Take it before clicking the spoiler. And don't look up any answers!
So, you probably did poorly, right?
I got 5 answers right...
Most people got somewhere in the range of 3 to 5 right, rather than the 8 to 10 that would be expected, if people were accurately gauging their accuracy.
Something like that. Remember, if you're not sure, you can always increase the size of the range. If you have no clue, you could make it very wide indeed. So on the question about the diameter of the moon, you could guess 100 miles to a 100k miles, and if you have any idea of scale, you would know that the answer has to be in there.
But you have a second chance! Another quiz, with different answers. Now that you know you're probably overconfident, you can make more conservative guesses.
I suck! Same on the first one, 5/10, and went for wider margins in the second one and still only got 4 right. Though that one had harder, less "easily guessable" answers.
6/10 and 9/10 (holy shit there are a lot of glaciers in Alaska
On the first one, I was pretty far off on 2 and only missed the other 2 by a very small amount. I think if I had given a little more thought to how much range I would need to be 90% sure I had the correct answer in my range, I would have gone for bigger ranges.
I feel that way too, Braces for Impact. Truth is I didn't read the part that said to pick a range wide enough to be 90% sure of being right. Fell into the classic mistake of not reading the instructions properly.
__________________
"freedom to differ is not limited to things that do not matter much. That would be a mere shadow of freedom. The test of its substance is the right to differ as to things that touch the heart of the existing order."
- Justice Robert Jackson, West Virginia State Board of Ed. v. Barnette
Heh, I got 9 out of 10, but I got it by giving rediculously wide margins. ...surprised I blew the one about the Nile. Knew I should have added an extra 0.
__________________
"...because everyone is ugly as sin, when you rip away their skin."
I did the whole not-reading-instructions thing the first time and just pulled numbers out of my ass. Got 1 right. Actually read the instructions on the second one and got 8 right. I missed the number of Mozart symphonies and the number of glaciers in Alaska, guess too low on both.
Neat tests -- very effective for inculcating a sense of epistemic humility. I think they don't really measure overconfidence in the normal sense of the word, though. They measure something more subtle, like how susceptible you are to anchoring effects (auto-stimulated ones, even). You could get zero correct for both of those tests -- and yet be someone who would make sure to look it up were any of those questions ever salient.
Being a poor estimator or exploiter of one's own implicit knowledge (skills relevant to those tests) is perfectly consistent with habitually seeking and deferring to expert sources rather than simply relying on one's unreliable estimating skills. But we wouldn't say that someone who, when uncertain, sought and deferred to expert authorities was overconfident.
I think the trick is the thing where it says to be as close as possible, which wins over the "90%". I had lost track of the "90%" and was trying to come up with "reasonable" ranges, which of course killed me a few times.
__________________ Hear me / and if I close my mind in fear / please pry it open See me / and if my face becomes sincere / beware Hold me / and when I start to come undone / stitch me together Save me / and when you see me strut / remind me of what left this outlaw torn
I think that's true. Most of the ones I got 'correct' were only because I entered ridiculously wide estimates. I would certainly not rely on my own powers of estimation if any of this was important, as I'm keenly aware of how shitty they are.
Ah, but the first test asks the kind of quantities we cannot estimate. I got 6.
Hm?
Why can't you estimate them?
It's not asking you to accurately guess the quantities, it's asking you to accurately evaluate how certain you are. If you don't have a single clue how many glaciers are in Alaska, guess 1 to a million, and you can be quite sure that you'll get the number right. I mean, you can obviously by a lot narrower than that, but you get the idea. I agree that it's difficult to estimate some of them because you don't (or I didn't) even know the magnitude I was talking about on a couple of them, so I was quite far off.
But, still... if you're still not very certain of your answer, you should just make the range larger. I could've made the range larger rather than assuming I even knew the magnitude of the number I was talking about.
I mean, I got 5/10 on the first one too, but I don't take issue with that aspect of the quiz.
I agree with Clutch tho that this particular type of task isn't necessarily all that indicative of how overconfident you are in real-life situations. A lot of people would rather guess something more specific and be wrong than say 1 to a million - either way it'll be obvious that you don't know what you're talking about... But that doesn't mean you undertake decisions without getting more accurate information. Ideally, you would want to quiz people on the types of things that they actually encounter - like, estimate how long it will take you to do something, or how much your dinner will cost if you go out to eat, that sort of thing. The type of thing where overconfidence might lead to problems, since most people aren't going to stake very much on being right about the gestation period of an elephant, or have any reason to need to know about it off the top of their heads. But those type of things also aren't really easily converted into a quiz for a general audience...
Or maybe you could try not to drag in shit from elsewhere.
It says to make guesses that you are 90% certain contain the answer. If you have no clue, then you need a very wide range to have 90% certainty, don't you?
If you have no clue, then you need a very wide range to have 90% certainty, don't you?
If you have no clue, whatever it is you are doing is not normally called 'estimation'.
You seem to be disputing anything I say now just for the sake of it, dude. Or maybe you could try not to drag in shit from elsewhere.
Is that a rule you want to impose on me? LOL There certainly are a lot of self-appointed policeman in this forum!
mickthinks, maybe an example to which you can relate would help.
Imagine the quiz asked you "How many Jews (or "mockies," as you say) were killed during the Holocaust?" You would "estimate" between 0 and 1. Of course the orthodoxy that you and Nick Griffin so bemoan would disagree, but it's an orthodoxy and you and Nick Griffin are heretics.
I just noticed that on the second link, after you submit your answers there's a link to yet another quiz. Which I got 9/10 on, albeit with some fairly wide ranges on a few.
Quote:
Originally Posted by mickthinks
Quote:
Originally Posted by erimir
If you have no clue, then you need a very wide range to have 90% certainty, don't you?
If you have no clue, whatever it is you are doing is not normally called 'estimation'.
Yes, which is why I used the word guesses in bold in that post, to emphasize that that is guessing. So yes, in that sort of situation you would be guessing.
It says to guess on both sites I linked to, and I said guess in the OP. So I don't see what relevance it had that you supposedly can't estimate those values since it didn't ask you to.
But even so, if you do have an idea, then it would seem to be estimation ("to form an approximate judgment or opinion regarding the worth, amount, size, weight, etc."). If you don't have any idea, then it's just guessing, true. But I never said that you had to be estimating everything, I said that it was possible to estimate them. Just because they're the sort of questions that most people would have to guess on some of the answers doesn't mean they can't be estimated.
But I don't see what the relevance is
Quote:
You seem to be disputing anything I say now just for the sake of it, dude.
In that case, why have I made posts supporting certain usages you made, or certain of your points?
Quote:
Is that a rule you want to impose on me? LOL
LOL
No, fuckhead, I was just saying that I don't give a shit about what's going on in the other threads. It's not a rule.
I was just trying to have a conversation, but you take any disagreement to be part of some malicious micksux agenda. I was being polite to you, but it's obvious that it doesnt matter how polite you are, if you disagree with mick, it's because all this other bullshit, not because you actually have a disagreement
Quote:
There certainly are a lot of self-appointed policeman in this forum!
So I don't see what relevance it had that you supposedly can't estimate those values since it didn't ask you to.
You don't see the relevance of estimation to Clutch's and Kael's comments about being a poor estimator, unreliable estimating skills, and powers of estimation? Then why didn't you say so instead of asking ...
Quote:
Originally Posted by erimir
Hm? Why can't you estimate them?
I thought you were disputing my contention that the first test asks the kind of quantities we cannot estimate. Am I being too 'literal' for you again?
Actually, yes you are being too literal. Speaking only for myself, I was not being terribly precise with the choice of the word 'estimate' instead of 'guess' or 'ass-pull'. I did not feel the need, and my post seems no less clear because of that imprecision. Clutch's doesn't seem any less clear either, though I'll not speak to his reason for using the word 'estimate'.
At any rate, your post-modern sally into this thread was to criticize erimir's use of the word 'estimate' and to claim that he was behaving in the manner you have been (or, as you see it, that you have been accused of). I reckon you owe him an apology.
The ideal target in a lot of these examples is an interval based on the (perfectly precise, in principle) idea of 90% confidence given your state of information. But you don't have (inter alia) the time or calculating power to construct a proper interval for 90% confidence; so you estimate it.
And even that's if you follow the instructions well. As erimir suggests, people may have strong preferences for accurate precision over mere accuracy that make it easy to give too little weight to the instructions. Which is quite understandable, since one probably doesn't possess enough evidence to make such an interval informative by normal conversational standards in any case. There may well be conversational norms that make following the instructions in this case feel even a bit communicatively uncooperative.