Maybe add Poland to the answers to the Country question -- last year it got more votes than France or New Zealand.
I'd use “Male”/“Female”/“Other” as answers to the Sex question and “Man”/“Woman”/“Other” as answer to the Gender question.
Why is “Number of Current Partners” write-in and “Children” multiple choice?
I'd split the “No” answer in “More Children” into “Not now” and “Never”.
I'd add “Apatheism” and “Ignosticism” to the Religious Views question, either as separate answers or as “Agnosticism/Apatheism/Ignosticism” (cf. “Don't know/Don't care/Don't understand”).
“Roman Catholic” and “Other Christian” is probably the weirdest possible way of cutting it. I'd have “Christian (Orthodox)”, “Christian (Catholic)”, “Christian (Protestant)” and maybe “Christian (other)”. (And maybe add “Atheist”, which got quite a few answers last year.)
Stylistically I'd prefer “None of the above” to “Other / no answer”, but whatever.
Would it be technically possible to have “%” displayed right after the write-in fields in the probability questions?
Maybe something more specific than “intelligent” in the “P(Aliens)” question -- would octopuses, crows, dolphins, or chimps count? A possibili
But LW is skewed more toward historically Protestant countries. I wouldn't be surprised if the breakdown actually has slightly more Protestant-background than Catholic-background.
I'm going to channel gwern from last year: give us a question that allows us to express disaproval about the handeling of the basilisk.
When I was interviewed about Friendship is Optimal, there was a minor side discussion in the comments on the interview. The comments were nonspecific enough that I think it's OK linking there; I'm pointing out that this is not going away since this came up with no prompting on something that only mentioned LessWrong. That interview is from 3 months ago, nearly a year after Yvain rejected having a basilisk question on the 2012 census.
This is still an issue. It will continue to be an issue. The way forward through this issue is to have something linkable that suggests that "XX% of LessWrongers (dis)agreed with the handling of the situation," so that the next time (Xixidu / RW / some internet rando) mentions the situation, we can point out that what the majority of LessWrongers actually think. (The phrasing there obviously suggests what I think, but if the results come back the other way, that too is useful information!)
I suppose it would be interesting to see if there is anyone left who does approve of how the basilisk was handled. I haven't been able to find anyone defending it recently, and Eliezer himself has implied that he now believes his response to the situation was counterproductive.
What would be the purpose of this question? It's too tempting to signal a contrarian "I am not in a cult" attitude by answering negatively. It is extremely hard to put oneself into Eliezer's shoes when he had to make a decision without knowing the repercussions, like Roko quitting with a bang, the resulting Streisand effect, etc. I suspect that Eliezer had to make similarly unpleasant decisions more than once, and most of them did not backfire as spectacularly. One recent example was handling eridu's posting on radical feminism, which had a potential to blow up but didn't.
What would be the purpose of this question? It's too tempting to signal a contrarian "I am not in a cult" attitude by answering negatively.
You don't really believe that this question's results would be meaningless. If we put the question in and the results were 100% 'I endorse Eliezer's handling of the basilisk', would you and everyone else simply ignore this, saying "it's too tempting to signal loyalty to prominent figures and willingness to make sacrifices"? No, of course not, you would make use of this evidence and cite it in future discussions.
And if one outcome is meaningful, then by conservation of evidence, the other outcomes (like, say, 90% polled expressing disapproval) are also evidence.
I suppose it would be interesting to see if there is anyone left who does approve of how the basilisk was handled.
As opposed to which other specific possible way of handling it?
For example I may think that there were both better choices and worse choices, and the Eliezer's choice wasn't optimal, but also wasn't obviously bad. Now do I agree or disagree?
Saying "I disagree" does not say what the person would prefer instead. It creates a non-natural cluster of people preferring various kinds of alternative solutions. A list of choices would give more information. For example "moderator should ignore it completely", "moderator should use a private message to suggest retracting the comment", "moderator should move all related comments to a separate discussion", etc.
In that way the people who think there should be a specific basilisk-related thread with trigger warnings don't end up in the same set as e.g. the people who think the site should be completely unmoderated. (And maybe we could get a result that most people think Eliezer should have done something else, but there is no general consensus about what specifically it should be, so it is likely that if Eliezer had actually done something else, he would still get a lot of criticism. You can't get this information by posing a dilemma of "I agree" and "I disagree".)
Alternatively, I'd like to have an answer: "I don't fucking care. Forever obsessing over a one-time event that happened years ago is more harmful than the event itself." Which is connotationally completely different from both "I agree" and "I disagree".
Can you add a question that asks if people self-identify with the effective altruist movement?
Repeating the CFAR questions again is going to be skewed because a lot of us will have prior exposure to the correct answers from reading about last year's survey.
Suggested new questions:
About how many hours of television have you watched in the past week? "Television" here means commercially-produced programs that are broadcast on some TV channel; regardless of how you personally access them.
Do you live with pets? Pets are nonhuman animals that live in or around your residence, are not livestock or wild animals, and that you or someone you live with takes care of.
Have you ever donated blood?
Think about the meals you have eaten in the past week. Whom did you eat with most often? (This is a plurality, not a majority.)
I'd like a question about how politically active people are. Tentatively suggested list of answers: vote, vote in primaries, do research before voting, involved with parties, tries to influence legislation by contact with people who can affect it directly, has run for office.
An option for probabilities to the effect of "this is so hard to estimate that I don't think much precision is possible"-- something equivalent to "revival from cryonics isn't going to happen next year, but I don't think there's a sensible way to talk about the odds for thirty years from now". In other words, the militant agnostic position: "I don't know, and you don't either".
Here's a question I've wanted to see about religion, but it would also work for rationality: Has rationality affected any of your major decisions about sex and/or money?
Has rationality been of practical use for you? Time, money, relationships, other.
Do you use rationality (not necessarily learned at LW/CFAR) as a filter for who you associate with?
"Vote in primaries" is a bit US-specific. How about "vote in national elections only" and "vote in other elections" (which would cover primaries, elections for local government, etc.)? Something like that, anyway.
Add a question about cryonics to distinguish technical feasibility estimate from total probability estimate (the current P(Cryonics) question). This distinction is important, as the results of past surveys are sometimes misleadingly cited as talking about technical feasibility. Something like this could work:
P(Cryonics | No external defeaters)
What is the probability that at some future time, it will become technically feasible to successfully restore to life an average person cryonically frozen today, conditional on no global catastrophe and on the storage facility remaining functional (in some form)?
(I replaced "will be restored" with "feasible", since I suspect it might be morally suboptimal to restore frozen humans as opposed to doing something else, which is a factor unrelated to technical feasibility. The "in some form" is intended to address hypothetical change in form of storage, such as plastination or uploading, taking place before the "restore to life" point.)
Family Religion
I suggest "None" as a distinct option.
SAT scores out of 1600
Would still like to know how to informatively fill this out given that, like many people, I took it when I was 12.
What is the probability that supernatural events, defined as those involving ontologically basic mental entities, have occurred since the beginning of the universe?
Maybe add a distinction between regularly attending meetups and ever having attended a meetup? E.g. change the "Meetups" question from Yes/No to three options.
Possible new question: “Who are you living with?” with answers “Alone”, “With parents (and/or siblings)”, “With roommates”, “With partner/spouse (and/or children)”, and “Other”.
(And maybe “Have you had sex in the last 30 days?” with answers “Yes”, “No”, and “Depends on what you mean by ‘sex’”.)
[De-jargonified in response to Xachariah's comment]
The 'Taboo 'sex'' option is somewhat obnoxious, isn't it? You're pretty much saying, "I anticipate that you think I asked a malformed question"
Add a "No political identity" or at least "It's complicated" or "Other" option to the "Political" question in "Views and Opinions". I'm just not interested in politics: it's hard to evaluate the implications of the elements included in its package deals (as opposed to fooling yourself or signaling tribal affiliation) and that effort ends up a waste of cognition. I don't want to study the listed options merely in order to pick one.
Last year a few people complained that they wanted to take the survey but it had been closed before they got around to it, so this time the closing date should be stated in advance when the poll is opened (and preferably be at least 30 days since the opening).
Modifications to existing questions:
Clarify what "looking for more relationship partners" means. Perhaps being open to more relationship partners (but not looking actively) should be distinguished from actively looking and from not being open to new partners.
"Planning on having more children" - what's the time frame on that? Am I planning to have more children at some point, or in the near future?
For "Religious Background", there should probably be a "Protestant" option before "Other Christian". It doesn't make much sense to lump Eastern Orthodoxy and Protestantism into one group while having Catholics in a separate group.
For "Moral Views", make these checkboxes (check as many as apply) rather than mutually exclusive selections, because it's possible to be both a virtue ethicist and a consequentialist, or alternatively a virtue ethicist and a deontologist.
For "Alternative Alternative Politics", these should also be checkboxes, as some of these aren't mutually exclusive.
I'm looking forward to the survey. Thanks for doing this again.
I'd like to see an option for referred by other fiction (Friendship is Optional, Three Worlds Collide etc.) as well as the one for HPMoR.
Suggested earlier, perhaps use score voting for politics? That is, ask people to give a number from 0 to 10 expressing how much they approve of the five political options. (I would do this along with asking people to radio button just one of them, so we can see how the radio button compares to the scores, and still compare with previous surveys.)
On the self-reported IQ question, please specify that the test is supposed to have a mean of 100 and stdev 15, and scores from tests with different scales should be rescaled.
It might be useful to have two different questions:
If you've taken an IQ test administered by a psychologist, and still have the scores, please tell us your score here. If your test was not normed to a mean of 100 and standard deviation of 15, please renorm it before reporting it.
What do you estimate your IQ to be? As with the previous question, please use a mean of 100 and standard deviation of 15.
[edit]We might want to give people an opportunity to say what age they were when they took their IQ test, or ask what test it was- that way we can separate out childhood IQ and adulthood IQ, which can diminish many of the criticisms of childhood IQ tests.
This is a minor nitpick, but the options "Man" and "Woman" to the question "What sex were you assigned at birth?" seem kind of off to me. No one refers to a baby as a man or woman! You could just make it "M" and "F" like in the subsequent question about gender.
The same problem as for the cryonics question is present in the anti-agathics question, and arguably it's worse for anti-agathics, as the total probability doesn't inform any decisions. The current question is as follows:
P(Anti-Agathics)
What is the probability that at least one person living at this moment will reach an age of one thousand years?
In my estimate, most of the factors influencing the answer have nothing to do with life prolonging technologies. For a person to live for 1000 years, there has to be no global catastrophe that kills all during all ...
I would like if there was personal religion question that listed main religions, like the family religious background one does.
I would also like Unitarian Universalism to be among the list of religions, as I would like to test whether or not being UU is correlated with being on Less Wrong.
Would it be a good plan to change the calibration question? People probably looked up the answer last year, or remember having got it wildly wrong last year... could any of that cause over- or under-confidence in a systematic way compared to a new question? It doesn't seem impossible.
On the sexual orientation question, I would change "Bisexual" to something like "Bisexual (or pansexual, etc.)" It's not something I feel strongly about, but I know some people do. And it got discussed on Dan Savage's podcast today.
When it comes to the SRS/Anki question, I would like an option that indicates that the person has used it in the past.
(I ctrl-F'ed this but couldn't find anything similar.)
Could you add a question or questions along these lines:
In a typical week, approximately how many minutes do you spend in moderately vigorous physical activity (at least as strenuous as brisk walking)?
If you lift weights, what is your (non-estimated) one rep max for bench press? Squat? Deadlift? Overhead press?
The cigarettes question needs a “Rarely / only tried a few times” answer too. (I second Username suggesting that all drugs questions have the same set of answers.)
The questions on Smoking and Nicotine distinctly lack a middle question "Do you use some kind of smokeless tobacco?" (eg I don't smoke but use snuff almost daily).
I want a mental illnesses question other than the Autism score one. I was going to come up with a wording that seems okay to me but surely you are more than qualified to decide what the more effective wording and answer choice is.
Separately, I will be interested in a question regarding drug use - both recreational and nootropics (perhaps as separate questions).
"Primary Language" clarified as "primary spoken language" is somewhat wrong in my case, as I speak in Russian, but read and write in English, and I read much more than I speak, so English is a better candidate for "primary language" overall. Maybe amend to something like this:
Primary Language
What language do you use most often (when reading, writing or speaking)?
Extra questions for politics. Policy questions might be better than identification labels.
The European welfare system is superior to the American system:
Strong Disagreement - Light Disagreement - Neutral - Light Approval - Strong Approval
Snowden did the right thing in making secret documents public:
Strong Disagreement - Light Disagreement - Neutral - Light Approval - Strong Approval
Capital gains should be taxed the same way as other income:
Strong Disagreement - Light Disagreement - Neutral - Light Approval - Strong Approval
Gun ownership should be legally ...
Since there were so many people from the US last year, maybe we should ask people from the US which state they're from (and maybe ask people from the UK whether they're from England, Scotland, Wales, or Northern Ireland).
1) If we are interested in autism quotient, I suggest adding RMET because, while not a direct measure, it's also not a self-report.
On that note, why all the focus on autism, and not generalized mental illness?
I suggest adding the most common Axis I and II disorders as listed here and asking people to specify [No] [Suspected but un-diagnosed] ][Yes, diagnosed]. If size is a constraint, I'd suggest just doing Axis I with a write-in option.
For ease of access:
Axis I: depression, anxiety disorders, bipolar disorder, ADHD, autism spectrum disorders, eating di...
Standard, pointless request, because it would break comparability, though it is still worth saying:
A new edition of the cryonics question became more misleading than before. The new version is as follows:
What is the probability that an average person cryonically frozen today will be successfully restored to life at some future time, conditional on no global catastrophe destroying civilization before then?
It factored out the probability of global catastrophe, but not other things, and as a result it now superficially looks even more like a question about technical feasibility, but it's still dominated by events unrelated to technical feasibility (mot...
P(Anti-Agathics) What is the probability that at least one person living at this moment will reach an age of one thousand years?
How is this to be interpreted? With or without the aid of cryonics?
As I think someone already mentioned, the Schrodinger Equation question tells more about whether we're good at differential equations than whether we're familiar with the ontology of quantum mechanics enough that we're qualified to judge the plausibility of Many Worlds or other interpretations. Assuming the latter is the point of the question, “... to prove Bell's Theorem” would probably be a much better test (though still not an excellent one).
You could ask a question on "How do you define rationality?" or "Which of the following do you think is the best definition of rationality?" My favorite definition is "Accurate beliefs and useful emotions."
(Inspired by this comment.)
What about a typing speed test? (And, while we're there, a question about alternative keyboard layouts such as Dvorak?)
As far as the standard political question goes I would drop the label communist as it had only 6 repondends last year.
I would split the label socialist into socialist and social democrat. Last year 41 of those who picked socialist also identified in the detailed question as social democrat and 60 of them did identify as socialist in both. 30 social demoracts also picked Liberal.
There are also 3 times as much people who identifiy as pragmatists than as conversatives.
New questions:
Have a Meta-Ethical Views question, with the options like "Moral Realism", "Moral Relativism", "Moral Nihilism", etc.
Since we have an Alternate Politics question, we could have an Alternate Ethics question, preferably with "check all that apply"-style checkboxes. Suggested answer choices: utilitarian, egoist, consequentialist, Objectivist, deontologist, Divine Command, eudaimonistic virtue ethicist, other virtue ethicist, intuitionist, moral relativist, moral nihilist. (Feel free to add to or modify these options.)
Have questions about tobacco use and alcohol consumption.
I have finally gotten the survey to a point where I'm pretty happy with it. I have no big changes I want to make this year. But as is the tradition, please take a week to discuss what minor changes you want to the survey (within the limits of what Google Docs and finite time can do) and I will try to comply. In particular, we can continue the tradition that any question you request can be added to the Extra Credit section unless it's illegal or horribly offensive.
You can find last year's survey results here and you can find the very preliminary version of this year's survey (so far exactly the same as last year's) here.
EDIT: I don't particularly like the IQ test or the Big Five test used last year. If you have any better replacements for either, tell me and I'll put them in.
EDIT2: CFAR, you added seven questions last year. Let me know what you want to do with those this year. Keep them? Remove them? Replace them?