It's that time of year again.

If you are reading this post and self-identify as a LWer, then you are the target population for the Less Wrong Census/Survey. Please take it. Doesn't matter if you don't post much. Doesn't matter if you're a lurker. Take the survey.

This year's census contains a "main survey" that should take about ten or fifteen minutes, as well as a bunch of "extra credit questions". You may do the extra credit questions if you want. You may skip all the extra credit questions if you want. They're pretty long and not all of them are very interesting. But it is very important that you not put off doing the survey or not do the survey at all because you're intimidated by the extra credit questions.

It also contains a chance at winning a MONETARY REWARD at the bottom. You do not need to fill in all the extra credit questions to get the MONETARY REWARD, just make an honest stab at as much of the survey as you can.

Please make things easier for my computer and by extension me by reading all the instructions and by answering any text questions in the simplest and most obvious possible way. For example, if it asks you "What language do you speak?" please answer "English" instead of "I speak English" or "It's English" or "English since I live in Canada" or "English (US)" or anything else. This will help me sort responses quickly and easily. Likewise, if a question asks for a number, please answer with a number such as "4", rather than "four".

The planned closing date for the survey is Friday, November 14. Instead of putting the survey off and then forgetting to do it, why not fill it out right now?

Okay! Enough preliminaries! Time to take the...




Thanks to everyone who suggested questions and ideas for the 2014 Less Wrong Census/Survey. I regret I was unable to take all of your suggestions into account, because of some limitations in Google Docs, concern about survey length, and contradictions/duplications among suggestions. The current survey is a mess and requires serious shortening and possibly a hard and fast rule that it will never get longer than it is right now.

By ancient tradition, if you take the survey you may comment saying you have done so here, and people will upvote you and you will get karma.

New to LessWrong?

New Comment
728 comments, sorted by Click to highlight new comments since: Today at 6:37 PM
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

Done. Too bad the basilisk question wasn't on it; I hope that will one day be possible.

There is no disagreement that only a small percentage of LWers believe in it (just as there's no disagreement that only a small % of scientologists are even aware of the more arcane aspects of their "religion"). But yeah if you had a survey the actual % may be worth listing on RW.
The coverage of the basilisk I've seen in the media does not include that, IIRC. Given the widespread mockery of those aspects & their failure to keep it under wraps, I'm not sure how ignorant the rank and file these days really are.
But then, media is not really focussing on the peripheral aspects of the movement. They're interested in what's being radiated from the core, that's why it is basilisk in the spotlight and not some random insane guy pondering the efficacy of shooting people rather than blowing them up, even though the latter is in principle more outrageous. There's just so many stupid details besides the big concepts like thetans...
The census provides a lot of ways to define core and most of them likely would give the answer that the majority of the LW core think differently than Eliezer about the basilisk.
Media is not writing about you. Or LW. It is writing about that FAI thing for which LW is just an online board that is in and of itself unremarkable.
The basilisk isn't a very central idea of that FAI thing. If you define remarkability the way Seth Godin does, LW might be remarkable in this case. I don't know whether FAI is central enough that a journalist get's tasked to do a story about FAI and then finds LW and writes a story about the basilisk. It might very well have been that the journalist heard about LW and then found that it makes a good publishable story. Lastly the article that could be said to be written about me because they write directly about my person are in the context of QS and not FAI.
What counts as a central idea? Does it have to be believed by a majority of the rank and file? Or is it sufficient that it is believed by the leader?
Central ideas are those that matter for the discourse about the idea. In academia that means ideas in academic papers. Ideas that are important enough that they get space in textbooks. Given low little academic papers MIRI publishes there might be central ideas that are unpublished and still important but I don't think the basilisk is among them by any reasonable count..
The most important tenet of the Catholic Church is probably something like "Jesus died for our sins". The tenets of the Catholic Church that critics pay the most attention to are their beliefs on abortion, contraceptives, homosexuality, etc. even though they are widely ignored by Catholics and certainly less central than "Jesus died for our sins". Why? Because even though the church would describe them as less important, they are the ones that get non-churchmembers the most worried. And they are consequences of the church's core beliefs, even if not (by church standards) the most important ones. And it's completely legitimate to criticize the church for its stand on abortion, and not criticize it for the matters that the church considers more important. The basilisk is every bit as central to criticsm of LW as abortion is to criticism of Catholicism, even if it's not central to the church and most LWers don't believe it. Most Catholics are fine with abortion too.
The Catholic Church spends a great deal of effort trying to influence the secular politics of abortion, with effects on the lives of non-Catholics. This is why non-Catholics criticize it. Less Wrong spends no time at all trying to influence the world at large with respect to Roko's basilisk. It is an amusing episode, but not likely to cause any problems for anybody.
Use the Galileo analogy then. The Pope's belief that Galileo, while right on the facts, shouldn't have challenged the church has pretty much no influence on anyone's lives, but still gets criticism. Also, LW has as its goal influencing the world about rationality and AI, and it seems that LW or at least Eliezer is unable to disentangle the Basilisk from tthe ideas it does want to spread. (Again, Eliezer doesn't believe in the Basilisk exactly as stated, but he does believe Basilisk-like ideas could be dangerous.)
If you ask a catholic priest whether the position of the church on abortion is important for him, I think he will say "yes". I you count official speeches and writing by popes I also think that abortion will come up from time to time. We are not talking about criticism of LW. LW as it stands is not important enough in society as a whole to warrant criticism from journalism. We are talking about centrality to FAI under the assumption that it's a topic that journalists want to write about. maybe in the background of Transcendence that raised the topic a bit in public awareness.
It's not an exact analogy, but it's close because it's much more important to outsiders than to insiders. The basilisk isn't directly a LW idea. but the basilisk follows from LW-style ideas and is close enough that Eliezer couldn't just say "nothing like the basilisk could possibly work". A closer analogy may be more like, oh, geocentricism. The church does not believe that geocentricism is true any more than LW believes in the basilisk. On the other hand, the man who became the current Pope has pretty much said that the church was right in its treatment of Galileo even if the church was wrong about geocentricism itself. And you still see this used to criticize the church. And I doubt that many priests would think that the way the church treated Galileo is very important compared to either abortion or Jesus dying for our sins.
Well, the point is, almost nobody cares about LW and where LW fits in. Few people have some ideas that are interesting due to the sheer ridiculousness, rest of the board is of no interest.
If you are dealing with media and the write partly about you, it's quite useful to understand what they do care about in more detail.
Define "believe in the basilisk". Even Eliezer doesn't believe in the basilisk exactly as stated. But he does believe that basilisk-like ideas could be dangerous for basilisk-like reasons.
There's no such thing as basilisk exactly as stated, because it has never been stated exactly.
It would have been a nice insurance agains possible future PR shitstorms. Was that your primary reason for suggesting it?
Geez, this might be the only issue so contentious that it can attract significant downvotes to a "did the survey" comment. Ironic that an alleged literal-mindkiller would become such a figurative-mindkiller.

Did the survey!

I took the survey. Out of curiosity (too late to change now) what should I have answered if I'm not my father's first child, but I'm the first child he had with my mom? (There are kids from my dad's first marriage, but I didn't grow up with them).

I went with "no older siblings" since I assumed this was a question about socialization (or maybe even about uterine environment) but not siring. But I'd like to know for next year.

This should be a warning to us all about how hard it is to frame a good queston.

I would also like to know for next year. I have four older siblings on my father's side, and two on my mother's, and only spent any home time with one (from my mother's side). So, I answered 6 for older, but depending on whether this was a socialization or uterine environment question, the best answer might have been either 1 or 2 for older.

I had the same situation. I was the oldest child my father had with my mom although I have siblings that are older that I didn't grow up with. I'm the only child of my step-father (they had no natural children) so I grew up as an old child and that is what I went with.

Taken! The way you were being so apologetic about the length, I thought it would be much more grueling - I found it quick and fun! :)

I completed the survey, huzzah!

Did the survey. Also, now I know my digit ratio!

Filled in, but did not do digit lengths because I have no access to a printer or scanner in the near future.

Completed the survey (arguably the first thing I've actually contributed to LW, though I've discussed it at some length offline; this is my first comment ever). I have some degree of access to a scanner but not conveniently (same goes for a ruler actually; at best I may have a measuring tape somewhere I could find in under an hour's search). I filled out all the rest, aside from the N/A questions. Some of my answers have very low confidence (calibration percentage?), though.

A tip for those who don't have the equipment to perform the actual test: if you can verify that the lengths of the fingers on your left and right hands are equal (align the crease in the skin at the bases of the same finger on each hand, palm-to-palm), you can use the same technique to compare the D2:D4 lengths (one hand against the other). My fingers are the same length regardless of which hand (to the limit of my ability to measure without mechanical aid), and my D2:D4 ratio is somewhere in the range 1.00 < D2:D4 < 1.05, probably under 1.02 but definitely in excess of 1.00. As a cisgendered male, I guess I'm weird?

Oh, and some feedback: Part Four's "Moral Views" section could have used links (L... (read more)

You should post this as a comment to the original post, not as as a reply to another comment! ;-)
Thanks! Yeah, I did add a top-level comment, with a link to this one, but I realize that was sort of the backward way to do it.


I'm a bit confused about the accuracy of my BSRI because my true answer was frequently 'only towards my SO', such that my score would be drastically different were I single.

I'm a bit confused about the accuracy of my BSRI because my true answer was frequently 'only towards my SO',

Same here. And in some cases it was ‘except towards my parents’ or ‘only when I'm very tired’. I still tried to take some kind of weighed average.

This is weird. I haven't noticed that until you've pointed it out, but I believe that my masculinity score was only a little lower than all the benchmarks and not extremely low only because I've considered how my partner would gauge BSRI questions. They seem to push me towards expressing masculine traits. Isn't it interesting that a sex-role inventory doesn't make allowances for situations priming different sex roles in people?
My true answers were also frequently "highly situation-dependent [in assorted ways]." I tried to give a weighted average too, but that weighted average would change significantly with the balance of situational contexts I experience.

Survey complete! I'd have answered the digit ratio question, but I don't have a ruler of all things at home. Ooh, now to go check my answers for the calibration questions.

Took the survey.

And yeah you should warn about the material needed for the digit ratio question in advance, so people don't start the survey if they aren't in the right conditions for it.

I'm done, but my ruler isn't good enough that I'm super confident in my digit ratios; I would have preferred one less significant digit (no pun intended, but I'll take it anyway).

Took the survey. I always feel like I did the last one only recently.

Done - and mildly disappointed that we won't be measuring the prevalence of transponyism this year.

Does this post appear on LW's Main or Discussion pages for anyone else? I only found it via an offsite reference. Edit: Nevermind, I had my Main set to 'Promoted' instead of 'New'.

Has that been actually suggested?
Yes, by the author of the grandparent.

Took the survey.

Taken the survey (would have loved to do digit ratio, but too difficult to get access to the equipment needed).

I did the survey. (Comments on specific aspects appear as replies.)

It's time to decouple sexual orientation from gender identity! If my gender is neither male nor female, but I'm primarily attracted to one of those, then I'm neither homosexual nor heterosexual (nor bisexual nor asexual). But neither am I some nebulous other; if only I had a binary gender identity, then suddenly I would have a binary sexual orientation too! Of course, some people identify specifically as homosexual or heterosexual (and some people even have prima-facie contradictory identifications such as both male and lesbian), and you could ask about that if you like, but you should also ask the more fundamental question of which genders one is attracted to.

... and that doesn't even get into the sexual-vs.-romantic issue. My girlfriend is cis and bisexual, but only andro-romantic (hetero). She identifies as bi, for purposes of broad categorization such as surveys like this, but has no interest in dating other women even though she is sexually attracted to them. In other words, yes, the better way to ask such a question would be something along the lines of "which gender(s) are you romantically attracted to?" and "which gender(s) are you sexually attracted to?" as different questions.
This strikes me as suspiciously like "she'd straight but identifies as bi because it's fashionable".

Out of curiosity, if I'd avoided mentioning how she self-identifies and had instead told you that "she has had sex with other women before and has asked me if it's OK if she sleeps with other women while we're dating (or brings them home for a threesome)... but has never shown or claimed any interest in actually dating another woman" (all of which is, incidentally, true), what would your response have been? Framed that way, one could assume that she's actually bi or even lesbian and the only reason she's dating me instead of one of those girls is because she wants to avoid the social or family stigma of homosexuality.

Or you could take me at my word. It's not like you're in any position to verify one way or the other, where she in particular is concerned, unless you're one of the handful of people who actually know who I am speaking of and know her preferences at least as well as I do.

It also doesn't matter for the point I raised (about how some people have different targets for sexual and romantic attraction) unless you intended to imply that not only is she personally actually neatly classifiable under the existing system but so is everybody else who would claim otherwis... (read more)

What would you expect it to look like if in fact she found both men and woman sexually attractive but only men romantically attractive, as she claims?
See also the OKCupid Trends post about The Big Lies People Tell In Online Dating.
That's a valid point. On the other hand, as a dating site, OKC messaging is probably going to be skewed towards the gender that one is interested in pursuing a relationship with (though maybe that's just the way I use it; as soon as I typed it I felt sure there were plenty of people just looking to hook up). When the topic is sexual orientation vs. romantic orientation, I'm not sure that OKC is the best source of data. I can't deny the specific claim that a large proportion of ostensibly bi people appear to not be both bisexual and biromantic.
The questions Family Religion and Religious Background seem to parallel the questions Religious Views and Religious Denomination, but they are phrased differently. The first is my family when I was growing up, while the second is simply my family. So as it happened, I was not thinking of the same families when answering them! Perhaps I should have paid more attention the name of the question Religious Background, which I really only noticed just now when I wanted to identify it for this comment. You did not in fact get information about my religious background in my answer to that question; you got information about the religious background of my spouse of less than 2 years (and my stepchild).
So I filled out the whole survey, and then I got to the part about the digit ratio, and I thought, OK, I'll do this! But I can't do it now (no photocopier at home, can't trust a measurement to 3 digits if I'm not doing it the same way as others). And I can't keep my answers up until I can do it (no battery in computer, must be turned off to transport, Lazarus plug-in has been problematic). So I put in a public and private key but no data. I will gladly supply the data to you tomorrow, using those keys to identify my survey.
Some countries hold elections but not major national ones; and sometimes a country has elections, but most people in them still can't vote. (Examples are Saudi Arabia and Kuwait, respectively.)
I'd be much more comfortable answering the probability sections if I knew what epsilon is. I usually say 0% when the value is less than 0.5% and 100% when the value is greater than 99.5%, rounding to the nearest whole percentage, on the grounds that the whole point of using percentages is to avoid explicit fractions (common or decimal). But then you ruin this by explicitly mentioning 0.5% and 99.99% as possible answers. If you had put a hard limit on the number of digits allowed, then I could have used that. In the end, since I saw no consistent guidance, I fell back on my usual practice. The result is that I had a lot of 0s and 100s; hopefully that won't mess up your algorithms. ETA: It is probably relevant here that I am a naturally lazy person.
I think it might have been better to ask people to estimate what are the odds that a given statement is true. If a probability of a statement is close to zero or close to one, it gives us better precision without having to worry about digits after the decimal point (however, if a probability is close to one half, it is probably better to ask for a probability). Although it is easy to convert odds to probabilities, how many people in this survey actually took the mental effort to remind themselves to calculate the odds first and only then to express them as probabilities? I might be wrong, but I guess that only a minority. An idea for the next year survey - it might be interesting to compare the answers of two groups, one of which would be asked to estimate probabilities, the other one to estimate the odds.
Are you using "odds" to refer to percentages and "probabilities" to refer to fractions? I don't think there is actually any difference in meaning between the two terms.
Colloquial language doesn't make this distinction, but by technical convention, they are different. Specifically, ‘odds’ refers to expressions like ‘5 to 3 against’; numerically, that's the fraction 5/3, or rather (because of the ‘against’) its reciprocal, 3/5. Thus odds run from 0 (impossible) to infinity (certain), with odds of 1 being perfectly balanced between Yes and No. In contrast, probabilities run only from 0 to 1. An event with odds of 5 to 3 against, or equivalently odds of 3/5, has a probability of 3/(3+5) = 3/8. So the numbers are different. The conversion formulas are O = P/(1 − P) and P = O/(1 + O). Then there are log-odds; this is log₂ O bits. (You can also use other bases than 2 and correspondingly other units than bits.) Now 0 indicates perfect balance between Yes and No; a positive number means more likely Yes than No, and a negative number means less likely Yes than No. Log-odds run from negative infinity (impossible) to infinity (certain).
Oh right, I forgot about that definition. The main probability conversions that I was aware of involved converting between fractions and percentages, sometimes expressed instead as probabilities between 0 and 1. Theoretically, it makes sense that odds can also be converted to or from probabilities, now that I think about it. Thanks for your explanation.
Yes, odds are good (and log-odds are even better), but people are bad at both dealing with very large absolute values and dealing with very fine precisions. I think that the survey is correct to put in a cut-off (whether an ϵ for probabilities, an N for log-odds, or one of each for odds); it should just tell us where. (Edit: put in stuff about log-odds properly.)
Epsilon is a minuscule amount. It's vanishingly small, but it's still there.
Yes, but which minuscule amount? To be more specific: If ϵ ≥ 5 × 10⁻ⁿ (which it must be for some n, if it is a positive real number), then I only need to figure out my probability to n + 1 digits. Upon doing so, if it's all 0s, then my probability is no more than ϵ, so I can enter 0. Otherwise, I should enter something larger. (And a similar thing holds on the other end.) Specifying ϵ serves the practical purpose of telling us how much work to put into estimating our probabilities. Since I had no guideline for that, I chose to default to ϵ = 1/2 (in percentage points), rather than try to additionally work out how small ϵ was supposed to be. If, instead of bringing up ϵ, the survey had instructed us to use as many decimals as we need to avoid ever answering either 0 or 100, then I probably would have done more work. (There are reasons why this is bad, since the results will be increasingly unreliable, but still it could have said that.) But since I knew that at some point my work would be ignored, I didn't do any. (Edits: minor grammar and precise phrasing of inequalities.)
I took epsilon to be simply 0.5, on the basis of "the survey can take decimals but I'm going to use whole numbers as suggested, so 0 means I rounded down anything less than 0.5". This is imprecise but gives me greater confidence in my answers, and (as you say), I have some tendency towards laziness.
Yes, that's what I did too (0.5%).
I don't think it will mess up the algorithms. My guess is that most people probably rounded most calibration answers to the tens place due to lack of enough confidence to be more precise, but since people are giving different values, the average across all respondents is unlikely to fall on an increment of ten, and should be a reasonably accurate measure of the respondents' collective assigned probability for a question.
It could mess them up, because in theory a single wrong answer with 100% confidence renders the entire series infinitely poorly calibrated. The survey says that this won't be done, that 100% will be treated as something slightly less than that. But how much less could depend on assumptions that the survey-makers made about how often people would answer this way, and maybe I did it too much. I doubt it, since I'm pretty sure that they know enough about these pitfalls to avoid them. But I felt that I answered 0 and 100 quite a lot, so I thought that some warning was in order.
Even though percentages are typically used for cases where precision is less important, I'd say that in this context it would be better to err on the side of precision.
I don't fit in well with any of the 5 answers to the Political question, and there was no Other, but skipping it also didn't seem right. (Several questions have explicit cases when they are to be skipped, but this was not one of them.) I eventually picked 1 of the 2 that seemed less wrong than the other 3; I would have preferred to pick some sort of non-moderated mixture of those 2. (Actually, that is how I usually describe my politics when asked for a response in the form of a political party: somewhere between the ___ Party and the ___ Party, only more extreme.) The Complex Affiliation was not a problem. (Actually, I was still torn between 2 answers, but this time I would have been happy with either of them!)
My public key is the same as my user name. Should it have been anonymous? (My private key was randomized and only identifies me if you know what format I use for general-purpose random strings.)
Assuming Yvain does the same thing as last year, both the public and private key will be released as part of the survey dataset if you checked the 'release my survey data' box.
Faith in Humanity moment: LW will not submit garbage poll responses using other LW-users as public keys.
If that's true I wish I'd known it before choosing keys.
The private key too!? Fortunately I used a one-time key for that. The public key is OK. I made sure that I was comfortable with people linking my answers to me before I used it. But then I thought that maybe I wasn't supposed to.
I hope that you'll publish the answers to the calibration questions, after the survey closes, of course.

I finished the survey.

Taken! Thanks as always for running it

Except for the digit lengths, survey taken!

I took it. If it's anything like last year, officially 2/5 of my karma will be from surveys.


Took the survey. My first one. Thanks for putting it together Yvain/Dan.

I took the survey. Started on the BSRI but abandoned it because I found the process of giving vague answers to vague questions distressing.

I'm missing something here, I filled in the public and private and keys, but saw no game theory problem. Are we being given equal chances of the monetary reward?

Anyway, fun survey.

Presumably. The idea is to incentivize participation in the more difficult digit-ratio section. (Although, of course, that does create a game-theory problem...)


Didn't have a scanner, so I traced my hand on a piece of paper with a pencil and measured that. Not sure I got enough accuracy to take seriously. Oh, well.

Given the ambiguity of the directions, you're probably as close as anyone else.
I'm confident you didn't.

About two hours ago, I submitted an incomplete census return -- it looks as if some keystroke produces an immediate submission, at least on my browser. I'll be submitting a complete one later today. Yvain, if you want to suppress the incomplete one and need help in identifying it then I can help. I was partway through the calibration questions when I accidentally submitted.

(I see TrE had the same problem.)

[EDITED to add: Complete return now submitted.]

Did it! I'm shocked that my digit ratio is so high. Like, I figured that it was pretty high, being a bisexual genderfluid "man" (assigned at birth, that is), but I didn't expect it to be greater than 1. Also, it was much shorter than I expected.

Taken. Wasn't bothered by the length -- could be even longer next time.

I exist in a quantifiable way! (I took the survey)


I think it is somewhat unrealistic to expect individual digit ratios to be accurate to three significant figures (although I understand that two significant figures might be too crude a measure to show effects of smaller size). One can hope that the errors are symmetric and it doesn't matter.

I don't think it's going to matter very much. 3 digits after the dot, with the understanding that the third digit is probably not very good, but the second probably is pretty good.

Suppose the actual length of a person's index finger is 80.5 mm and the actual length of his/her ring finger is 83.5 mm. Then the 2D:4D ratio is 0.964. A measurement error of 0.5 mm is very easy to make, e.g. due to inaccuracy of a photocopier, inaccuracy of a ruler, inexactness of where a finger joins the hand (and even if it wasn't a vague concept it would still be a problem to pinpoint the precise location of it with a great accuracy) and even differences in muscle tension in fingers at the particular moment of placing a hand in a photocopier. If a person measures his/her index finger as being 80 mm long (0.5 mm shorter) and her/his ring finger as being 84 mm long (0.5 mm longer), then they would obtain 2D:4D ratio of 0.952. Whereas if the length of the index finger is measured to be 81 mm, and the length of the ring finger is 83 mm, then 2D:4D ratio is 0.976. Therefore, the first digit after the decimal point does not vary that much (in the vast majority of all cases it is 9), the third one is basically noise, and even the second one is not that reliable (in an individual case). However, that might still be enough to notice some interesting correlations and if the errors are symmetric it might not even matter that much when all data will be aggregated.

For me its not even the second digit. Even left and right hand differ significantly. Copire doesn't make things really better (OK, the copier quality was low, much too dark).
Agreed. Most rulers don't give measurements more precise than millimeters.
This is why a scanner might make sense. Even 300 DPI is less than 0.1mm resolution, so just scan it in and measure with an electronic ruler in your image-editing software of choice.
There's no point in measuring something more precisely than the thing itself exists. Which pixel "is" the base of the finger? Has anyone tried repeating the measurement, following the same procedure each time? Do this on occasions at least a day apart, to avoid unconsciously imitating the second time exactly what you did the first time. I have. Reproducibility was no better than 1mm, and pretty much independent between fingers. You could also try under different conditions, e.g. when your hands are cold and when they're warm, when you've just been exerting them heavily and when you haven't, etc. I doubt that digit ratio exists as an entity to better than 1 or 2% accuracy.
I suspect it does, but you need to be a lot more precise with your instructions than Yvain was in this case.

Completed. Very excited to see the digit ratio data.

I did the survey in all its parts, and upvoted every top level comment to promote LW's census partecipation.

It was fun and not particularly long, although I miss the 'global prisoner dilemma' of the last survey.

I completed the survey (and learned surprising things about my digit ratio)


I took the survey! This is my third survey.

Answered. WRT Type of Global Catastrophic Risk, I answered conditioned on greater than 90% of humanity being wiped out before 2100, which I assume is what you meant. If it wasn't, well, I ruined everything, then.

I wondered about that too, but for me "wiping out civilization" includes the possibility that some disaster leaves half of humanity alive, but smashes all our tech, knocking us back to the stone age. Intelligence forbid!

Answered all I could except the digit one because of no access to scanner. Looking forward to the results!

In-group fuzzes acquired, for science!


I took it. A bit sad that it's shorter than the last one.

I took it. A bit happy that it's shorter than the last one. Last time I didn't find time to do any of the optional questions, but this time I did all but a couple.

Done. Fairly high confidence that I'm still the lone Filipino LessWronger.

Survey done, awesome as usual, Yvain. Can't wait for the results.

Survey finished- erred on the side of not screwing up Yvain's numbers where possible, but I'm curious what the ideal way to mark down Religious Background for results of families that divorced over religious disagreement is. Also had a really strong desire (thwarted, but present) to put a SQL injection into the question about whether the universe is a simulation, which is a bad idea no matter what the answer turns out to be or whether I could conceivably affect the simulation. It's like a pascal's wager mixed with a Russian roulette, only the gun is fully loaded. Either I screw up the numbers, I tank the survey, or I crash the simulation. Dear brain, we were reading about akrasia just recently, were you paying attention?

Why would the universe be particularly likely to run an SQL statement in a form question about whether the universe is a simulation? All you have to do is think the attack and


Somehow this made me think of Larry Niven's "Rotating Cylinders and the Possibility of Global Causality Violation".

Just completed my first survey!

Did the survey!

Took it.

EDIT: I was surprised to find the BEM test in it. I took it some time ago and it resulted in 65-70% F and 50-60% M (as far as I can see largely because of my strong and caring relationship to my children).

I didn't determine my digit-ratio during the test but did right now. I arrive at totally different values (between 0.91 and 1.05) depending and hand and exact points and the copier print reading gives still different values. My best guess is that it is somewhere around 0.96.

I think you're supposed to measure from the middle of the bottom crease to the middle of the tip. Also, since the bottom crease itself can be about a millimeter or two wide, I measured from the middle of that crease by its width in addition to its length. When I do that I get consistent results even on repeated measurements.
Sure. If I pick the same spot I get the same results. Esp. with a photocopy. But at least the significant difference between left and right hand remains. Even when photocopying it makes a difference how strong you press your hand against the plate and how much contrast the scan has (for me it was too dark to make out the creases clearly).
It is common for them to be different; that's why the survey asks for each hand separately. Inter-rater reliability (given the same scan) for this measurement is in the r=.9 range, if I remember correctly, so don't feel that bad about it being variable; the underlying quantity is actually difficult to measure (but meaningful nonetheless).
I don't doubt that the measure is meaningful. The influence is surely real. The question is rather whether anything meaningful can be derived on the individual level.
Maybe the corresponding fingers on your other hand really are different in length. Mine are. Whenever I press my fingers against each other such as to line up their bottom creases (keeping the orientation of the fingers as straight as possible), the middle and upper creases and fingertips don't line up. My right fingers are slightly shorter. Good point about the photocopier. Hopefully these issues won't add too much noise to the results and obscure any significant results.

Survey taken!



I did the survey.

I hope you don't count fanfiction as "books", because otherwise my response is off by at least two bullet points.

I took the survey.

Survey surveilled!

Nope. You've been surveilled, by the survey.

I think you've been surveyed, rather. (Although undoubtedly surveilled as well, given the current political climate...)
Well, in all fairness, Rubix presumably did also survey the survey. And hopefully perused it as well, maybe even filled it out!

Tooken. My scanner was being evil today so I only had low-res overview scans, and could only get to within a tenth of a centimeter, but I think my results are dramatic enough that it's not wildly incorrect to use my guess? Drop me if I'm wrong, I should be easy to pick out of the crowd via karma.

Done! Wish I had had a scanner handy going in, I'm curious about the digit ratio.

I think it should be fine to just hold a ruler up to your finger. The only potential problem might be that the highest tip of your finger wouldn't actually touch the ruler, but if you don't want to estimate by sight you can hold another flat surface perpendicular to it to see where that touches the ruler. I get consistent measurements this way.


Can anyone explain the Bem Sex Roles thing and why its relevant? I scored slightly more masculine and less feminine then average which confused me slightly. Its all self reporting though so I'm not sure how much it will express m nature vs what I value (like to think about myself)

Done, though sadly without the digit ratio due to lack of equipment. I'm a newbie and I just thought that was really cool.

I did the survey! This is the second time I've completed an iteration of this survey, but this year was the first time I answered all the questions. I also did all the extra credit except for the digit ratio question.

Took the survey.

Finished the survey. Didn't answer the SSC question even though I read it regularly because I plan to take the edited version when it's posted there, and I also didn't answer the digit ratio question.

Did it, that was fun! Can't wait for the results.

Did the survey! I think i gave highly contradictory answers.

Survey done, including digit ratio. And I learned something new.

But not particularly confident in the accuracy of my measurement.

Took the survey, except for the digit ratio part.



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

The entire community is extremely insular and is weighed down with it's own established ideas. Most of the writers speak with total conviction, absolutely convinced of their own conclusions, despite the entire point of the endeavor being the pursuit of ever increasing amounts of correctness, thus making them 'less wrong'.

It consists mostly of extremely narrow demographics, cutting it's objectivity off at the knees by creating a culture that is perfect for serving as echo chambers despite their criticism of one another. It has also engaged in censorship of ideas, something that CANNOT be allowed in a group that is trying to further rational thought.

Aside from that there is also the personality cult surrounding Eliezer Yudkowsky. Objectivity is impossible if people weight the merit of your arguments by your popularity, which is inevitable in such a situation.


Took the survey. Skipped the digit ratio - I could have done it but didn't feel like walking to the copier or finding a ruler.

Next year I want to see an independent measure of conscientiousness, and compare this between people who bother to answer the digit ratio question and those who don't...

The conscientiousness/akrasia interactions are also fascinating, but even harder to measure. There's a serious missing-not-at-random censoring effect going on for people too conscientious to leave off digit ratio but too akrasic to do the measurement. I nearly fell into this bucket.

Survey done, except for the digits ratio question!

I would have given a response for digit ratio if I'd known about the steps to take the measurement before opening the survey, or if it were at the top of the survey, or if I could answer on a separate form after submitting the main survey. I didn't answer because I was afraid that if I took the time to do so, the survey form, or my https connection to it, or something else would time out, and I would lose all the answers I had entered.

It's a Google-forms survey. I'm pretty sure they don't do that. Can't blame you for being cautious, though.

Did the survey.


Done, except the digit ratio thing.

I filled in the survey! Like many people I didn't have a ruler to use for the digit ratio question.

I have taken the survey, and to signal my cooperation I have upvoted every existing top-level comment here. Do unto others...

Survey complete!

I'm kind of surprised at how much better I feel like I've gotten about reasoning about these really fuzzy estimates. One of my big goals last year was "get better at reasoning about really fuzzy things" and I feel like I've actually made big progress on that?

I'm really excited to see what the survey results look like this year. I'm hoping we've gotten better at overconfidence!

The gender default thing took me by surprise. I'm guessing that a lot of people answer yes to having a strong gender identity?

This has seen a lot of discussion over at Slate Star Codex. Judging from the anecdotes I've seen in the comments there, there doesn't seem to be an obviously dominant answer, although of course there are self-selection issues in that context; I'll be interested to see what the survey turns up.
I definitely don't have a strong identity in this sense; like, I suspect I'd be pretty okay if an alien teenager swooped by and pushed the "swap sex!" button on me, and the result was substantially functional and not horrible to the eye. Like, obviously I'd be upset about having been abused by an outside force, but I don't think the result itself is inherently distasteful or anything like that. I'm really curious to see how this and related stuff (male/female traits, fingers) relate.

Submitted. (Yvain, if you're reading this, you might want to see my note about an accidental incomplete submission.)

I am somewhat disappointed to be asked about favorability with a movement without allowing me to distinguish between the ideals of that movement and the movement as it exists (see: feminism and social justice, which, as phenomenon in reality appear to be ways to generate indignation on tumblr -- I love equality but do not use tumblr and I don't see any purpose in being indignant on the internet).

Also, as regards a "Great Stagnation": Strongly Doubt is not the opposite of Strongly Believe. So I have strong doubts where the balance of my estimation is that Cowen is incorrect -- my radio button does not exist, it is too far to one end of the spectrum, despite not being a hyper-radicalized opinion.

There's the movement as it exists, and there's one facet of the movement as it exists. For example (and not to push any particular point of view here, it's just an example), I'm involved in the feminist movement. But I spend no time on Tumblr. Sometimes I read things that reference Tumblr, and my impression is that to get involved on Tumblr would be a colossal waste of time, so I don't do it. (Once in a while somebody links to something on Tumblr, basically saying "Look at this thing that I saw on Tumblr.", and I look at that one thing, but I never feel the urge to do more.) I also make it a point not to get indignant on the Internet, even when discussing feminism. (Occasionally I get indignant in face-to-face contact, but I have time to edit myself on the Internet.) Most of the feminism that I do on the Internet is arranging face-to-face meetings of feminists, so there's not much to get indignant about. But occasionally I expand my focus to commenting on posts where a discussion, or even an argument, is taking place. The last time that I did that, one person private-messaged me to call me "diplomatic" and another person agreed that I was right after all; both of them had gotten indignant before this, but I hadn't. (To be honest, this foray was more successful than usual, but the usual is neutral, not disastrous.) So I do not use Tumblr, and I very rarely get indignant on the Internet, but here I am, in the feminist movement as it exists.
I think you should average over your meta-uncertainty and answer according to your overall probability.
You may have misunderstood me. I have high levels of doubt but some certainty. Let's say I'm 80% unsure but have information that leads me to be 20% sure (or, in other words, the probability I would assign to my analysis being correct is only a bit better than guessing). So I'd want something maybe 1/5th away from "Strongly Doubt". But I am not 1/5th closer to "Strongly Believe". I am 1/5 closer to "Strongly Disbelieve" or "Strongly Disagree", perhaps.
Sorry, I hadn't noticed that the leftmost option was labeled “Strongly doubt” -- I think I must have seen the “Strongly d” part plus the “Strongly believe” label on the rightmost option and my brain must've autocompleted the former to “Strongly disbelieve”. I would have picked the third radio button rather than the fourth if I had noticed that in time.

Done, except the digit ratio thing. I still picked a public key and a private key, so that if I get near a scanner or photocopier before November 14 I will submit an otherwise empty survey response with my digit ratios and the same public key and private key as today. Is that OK?

In Political, going only by the descriptions after the colons it looks like Liberal is halfway between Social democratic and Libertarian, and I picked it based on those, but... note that Moldbug also is socially permissive in most all the senses I care about (besides the post I linked, he also supports gay rights) and yet his position doesn't resemble that of the US Democratic Party or the UK Labour Party.

In Less Wrong Use, I rounded my top-level posts down to zero.

In Time on LW and Hours Online, thanks to LeechBlock, I didn't have to pull numbers out of my ass! Likewise for Meditate thanks to Beeminder. OTOH, I answered Books by counting the books I can remember reading and dividing by an anally extracted estimate of the fraction of books I read that I remember.

In the second part of the Calibration questions, does “correct” imply ‘correctly spelled’? My answers are P(correct and correctly spelled) + P(rec... (read more)

I answered that I'm cis by default, but I would freak out if I woke up in a woman's body.

I think it's totally reasonable to consider that freaky for reasons other than that you now have to live as a woman. I think the spirit of the question was more, "If you were a woman but had the same personality, would you be okay with that?"

This seems like a contradiction to me.
You can make your own soylent. I do so, and it's pretty tasty.
Wikipedia describes its origin. The items on the test are based on the opinions of 100 Stanford undergraduates in the 1970s about what traits of behaviour and personality in each sex are socially desirable, and the norming of the test was done with a total of about 1500 Stanford undergraduates. Here is a review article about the BSRI from 2001.
LOL. Oh, boy...
Cue the jokes about the 1970s … but the fact that they were all Stanford undergrads (very W.E.I.R.D.) is probably even more significant.
I see liberal vs. libertarian as a two dimensional thing as depicted here.
I'm familiar with a similar thing from Political Compass. Going from the descriptions after the colons only, Yvain divided the upper half plane into "Communist" and "Conservative" and the lower half plane into "Social democratic", "Liberal" and "Libertarian".

Done! The length is fine; the questions are interesting and fun to consider.

EDIT: removed concerns about "cryivf" if. "srzhe" nf ynetrfg obar (znff if. yratgu); gur cryivf nccneragyl vfa'g n "fvatyr obar".

You should probably Rot13 this. I scanned the comments before I did the survey, and I couldn't remember why I was so confident in the correct answer, but I was.

Survey completed. Account created to get starting karma and increase likelihood/amount of future participation.

I'd like to note that the current formulation of sex/gender/sexual orientation questions forced me to misrepresent myself because the technically correct answers seemed to cause an even greater misrepresentation. I would like extra options to the "sex assigned at birth" question, perhaps "male, now transitioned to female/other" and vice versa, to account for other-gendered transitioners; but I'll be the first to admit that this probably isn't a major issue.

I'm confused. If you were male at birth and transitioned to female, can't you just answer the "sex assigned at birth" question male, and the gender question with "transgender m -> f" ?

Well, that's how I answered, but "other" would have been a more honest description of my gender. The question asked: "With what gender do you primarily identify?" and I don't have a female identity, only what I can describe as a femininely androgynous body image (prompting transition treatments) and much heavier social dysphoria about being male'd than female'd, although the optimal no-mental-suffering-causing option would be to be recognized as non-binary. Answering "AMAB other homosexual" probably wouldn't have had a statistically relevant effect but the possibility of being interpreted (even though realistically nobody would have cared about it in the anonymized answers) as a "male" genderqueer attracted to men was psychologically too painful.

I completed most things except for the digit ratio. Thanks for putting this together, the results are always very interesting to see. Now to see how many of the trivia I got right.

Carrier has arrived ... what? huh? Where am ... oh right, yea, did the survey.

We did it. Mission accomplished.

y'all a bunch of paranoid delusional mentally-disabled freaks who can't get laid

Taken. Looking forward to seeing the results!

Took the survey. However, my answer for the probability of MWI is "Since MWI makes the same predictions as the standard interpretation, asking for the probability of MWI is meaningless. It is like asking "this glass is 50% full of water. What is the probability that it is half empty? What is the probability that it is half full?" I put 0 for the MWI question, but I'm not sure what you want for that.

For some of the other probability questions, my answer is "I don't have enough information to come up with a good estimate, and I also don't have enough information to come up with a probability that takes into account my inability to come up with a good estimate". Again, I put 0.

Also, after the test, I'm starting to get worried how you anonymize the questions. Releasing the data without a name attached is not anonymization, if the answers people give are enough to identify them.

Both are 100%. Duh. Likewise, I interpreted MWI to include any interpretation that makes the same predictions as it. I still gave an answer less than 100% because I wouldn't completely rule out all objective collapse theories just yet.
I was especially bothered by the MWI question because it asked whether it was "more or less" correct. Of course it's more or less correct since its math works! But since I assumed the question was intended to find out whether or not I thought favorably of the theory, I just skipped it.
Um, nothing complicated is required here. Consult your gut feeling and put whatever number comes up. Turning this into an actual probability is why we have calibration questions. Putting 0 is definitely a wrong way of refusing to answer the question (you can already leave the box blank, if you really want, and thereby avoid polluting other results). As an aside, given we had at least three people with such strange interpretations of the MWI question so far, perhaps the survey should include a question asking one's opinion on map-vs-territory... I believe that effective anonymization (apart from removing names) is in general impossible. Although, for specific analyses a reduced dataset can be effective, such as all pairwise joint histograms, which I think preserves a certain amount of anonymity depending how unique each person's answers are.
I also put 0 for MWI, although I feel pretty good about that. (For reasons explained in this comment, a 0 means that my answer is less than 0.5%.) I am the kind of Bayesian who strictly speaking only speaks of probabilities of potentially observable events. (This is a kind of logical-positivist Bayesianism, I guess.) It doesn't do to be too strict about this sort of thing (I don't want to just wall off entire subjects as unspeakable, which is the classic failure mode of logical positivism), but it does mean that I have to think about what other statements really mean in practical terms. So I interpreted this to mean, assuming that I learn much much more about the nature of the world than I know now, would I think that the MWI is a useful way for people today to think about things? (That's pretty much how I always interpret questions about interpretations.) And no matter how much learning I contemplate, the log-odds are never as good as 8 bits against, so that's a 0.
Putting 0 is misleading. It implies that you're confident there is no chance at all. If you're really not leaning one way or another, your best bet is to just put 50, or perhaps even skip the question if you really don't want to give a probability.
Choosing 50% is availability bias. Just because the question is presented as a choice between MWI and everything else doesn't mean there are only two choices. There are zillions of choices; MWI is just the one mentioned on the screen in front of me.
I assumed you'd already factored in those other choices and still weren't leaning more for or against it relative to all the other possibilities combined. By "leaning one way or another", I meant along a hypothetical axis of "strongly believe" or "strongly disbelieve" for the given proposition. You have a good point about availability bias though. You can self-correct for that to some extent by decreasing your assigned probabilities, and we'd have to take availability bias into account while interpreting the probabilities given by other people.

Done. The basilisk question was really interesting.

Welp, gotta go and destroy all humans now...

I don't think I saw such a question? Spooky.
I suspect it was a joke, but god only knows wrt that clusterflip...
Can I ask what sentiment you were going for with "clusterflip"?
A more polite form of "clusterfuck". It was the least ridiculous-sounding form I could think of.

Survey complete!


Hopefully, i'll be able to change a few of my answers regarding the LW meetup frequency by next year. And the answers regarding donations should change drastically within 3 years.

Was pretty happy that I knew a bunch of the answers wrt the calibration section. :)

Now hand over them Karma points.

Took the survey!

Done, except for the digit ratio, because I do not have access to a photocopier or scanner.

I took the survey.

I have filled in the survey (I wouldn't have minded if it was longer!)

Survey completed in full, reporting in for karma as per ancient tradition.

Thanks to Scott and Dan for all the work they put into this!



I did everything but finger length. I am shame.


Did the survey!

Did it, including the digit ratio.

I may have found a problem-- if I didn't click on the background after answering a radio button question, then using the down arrow marked a lower radio button. I think I cleared up all the resulting errors, but it took two passes, and I may not have caught all the errors.

I didn't express how serious I think the down-arrow problem is, though perhaps my computer habits are unusual enough that no one else had it. I think it led to at least ten wrong answers, and some of them showed up on the second pass when I was correcting the first batch. Did anyone else notice this problem?
Always an annoying thing about radio-buttons on scrolling pages. I adjusted to clicking on the background in such situations a long time ago (for the most part), but it still really annoys me.
I was hyper-aware of this problem, since it's happened to me on internet surveys in this past, so I nervously clicked on the background at least once before scrolling down each time. If there was some way to know that this wouldn't happen it would be a bit of a weight off my mind :p
I usually scroll using the PageDown key and/or the scrollwheel on my mouse (I mean, the side bar on my touchpad, but I'm too used to using the word “mouse” to refer to any pointing device).
I've learned to use the mouse, and not the keyboard, when answering this sort of thing. YMMV.
Nthed. I had to scroll using only the mouse.
Since you can move through that with the tab key, I don't think it's a significant problem.


Is it deliberate that the size of the MONETARY REWARD is not stated anywhere?

Finished it. I can't wait to read the post that talks about how bad people are at following directions.

I can already tell you that...well, you remember the preview thread. The one where I posted a version of the survey saying in big letters on the top "DO NOT TAKE THIS, IT IS NOT OPEN" and the first question was "You are not supposed to take the survey now" and the only answer was "Okay, I'll stop"?

Four people took it. Obviously they won't be counted.

Did those people get the coin probability question right?

If there were more of them, I'd be interested to know if there were significant differences in the survey responses between the people who did and didn't.
(Reply to somewhat old post) In "The Design of Everyday Things", Don Norman points out that having a "push" sign on a door where the most natural action is to pull the door is a bad design. Having a sign saying "do not take this survey" on a survey that would otherwise look like you should take it seems to be an extension of that.

Done! The survey has been a progressively smoother experience each of the past three years. And it's nice to have a time to think about the past month's habits in a structured way during the school year.

My first comment is to say I did the survey.

I completed the survey. (Did not do the digit ratio questions due to lack of available precise tools.)

Submitted, answering almost all questions.

The hardest question was choosing a single favorite LW post.

Also, I wasn't sure if Worm should count as more than one book. (It didn't end up mattering.)

A scanner + Photoshop makes it significantly easier to measure digit ratios.

Was that question not there yesterday?
I'm starting to feel rather disappointed that I took the survey so early. Should have waited around for Scott to add the interesting questions...

Surveyed --- I feel somewhat unconfident about my calibration.

Did it just now. Lone portuguese (from Portugal) here with high certainty.



Took the survey! Now to upvote everyone who took it.

For those that have mentioned a lack of a ruler, I used this one online:

Might be worth it to link in the survey, if it's still editable.

First thing I thought was ‘I'm not sure it's accurately calibrated’, but since we're measuring ratios it doesn't matter.
You can click "select your monitor dimension" to resize the ruler. The default they gave me was wrong. I actually suggest making the ruler even smaller than the authentic size, so that the distance between millimeters will be shorter and thus the ratio will be more precise.

I'll be interested to compare the results to the 2014 Effective Altruists Survey from earlier this year. Peter Hurford will be presenting its results soon, and I believe he's cross-tabulating them with those on the 2013 LW Census (including figures like the gender ratio and how much people donate).

Also: I've now taken the survey. There were some interesting questions there.

I took the survey.

I have a few suggestions though.

For the race question, I recommend allowing people to pick more than one option, or creating an extra option saying "I don't primarily identify with one race".

For profession, I feel like it was unclear what people who aren't currently students or employed are supposed to pick. What they most recently worked in or studied in a formal setting? What about students who haven't declared a major yet? The field of study they're leaning toward?

For the time in community question, I suggest clarifying whether that includes lurking. My guess was no, but I think it was sufficiently vague to where a significant number of people wouldn't have guessed that.

I would also be interested in seeing a question relating to use of artificial cognitive enhancement techniques such as tDCS and nootropics.

Thanks for working on the survey. :)

Amen. Though maybe in terms of analysis "I do not identify with any race," which I imagine may be more common here than other places since people choose not to identify with other variables for which it is a more radical statement, is uninteresting to the survey. I that case, "I do not identify with a race" or "I identify with more than one race" could be usefully lumped in with "other." If we're the only two a racial people on the site I'm not sure it's worth the effort.
It might be uninteresting from the standpoint of someone who only wants specific racial information, but it still might be interesting for other reasons to see what other qualities correlate with someone who picks that kind of answer. The thing is, I wasn't sure Yvain had the capability to create checkboxes that allow selecting more than one answer choice, as I didn't see them anywhere on the survey. The "I don't primarily identify with one race" was meant to be a catch-all for mixed-race people who don't want to pick sides between their races, but I agree it would be more useful to subdivide that even further to "I identify with more than one race" and "I do not identify with a race". I personally got around this by selecting "other" on the grounds that I identify with the human race. It looks like for the 2013 survey, 51 people answered "other" and 22 people left it blank, so I think there are enough people for further distinctions to be worthwhile. There were other race options that even fewer people selected. I feel like "other" is best reserved for people who do identify with an ethnicity that wasn't represented in the answer choices, and leaving the question blank is best reserved for people who dislike the question/answers, want to be more anonymous, etc.
I agree. Selecting other felt like I was shoehorning in. I just hate asking for special treatment. Growing up, a whiner was the worst legal thing one could be, and it stuck with me a little. But 73 people does seem significant .


I tried doing it on my phone earlier, but was having "issues" and decided to wait until I could do so on a laptop. In the mean time, I read the digit ratio comments and decided to try and measure mine.

I measured wrong, and the ruler (which is no more precise than half centimeters) did not come with me to my current location. *is sad*

I have submitted the survey, AND for the first time realized I'm not sure the example lifespan in the anti-agathics question should be understood as continuous. And I learned about natural law!

Done did the survey!

Thanks, I did the survey. I had been lurking some multiple months in irc and reading bits of sequences and now made an account after the survey.

I would be interested to work with the organizers to include an actual IQ testlet in a future survey.

My worry is that taking an IQ test online (even timed with reliable questions) cannot duplicate the exact same experience as taking an IQ test in a proctored setting. There are likely to be more confounding factors that throw off the scores relative to proctored tests, since the environments cannot be as strictly controlled.
Well yes. Mainly including a couple of testlets would alleviate the self-test worry. We could infer the population average IQ relative to those testlets' hardness, which could confirm or disprove the self-reported IQ accuracy. I have understood that there has been some amount of doubt related to self reporting of IQ on the census here.
Sure, if you gave the same test to a representative sample of LWers and to a representative sample of the general population, you could calibrate IQ scores across them. I still expect it to be less reliable than proctored IQ tests though, not because I'm worried about people lying about their scores, but because of a higher incidence of confounding factors such as distracting noises, internet connection failures, and even the presence of daylight from a nearby window. I suppose it might be interesting to include some IQ questions anyway, as it might still turn up some interesting results. We'd just have to keep the limitations in mind while analyzing the results.

Hmm, I did worse on those calibration questions than I would've expected.

Most people do worse at calibration than they expect, but you can improve with practice.

All done.

  1. I put an estimate on one calibration question that I knew was wrong. In hindsight I shouldn't have done that. The mistake: I don't know what bone is the longest in the body, but I knew that. So I put down a random answer for that question. But then I felt like it would be cheating on the calibration to put 0% after an intentionally wrong answer, so I put a higher number that wasn't accurate. My mistake, but other people might have done something similar.

  2. I want the political questions to measure the importance of an issue on next year's survey.

If you put down a random answer and know you did, then it seems like the correct estimate for your calibration would be 1 over the size of the sample space. Google tells me there are 206 bones in the adult human body, but a lot them are mirrored left to right, so maybe you'd be looking at something just south of 1%? Probably higher, though, if you filtered out the many small bones in e.g. the fingers and toes, or the vertebrae.
You're assuming the answer I wrote down was an accurate name of a bone.
Even then your subjective probability wouldn't have been exactly 0. You could have put 0.00000000001 or something like that. The instructions didn't forbid you from using long decimals. Even so, I think it would have been fine to put 0 if your subjective probability really was 0 or you felt like rounding down to it.
The question was about the largest bone, not the longest bone.
Tomayto, tomahto. Comes out to the same. Which is good, since the question would be ambiguous otherwise. Wasn't sure whether to round to 100 or to 99. After all, we could all have been lied to.
If you think that the probability that we've all been lied to about this (the relative sizes of the bones in the human body) might be as high as 0.5%, then you live in a more interesting world than I do. Unless you just mean that you checked Wikipedia, and somebody who knowingly puts a false statement on Wikipedia (a public website) is technically lying to everybody, and you didn't check the references or even the edit history, so you were unsure whether the probability of having found such a false statement was higher or lower than 0.5%, then … well then I still think that that's much too high! (Edits: precise phrasing of stuff about technical lying.)
Well it gets really murky as to what constitutes lying if we're in a simulation, which is more probable than 0.005 by far. What if there were historic humans, but you're just a virtual facsimile of one? Is that a "we were lied to about our bones"-scenario? And so on. That's mostly what I was pondering.
Well, the statement could still be true in the context of the simulation. You may not have bones that exist in the universe outside the simulation, but you still have "bones" within the simulation. The name "bone" as well as the names for specific bones would be accurate if those are the agreed-upon names within your simulated culture. Whether the bones need to physically exist in the most fundamental level of reality in order to be considered bones seems like an argument over semantics. They still possess the other typical characteristics of bones that our culture has decided bones are supposed to possess. In everyday practice, people assign objects to linguistic categories based on resemblance to a prototypical example, not by making sure they fulfill a list of necessary criteria.
Oh, I agree that "the statement could still be true in the context of the simulation". Likely so, in fact, which is why we go down all the way to 0.005 from P(we all live in a grand ol' simulation, in a simulation, in a simulation). The whole survey was full of definitional quibbles. What is 'supernatural' etc.
Largest is ambiguous. It could mean longest, or largest volume (with or without counting the volume enclosed, if we're talking about the skull), or even heaviest.
Not that i knew the answer, but I assumed that of course it meant the heaviest. I don't seem to have much company in this!
I think it means largest volume without counting the volume enclosed.
That's what I thought too, and apparently I was wrong...

I accidentally pressed enter and the form was sent away - half-filled.

This is stupid. I sent another form with only the second half of the survey filled out. Dividing line is the population question, which I incorrectly answered with Rot13(Ehffvn).

You might want to rot13 that.

Done. I accidentally hit enter when I had everything done except for the digit question, so It submitted my entry and I was not able to answer that question. :(

Completed. I'll be fascinated to see how digit length correlates to gender default. It would imply some very interesting things about sexuality.

I answered every question except the last one (I don't have a scanner set up).

Took the survey. I think I've mentioned this last year: I'd like more clarity about the distinction between a "supernatural" God and living in a simulation.

Thirded. I was momentarily stumped by that question, not being sure whether a simulator living in a universe with different natural laws than our own counted as "supernatural". I ended up deciding no. The simulator's universe might be a different kind of natural, but not "supernatural". Still, including a clarification in the question would have reduced errors due to misunderstanding, not to mention saved us time. The survey is already quite long as it is.
I wouldn't mind the survey being twice as long if it allowed it to handle these can't-answer situations, though I would expect it to be the same length but just have a button or two to the right of each entry blank.
That would seem kind of redundant as it's already not necessary to answer every question, even the ones that don't say they're extra credit or skippable. Maybe Yvain could have made that clearer at the beginning? I personally wouldn't have minded a longer survey either. I'm just worried that making it longer would deter others from completing as many questions or even taking the survey in the first place. It might be a good idea to have a poll (perhaps within the survey itself) asking for the amount of time we'd be willing to spend on such a survey.


Thank you for continuing to run it.

Given the decision on a cap in length I think it might be worthwhile to do a second LW Lifestyle and Values survey in addition to the census. At best with half a year of distance to the census.

I made this suggestion in the Yvain's call for critiques on the census. Who'd do this. If nobody else is willing, I'd help create the Google Form, but I don't know how to do statistical analysis. Otherwise, I am a generic volunteer for this project, to be assigned tasks.

I am curious what kind of analysis you plan to run on the calibration questions. Obvious things to do:

For each user, compute the correlation between their probabilities and the 0-1 vector of right and wrong answers. Then display the correlations in some way (a histogram?).

For each question, compute the mean (or median) of the probability for the correct answers and for the wrong answers, and see how separated they are.

But neither of those feels like a really satisfactory measure of calibration.

At the very least, I suspect one of the analyses will be 'bucketize corresponding to certainty, then plot "what % of responses in bucket were right?"' - something that was done last year (see 2013 LessWrong Survey Results) Last year it was broken down into "elite" and "typical" LW-er groups, which presumably would tell you if hanging out here made you better at overconfidence, or something similar in that general vicinity.

The question called P(Global Catastrophic Risk) should really be called something more like P(not Global Catastrophic Risk). (Or else the question itself should be inverted, but that would be a Bad Idea since some people have now filled in the survey.)

Some US states do not have partisan voter registration, so choosing "no party" does not necessarily mean someone would not register by party if that option were available.

Did the survey! ...And now to upvote everything.


I left the HBD (human bio-diversity) question blank, due to having misplaced my barge-pole.

Took the survey!

i did it yay me

Done. Skipped the digit ration questions to not put off answering the rest.

I've gone back, sorted the comments by 'new', and upvoted everyone who commented they did the survey since I took it, and upvoted everyone who did it before me. This way I've upvoted everyone, and they got more karma. It took me three minutes. If you spend a substantial amount of spare time on Less Wrong, it might be worth it for others for you to do the same. The more people who do this, the more karma everyone gets. Also, it can act as an incentive for people to take the survey for karma even if they're late to the game.

Did so too.


Oh, right. Alternatively, just noticing comments on this post in the 'recent comments' sidebar might suffice.

Me, survey, did, etc.

EDIT: I do not self-identify as a LWer (and am a bit surprised other people here would do that), but I would expect to be in the survey target demographic none the less.

People who frequently play chess are chess players. People who frequently spent time on LW can be seen as LWers. With >1000 karma you simply fit in that category.

Undoubtedly a point of controversy. Examples: * In some societies, a great many people play games, if only mobile/phone/web games. Yet only a fraction of them would "identify as gamers". * Birth genders vs. "identification". Or, myself: I identify as an LWer but only made an account today, and certainly haven't yet finished all of the sequences. I could feel like a bit of a poser, or worry others would call me "fake", but that's not actually relevant to my own self-identification.

Wait, what other people?

I took the survey. No scanner available, alas.