It's that time of year again. Actually, a little earlier than that time of year, but I'm pushing it ahead a little to match when Ozy and I expect to have more free time to process the results.

The first draft of the 2014 Less Wrong Census/Survey is complete (see 2013 results here) .

You can see the survey below if you promise not to try to take the survey because it's not done yet and this is just an example!

2014 Less Wrong Census/Survey Draft

I want two things from you.

First, please critique this draft (it's much the same as last year's). Tell me if any questions are unclear, misleading, offensive, confusing, or stupid. Tell me if the survey is so unbearably long that you would never possibly take it. Tell me if anything needs to be rephrased.

Second, I am willing to include any question you want in the Super Extra Bonus Questions section, as long as it is not offensive, super-long-and-involved, or really dumb. Please post any questions you want there. Please be specific - not "Ask something about taxes" but give the exact question you want me to ask as well as all answer choices.

Try not to add more than a few questions per person, unless you're sure yours are really interesting. Please also don't add any questions that aren't very easily sort-able by a computer program like SPSS unless you can commit to sorting the answers yourself.

I will probably post the survey to Main and officially open it for responses sometime early next week.

New Comment
274 comments, sorted by Click to highlight new comments since:
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings
  • I suggest a basilisk question, as usual.
  • I'd also like OCEAN/Big Five restored. The results from last time weren't very definitive and more data would be helpful.
  • Given the recent SSC debate on how effective MIRI is, it might be interesting to ask something about opinions of MIRI's effectiveness in pursuing its mission: just asking how much people donated is insufficient since one might donate faute de mieux, and IIRC the modal donation is $0 in the first place.
  • Perhaps the AI projection questions could be replaced by the same question-set used in "Future progress in artificial intelligence: A poll among experts", VC Müller, N Bostrom? Buys one comparability with their survey responses; it would be very interesting and intriguing if LWers turned out to be less extreme.
7Scott Alexander
Any particular implementation details on OCEAN? Exact same as last time?
I suggest including the Big Five Inventory (BFI) in the survey itself. I've created an example of this on Google Forms. (I've reordered the inventory such that the first 11 items constitute the BFI-10, so that respondents can choose between the 44-item and 11-item versions). The BFI is the inventory that was used in the online test to which the 2012 LW census linked. See also my comment about this in the 2012 LW census thread.
2Peter Wildeford
Why not directly include the 10-item Big Five in the survey itself?
I'm not sure. Didn't we have some odd results last time? It may be better to use a different website to see if the aggregate results or the results from retakers differ with last time.
Agree with this. Big 5 is worth having.
I strongly agree with the basilisk suggestion; have you provided Yvain with a specific question and set of answers to use, per his request, or can you do so? Or would you not be satisfied with including it as a Super Extra Bonus Question, wanting it to be in the main body of the survey?
I believe I did so, but Yvain is a smart enough cookie that I don't really need to draw up a list of exact phrasings. The question is not how exactly the question will be written, the question is whether such a question will be allowed at all. If it is to serve its intended purpose, it would be much better to get it in the main body of the survey to defang the sample size objection.
He said “Please be specific - not "Ask something about taxes" but give the exact question you want me to ask as well as all answer choices.”
The basilisk question is an exception.
No it isn't
At best I think we should directly the corresponding test questions into the census. Is there a particular set of questions that would be ideal for this purpose? I also oppened a question on cogsci.stackexchange for that purpose.
It would also be very interesting to see whether those who think that MIRI is more effective actually donate more.
A basilisk question isn't necessary! What could we learn from it? However, the MIRI question is a good one.
That most people on LW don't believe in the basilisk. That could be useful when facing outsiders who write articles about the basilisk.
I guess when you put it that way I can see its use.

I would personally be interested in more detailed drug use questions. Examplae. In the Past year have you taken:


-Amphetamine (of any sort including Adderall)

-Heroine/Other Opiates (not prescribed by a doctor as a painkiller) -Marijuanna

-Hallucinogens (LSD/DMT/Psychobillin)

-Testosterone/HGH (if you took testosterone because you are transexual do not click yes)

Maybe there is a better list of drugs?

I had this last time, and several people told me to take it off because it was bad to make people admit to illegal activities.

Also, for complicated reasons I can't do "Check as many as apply" questions, so this would take forever.

Well, nicotine, caffeine and alcohol are legal in jurisdictions most LWers come from, so you could at least ask about those.
You could use randomized response method for this question or any other controversial question (for the type of questions that are technically feasible, of course).
Wouldn't that make it pretty much useless for anything other than estimating the number of people who would answer in a certain way?
Alcohol too.
Let's see ... * Caffeine * Alcohol * Nicotine * Cannabis (marijuana, hashish, hash oil, Marinol, etc.) * Cocaine * Amphetamines (including Adderall) * Eugeroics (wakefulness drugs — modafinil and its ilk) * Sedatives (e.g. benzodiazepines including Valium, Klonopin) * Hypnotics (sleep drugs, e.g. Ambien, Lunesta) * Opiates (heroin, morphine, oxycodone) * Psychedelics (LSD, psilocybin, mescaline, DMT) * Dissociatives (DXM, ketamine, nitrous oxide) * Deliriants (belladonna, brugmansia, datura) * Entactogens (MDMA, etc.) * Medical antidepressants (SSRIs, etc.) * Herbal antidepressants (St. John's wort, ashwagandha, rhodiola) * Serotonin precursors (L-tryptophan, 5-HTP) * GABA-β agonists (phenibut, gabapentin) * Antipsychotics * Aphrodisiacs (yohimbine, etc.) * Vasodilators (Viagra, amyl nitrite, etc.)
Admitting to taking medication for treatment of a mental condition can be very bad for your career. If these questions are asked, the results should be segregated from other columns to prevent deanonymization.
Admitting to taking illegal drugs can be, too. Some forms of prejudice are even more institutionalized than others.
Idk how long of a question Scott Wants this to be. If Scott is ok with a very long list than I too would prefer that. You list looks reasonably good. Though I would include testosterone. People take this for reasons other than muscle growth.

The question about "Country" should clarify whether you are asking about nationality or residence.

The "currently looking for more relationship partners" feels to me ambiguous -- what if someone is open to having more relationship partners, but not actually looking for more? Perhaps make this a three option issue "actively looking" vs "open to having more relationship partners but not actively looking for more" vs "not open to having more relationship partners"?


I'd add: "Do you use/have you used a Spaced Repetition program, like Anki or Supermemo?"

  • No
  • Yes, regularly
  • I tried it, but stopped

In the politics I'd add a "how strongly do you care about this political affiliation" (scaled on 1-5, from "not at all" to "it's an important part of my identity")

In the politics I'd add a "how strongly do you care about this political affiliation" (scaled on 1-5, from "not at all" to "it's an important part of my identity")

A slightly longer option is to ask people to rate every political affiliation from 1 to 5. This lets us identify clusters better, as well as total political identification (if someone gives 1s to everything, that's the 'apolitical' option).

I like this idea. If implemented, it might be wise to retain the old-style political question (maybe explicitly labelling it as such) as well, for backward compatibility with past surveys.

A couple (more) questions I'd find interesting:

  • How many times do you exercise per week, on average?

  • How many nonfiction books do you read per month, on average?

  • How knowledgeable would you consider yourself in the following fields? (on a scale of 0 to 5, where 3 is about "studied it at university" and 5 "I'm a publically recognized expert")

    • Psychology

    • Economics

    • Artificial Intelligence

    • Statistics

For the first one, it might be better to ask how many hours / minutes rather than how many times. Otherwise somebody's 10-minute cycle to work is counted with as much weight as somebody else's 2 hours in the gym.
(agreed, and the question should probably be formulated to distinguish deliberate exercise from things like cycling to work)
I prefer asking for the times of exercise per week and a question for when the person last exercised to check whether certain people overestimate the amount they exercise. It would be interesting to see whether people who are better calibrated on the calibration question do better at this task then people who aren't calibrated. I also don't think that minutes are very meaningful. Intensity matters as well. It's easier to stay with the simple question for the amount of exercise. I don't think there a strong systematic bias at play.
I like the calibration check idea, and it's a fair point about intensity. The last survey I took that included this kind of question asked about "moderate exercise (eg brisk walking)" and "intense exercise", or some similar wording, which I thought was a reasonable split. These might all be details we don't care about though.
Do you think the survey should also take into account BMI + bodyfat % if it includes fitness questions?
Could be a good addition! I don't really know what the purpose of these questions is other than vague general interest... is there some hypothesis like "people who think there will be a Singularity soon are more active/healthier than people who think it will be less soon / never"??
I would think it would be the other way around - no sense preserving your health if you'll be uploaded in a few years anyway, right?
I think it would depend on the probabilities and payoffs: if surviving to the Singularity is sufficiently worthless, or taking care of your health is sufficiently expensive, or the Singularity is near with extreme confidence, then you might conclude that you can neglect your health, stop brushing your teeth, etc. But very few people expect any singularity to be near or sooner than a decade or two, or with extreme confidence, so I think for any plausible values you'd wind up wanting to pay attention to your health. Cancer can strike at any time, among other threats.
What does it mean "studied it at university"? Do you mean something like "Took econ 101 and 102 as part of gen ed requirements" or "majored in economics"?
The goal is to give reference points for the jumerical scale so the numbers can be more meaningfully compared, not necessarily to have a detailed reference that will annoy some people (too much to read) and confuse others (don't fit in). I was thinking of something intermediate, like "took specialized classes, not 101", but I'm afraid being too specific about education means people will not take into account other things like how much they forgot or how much they learnt in the meantime. Someone who took econ 101 but works in a bank and argues about economic policy on the internet everyday and read a few good books is probably more knowledgeale than someone who majored in economics and then said "screw it all" and became an actor.
Then make this clear in the prompt.
Good questions. Especially about excercize. Studies allways seem to show big correlations with that so its worth asking about.

Formidable work. Given recent discussions, might it be worth adding "vegan" to "vegetarian"? (And perhaps even "pescetarian" or "flexitarian" but I suppose one can get lost in small distinctions.)

I'd propose the following three:

  • vegan
  • vegetarian
  • reduced meat intake

The latter, 'reduced meat intake' is intended to represent pescetarian, flexitarian, and meat reduction all in one.

This is a good idea
Alternatively, there could just be the single question 'Are you a vegetarian', and below, a text box next to the question 'if you answered 'yes', why are you a vegetarian?
I don't think a text box is called for, but radio buttons of Ethics/Health/Other might be good.
Not sure whether "aethetics/personal taste" is common enough to be worth breaking out of "Other" but I'm vegetarian because I'm a picky eater and hate the texture of meat.

I'd been interested in seeing how many people participate in systematic self-improvement efforts outside the LW/CFAR space. Some examples in rough increasing order of weirdness: diet, fitness, psychiastric help, life coaching etc., social skills programs of various kinds, nootropics. This covers a lot of ground and there's probably some options I haven't thought of, but these two questions would be a good start:

  • Do you follow a limited or unusual diet for health (not ethical) reasons?

(this would probably need "I tried and it didn't work" and "I tried but found it too hard" options. We'd probably also want some way of distinguishing between "I'm allergic to onions" and "I kind of try not to eat red meat" and "I'm full-blown hardcore paleo".)

  • How many hours a week do you spend exercising, on average?

(for example, 0 if you are single, 1 if you are in a monogamous relationship, higher numbers for polyamorous relationships)

This implies that you are only poly if you're in more than 1 relationship. I suggest simply

(for example, 0 if you are single, 1 if you are in a monogamous relationship)


  • If you're poly, the idea of counting relationships will be familiar to you already; it's only those who aren't who might need it set out explicitly
  • I hate saying "you have to add more text here to cater to my itch" - if I can fix what I perceive to be a problem by removing text, I much prefer that.

I feel like using Scandanavia as an example of "socialism" is not really accurate - they're capitalist welfare states with slightly higher taxes than other capitalist welfare states.

Agree. Here is the Heritage foundation ranking of countries by economic freedom. The Heritage foundation's libertarian perspective views Scandanavia reasonably favorably. And they include tax rates in their analysis:
3Scott Alexander
How would you handle this?
Socialism - workers own the means of production. (Although according to Wikipedia, "There are many varieties of socialism and there is no single definition encapsulating all of them.") Communism - No markets/the state determines what is produced.

That definition of "socialism" conflicts with the platforms of parties caling themselves "socialist", which is probably what most people using that label have in mind.

There are other confusing things in the survey draft, for example, UK Labour party, which is used as example of liberalism (in the American sense), is affiliated with the Party of European Socialists, the exact same group the Swedish Social Democratic Party (which, I suppose, is used as an example of socialism) also belongs to. Which group do other PES-affiliated parties belong to? Am I correct that by making categories "liberalism" and "socialism", Scott tried to capture the distinction between "mainstream centre-left" and "those that are further to the left than mainstream centre-left" (e.g. GUE/NGL group)? Although I might be unfamiliar with whether left wing people themselves perceive this as an important distinction. I would be very surprised if ~30 percent of LWers self identified as socialists in "workers own the means of production" sense rather than "voting for a PES-affiliated (or equivalent) party" sense.
Just because two parties affiliate doesn't mean they are pursuing identical policies.
If this is intended as a question about existing political clusters, then I think you should merge the "Liberal" and "Socialist" options; I don't think people see them as separate positions (at least in my country) in the same way as Conservative/Libertarian. If this is really just a question about the left-right economic axis, then let's make it one, viz: "Which of the following most closely describes your position on economic freedom vs. wealth redistribution?" * Extreme economic liberal, e.g. US Libertarian Party: minimal/no taxes, no redistribution of wealth * Moderate economic liberal, e.g. US Republican Party or UK Tories: low taxes, low redistribution of wealth * Moderate economic progressive, e.g. US Democrats or UK Labour party: high taxes, high redistribution of wealth * Extreme economic progressive, e.g. (I don't know, UK Lib Dem?): high taxes, major redistribution of wealth
Here in France, "liberal" is a slur used by left-wingers against anybody suspected of liking the free market (there used to be right-wing politicians claiming that label, I haven't seen any in the recent elections)
Scandanavia is pretty much what people using the word socialiam in the modern world refer to. Although it have have a different historical definition it would probably cause confusion to use it.
Even though the tax rates in America and Scandinavia are similar, the way they're spent is very different: IIUC in the US more money is spent on means-tested transfer payments and in Scandinavia more money is spent on public services like healthcare and education.

For relationship status, a polyamorous person can be married and in a relationship at the same time, which is a problem. Similarly, someone can be living with their partner/spouse and additional roommates. Also, "Liberal" in the Political section should probably be renamed to "Progressive", to avoid collisions with how "liberal" is used in Europe and in political philosophy.

Please change the "referred by a link on another blog" option to "referred by a link on another blog or website". It's been bugging me for years.

Again, your relationship style question conflates very different clusters. There is a big difference between traditional monogamy and serial monogamy, or between the ideal of polyamory preached in rationalist circles and the harem type of polyamory (polygyny, mistresses, etc...) which is traditional for a sufficiently high-status man to have. The obvious way to solve this is to split the "prefer monogamous" and "prefer polyamorous" answers into two answers each, each of which describes the relevant clusters in a short sentence similar to the answer choices in the political question.

You are asking too many calibration questions for me. I might answer 1, 2, or, at a stretch, 3. 10 questions is gonna cause me to leave that whole section blank.

Likewise, I don't think I'm gonna answer the probabilities section unless you switch to radio buttons, like you use in the extra credit political section.

Per your own advice, you may want to add a completely implausible question to the survey so you can measure ou... (read more)

Per your own advice, you may want to add a completely implausible question to the survey so you can measure our Lizardman's constant and subtract it from the other results.


Wait. There's other "rationalist fan fiction" other than HPMoR?! Will someone link me to a list? (Also, how do I decode the gibberish?)
Of course there is. Why, there's _Luminosity_, Friendship is Optimal and the wider Optimalverse, Myou've Gotta be Kidding Me, "Mortality Report", "Mortal" and "Mother of Nations", "Good Night", Hamlet and the Philosopher's Stone, Branches on the Tree of Time, The Metropolitan Man, "The Amazing Peter Parker"... Or you can just look at the TVTropes page. Gibberish is in ROT13.

I shall repeat my request for a second question under "Degree" that asks about one's highest degree attempted or in progress.

It seems to me that the distinction between "started work on degree X, but abandoned it" and "currently working towards degree X" is almost as large as that between either of those and "never attempted degree X". (Likewise for qualifications that aren't degrees.) So if we really want more resolution than a single question will get us, it seems like we need three questions to capture all the information. If so, maybe they should be (1) highest degree attained, (2) highest degree attempted or in progress, (3) estimated probability of completing #2 if different from #1. So if you've abandoned a degree, you'd put 0 or something close to it for #3; if you're still working on it and fully expect to complete it, you'd put something close to 100% for #3; if you're struggling through a PhD and uncertain of the future, it might be more like 50%. This gets (aside from some unusual cases where someone has both abandoned a degree and started on a new one that they haven't finished yet) strictly more information than "highest attained; highest attempted; highest in progress" -- but at some cost in comprehensibility. If having three questions is excessive but having two isn't, perhaps the following would be a good balance: (1) highest degree earned or in progress and confidently expected to be completed, (2) highest degree abandoned or in progress but quite likely not to be completed. If only one question, I think asking for highest degree actually earned is the best option.
The standard on psych studies seems to be "Level of education: (1) Some high school (2) High school graduate (3) Some college (4) College graduate (5) Some post-graduate (6) Graduate degree." This is pretty simple and does not warrant more than one question.
I personally have no problem with that -- but komponisto wants to make more detailed distinctions, and was originally (i.e., at the other end of the link in the great-grandparent of this comment) responding to someone else who wanted to count courses currently in progress as well as ones already completed. I'm sure both of them have reasons (indeed, it's not hard to guess some) and I bet they're both aware that it's usual simply to ask for highest qualification actually attained.
Kalium's suggestion would in fact satisfy me -- it captures the distinction between someone who went to college/graduate school but didn't finish and someone who never went in the first place. I disagree with your comment above that For a completely hypothetical example, let A = someone with a B.A. who spent 6 years in a Ph.D. program and left without finishing B = someone with a B.A. currently in a Ph.D. program who hasn't finished yet C = someone who completed a Ph.D and D = someone with a B.A. who never entered a Ph.D. program.. My view is that there should be some notion of "education level" that clusters together A,B, and C, as distinct from D; but "highest degree attained" clusters A,B, and D as distinct from C, and your proposal seems to put A, D in one cluster and B, C in another.
  • Neither of "...and currently looking for more relationship partners" or "...and not currently looking for more relationship partners" seem to capture what I feel is a common poly sentiment, "...and I'll figure out what I want with new people as I get to know them".
  • Neither of "Atheist but spiritual" and "Atheist but not spiritual" seem to capture "Atheist who doesn't know what 'spiritual' means in this context and therefore doesn't want to come down either for or against it, because there are both int
... (read more)
That's not a terribly rare sentiment among non-poly single people either.
P(Aliens) is P(Aliens) and not P(Aliens|Simulation=False)
It seems to me that unless a new person shows up and you say "sorry, not looking for more right now" you're in the first category. But separating out "looking and receptive," "not looking but receptive," and "neither looking nor receptive" seems potentially useful.

(hint: if you've been here since the start of the community in November 2007, put 6)

This should presumably be 7 now.

Consider asking people in what year they joined the community? So if you've been here since the start, put 2007.

I would much rather just have to remember a community start date than recalculate this every year. The primary issue is it makes the data a little less clean, especially if you want month-level accuracy.

If there's no good way to ask about specific supplements in the survey, would anyone mind if I raised it as a question in Discussion?

How about a survey question about how much people care about their LW karma?

For Super Extra Bonus Questions: (feel free to modify the answer choices)

With which of these metaethical positions do you most identify?

  • Non-cognitivism: Moral statements don't express propositions and can neither be true nor false. "Murder is wrong" means something like "Boo murder!".
  • Error theory: Moral statements have a truth-value, but attempt to describe features of the world that don't exist. "Murder is wrong" and "Murder is right" are both false statements because moral rightness and wrongness aren't features
... (read more)
Could you have less fine-grained answers, so that I don't have to spend a week on the SEP just to know what the answers mean?
If you want less fine-grained answers, there's the consequentialism/deontology/virtue ethics question in the earlier part of the survey.
Or “Let's please not murder”, which doesn't express a proposition either.
"Moral nihilism is the meta-ethical view that nothing is intrinsically moral or immoral." ( Utility functions (aka morality) are (is) in the mind, not in Nature. That would probably be the answer of most LWers. Otherwise, you'll have to tell me what you mean by morality. Is utilitarianism used as "maximizing happiness" or "maximizing utility". If it's "maximizing utility", well isn't that everyone's position? What differs is simply what counts as "utility".
That is an inaccurate definition of nihilism because it doesn't match what nihilists actually believe. Not only do they reject intrinsic morality, they reject all forms of morality altogether. Someone who believes in any kind of moral normativity (e.g. a utilitarian) cannot be a nihilist. Utilitarianism is used as "the normative ethical theory that one ought to maximize the utility of the world". This is in contrast to something like egoism ("the normative ethical theory that one ought to maximize one's own utility") and other forms of consequentialism.
Thank you for your answer. Do nihilists think they have no goals (aka terminal values) or do nihilists think they don't have goals about fulfilling others' goals or is it something else? Ok so would that be right to say this?: Utilitarianism is giving equal weight to everyone's utility function (including yours) in your "meta" utility function. Egoism means you don't consider others' utility function in your utility function. And then there is everything in-between (meaning giving more weight to your utility function than to other's utility function in your "meta" utility function).
I am not a nihilist, and I don't know if I'd be able to pass an Ideological Turing Test as one, but to give my best answer to this, the nihilist would say that there are no moral oughts. How they connect this to terminal goals varies depending on the nihilist. The first part, kind of, the second part, no. The utilitarian holds that the right thing to do is determined by what maximizes world utility, which is produced by utility functions. All utility, including your own, is given equal weight in the "moral decision" function. As for egoism, it simply means that you consider others' utility functions to the degree that they're a part of your utility function. It doesn't mean that you disregard them altogether.
Ok thanks for your answers!

Thanks so much for running this again


-Items should either be randomized or if sorted, they should be sorted based on last years prevelency. For example in politics libertarian is the top listed item despite not being alphabetically first or being a majorty / plurality view.

-The religious denomination shouldn't ask atheists to skip the question. One is about what you believe the other about what you do. Plenty of atheists fast on Yom Kippur or go to church on Christmas.

-Less wrong use should have two items on comments that differentiate by frequen... (read more)

That should be Yom Kippur.
Oops. Thanks for pointing that out. I'll correct it.

If this survey generates interesting psychometric research, someone might try to get a journal article out of it. If so, we will need your explicit consent to have (an aggregate of) your anonymous data published.

  • Someone? Who is someone? Honestly, I'm curious, because I can't think of who these someones would be. What research psychologists who aren't already members of Less Wrong pay attention to it? I doubt it would be Scott using it to publish something(?) Maybe he shares the results on Slate Star Codex, and a psychiatrist friend of his considers Les

... (read more)

In the mental health category, I'd love to see (adult) ADHD there as well. I'm less directly interested in substance abuse disorder and learning disabilities (in the US sense) / non-autism developmental disabilities, but those would be interesting additions too.

Some thoughts from my notes when I did the LWSH survey. I mostly plagiarized it from you, so they should be relevant:

Consider adding an option under work and/or profession for "Homemaker." We have at least one.

Consider an option under the relationship goals section for "not looking, but open to the possibility."

There used to be a whole section on akrasia. I note that it is now missing, but it had questions about use of medication and supplements. Those questions didn't specify whether the meds were taken specifically to combat akrasia.... (read more)

How would you describe your opinion on immigration?

Immigration into one's own country (as for the following two questions) or in general?

  • Given that you already ask the Sex question, the Gender question can have just three answers “Man”, “Woman”, and “Neither/Other”.
  • As suggested by people on SSC, you might want to add a question “Is your gender an important part of your identity?” or something, to find out how many of us are cis-by-default.
  • I'd have three answers to Relationship Goals, “actively looking”, “open but not actively looking”, and “not open”.
  • Split the “no” answer to Children into “not now” and “never”.
  • Selfishly, I'm glad you didn't split physics in the Profession question into
... (read more)

Anything I do with gender and sex is going to have lots of people yell at me. But if I keep it the same, it will be the same people as last year and I won't make new enemies.

But to all of us perched on the back of Cthulhu, who is forever swimming left, is it the survey that will seem fixed and unchanging from our moving point of view?
Trans people overwhelmingly dislike things that imply sex/gender to be distinct. I find this somewhat silly and long-term counterproductive, but it causes them psychological distress (often severe) to disregard it, so keeping things very clearly defined is probably a good thing.
Distinct? Regardless, I take issue with your comment. It reads insultingly.
As far as most (I think all) trans people I know are concerned, sex and gender are synonyms, because as far as triggering them is concerned, they are synonyms. In the medium-to-long term, making their be very clear distinctions between sex (biological) and gender (social) would be a very good thing to have fixed in the broader culture, as it would normalize the distinction and have the language reinforce lack of transphobia, but in the short run it would pick at existing emotional wounds, so at least unless and until there's an organized effort to establish the distinction in common language, reflecting their strong personal preferences seems like the right thing to do.
Hm. Would an example w/r/t triggers be that many trans men don't like being called female-bodied? This doesn't stem from seeing sex and gender as synonymous, but is just due to the terminology rubbing the wrong way.* AFAB (assigned female at birth) and AMAB are the generally preferred terms. *I'd speculate that "female-bodied" implies that the word "female" on its own has something to do with gender (why else add "-bodied"?) which in turn makes "female-bodied"'s implied meaning "body belonging to one of the female gender". Also, not merely referring to birth assignment means that the term breaks down during medical transition as the body changes.
Things like that, yes, but it applies pretty broadly. In my experience (and I've talked about this with a couple close friends to get a sense of the boundaries), anything that generalizes beyond very specific traits tends to trigger dysphoria, and that includes stuff like "you still have a typically-female muscle distribution, so you should probably use workout advice for women initially", or 'well, female bodies are usually more sensitive to extreme temperatures'; the kind of things that are talking about how their body is shaped right now, where it's the growth pattern, rather than assignment at birth, which is important. This is especially relevant because for a lot of this kind of thing, it doesn't change even when going through hormones; hip bones, larynx, and sex organs are the most obvious ones, but a lot of stuff about bone structure will only change on the scale of decades. A lot of this will become less relevant if puberty-blockers become standard, because then a lot of the generalizations will only apply to people who first considered the possibility later in life or were blocked for family reasons. Some people I know are optimistic that standard will happen soon, and if they're right, then that probably will happen before the cultural payoff of separating the terms would show up. I don't think that's terribly likely, myself.
If they insist on claiming to be negative utility monsters, I don't see why we should be indulging them. I find this sentence, and the whole tone of this comment, rather amusing because of how vividly it illustrates the contradictions of their claims. Namely, that people have an "inherent gender" that is independent of their biological sex and not subject to change by any means even as, they hope, all other sex-related traits are changeable.
What do you think is going on with trans people? Many of them spend considerable amounts of money and effort to change their genders. (At least in the US, sex change surgery isn't covered by insurance, so far as I know.) They're also generally up against a lot of stigma. They risk their marriages. These revealed preferences suggest that they really don't like their assigned at birth genders.
People do highly irrational things all the time. Probably a combination of things. A lot of MtoF's are cases of autogynophilia. Some are men who realized that by putting on a dress they transform from yet another white male CEO to a twofer minority CEO. Now that some municipalities are passing laws requiring bathrooms to be based on self-identified "gender", I predict creepy guys suddenly claiming to be trans-women as an excuse to hang out in women's bathrooms. A few articles on by Steve Sailer on some of what else may be going on here and here. What do you think is going on with this guy?
"Claiming" is an odd word choice, since we don't have any reason to expect that they're lying. Even if we reject innate dysphoria or whatever, it's ridiculously easy to self-modify into being more easily upset (even unintentionally).
I know, I still don't see the point of humoring the people who do so.
I just thought it was important to point out, because there's a significant difference between humoring people who are pretending to be easily upset and handling people who are actually easily upset.
If you refuse to humor someone who is easily upset, he will (possibly after a tantrum) become less easily upset. If you humor someone who is easily upset, he will become even more easily upset.
Interesting. My experience dealing with easily-upset people is limited, so this is good to know.
The amusing thing is that even liberals apply this knowledge when dealing with easily upset people who are in their actual outgroup.
There is no contradiction there, and I never addressed your ridiculous claim to the contrary. Here's a short, basic developmental human biology excerpt: Secondary and tertiary sexual characteristics are shaped during devlopment by simple hormone balance and reshaped during puberty by hormone floods, which are simple to change at the correct time. A person's gender, along with many other idiosyncratic features like their preferences about sex, food, etc., are properties of their mind and brain, and brain structure is both vastly more complex, and much slower to change.
And your claim is that there are significant numbers of people where nearly all those secondary characteristics associated with the brain and only those are one way whereas all the other ones are the other. Given the type of code evolution tends to write this is highly improbable.
Citation needed; evolution writes shitty spaghetti code that's highly vulnerable to minor changes in the runtime environment during the build process, and has no error-checking or error-correcting short of a crash. Since the brain structure develops at a totally different stage from the secondary sexual characteristics (continuously from birth through childhood vs. puberty), this is totally plausible and there's not much reason to think they should happen in the same direction in all cases. Also, calling anything about the brain a "secondary sexual characteristic" is highly specious.
What do you mean with that statement? There are various DNA repair mechanisms that do error checking. Evolution frequently copies genes and then changes one of those copies. You could see that as a way to produce redundancy against errors. We have two copies of every chromosome to provide for some error correction and survive one of the two being broken.
Precisely, it also has no reason to neatly compartment mental from physical. However, most aspects of brain structure do in fact develop either pre-birth or during puberty, same as the other secondary sexual characteristics, and the primary sexual characteristics for that matter.
While 'inherent' might be a little strong, 'fixed' is a totally accurate description. I suggest you find some actual trans friends to talk about this with before you dismiss them. Also, I'm not sure where you're getting the idea that they're in any sense negative utility monsters. If you're cis and don't have any trans people you interact with frequently, this won't affect you at all, except insofar as it might gradually move the words 'sex' and 'gender' further apart in conceptspace. It's still good practice to do it anyway, because you might regularly interact with a closeted trans person who you are harming, but unless you choose to do that it had no impact either way.

More Children

Please make the options:

  1. Yes
  2. Yes, but not now
  3. No
  4. Uncertain

Is anyone interested in asking whether people are cosmopolitan?

How about natalist versus antinatalist?

[Added] I don't care about this that much, but here's a question text for cosmopolitanism: "In prioritizing political or charitable causes, would you weight the interests of a person of another nationality equally with those of a fellow citizen?" (Yes or No) A link to Chris's post might also help.

A Pro- and Anti-Natalism question seems like it could be high value. I'm not sure what the best wording might be.

Not exactly a question per se, but I remember Yvain complaining about invalid entries, such as words in the number field etc. etc.

This sounds like something that simple Regular Expressions could speed up. Perhaps we could ask him what his current survey workflow is and try to see where they can fit?

Although this advice seems a bit TOO obvious not to have been mentioned before. Apologies if so.

The "with family" option in the "living with" question is ambiguous for those of us with children. I suggest changing it to "with parents or guardians[1]", changing the partner/spouse option to "with partner/spouse (and children if applicable)", and adding an "other" option for less traditional living arrangements.

Questions in the mental health section are inconsistent about whether they're referring to whether you have ever suffered from a condition ("have you ever been diagnosed...") or whether ... (read more)

I live my with my children but not with a partner or spouse, so I'd want to see even more family arrangements, since I don't think single parenthood is unusual enough to be lumped in with "other."
I'd propose “with parent(s) and/or sibling(s)”, “with partner(s)/spouse(s) and/or child(ren)”, “with roommate(s)”, “alone” and “other/I'd rather not say”. Maybe also “in communal sleeping quarters, e.g. students' dorm or nursing home”.

I'll write down a long list of questions. I think they are useful. If you don't like them all because they are too much I would certainly like to keep data about spaced repetition software usage. The question from last year was good.

I would specifically like to have the bolted questions included. With them we could cut the question that asks for self identification as vegetarian.

Given the absence of perfect information our justice system has to convict a few innocent people. Where the optimum of the ratio between guilty going free and innocent getting se... (read more)

I like the moral rating questions, but I think they should have a scale that includes both positive and negative values. Asking "how good or bad is X?" will get more honest answers than "how good is X?" or "how bad is X?".
You are probably right. The core things I care about isn't direct honesty. I want to see how the difference in judging eating animals translates into rating of how much animals the person eats. That means I need a big scale where a lot of people can answer that eating animals isn't ideal but not that big of a deal. I also don't know that anyone argues that humans have a moral obligation to eat animals. On the other hand in the case of rerouting trolley cars a two sided scale seems important.
"flesh" feels like a gratuitously loaded term. Is there any reason for that question not to simply be "when was the last time you ate meat?"
I have no problem with changing it to meat.
That depends on which crime we're talking about -- somewhere around 10^6 for illicit drug use, 10^3 for shoplifting, and 10^0 for murder. Does Duolingo count? Given that you also ask about multivitamins, I guess you mean ‘a vitamin D supplement not as part of a multivitamin’, but you might want to say it explicitly. Do e-cigs count? Language nitpicks: “eat meat [not flesh]”, “ride [not drive] a bike”.
I would prefer not counting it. Whether it's bundled isn't that important. On the other hand what's interesting would be the dose. But that add another question.
FYI, when I read your abortion questions I was unsure whether you were counting months post conception or post birth. Timing of pregnancy is more often discussed as trimesters or weeks.
I don't think abortion post birth is a thing. If it's easier to understand I have no issue with changing it to 12 weeks and 28 weeks.
"After-birth abortion" turns up sometimes in thought experiments and philosophy papers, and my prior on weird thought experiments turning up on lw is pretty high.
Actually, I remember hearing someone mention that in general LWers are more okay with Infanticide (or post-birth abortion if you prefer) than the average population. The reasoning, I assume, being that their self-awareness is more similar to animals than to a human adult and that you aren't really destroying a full human consciousness. I don't remember where it was posted, but i do remember it sounding like someone was summarizing LW survey results. Basically, what I'm saying is that I wasn't sure either given the phrasing, since I've heard this kind of thing discussed before on LW.
Yeah, the infanticide thing is a classic Peter Singer bit which I imagine a lot of people on Less Wrong have heard and considered before. I think the standard counterargument is that we need a good Schelling fence for when it's okay to kill people - pro-life folks would argue that it should be at conception or something, pro-choice folks would say it should be at birth, but so far nobody's come up with any reasonable one that comes after that. So infanticide should be disallowed for societal reasons, even if we allow that it might be acceptable in various hypothetical scenarios.
That's very much a description of the US landscape of memes. In the EU the situation is different as our abortions laws are made by elected parliaments. In most EU states an abortion at 7 months (or 28 weeks) is illegal in the US it's legal. I haven't meet anyone in Germany who argued that Germany should allow abortion in more cases and have birth as the schelling fence.
Curiously, in some rare cases and under certain very specific circumstances it is more than just a thought experiment (the paper you have linked also briefly mentions it), although it is thought of as a special case of euthanasia rather than an unusual kind of abortion. [...]
Well Peter Singer supports it and for some bizarre reason he is considered a moral authority by many lesswrongers.
Hey, Peter Singer is pretty good. Even if you don't agree with all his ideas, his reasoning is refreshingly clear and concrete for a philosopher. If you haven't read Practical Ethics, I strongly recommend it.
I'm aware that Singer basically supports the policy. I'm however not aware that he choses those semantics.
It might be one of those Europe vs US (or maybe continental Europe vs Anglosphere) things -- pregnancies are usually measured in months here in Italy too.
  • Order the countries by their frequency in previous surveys.
  • As ever, the political labels used are misleading. The Scandinavian countries are not socialist, and the UK labour party are in many ways extremely illiberal.
  • You might want to break the cryonics question up a little more - some people are 'just finishing up the paperwork' for a very long time.
  • The immigration question is presumably actually asking about immigration controls, not immigration itself.

You have the "Slate Star Codex" question under the "privacy" section. You may want to retitle the section to something like "top matter".

I'd like an "apolitical" choice for my political views. Especially here, there's likely a significant slice of the population who choose to not have any politics in the manner of Paul Graham, and thus equally don't identify at all, not even a little bit, with any of the given political options.

"If you were referred here by another blog—" "another" is needless.

Con... (read more)

Possible drawback: this will make it easier to work out who any given response is from, which some people might be uneasy about.
I think of Paul Graham as having very strong political opinions.
Instead of an specific apolitical choice, simply "others" might be useful.
I think that "None" and "Other" should probably be kept separate options.

Formatting issues:

  • The title "Part Eight: Slightly More Complicated Questions" appears twice.

Question requests:

  • Ability to solve the Schrodinger equation for the hydrogen atom.
  • OCEAN personality test results
  • Split "no" option in meetups into "no, because there are no meetups near where I live" and "no, there may be meetups near where I live but I don't want to go to them"

Other comments:

  • I like the multiple calibration questions

Ability to solve the Schrodinger equation for the hydrogen atom.

In case you care about that in order to know which respondents know what they're talking about when answering the MWI question, that's a very poor choice (as I mentioned two years ago IIRC). It basically mostly only checks whether people took QM classes (in many of which interpretational issues are discussed hardly at all) and can remember the tricks to solve second-order differential equations in spherical coordinates. Asking whether people can prove Bell's theorem would be a much better choice. (You should weigh Scott Aaronson's opinion about MWI over mine even though I'm a physicist and he isn't.) Having read How the Hippies Saved Physics, I'd guess that if anything ability to solve the SE for the H atom would anti-correlate with trustworthiness about interpretations of QM when controlling for work status, profession and degree.

OCEAN personality test results


Fair enough. In that case, I'll request a question as to whether you can prove Bell's theorem. I guess I was lucky that in my university, interpretational issues were discussed a fair bit in later-year theoretical physics classes.
I think both questions are informative, they just test a different thing. To give an analogy from copmputer science, the question about hydrogen atom is similar in spirit to, "Would you be able to implement quicksort?", whereas the one about Bell theorem is more like, "Would you be able to reconstruct the halting problem proof?" The latter seems like a much higher bar. I'm curious, do you think there exist many people who can actually reconstruct the proof of Bell's theorem, but who can't solve the Schrodinger equation for the hydrogen atom? (I'm assuming that by solving the Schrodinger equation for the hydrogen atom, Daniel meant deriving the energy levels of a hydrogen atom from SE, as opposed to say providing the full basis of eigenfunctions including these for E > 0; the latter is much harder and I wouldn't expect most people who took even advanced Quantum Mechanics to be able to do it without looking things up).
I had assumed something in between -- deriving the energy levels, and the eigenfunctions for the levels with E < 0.
Maybe the answers given (either in the Bell theorem case or in the Schrödinger equation case) could be “Yes, right now”, “Not off the top of my head, but I know where to look stuff up” and “Can I prove the what?”
4Scott Alexander
There was a Schrodinger atom question a couple years ago. I'm trying not to keep all questions lest the survey just grow and grow forever. Any particular reason you want to know whether the Schrodinger solving percent has changed since last time?
I mainly just continue to be curious about the correlation between knowing about interpretations of QM and P(MWI). As stated above, I now think that asking about Bell's theorem is a better way of doing that (see army1987's comment and my response to it), which (to my knowledge) hasn't been asked before.

What is your preferred relationship style?

No explicit option for asexual and/or aromantic? (Or whatever the term would be for preferring not to be in a relationship, as opposed to not having a preference.)

Please answer on a scale from 0% (definitely false) to 100% (definitely true). For your convenience, 0% will be interpreted as "epsilon" and 100% as "100 minus epsilon".

For estimated probabilities that are greater than 99%, or less than 1%, how many digits of accuracy do you wish the answer to be? Just jump straight from 99% to ... (read more)

I think the whole point of the world "revealed" is to avoid referring to something like "atheistic Buddhism".
Probably not. "Revealed religion" means something like "religion whose teachings are based on revelations from the divine to mankind". If you think that your religion was figured out, discovered, or intuited by a human being rather than delivered to him or her by the gods or spirits, it's not a "revealed religion". Yes, this means Scientology is not a "revealed religion", because they think Hubbard discovered it. (Scientology is, however, a mystery religion, meaning that it has inner and outer teachings.)
I would be quite surprised if any of us could justifiably estimate any of those questions to an accuracy better than the nearest 1%/epsilon offset.
I regularly think in terms of decibans, where 10 decibans = 90% confidence, 20 = 99%, 30 = 99.9%, 40 = 99.99%, etc. From that perspective, there's as much difference between 99% and 99.9% as there is between 99.9% and 99.99%. Put another way, if 99% is the cut-off, then anything more than 20 decibans of evidence is ignored, even if the answerer has at least 40 decibans.
And yet every year lots and lots of respondents answer such questions with 0 or 100.

way fewer people use LW now, it would be interesting to track immigration and emmigration. Add some questions to see what time period people report their peak use being / what people think changed ?

Citation needed.

You should add a question that asks what model of population ethics they support (Additive utility, Average utility, Other, Non-utilitarian, Don't know).

That's a good thought. But many versions of utilitarianism reach wild conclusions, so I like better "Do you support actions or policies that lower the global population growth rate, increase it, or neither?" [Added] But I don't care much.
I think the theory question would be much more interesting, honestly. But if it's all going into the super bonus question section anyway, I wouldn't mind having both.

Question suggestions:

Only answer if you think the chance of a singularity in the 21st century is over 1%. Do you think affecting the singularity's outcome is tractable enough to be worth any of your time or money?

Are you a narcissist? (rationale)

(This is only for people who had an SAT score out of 1600) Did you take the SAT after (or during) April 1995?

And actually, the narcissism question is not yes or no. Here it is: To what extent do you agree with this statement: "I am a narcissist." (Note: The word “narcissist” means egotistical, self-focused, and vain.) 1=not very true of me 2 3 4 5 6 7=very true of me
"Only answer if" is bad design in this case. It makes more sense to ask for the chance of a singularity and ask the second question at the same time.
If Yvain wants to include another question then I'd agree. E.g., * What do you think the chance of a technological singularity in the 21st century is? * Do you think affecting a singularity's outcome is tractable enough to be worth your time or money? Also I realized that "after (or during) April 1995" is better since the recentered SAT started in April.

May want to add Slate Star Codex as an exception to the referrals question.

Time in community question needs to be updated to 7 years for the start of the community.

Might be worth it to specify aggregate Karma if you have multiple accounts. (This is an account that I started using after I decided I no longer wanted to use my real name. I mostly lurk anyway, though.)

It would be worth it to add a "no meetups in my area" option to the meetups question.

The header for part eight is listed twice.

I've been reading the thread, and it appears there are some categories of questions Less Wrong wants explored in more detail, with more questions. These subjects include:

  • rates of vegetarianism/paleo/alt-diets
  • what drugs do you take without a prescription, including caffeine/nicotine/alcohol, nootropics, and illicit recreational drugs.
  • a wide diversity of philosophical moral/meta-ethical positions
  • opinions of certain policies
  • wider variety of sexual/romantic/gender orientation options.

This is all very interesting. However, with upvotes, below all these ... (read more)

Somebody upvoted my above comment. At least one person liked this comment. I grouped all my general suggestions for changes in one comment because I thought I shouldn't clutter the thread with all my ideas in their own separate comments. I figure people skimmed over the long comment to upvote terser, better suggestions, which makes sense. Whether you upvoted the above, or not, if an single example is an excelletn suggestion, build upon it in your own new comment, or let me know. Maybe a put up a bunch of noise above, and it's difficult for us to send the signals. I just want to send my best signals through as suggestions to the survey like everyone else.

Suggestions for Questions Specific to the Center for Applied Rationality

Maybe take the opportunity to ask more questions about CFAR, and ask the CFAR staff you know if they want questions answered, because they love data. As a CFAR workshop alumnus, I'm interested in seeing where that community is headed. Here are some suggestions:

  • If you're not one of the CFAR alumni, what's your estimate that you will attend one of their workshops in the next [six months, one year, two year, insert timescale here]?

  • If you haven't attended a CFAR workshop, do you ever i

... (read more)

A question on romantic orientation would be good.

I'm unclear on the distinction between "sexual orientation" and "romantic orientation". I can understand the distinction between "sex" and "romance", but the two are strongly connected to each other. Are there people who want sex with one gender, but romance (whatever that means without sex) with the other? The Wikipedia article, um, isn't helping me.
My impression is that a person's romantic orientation is almost always a sub- or superset of their sexual orientation. At any rate I don't recall hearing of anyone who identified otherwise. But the inclusion can go either way (e.g. asexual but homororomantic, or bisexual but heteroromantic). They're strongly correlated but distinct.
There are also non-asexual people who are a- or demi-romantic.
What do you mean with "romantic orientation" as opposed to sexual orientation? I don't think the term is well known and we already have enough questions about that domain
What gender/s you are romantically attracted to, and also how strongly you feel that attraction, see the Wikipedia page. It is mainly useful for asexuals (and also, I imagine, people who answer 'other'), but it's certainly possible to have a romantic orientation that doesn't match your sexual orientation. Maybe it could be included as an optional write-in box, or at the end?
Could you provide a source for that claim that there is a sufficient number of people for which that distinction is useful?
In the last survey, there were 47 asexuals and 39 'other's. It is a useful distinction for asexuals and I imagine it would be useful for many 'other's. Furthermore, as per philh's reply to RichardKennaway's comment, the distinction is probably useful for some non-asexual people. Whether this is a sufficient number of people to add an extra question is a bit of a more thorny question. For comparison, there were 25 trans people in the last survey, fewer than the number of asexuals, and there are options for them in the gender question. Even if it's too onerous to add to the main sex/gender/relationship section is too onerous, I think that it could find a happy home in an extra credit section.
Okay, 86 people seems like enough to be able to separate them into smaller chunks.

All the time, people put DO NOT CITE on their papers, but that digit ratio web page had a link, explaining why you shouldn't cite: because it's just a copy of wikipedia.

Also, looks really sketchy to me. Not only is it no longer nor, but now, but it doesn't have instructions on measuring digit ratio.

Typos in digit ratio question: "Use a ruler of calipers", "The result is 2D:4D" (should that be 2D/4D)? This question could probably use an image.

1a) If you're planning on comparing IQ scores to SAT scores, you should include age at which the respondent took the SAT. I know of at least one international gifted program, run by Johns Hopkins University, which administers the SAT to 12-year-olds.

1b) Scores on the different sections of the SAT might be interesting. E.g. SAT Score out of 1600; SAT-Verbal score out of 800; SAT-Math score out of 800 and SAT Score out of 2400; SAT-Reading score out of 800; SAT-Writing score out of 800; SAT-Math score out of 800.

2) A mere typo: "Do you attend Less Wrong... (read more)

First of all, the survey is excellent as it is. So I, too, merely suggest an addition.

I'd like to see added something about teaching / spreading rationality / proselytizing. My first idea would be something like: "Do you attempt to make others think more rationally?" with options on a spectrum from "no" and "I generally try to be the voice of reason in discussions" to something like "Many know me as an advocate for rationality". Not sure about how to adress an online/offline divide.

"None of your family / friends / acquaintances will suffer from any form of cancer anymore, nor will any of their descendants! How many years of your remaining lifespan would you be willing to give up in exchange for the whole of humanity enjoying the same protection, given your sacrifice would remain anonymous? (If you provide a large number, no life extension would be entailed.)" -- (Requires numerical answer: number of years).

A question on a similar vein I once thought up but forgot to ask on LW: Omega offers to give you $X, but he will kill you one year from now. What's the minimum value of X for which you'd accept? What if it's one month? Ten years? Thirty years? (No, he isn't going to tell you anything about whether cryonics will work. Also, he promises not to tell anyone about your deal (when you die it'll look like you had a heart attack or something) unless you try to buy life insurance.)

How about:

What country is the most successful at providing a good quality of life for its citizens?


What historical society emphasized the moral virtues with which you most strongly agree?

Should I take into account the fact that certain countries make tons of money exporting oil? (Anyway, it's very hard to answer such a question without doing lots and lots of research, even in you only restrict it to countries with say > 40M inhabitants.)

Typo in HBD question: "are in fact scientiically justified"


Extra Credit: Politics, Question about the Great Stagnation: the link is too large Screenshot

To the Super Bonus Questions, could you add:

Is there a number which you have considered your "favorite number" for some period of time (i.e., from before you read this question)? If so, what is it?

Maybe you could split “Yes” in the Blood question into “Yes, in the last 12 months” and “Yes, longer ago”.

Before the professional intelligence measurement question, please add something like:

Please ESTIMATE your IQ, using a standard average of 100 and stdev of 15.

It might make sense to add on something like:

You may want to check a normal distribution calculator (such as ) to ensure you aren't overestimating.

Or, alternatively, we could get at the same sort of question by asking this:

Please estimate how many people would have to be in a group of adults, chosen uniformly at random, for you to expect the

... (read more)
That depends a lot on the group of adults you use as reference. World population? The population of the country in which you are living? City? Your facebook friends? I think IQ is much better because there are objective standards. "Smart" is also a word that different people interpret slightly differently. IQ is much more precise.
Ideally, ethnic Brits, since I believe that's what's typically done for IQ reference distributions. Yes, but people may be better at imagining groups than normal distributions. Just how much unlikelier is 150 than 145?
That would also be interesting to test. Ask both for the IQ and ask for the relative intelligence. Do well calibrated people perform better on this task?
I like the group-of-adults formulation, it's clever and I think it's more amenable to getting at people's real estimates.
I expect people to have a better idea of group sizes than normal distributions, but the worry is with 'chosen uniformly at random,' since people tend to be in IQ bubbles. If you can think of 100 people you know dumber than you, and 10 people you know smarter than you, then you might say "well, that implies a group size of about 10," without realizing that half of those people you know from graduate school where the average IQ is in the top percentile. But it seems like we can compare that to SAT scores and measured IQ (if provided) and that'll give us useful information. I've also edited it so it's a little less ambiguous (were you randomly selected before or not?) but I want to add the super formal mathematical version to make it totally unambiguous, which may or may not be a good idea.
Yeah, I'd expect some pretty serious regression to the mean, or more accurately to slightly above the mean, going on with the group-of-adults method. People of low intelligence hang out with other people of low intelligence and probably have a slightly inflated opinion of their relative capabilities; likewise for people of high intelligence. But people outside your social circle are totally invisible. At least, assuming people actually use the group-of-adults method, which I don't really expect them to.
That is, you would expect people to say "well, I want my IQ to be 145, so let's calculate the percentile for that, and then use that percentile to calculate the group size"? Or that they'd just leave the question blank?
I'd expect a lot of people to leave the question blank if there are other IQ questions. I'd also expect a lot of people to work backwards from scores, more or less formally -- I don't think I'd expect many people to actually do the math or track down a normal distribution calculator, but thinking like "well, I got 99th percentile on that standardized test in high school, so I'll say 100" would probably be common. To be fair, that's probably more accurate than what you'd get by counting up the number of people you know who're smarter than you.
Meh. If I was asked that question, I'd try to guesstimate how many people smarter than me there are in my town and divide by its population.

Some questions I'd personally put in if I was writing this survey (you don't need to put these in if you don't want to, just suggestions):

  • To what extent have your beliefs about the world and your reasoning processes changed since you became a member of this community?
  • To what extent do you think that MIRI's mission is a worthy one (i.e. do you think that AI is a major existential risk, do you think it's important to ensure the development of friendly AI, do you think that MIRI is capable of advancing that goal, etc.)
  • To what extend do you identify with
... (read more)

Why not use the Kinsey sex scale for sexual orientation? (really depends on what you want to gather out of this information)

A split question of whether your children are biologically related to you or just children you are raising.

Allow multiple options (checkboxes) for work-status, and profession.

I have seen "education level" as highest level completed; but it doesn't take into account highest level incomplete. so where I might be part of a bachelors, I could not say so because I have not yet completed it. Perhaps a level of "highest edu... (read more)

Good idea. There are even more detailed scales, such as the Klein grid.
If you are currently working towards a bachelor's, you'll have answered “Student” on the Working Status question.
That makes the data harder to break down; but also someone could be a student + working. As I suggested checkboxes for that question. Also what if someone got partway through a bachelors then got a better offer and is now working. the highest level of education /completed/ would not reflect their knowledge/skill.

On "Profession," the field label "Art" is vague. Better would be "Arts and humanities."

"Humanities" includes philosophy, language, and religion; and sometimes history and law, too! I think what is meant in the survey is specifically the creation of art, e.g. design, sculpture, music, theater, fiction, and so on. It should be clarified, though.
Yes, it should be clarified. The main ambiguity that I was reacting to is that "art" can mean specifically visual arts or it can mean "the arts," extending to performing and literary arts. As it is, I'm not sure if my profession (scholarship concerning music) is "art" or "other." In fact (now addressing Yvain again), why is this category called "Profession" instead of "field"? It creates some odd overlap with the previous category of "Work status" which produces a little bit of confusion per my original suggestion and fubarobfusco's reply.


This question bugged me from last year's survey, but I'm not sure if my objections are valid. Consider a subset of this question: "What is the probability that Bayesian mathematical reasoning was created by a supranatural entity?" It seems... outside the domain, a little bit? Like a question, "What is the probability that Bayesian mathematics is true?"- what would it mean for this to be false? Which is to say, I don't know whether probability estimates are capable of a certain degree of recursive self-analysis, and asking about the origins of Bayesianism may be in that class.

any question you want in the Super Extra Bonus Questions section

Do you like the cartoon, "My Little Pony: Friendship is Magic"?

  • Yes, and I consider myself a brony/pegasister
  • Yes
  • No
  • No opinion / unsure

(And to compare that with a fandom that seems to be less represented amongst LWers:)

Are you a furry fan? (Reference link: )

  • Yes
  • No
  • No opinion / unsure
It doesn't seem useful to compare MLP to furries without knowing the prior probability that a particular member of the population is a member of either fandom, and to get that prior probability, we need to have reasonably accurate data about the sizes of those communities.
The most prominent social media site for bronies is, arguably,; its counterpart for furries is likely (And both sites are similar in that they have a focus on the creation of creative content, while also allowing for journals, commentary, and similar interactions.) To the best of my knowledge, FimFiction currently has about 160,000 user accounts; FurAffinity has around 750,000. If the results of the two questions turn out to be that there are significantly more than 1/4 the number of bronies as furries, then it's at least arguable that bronies are more heavily concentrated in the LW community than would be expected if there was no correlation between rationality and ponydom. It may not be /strong/ evidence, possibly even only on the order of a couple of decibans, but it seems most likely to be above zero decibans.
The number of furries seems dwarfed by the number of bronies; most estimates seem to suggest at least ten times as many bronies as furries. Compare attendance at the largest furcons and the largest bronycons (it doesn't match the 10x number for a number of reasons, but it suggests the direction).
Okay, we have the largest creative sites being tilted around 4:1 in favor of furries, and the largest cons' latest attendances being 2:1 in favor of bronies; does LW's furry-to-brony ratio fall within that range, or outside it?
So, the ratio might have problems when people fall in both groups; looking at the Brony Herd Census it looks like 17% of respondents were furries, which makes the 10 to 1 estimate calculated other ways seem less likely. (I think most of the estimates I saw for furries predated MLP growing, and it seems likely that the communities grow off each other.)

"How frequently do you meditate?"

"How many books did you read last year?"

More graphs in the reported analysis? I like pictures over numbers when I can get them.

I request that you add an intensity of importance question for each item on the politics section. I might think that abortion is terrible but not at all politically important, for example, and both of those seem like worthwhile information.

The survey is long enough.that I'm unlikely to put genuine thought into each of the questions. If none of the answers us closer enough, I'm likely to skip the question. I'm likely to pick numbers it off thin air on any questions asking for a probability. I'm unlikely to actually dig up my SAT scores etc and will guess from memory. I will probably skip half of the bonus questions that Seth uninteresting because there are so many of them.

There are probably other people who are willing to run through the survey to help out but will spend a fixed and not too large amount of time on it. This is likely to introduce some bias and inaccuracy of answers given its current size. Just noting this as a set of trade offs to keep in mind.

I was wondering (though this might be missing the point) about some options if you don't want to answer a question. The options don't cover what I believe is true. I don't know. it's more work than I want to put in to figure out an answer. I'm getting tired of the survey.
Good suggestion - having these options world be nice, but they would visually bloat the survey if added to every question, making it harder to scroll down. Not sure that would be a huge problem, though, given the length already.

If it's not yet too late, I should like to see:

How often do you use a physical library? Often Rarely Never Online Resources Only

I approve! I think you should put in High school tests in regards to other countries (maybe at least the UK and Australia). It would also be helpful to track GRE scores and similar graduated standardized testing. I think it would be also interesting to see what people we have on the Myers-Briggs testing as well as an optional question. I think that would be interesting to look at on the data side.

In the Religious Denomination question, "If atheist, please skip this question." should be replaced with "If non-religious, please skip this question."

Maybe even remove this sentence and add "non-religious" or "not applicable" as an option.

I think this is a very important change. I think there are many people who identify as Jewish Atheist, Buddhist Atheist, or Unitarian Universalist Atheist, (and maybe some others) and right now you are leaving it up to them to choose how to interpret the question. No information is lost by implementing this change, as there was already a question about theism.

Am I doing the digit thingy right? I scanned my hands and used MS Paint to make a line through each finger, then divided the number of pixels in each line.


Mental Health:

  • The questions are currently mixing point and lifetime prevalence: "have you ever been diagnosed" with "I believe I have it".

  • Can we give social anxiety its own question?

A random comment.

This is the first time I've seen "anti-agathics". Based on what I know of biblical Greek, I read this as something that would be like "anti-good". If I had been in charge of making up an anti-aging drug, I would have called it something like anti-presbycs (maybe that wasn't chosen because it looks too much like "presbyterian"? Presbyterian does derive from the world meaning "elder"...).

This isn't a request to change the wording if that's what people who will be taking the survey are familiar with BTW... (read more)


(double-post, sorry)

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

(double-post, sorry)

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

Choice suggestions:

Some ideal world would use the empirically valid Eysenck model of left vs. right and authoritarian vs. libertarian for the political section. Oh wait, you basically did. Good job.

Complex Affiliation choice - Inscrutably Idiosyncratic

Autism Spectrum choice - I don't know but my Autism Spectrum Quotient was [blank]

Question suggestion 1:

Only answer this if you think the chance of a singularity before 2100 is over 1%. Do you feel that influencing the singularity's outcome is tractable enough to be worth your time or money?

Question sugg... (read more)

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

I would like that the question for previous surveys asks for the private key that was used in the last survey if the person remembers.

Bring back the Unreasonably Long Questions section (the one with the personality tests etc.), and maybe add the BSRI to it.

Maybe bring back mother's age and father's age in the Super Bonus Questions.

What about a question about PUAs in Extra Credit: Politics?

How would you word that question?
By analogy to the other questions, it should be something like "How would you describe your opinion of pick-up artistry, as you understand the term?" and either a link to an external page describing PUA or a short inline explanation as with the HBD question. You could try to tease apart descriptive and normative claims, but note that no such distinction is attempted with e.g. the feminism question.
I don't think that's good. PUA isn't simply a political view. It's a set of behaviors and if you want to know whether someone engages in them, the label isn't the core issue.

I oppose the Birth month question. It costs a lot of anonymity but provides no good return.

6Scott Alexander
See for example
In that case it needs to clarify between northern and southern hemisphere (although I guess you could do that via the country question with some effort).

Is the "Birth Month" bonus question just to sort people arbitrarily into groups to do statistics, or to find correlations between birth month and other traits? If the latter, since the causal mechanism is almost certainly seasonal weather, the question should ask directly for seasonal weather at birth to avoid South Hemisphere confounders.

I have no idea what the seasonal weather was like when I was born - but I know which hemisphere it happened in.
Not the only possible one. If Alice was born in January and Bob was born in December, she will be 11 months older than him when they start going to school (and their classmates will be in average 5.5 months younger than her and 5.5 months older than him), which I hear can make a difference. The survey already asks for country. Sure, some people will have been born and grown up in a country in a hemisphere other than that they “most identify with” today, but they'll probably be a small enough minority that they wouldn't screw up the statistics too much.
I think this way of sorting classes by calendar year of birth might also be six months shifted in different hemispheres (or perhaps vary with country in more capricious ways). IIRC, in Argentina my classes had people born from one July to the following June, not from one January to the following December.
In Massachusetts, USA, the classes of people were from one September to the following August, being based on the day classes started.

I would really like to see these questions in the survey:

For the questions:

  • Give the time from your birth to your death in subjective years (so years where you are cryonically preserved don't count)
  • Give the estimate where you are 50% sure it would be less than your answer and 50% sure it would be more than your answer)

The questions are:

  • 1) How long do you think you will live?
  • 2) If your only way to die was by really wanting to die, when do you think you would die...
  • a) if you could control your aging process but the world would be otherwise unchanged?
... (read more)
I think most people's desired life span has a lot to do with how healthy they expect to be.
Good point, I edited the post to make that clear.
I understand that it's part of the framing of the question, but I still think that a lot of people would take issue with this part.
Is it because a lot of people think that continuing to live as a clone or a simulation is just as good as continuing to live as the original? If so, then I don't mind rephrasing what I mean by death. The important point is that I don't mean the death of the body, but rather the death of the mind.

Don't you want devoted followers?

Who leave their families for you

Give their money to you

Give their bodies to you

Give up their lives for you

Consider you God, and will kill for you

Don't you want to become a cult leader?

Since the death of God there has been a vacancy open

You can fill that void, here is how

  • I'd like to split the Gender:Other option into "I identify as a gender not listed here" and "I don't identify as any particular gender", as I'm curious how many other people have the latter view.
  • "roommates" is a bit of an americanism; I might phrase that as "with other people". I'm kind of curious about the split between "with friends" and "with people I didn't otherwise know", but I can't think of anything terribly useful to do with that.
  • I'd like a referral option for "Referred by other r
... (read more)

Question about politics is lacking a "none of the above" option.

That's how it must be. The question is about which you most identify with. "None of the above" would be confusing in the context of that question.

I disagree even a question of what you "most identify with" carries a connotation of a non trivial identification with what you would answer. At least it does to me.