at any given point, I should expect to be in the not-interfered-with simulation. "God", if you can call it that, but not "Supernatural" because this prime mover cannot affect the world.

2NancyLebovitz11y

Why would they converge?

0FeepingCreature11y

I probably used the wrong word; rather, they don't diverge, they end up looking the same. If initial state is the same, and physics are the same, then calculation will end up the same likewise. In that sense, every interaction by simulation Gods with the sim is increases the bit count of the description of the world you find yourself in. (Unless the world of our simulation God is so much simpler that it's easier to describe our world by looking at their world. But that seems implausible.)

[-]Conscience9y00

Definition of Rational Atheist - considers probability of God at ~5%.

Row Labels Average of PGod StdDev of PGod Count of PGod

Agnostic 17,20 22,13 130
Atheist and not spiritual 3,27 10,15 958
Atheist but spiritual 7,95 19,66 129
Committed theist 75,07 34,40 47
Deist/Pantheist/etc. 41,24 36,04 19
Lukewarm theist 48,09 34,61 39
(blank) 3,67 5,51 3
Grand Total 9,96 22,90 1 325

[-]Conscience9y00

I don't understand:

393 answer “yes” to at least one mental illness question. Lower bound: 26.1% of the LW population is mentally ill. Gosh, we have a lot of self-diagnosers.

Depression Yes, I was formally diagnosed: 273, 18.2% Yes, I self-diagnosed: 383, 25.5% No: 759, 50.5% *

The question above is one option, so for depression only self+notself 383+273 = 656, how come 393 lower bound?

[-]private_messaging10y00

I think it's interesting to note the lack of significant correlation between either IQ or calibration(as a proxy for rationality and/or sanity) and various beliefs such as many worlds. It's a common sentiment here that beliefs are a gauge of intelligence and rationality, but that doesn't seem to be true.

It would be interesting to include a small set of IQ test like questions, to confirm that there is a huge correlation between IQ and correct answers in general.

[-]taryneast10y00

In retrospect, I'm really really sad we didn't have a "city" question (along with country) this year. In talking to other Australian organisers, we occasionally wonder about which cities to target with "come to your local LessWrong meetup" notifications... that time we had city in the survey was useful for knowing that there was a body of local people who were interested, but we hadn't reached.

So I'd like to request that next year we add that question back, please :)

1ChristianKl9y

The problem with a city question is that it allows you to look up the results of individual people quite easily.

1taryneast9y

This can be solved by only using it in aggregate (ie not releasing it in the final CSV)

[-]Good_Burning_Plastic11y00

What's the correlation between the left- and the right-hand digit ratio?

[-]jbay11y00

A note about calibration...

A poor showing on some questions, such as "what is the best-selling computer game", doesn't necessarily reflect poor calibration. It might instead simply mean that "the best-selling game is Minecraft" is a surprising fact to this particular community.

For example, I may have never heard of Minecraft, but had a large amount of exposure to evidence that "the best-selling game is Starcraft", or "the best-selling game is Tetris". During the time when I did play computer games, which was before t... (read more)

7Vaniver11y

Well... this can get into philosophical territory. It seems cleanest to describe confidence in something wrong as "false," although it may have been the best possible answer for you to give at the time. Saying "the confidence wasn't false because it was from a systematic bias" seems like it opens up a host of problems. An idea that seems useful here is "being predictably surprised." You can't stop yourself from being surprised without omniscience, but you probably can stop yourself from being predictably surprised with a well-trained outside view. If the question had asked about, say, the median price of a meal at a restaurant, and you hadn't frequented restaurants in 10 years, you would probably have very low confidence in the specific price you remember, since you expect things to have changed in 10 years; a similar approach would have worked here. ("Well, Starcraft was super popular, but it's been a while; let's say Starcraft with, say, 10% probability.")

[-]Unknowns11y-10

Strange, someone downvoted this summary post.

[-]skeptical_lurker11y-20

Has the digit ratio/feminism/immigration thing been published elsewhere? If not, this is an interesting novel result.

Its also kinda worrying IMO, as it seems to indicate that (surprise, surprise) political reasoning is far more emotional than logical, logic being independent of the hormone balance of the person doing the reasoning. This also explains how open borders and basic income are both really popular, despite being mutually exclusive AFAICT (if one country has both these policies, then loads of poor people would move there, and the system would coll... (read more)

5Kaj_Sotala11y

I don't disagree with the conclusion of political reasoning being more emotional than logical, but "hormone balance affects one's political stance" doesn't necessarily imply that conclusion. Hormone balance could also affect your values. So even if everyone was a perfectly logical reasoner, hormone balance could still put them in different political camps since their values would be different. (The most stereotypical hormone -> values pathway that comes to mind would be that both women and left-wing groups are generally thought to put a higher value on the care/harm axis of the MFT than men/right-wing groups do.)

5skeptical_lurker11y

Yes, you are quite right. My mistake.

LESSWRONG
LW

128

2014 Survey Results

128

128

I. Population

II. Demographics

III. Mental Illness

IV. Politics, Religion, Ethics

V. Community Participation

VI. Other Categorical Data

VII. Numeric Data

VIII. Fishing Expeditions

IX. Digit Ratios

X. Calibration

XI. Wrapping Up