# Bell Curve Question **URL:** \ **Category:** Factual Questions\ **Created:** [December 25, 2008, 2:57pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653 "2008-12-25T14:57:34Z")\ **Posts on this page:** 19\ **Page:** 1
**Author:** ![greatshakes](https://avatars.discourse-cdn.com/v4/letter/g/58956e/32.png) [@greatshakes](https://boards.straightdope.com/u/greatshakes)\ **Post date:** [December 25, 2008, 2:57pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/1 "2008-12-25T14:57:34Z")
I have had one semester of Statistics Light. I was, recently, trying to impress some babe (well, not impress, just let her know that I ain’t no hick) and I mentioned something about a bell curve. I can’t even remember what I said, because as soon as it was out, she bellowed that what I said was wrong, because a bell curve automatically, by definition, assumes that somebody is at the 1% mark, and also that somebody else from the population is at the 100% level. I mumbled something and skulked off. I didn’t cry, but I got angry, later, because I don’t think she was right, and I wouldn’t have had to mumble before I skulked off. Anyway, my whole concept is that for a ‘true’ bell curve, her assumption is wrong, because it depends on what the sample is. As I said, I only had one semester, and that was 2 years ago, and the class was something like Stats for non-business/non-math majors. So, am I misguided, or is this Jezebel not only wretched, but ignorant as well? ---
**Author:** ![beowulff](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/beowulff/32/542_2.png) [@beowulff](https://boards.straightdope.com/u/beowulff)\ **Post date:** [December 25, 2008, 3:02pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/2 "2008-12-25T15:02:30Z")
For a sufficiently large population, she is correct. ---
**Author:** ![Shagnasty](https://avatars.discourse-cdn.com/v4/letter/s/9dc877/32.png) [@Shagnasty](https://boards.straightdope.com/u/Shagnasty)\ **Post date:** [December 25, 2008, 3:06pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/3 "2008-12-25T15:06:49Z")
I am not sure that I am following but at least one thing she said is immediately wrong. The 1st percentile should match the 99th percentile and not the 100th percentile. There is no 100th percentile in a true bell curve but you could have the 99.99th percentile with a large enough sample size. You didn’t give enough other relevant information to comment on. The Bell Curve is a mathematical construct with known theoretical properties but it also tends to match real-world results very well whether it is test results or the results of a running race among randomly selected people. Some of this is circular logic because many tests are designed so that they produce results that approximate a Bell Curve (normal curve). This isn’t always true however. The GRE analytical test produces a bimodal distribution. Lots of engineers and math majors ace it and the liberal arts students tend to have a curve of their own further down. An actual normal curve is a very well defined property so she is (almost?) right if I read you correctly. There are other types of defined statistical curves however so, if you are calling all of them a Bell Curve, you are very much incorrect. ---
**Author:** ![Wendell\_Wagner](https://avatars.discourse-cdn.com/v4/letter/w/8491ac/32.png) [@Wendell\_Wagner](https://boards.straightdope.com/u/Wendell_Wagner)\ **Post date:** [December 25, 2008, 3:20pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/4 "2008-12-25T15:20:51Z")
greatshakes writes: \> I was, recently, trying to impress some babe (well, not impress, just let her \> know that I ain’t no hick) and I mentioned something about a bell curve. I would be happy if just once a woman was impressed by my ability at math. Do you know a lot of such women? Would you introduce them to me? ---
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [December 25, 2008, 4:52pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/5 "2008-12-25T16:52:01Z")
First of all, the o.p. needs to refrain from attempting to impress people using a barely understood area of knowledge from a half-remembered snoozemath class and then getting pissy when called out on his ignorance. It is bad form and unimpressive to the ladies, as it combines ignorance and insecurity. To expand on what **Shagnasty** said, the so-called “bell curve” is simply a graphical plot of the probability density of a [normal distribution](http://en.wikipedia.org/wiki/Normal_distribution). The normal distribution is so named because it takes a population distribution which is indexed by a single variable or (set of independent variables) and “normalizes” it (i.e. it is divided by some factor, in probability distribution the sum total) and shifted to the mean, about which the population is symmetrically distributed on an exponential function. This distribution represents a wide number of phenomena that can be characterized by variation about a point on a defined interval. For instance, if you were to stand on the end of a crane and drop marbles onto hot asphalt or clay (so that they do not bounce, and the only variations are slight deviations in how you release the marble) you’d find the resulting distribution about the radial distance from the target point to be a normal distribution. Most single variable populations studies about any arbitrary characteristic of a roughly homogeneous population–say, height, IQ, or propensity to fall for random bullshit guys throw out in a bar conversation–fall nicely into a normal distribution for a sufficiently large population. Errors in measurement also fall into a normal distribution, and this is therefore used on otherwise discrete measurements to capture the difference between what is and what is measured. The normal distribution also lends itself nicely to being able to distinguish between modifications to a variable that will have a significant effect (within a few standard distributions) to those which will have a very small effect (many standard distributions out, toward the “swan’s tail”). This is the basis for business and process improvement methodologies like Statistical Process Control and Six Sigma; unfortunately, many business-type people don’t seem to understand that these methods only apply to processes that can be quantified in terms of a few discrete variables and measured on a statistical basis, and do not apply to low volume, highly variable business practices, and so have been much maligned in their misapplication. To address the question of the o.p.:for a sufficiently large population (~1000) there is very likely some member in the 1st percentile region, and someone in the 99th percentile. For a much smaller sample size, the odds that someone is near the tail ends of the curves are much lower, but still non-zero. From the o.p.'s description of the interaction it would seem that both parties were operating from barely remembered knowledge further compounded by the acoustic environment and doubtless haze of intoxicating beverages, and so little in the way of comprehensible discussion did or was likely to ensue. **Wendell Wagner** , I find that quantum mechanics makes for a more engaging topic to be wizardly about, though Jearl Walker’s _The Flying Circus of Physics_ is really the go-to book for physics anecdotes to use in bars to impress women. Mind you, none of this has ever gotten me laid, but nor has a knowledge of Russian literature, Eastern European history, or a comprehensive reviewing of the films written by Charlie Kaufman. Having a good line of utterly meaningless patter and a ready supply of minor insults seems to be the most effective strategy with picking up women. Stranger ---
**Author:** ![Napier](https://avatars.discourse-cdn.com/v4/letter/n/ce73a5/32.png) [@Napier](https://boards.straightdope.com/u/Napier)\ **Post date:** [December 25, 2008, 10:19pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/6 "2008-12-25T22:19:15Z")
This snippet of conversation is weird in a couple of ways. Glaringly, there isn’t anybody at the 100th percentile, as Shagnasty was quick to observe. I think it is a bit inaccurate to say that somebody will be at the nth percentile, if we are discussing a continuous variable, because all the members of the population and all the integer percentiles will have unique values. I mean, if the 99th percentile is 72.284662894658267347267834 inches, you aren’t going to find that exact value in any of the members of the population. Also, some of us prefer not to use the terms “bell curve” or “normal distribution”. Many distribution curves have a shape like a bell, and what distribution is normal in some context depends on what the context is. I do a lot of particle counting, for example, and what is normal for particle counting is Poisson distributed variables. So I prefer the term “Gaussian” for what is often called the bell or normal curve. However, it does seem most people are happy using “bell” or “normal” - I only offer what I hope might be a clearer and more explicit refinement, not a correction. ---
**Author:** ![ultrafilter](https://avatars.discourse-cdn.com/v4/letter/u/3d9bf3/32.png) [@ultrafilter](https://boards.straightdope.com/u/ultrafilter)\ **Post date:** [December 26, 2008, 3:03am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/7 "2008-12-26T03:03:22Z")
Anybody who says “normal distribution” and means something other than a Gaussian deserves to have their knuckles smacked with a ruler. ---
**Author:** ![Santo\_Rugger](https://avatars.discourse-cdn.com/v4/letter/s/e95f7d/32.png) [@Santo\_Rugger](https://boards.straightdope.com/u/Santo_Rugger)\ **Post date:** [December 26, 2008, 4:21am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/8 "2008-12-26T04:21:44Z")
> [@Wendell\_Wagner](#): > > I would be happy if just once a woman was impressed by my ability at math. Do you know a lot of such women? Would you introduce them to me? The actual math is harder to impress with; it’s application much easier. I’d also venture to say it’s easier to keep 'em with it than it is to lure them with it. For example, night before last, while laying in a post-coital heap, the gal I was with turned to me, and said, “tell me something smart.” Most of the gals I tend to associate with tend to be impressed by intelligence, and last time I was asked to introduce somebody, I ended up getting a date out of it, so come on over, we’ll see what we can do with you. On the other hand, one can go too far. For example: > [@Stranger\_On\_A\_Train](#): > > **Wendell Wagner** , I find that quantum mechanics makes for a more engaging topic to be wizardly about, though Jearl Walker’s _The Flying Circus of Physics_ is really the go-to book for physics anecdotes to use in bars to impress women. Mind you, none of this has ever gotten me laid, but nor has a knowledge of Russian literature, Eastern European history, or a comprehensive reviewing of the films written by Charlie Kaufman. To which the reply I would expect would be: “Huh?” ---
**Author:** ![susan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/susan/32/17537_2.png) [@susan](https://boards.straightdope.com/u/susan)\ **Post date:** [December 26, 2008, 5:50am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/9 "2008-12-26T05:50:24Z")
Was her concern about the math part, or about how “somebody’s being oppressed (by your damn-fool assertion of a normal distribution)”? In other words, what were you talking about that made you refer to the bell curve? By the way, gentlemen, I’m a sucker for a fella with a slide rule. ---
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [December 26, 2008, 4:40pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/10 "2008-12-26T16:40:32Z")
> [@Santo\_Rugger](#): > > To which the reply I would expect would be: “Huh?” Yeah, but when I find the woman who can fluently discourse in the majority of those topics I’ll know that she’s the one I want to kidnap and chain in my basement. > [@susan](#): > > By the way, gentlemen, I’m a sucker for a fella with a slide rule. I have both a linear duplex and a circular slide rule. So, next time I’m in your neck of the woods, perhaps we can get together and calculate some cube roots… Stranger ---
**Author:** ![The\_Hamster\_King](https://avatars.discourse-cdn.com/v4/letter/t/8edcca/32.png) [@The\_Hamster\_King](https://boards.straightdope.com/u/The_Hamster_King)\ **Post date:** [December 26, 2008, 4:58pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/11 "2008-12-26T16:58:21Z")
Many people who don’t know much about statistics may nevertheless heard about [The Bell Curve](http://en.wikipedia.org/wiki/The_Bell_Curve), a 1994 book that argued that blacks are genetically less intelligent than whites. They may not realize that the problem with the book is not the idea of a bell curve in and of itself, but the particular methodology used to determine the relative “intelligence levels” of different “races”. It’s possible this is the source of the OP’s problem. ---
**Author:** ![Napier](https://avatars.discourse-cdn.com/v4/letter/n/ce73a5/32.png) [@Napier](https://boards.straightdope.com/u/Napier)\ **Post date:** [December 26, 2008, 10:23pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/12 "2008-12-26T22:23:13Z")
\>By the way, gentlemen, I’m a sucker for a fella with a slide rule. Ooooo, baby, baby. Ignore Stranger. I have several log log duplex decitrigs, several large Pickets, wooden slide rules, aluminum slide rules, slide rules inlaid with ivory, and a slide rule tie tack that actually functions. Besides that, I have actually designed and had manufactured a slide rule - a real one with log scales and special scales that calculate a kind of purity and figure of merit, plus several handy unit conversions. And a program I wrote to generate nomographs. I’m your man. I even have the name for it. That is, unless I don’t quite catch your drift… ---
**Author:** ![John\_Mace](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/john_mace/32/185_2.png) [@John\_Mace](https://boards.straightdope.com/u/John_Mace)\ **Post date:** [December 26, 2008, 10:27pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/13 "2008-12-26T22:27:57Z")
> [@ultrafilter](#): > > Anybody who says “normal distribution” and means something other than a Gaussian deserves to have their knuckles smacked with a ruler. Whatever you say, Sister Mary Ultrafilter! 🙂 ---
**Author:** ![susan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/susan/32/17537_2.png) [@susan](https://boards.straightdope.com/u/susan)\ **Post date:** [December 26, 2008, 11:12pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/14 "2008-12-26T23:12:02Z")
> [@](#): > > a slide rule tie tack that actually functions Ooh. But isn’t it a little… small? ---
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [December 27, 2008, 2:03am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/15 "2008-12-27T02:03:10Z")
> [@susan](#): > > Ooh. But isn’t it a little… small? It’s not the size, it’s how many different types of operations the man can perform with it that counts. Stranger ---
**Author:** ![Napier](https://avatars.discourse-cdn.com/v4/letter/n/ce73a5/32.png) [@Napier](https://boards.straightdope.com/u/Napier)\ **Post date:** [December 27, 2008, 11:48pm UTC](https://boards.straightdope.com/t/bell-curve-question/478653/16 "2008-12-27T23:48:22Z")
\>Ooh. But isn’t it a little… small? Are you assuming it runs across the tie, and not along it? ---
**Author:** ![susan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/susan/32/17537_2.png) [@susan](https://boards.straightdope.com/u/susan)\ **Post date:** [December 28, 2008, 1:29am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/17 "2008-12-28T01:29:53Z")
> [@Napier](#): > > \>Ooh. But isn’t it a little… small? > > Are you assuming it runs across the tie, and not along it? 😃 ---
**Author:** ![Rand\_Rover](https://avatars.discourse-cdn.com/v4/letter/r/db5fbb/32.png) [@Rand\_Rover](https://boards.straightdope.com/u/Rand_Rover)\ **Post date:** [December 28, 2008, 3:13am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/18 "2008-12-28T03:13:51Z")
> [@Pochacco](#): > > Many people who don’t know much about statistics may nevertheless heard about [The Bell Curve](http://en.wikipedia.org/wiki/The_Bell_Curve), a 1994 book that argued that blacks are genetically less intelligent than whites. Well that’s a weird way to summarize the book in one sentence. A better summary (that I just wrote in ten seconds) would be as follows: The book argued that the average IQ of the members of a race is higher or lower than the average IQ of all humans (depending on the race). And if you simply must spoil the big surprise ending, I’m not sure why you’d mention blacks and whites since blacks were not on the bottom and whites were not on the top. ---
**Author:** ![Olentzero](https://avatars.discourse-cdn.com/v4/letter/o/90ced4/32.png) [@Olentzero](https://boards.straightdope.com/u/Olentzero)\ **Post date:** [December 28, 2008, 3:49am UTC](https://boards.straightdope.com/t/bell-curve-question/478653/19 "2008-12-28T03:49:56Z")
> [@susan](#): > > By the way, gentlemen, I’m a sucker for a fella with a slide rule. Pffft. I got an abacus and I know how to use it.