Are you a woman or a man?

February 24, 2001 12:22 AM

Are you a woman or a man? No fair peeking. (As silly as it is, it guessed me correctly.)
posted by Steven Den Beste (75 comments total)

Oh, by the way: link courtesy of JC's PC News'n'links.
posted by Steven Den Beste at 12:25 AM on February 24, 2001

Apparently, I'm 80% woman. I guess that confirms everyone's suspicions then, doesn't it?
posted by endquote at 12:34 AM on February 24, 2001

I tested as a woman (straight man in reality), and spent the rest of the day deciding if that meant I failed the test or passed it.
posted by david hedge at 12:47 AM on February 24, 2001

So are you a man or a woman?
posted by asamee at 12:47 AM on February 24, 2001

A long time ago, I learned a different test which correctly differentiates men from women. It goes like this. You kneel on the ground, and place one elbow against one knee with the hand extending directly away, palm up and fingers outstretched. At the tip of the middle finger, you place a dime (a very small coin) on the floor.

Now, clasp your hands behind your back, and without moving your knees stretch out and try to touch the dime with your nose. Men will fall over; women won't. A man can get his nose on the dime but can't lift off afterwards without using his arms. A woman can lift back off the dime while still clasping her hands behind her back.

There are two reasons for this. First, men's arms are proportioned differently and the dime is further away. Second, (ahem) weight distribution is different, what with men having heavier shoulders and women being heavier lower down.
posted by Steven Den Beste at 1:14 AM on February 24, 2001

I am male as well.
Kinda middle of the road kind too.
posted by tamim at 1:20 AM on February 24, 2001

I tested so female. So wrong.

Dude! I even use the word "dude" in conversation?

I gots some 'splaining to do.
posted by thebigpoop at 1:26 AM on February 24, 2001

I tested 86% male, and believe me when I say, I'm 100% not male. So judging by the other results in this thread, I'm thinking they may be switching the genders intentionally to screw with us.
posted by fiery at 1:27 AM on February 24, 2001

A long time ago, I learned a different test which correctly differentiates men from women.

I think this is more of a psychographic test which is supposedly based on the notion that men and women think differently.

While it could be accurate you can never make out because for every deviation they would tell you that you are a borderline case with a mind of the opposite sex. And if they are right they get more believers.

Though it is good fun, I do not think it would hold water to a scientific evaluation.(maybe because a scientific evaluation just can not be done, since you can not find out the number of people who would have a mind of the opposite sex.)
posted by asamee at 1:29 AM on February 24, 2001

It guessed 86% confidence in the correct gender.

I almost backed out of it because it didn't state the number of questions upfront (at least I don't think it did). Argh, annoying. Same with CBS not indicating how many pages are in the Survivor episode summaries.
posted by gluechunk at 1:38 AM on February 24, 2001

Hmm, I took that test a few weeks ago, and I think I got something like 93% male... I guess I'm just a real man! :)
posted by swank6 at 2:29 AM on February 24, 2001

That was alright, Steven.
posted by xtrmntr at 2:57 AM on February 24, 2001

I did it and I was middle of the road and it still guessed wrong, oh well.
posted by john at 3:04 AM on February 24, 2001

Mine was exactly at 50%

Anyone see the 2nd picture of the cartoon woman? Damned ugly...looks like someone drew the head of a man on a woman's body.
posted by Amish Retard at 3:50 AM on February 24, 2001

I am every woman, it's all in me. Well, I don't believe in sexuality much - so perhaps that's why i'm 86% chickadee. Natch.
posted by holloway at 3:52 AM on February 24, 2001

It guessed me incorrectly. Oh well, so does everyone else. (errm, not people who see me, I mean stupid anti-tobacco activists who e-mail me about my website).
posted by dagnyscott at 6:07 AM on February 24, 2001

Italics no more.

I was more amused by the pop-up ad for a sweepstakes: "Free Pizza for a Year!" I imagine that if I had been determined to be female, it would have been "Free Salad for a Year!"
posted by dhartung at 6:29 AM on February 24, 2001

No no. Free dessert. ("I'm being bad but I'll exercise more this week.")
posted by argybarg at 6:35 AM on February 24, 2001

Though my anatomy dictates otherwise, I tested as 86% female.

What's interesting is that the test never claims to be a test of 'genetic' gender. It has nothing to do with X and Y chromosomes, merely a certain state of mind.

And when I think about it that way, I believe I'd much rather be considered female - I've no tolerance for monster truck rallies, prefessional wrestling and the like.

Though I have been known to kick back with some chicken wings and hardcore porn every now and again . . .
posted by aladfar at 8:56 AM on February 24, 2001

Okay -- that's one too many "86% female" responses even in this small sample. That's what I got, too. Could it be that there's no actual algorithm behind this page and it's just programmed to respond "86% female" most of the time? After all, it looks like the login count is just about 86% female ...
posted by argybarg at 9:07 AM on February 24, 2001

I was 80% male when I took it.

Sapphireblue flexes.
Sapphireblue ripples.
posted by Sapphireblue at 9:16 AM on February 24, 2001

i think that for most people it could all be determined with the single question "do you carry stuff in your pockets?"

funny, this test correctly guessed me as female, but this test said i was completely androgynous.
posted by bluishorange at 9:17 AM on February 24, 2001

Here at TheSpark, Privacy Invasion is a competitive sport. In fact, we're so good at unwelcome probing, you needn't tell us your personal information. We can guess it every time.

This little tidbit commented out on the first page of the test. Wonder what they mean?
posted by xiffix at 9:18 AM on February 24, 2001

The test seems to think that I'm 93% male. I was under the impression that I was batting closser to .1000 on that particular issue, but I'll take what I can get. But it does bring up an interesting question. (Ok, interesting by my (admittedly very low) standards.) To wit:

So the computer rates me 7% Nancy and 93% Boy. I wonder where the girlie 7% is hiding. Rather worrisome question, really, as that 7% could make itself known at any time. I could be addressing the Board of Directors (let's pretend I have a job, and at this mythical job, my mythical boss is dumb enough to let me talk to this mythical the Board of Directors) when, lo and behold, that part of me which is pure woman picks that very moment to roar! Next thing you know, my professional life has devolved into a scene from All of Me.

If someone buys me a pony, I'll promise never to post anything this dumb again.
posted by Optamystic at 10:02 AM on February 24, 2001

According to the test, most men think the test is retarded. I think the test is highly suspicious- mainly because I realised that I could enter the test without selecting any checkboxes whatsoever and it would still make a decision as to whether I'm male or female.

Could it be that there's no calculation involved at all and it just makes a guess? I thought it claimed 100% accuracy! It got me wrong, by the way, saying that I'm a girl when I'm not. Or at least I don't think I am.
posted by tobyslater at 11:11 AM on February 24, 2001

80% male here, which is no surprise. I was just telling a friend yesterday that despite the fact my husband and I seem like a perfectly normal heterosexual couple, we're more like a butch/femme couple. I've been told many times that I look like a girl, but I don't think like one.

And what's the deal with the "Canada Sucks" question?
posted by likorish at 11:54 AM on February 24, 2001

Okay, so their questions are wrong. Here is the definitive list of questions that separates men from women:

1) Do you prefer marmalade or jelly?

2) As a child did you like The Muppets or The Electric Company more?

3) Do you like wheat or white bread?

4) Were you born in February or a month beginning with 'J' or 'A'?

5) Of the numbers between 1 and 10, do you like the even numbers more than the number 3?

6) Who do you like more, Ponch or John?

7) Starsky or Hutch?

8) When flipping a coin, do you prefer heads or tails?

9) Were you born in an odd or even year?

10) Is your favorite color blue or some other lesser color?


posted by muppetboy at 12:50 PM on February 24, 2001

The spark says I am also 80% woman. Thank you Steven, for helping reaffirm my gender (although I almost broke my nose).

Before I took the test, I was pretty sure it would tell me I'm a woman. I'm another one who doesn't enjoy monster truck rallies, professional wrestling and the like in the least bit.
posted by tomorama at 3:48 PM on February 24, 2001

Hey optamystic- I think you mean you were closer to 1.000, unless you meant that you're ten percent male...
I got 86% male, by the way.
posted by dogwelder at 3:51 PM on February 24, 2001

oh, you're right...1.000 makes much more sense :)
posted by Optamystic at 6:20 PM on February 24, 2001

*sigh*

I tested 86% female... and though it may speak volumes of my mindset, it doesn't bid well for my ever finding a wife.

What I want to know is what on earth was up with the "Which one will win" question with the two different monkeys: the one menacing standing up or the one menacing while jumping.

I chose the standing one (I thought it looked bulkier/stronger).

PS... I spell checked this one Mathowie, and the program choked on the percent symbol.
posted by silusGROK at 6:37 PM on February 24, 2001

Hmm. Saw this a while ago (via plasticbag) and just could not answer the first few questions. I could not make a decision.
Maybe someone should provide a less accurate means of definition:
8 ways to tell whether you're a man or a woman.
posted by davidgentle at 7:06 PM on February 24, 2001

Got me right with 80% certainty. I'm working on that other 20%.
posted by kindall at 7:32 PM on February 24, 2001

Okay, I got 86% certainty too. For some reason, the spark test really likes spitting out that number.

It was right though. I am mostly male. But that other 14% does come in useful every now and then...

Like when I have to cook, clean, dress myself, write, speak sensibly, and anything else that doesn't fall under the categories of sex or violence.
posted by jia at 8:56 PM on February 24, 2001

I was surprised that no one mentioned ye olde playground gender test. Tell someone to look at their fingers (or maybe it was fingernails?). Supposedly, most women immediately spread their hands out in front of them (picture someone displaying a ring), men supposedly curl their fingers into their palm and turn their hand around.
posted by GaelFC at 9:19 PM on February 24, 2001

By the way, if anyone (particularly matthowie himself) wants me to stop making up lists of 8 facts inspired by Metafilter threads just say so and I will.
posted by davidgentle at 9:57 PM on February 24, 2001

93% man. This is why none of you should even think of looking @ me crosseyed.
posted by thirteen at 10:01 PM on February 24, 2001

Hmm. A lot of 93%'s too. I wonder if it is rigged, or if we fall into familiar patterns. Does prefering to fall to you death make you more willing to be lonely for all eternity?
posted by thirteen at 10:08 PM on February 24, 2001

David. Please. Stop. Thank you.
100% female Lynsey here. I think it was the colors lesser than blue thing (YES!) that pegged me. My question is, what took you all so long to discover thespark.com and the gender test? Zannah posted it a month ago....
posted by Lynsey at 10:12 PM on February 24, 2001

Minds aren't male or female, bodies usually are. You can learn to think one way or the other, but that's all based on your perception. How many men are *really* like OJ? Not that many, init?

I got that magic 86% too ... but then I was trying to "think the other way" by guessing what *they* thought the answers would be.

Anyway, Canada doesn't suck, Canada blows.


posted by Twang at 11:18 PM on February 24, 2001

Erm, isn't there a world of difference between "predicting something with 86% accuracy" and "projecting that some trait is 86% exhibited"? Same way I could be, say, 90% certain that Monday will be an aggravating day, though that doesn't really imply that 90% of all Monday's events will be aggravating?

So far among me and mine, the gend-o-matic indicator beneath the man/woman image has seemed to be a far more interesting metric....
posted by youhas at 12:00 AM on February 25, 2001

It got me wrong. I was right down the middle, but I'm also prone to defying these tests. Depending on my mood, I'm either an introvert or an extrovert (an introverted extrovert or is it an extroverted introvert?) on Myers-Briggs tests.
posted by shackbar at 12:33 AM on February 25, 2001

shackbar-The term I like to use is 'closet extrovert'.
posted by fiery at 12:55 AM on February 25, 2001

Twang, there's increasing evidence that minds really are male or female. In particular there's now irrefutable evidence that testosterone alters the way that a man's brain works -- and indeed that a man's brain doesn't operate properly without it.

It's more than just sexual preference. On a deep level men and women really do think differently (in main; there is considerable overlap). Unfortunately, in an age where equality has become nearly an article of religion, this is an uncomfortable thing to face, and the researchers in this area have been keeping their heads down because they understand the political consequences of their work. They risk backlash and loss of funding, and they know it. (Which is sad.)

Now it's important to emphasize that so far as anyone knows neither is superior. But they are different. Orthodoxy has decided that all these differences are cultural, none are biological, and a lot of people seem to accept this biological statement for political reasons. Unfortunately, research has demonstrated again and again that some of it really is biochemical.

It's known that men who, for various medical reasons (e.g. cancer), lose their testicles and no longer have that supply of testosterone exhibit rather profound changes in behavior. In particular, they become unaggressive, withdrawn, uncertain of themselves, and seriously depressed. This condition is known as hypogonadism. The treatment is hormone replacement therapy, which solves the problem.

Testosterone is made by a man's testicles but also by the adrenal glands in all humans. But the adrenal glands produce far less (on the order of 1/500) and these men have this problem because their serum testosterone level drops to what women (or boys!) normally have. Yet women, girls and boys don't suffer from those mental effects. That's because at puberty, when a boy's testosterone levels skyrocket, this causes profound and permanent changes in his brain, and thereafter testosterone is required for proper mental behavior. No-one has identified on a detailed level exactly what those changes are, but the clinical evidence for this is beyond doubt.
posted by Steven Den Beste at 6:58 AM on February 25, 2001

One may wish to compare the following:
chart_1
chart_2
chart_3
chart_4
chart_5
chart_6
...etc.

Seeing as how there are only 5 "confidence levels" per gender, and two of them are identical (80, 86, 86, 93, 98), I'm a little skeptical of any claims to be dynamically generating anything.
posted by dhartung at 8:52 AM on February 25, 2001

Lynsey: Why stop now?
posted by davidgentle at 1:35 PM on February 25, 2001

Steven: I was about to lambast you for your post when it occured to me that you might not be serious. And are you really in your mid 60's?
posted by davidgentle at 1:38 PM on February 25, 2001

I'm too lazy to look this up in a table of Z-scores, but those numbers look like the likelihoods generated by multiples-of-X standard deviations from the mean, for some X.
posted by rodii at 2:19 PM on February 25, 2001

In particular there's now irrefutable evidence that testosterone alters the way that a man's brain works -- and indeed that a man's brain doesn't operate properly without it.

Hogwash.

Levels of testosterone vary for person to person just as estrogen levels do. Yes, I think that they do affect a person's personality and drive but how that relates to "proper" function is so very debateable.
posted by amanda at 2:37 PM on February 25, 2001

This is the second time in a week that someone thought I was writing an extended hoax. Believe me, when I am joking about something you'll know.

I am 47 (though I don't understand why my age matters). What I described is a medical fact. Hypogonadism is a recognized medical phenomenon in some men, and changes in brain function are part of the syndrome. Here's some references:

"On psychologic testing, the men with untreated hypogonadism tended to score high on depression, anger, fatigue, and confusion scales." (link)

"Administration of androgens to hypogonadal men has clear-cut effects on mood: increased friendliness, energy and well-being, with decreased nervousness, irritability, sadness and anger." (link)

"The best-known consequences of hypogonadism in men are impotence and dwindling libido, but melancholia and psychiatric disturbances can also occur in association with testosterone deficiency." (link)

The human body is not well designed. When human engineers design things, we try to modularize and try to limit cross-system interactions. Evolution was not so clean, and things have effects in many places where you wouldn't expect them. Sexual hormones affect a lot of things not related to sex. This is a medical fact.
posted by Steven Den Beste at 5:37 PM on February 25, 2001

>This is a medical fact.

Impossible-- I don't like its implications.
posted by WalleySegap at 5:55 PM on February 25, 2001

Steven: I asked about your age because I assumed from your posts that you were about 16. And then you implied you were in you 60's (I think I missread another post). Thanks for clarifying.
If I were even remotely normal (not sure that I am) and I lived in a society where masculinity was associated with a given organ, removal of that organ would certainly depress me and would probably have other psychological effects.
posted by davidgentle at 7:44 PM on February 25, 2001

I still think you're joing about that gravity thing. ;)
posted by rodii at 8:44 PM on February 25, 2001

The depression is not caused by knowledge; it's a very real biochemical effect directly caused by the lack of testosterone. The point is that testosterone is necessary for the proper functioning of a man's brain, though no-one knows why.

But women's brains work fine with almost no testosterone at all. Clearly there's a difference between them.

However, the difference is not genetic, directly. There exists a rather strange and extremely rare condition where a chromosomal male has a mutation such that the testosterone receptors in the body don't work -- and such an individual, with an X and a Y chromosome, develops with the body pattern of a woman. Indeed, the reality of such a person's condition only becomes apparent when it becomes obvious that they're not fertile. (I don't know if they have periods. It seems unlikely.) But from direct external examination of such a person's body, you can't tell that it's not the body of a human female. And such people's brains work properly without testosterone (which is never in great supply because there are no testicles, and which in any case has no effect because of malfunctioning receptors). People with this condition are legally considered women, despite the Y chromosome.

The changes in a man's brain are caused by exposure to large quantities of testosterone at puberty (when the blood level rises by a factor of about , but once that happens the changes are real and permanent, and his brain requires a high blood level thereafter to operate properly. This has been established beyond all doubt. It's part of normal clinical practice and endocrinologists and psychiatrists deal with it all the time; it's been such for a couple of decades.

I don't understand why this is so difficult to believe.
posted by Steven Den Beste at 9:25 PM on February 25, 2001

Because we just don't understand how the brain works well enough to be so certain.
So what you're saying is that men are effectively testosterone addicts? So what function does the testosterone serve once it's insinuated itself into the brain?
posted by davidgentle at 9:50 PM on February 25, 2001

rodii, each one of those charts places "your" result in a cloud chart. I don't think, with ~50 questions, that there are 10 possible locations in that cloud chart. In other words, I think the cloud chart is so much smoke.

Unless ... the final screen says that they've determined certain things from the questionnaire, e.g. that men are more likely to term the test retarded. If they are only using certain questions to determine the result, rather than all ~50 ... i.e. burying the real questions amongst a lot of chaff.
posted by dhartung at 9:54 PM on February 25, 2001

David, evolutionary systems are not as easily susceptible to "why" analyses because there's no guiding intelligence behind the design. Things happen because they work, whether they make sense or not. Testosterone was already present in men (and other adult male mammals) and apparently some mutation back there took advantage of it to do, well, something -- since it was handy.

Exactly what testosterone does to a man's brain is unknown, just as a thousand other aspects of the brain are unknown. The human brain is the most complex structure known to us and the least well understood. They haven't even solved the problem of what memory is yet, let alone what thought is. (One wag said "If the human brain was simple enough for us to understand, we'd be too simple to understand it.")

Treatment of mental illness is as yet far more primitive than most people realize. Most modern psychiatric treatments for depression, for instance, work by influencing the background level of the monoamine neurotransmitters (serotonin, norepinephrine, dopamine), but no-one knows why that's important. No-one has the slightest idea how any of the antidepressants work, or why they don't work for everyone. (But that's not surprising, since no-one knows what causes depression. Indeed, by figuring out what these drugs do, it may be possible to figure out what depression really is.)

However, it's not necessary to explain something to be certain that it's true. Evidence doesn't require explanation. The antidepressants do work for many people even though no-one knows how. Equally, testosterone unquestionably has effects on adult male brains, but no-one knows exactly what the effects are in detail.

Truth of some statements can be demonstrated absent any explanation of a reason why. This dependency of adult male brains on testosterone is a well-established fact. No-one has the slightest idea of what the actual effect is, though, except for a clinical description of the effects caused by its removal.
posted by Steven Den Beste at 10:13 PM on February 25, 2001

i think it's pretty obvious that not all the questions are for real. "are clams alive?"

davidgentle and other nonbelievers: i'm with steven on this one. and why in the name of god did you think he was 16? he's one of the only really informative posters on here. i don't always agree with him, but at least he backs himself up instead of just stating opinion and getting defensive.
posted by pikachulolita at 10:38 PM on February 25, 2001

Steven, I have a question:

How do you know that the testosterone levels only effect men? If a man loses a testicle then his testosterone drops and he becomes depressed etc...but if a woman's small amount of testosterone were to drop how do you know she wouldn't be victim to the same effects, or at least similar ones? Is it really proof that men and women are biologically different or is it just proof as to what testosterone controls? (and no, I'm not attempting to say that men and women are the exact same biologically.)
posted by crushed at 11:08 PM on February 25, 2001

Testosterone in a human body is created by the adrenal glands in everyone and by the testicles in men (but not boys). When a man's natural supply from his testicles goes away (for any of a number of reasons, not all of which involve physical removal), his testosterone level drops to the level maintained by the adrenal glands, which is the level present in women, girls and boys -- and that's when he suffers ill effects. The level doesn't drop to zero, but it drops by a factor of something like 50. As to whether testosterone affects the brains of non-men, I was extremely careful to not say that, because I don't know and I don't think anyone else knows.

Both testosterone and estrogen are present in at least small quantities in boys, girls, men and women and all of them need both. Estrogen (specifically, estradiol) is also created in small quantities by the adrenal glands. In boys, girls and men their total supply comes from the adrenal glands (which, interestingly, produce it by conversion of testosterone, which is produced from something called DHEA). Of course, in men testosterone is also created in much larger quantities by the testicles and in women estrogen is created in much larger quantities by the ovaries. But the big difference is that a woman's ovaries stop producing estrogen in large quantities at menopause, whereas a man's testicles continue to produce testosterone as long as he's alive and healthy.

What is known about the relationship of testosterone to brains is that men's brains depend on high levels of testosterone, and that when a man's serum testosterone level drops to that of a normal healthy woman (about 2% of what's healthy in a man) then his brain starts to malfunction.

Losing function of a single testicle doesn't change the serum testosterone level by a factor of two, if the other one continues to work. Indeed, it hardly affects it at all in a man. The reason is that testosterone level is maintained by a feedback loop. The hypothalamus (a part of the brain) monitors the testosterone level in a man, and generates something called "Gonadotropin Releasing Hormone" (GNRH) with an inverse relationship to the testosterone level. GNRH is sensed by the pituitary, and in a linear relationship produces Follicle Stimulating Hormone (FSH) and Leutinizing Hormone (LH). FSH is sensed by the testicles and causes production of sperm. LH is sensed by the testicles and causes production of testosterone. (LH and FSH also figure highly in the control of a woman's system. FSH causes ovulation; LH causes menstruation. In a man they're present all the time; in a woman only about ten days out of every thirty.)

If the level of testosterone in a man drops due to loss of one testicle, then the hypothalamus generates more GNRH, which causes the pituitary to generate more LH, which causes the remaining testicle to create more testosterone than one would ordinarily create, and the level remains about the same. (Which is why one cause of hypogonadism in a man is brain damage; if the pituitary malfunctions then no LH or FSH is created, and a man's healthy testicles will shut down.)

On the other hand, the adrenal glands don't pay attention to LH; they just generate low levels of both testosterone and estrogen at all times, in everyone. But though that level is sufficient to maintain health in girls, boys and women it is not enough for either physical or mental health in a man.
posted by Steven Den Beste at 11:45 PM on February 25, 2001

Why is it so hard to contemplate the possibility that men and women are different? Whether Steven's extensive medical knowledge is correct or not, just by observing the real world, we can make fairly accurate generalizations. Women are better nurturers; men are more aggressive physically; women prefer cooperation; men prefer competition. Let me point out that these are just generalizations. They don't apply to every individual, because we are all unique.

If you don't think men and women have different natures, then why do we have such problems with relationships?
posted by shackbar at 12:51 AM on February 26, 2001

shackbar, I personally think that yes men and women have different natures in a few large categories. They are most definitely different biologically, at least in several ways,and to say otherwise, in my opinion, is fairly ignorant. To say that that's the reason for most relationship problems is a whole other topic that you don't want to get me started on. :)


Steven, thank you for your very well thought out and obviously well researched reply. I like seeing MeFi comments that are high in fact content as well as when someone can tell you something without just linking to it. Women don't all have the same amount of testosterone or estrogen, the same goes for men. Since women, girls, and especially boys function fine with low amounts of testosterone, I still have to wonder if it's merely a reaction to a loss of the testosterone that your body has become used to, therefore possibly experienced by both sexes, or if indeed men just need high amounts of testosterone to function. Either way it's definitely worth looking into.

But another question: FSH causes ovulation; LH causes menstruation. In a man they're present all the time; in a woman only about ten days out of every thirty Does that mean it's true that women with PMS are actually just acting like men do all the time? ;)
posted by crushed at 3:24 AM on February 26, 2001

I was debating whether to mention PMS, but decided not to because I don't really have any direct knowledge of it. No girl friend I ever had suffered from it.

I suspect PMS is not a reaction to FSH and LH, but rather a reaction to estrogen deprivation. Another difference between men and women is that while lots of testosterone is present all the time in a healthy man, large quantities of estrogen are only present in a woman part of the time. That's because of the interplay between the hypothalamus in a woman and how her ovaries work. (I could go into more detail but it would be another long essay.) During the interval of a few days just before menstruation, estrogen levels in a woman drops to the levels present in boys, girls and men (that is, the level produced by the adrenal glands). A woman's hormone levels are controlled more-or-less digitally, and hormone levels in a woman plotted over a couple of months are step functions, more or less square waves with mushy transitions and non-equal duty cycles. Estrogen in a woman is on about 20 days and off about 10; the "on" transition takes maybe a couple of hours and the "off" transition maybe a day.

The feedback mechanism in a woman runs "to the rail", saturating on each swing, because estrogen production in a woman doesn't directly respond to FSH or LH. Whereas testosterone control in a man is analog, trying to maintain a steady level all the time with proportional feedback. (Not quite; it varies daily by a factor of two with the peak being in the morning and the trough being in the evening. It swings from perhaps 60 times a woman's level to 30. However, it varies on a continuous curve and not as a square wave. No-one knows why, and given that some augmentation therapies cause flat-line blood levels with no apparent problems it may in fact not be important at all.)

The only time a woman naturally is exposed to continuous estrogen is during pregnancy; the reason is that one of the primary purposes of estrogen is to suppress ovulation.

So even before menopause women are not accustomed to continuous presence of any of their sex hormones in large quantities except rarely (pregnancy). As such, it doesn't have the "It was just lying there so I used it" attraction to evolution that testosterone has for men.
posted by Steven Den Beste at 6:09 AM on February 26, 2001

Dan--I think they are almost certainly using *most* of the questions as distractors.

The "cloud chart" is really just a one-dimensional plot, with the y-axis added as "jitter" so all the dots aren't on top of one another (actually, looking at it, there's probably some x-axis jitter too). Visually, though, it looks like a pretty good bimodal distribution composed of two normal distributions with distinct means, one for women and one for men. If that's true, that's all that needed for this test to "work" fine. It doesn't matter, past some fairly low number, how many questions you've got, as long as you have a big enough sample. Even that doesn't have to be very large.

If your data is (roughly) normal, you can characterize any point in terms of its Z-score--how many standard deviations from the mean it is (Z = [score - mean]/SD). Each value of Z corresponds to a certain percentage of the sample that is closer to the mean. A case with a Z of 2, for example, would mean that 99% (or something) of the men (for example) scored lower (more toward the mushy middle). So I'm guessing that the 86%, 90%, etc. figures are just the result of picking some standard set of Z-scores (or maybe quartiles of some other equivalent measure) to report how far out on the tail of the distribution your score is.

The question isn't whether the test is valid--that is, whether it measures anything "real" so much as whether it's reliable. Out of the 50 questions, as more and more men and women answer them and then identify themselves as men or women, we'll find that some question will emerge as more or less reliable predictors. The unreliable ones become distractors, and the reliable ones are the ones that are actually scored. (The distractors are in there to keep subjects from second-guessing the "right" answers, which would undermine the test, as you suspect. "Are clams alive?" probably isn't counted in scoring. Even if there are only 10 questions that are actually counted, that gives us 11 possible distinct scores, more than enough to establish a normal distribution.)

So the likelihood scores probably don't reflect the number of "bins" or possible scores directly at all--they're derived from the Z-score or quartile score, which represents an idealized (normalized) picture of the scores as a normal distribution. This is what happens when a (statistically competent) professor grades a test "on the curve".

Anyway, using your 11 "bins" you can test to see if there are significant differences between the two populations (using a t-test or something similar). If there are, that's all you need to make predictions. For any score, you can ascribe a definite level of confidence to a prediction that they are a man and another level of confidence that they are a woman. If a score is to the left or right of the two means, then the likelihood that they are a man (for example) is much higher (thus 86%, 90%, etc.). If the score is in the valley between the two means/modes, than the likelihoods are close to equal (thus 50%). A score of 50% doesn't mean "half man, half woman" or "girly-man", it just means that score is about equally likely to be a man or a woman.
posted by rodii at 6:36 AM on February 26, 2001

It's not clear that the distractors have the purpose you describe. Going in, the people creating the test may not have known which questions would be important and which not. As you say, the test is self-tuning.
posted by Steven Den Beste at 7:06 AM on February 26, 2001

Hm? That's what the purpose of distractors *is*. Anyone designing a test like this would include distractors, so I would guess that they deliberately threw in some questions that they knew wouldn't work as predictors. I'd be surprised if "Are clams alive?" was *ever* expected to predict anything.

I'd like to know more about the self-tuningness. It's also possible that they're letting the weights evolve, neural-network style, though that would make any validation trickier, I would think. You *could* do a test like this by simply throwing in lots of random questions and training it based on people's responses to the gender question, but I bet the confidence figures would be a lot harder to come by. Of course, as Dan says, they might be more or less made up anyway.
posted by rodii at 8:15 AM on February 26, 2001

I think what happens is that the engine checks the answers for each question for correlation to the result, and assigns a weight to it based on how strong its correlation is. Probably they rerun the calculation on a regular basis (i.e. once a day) rather than doing so on each individual visit. That's what I meant by "self tuning". What that would mean is that those who set up the questions ahead of time didn't necessarily actually know which questions would turn out to be significant (though they may have had suspicions). Such simple things as the difference between "yes" and "yeah" might actually have turned out to be important. It's even entirely possible that questions that they didn't think would be significant turned out to be.
posted by Steven Den Beste at 8:27 AM on February 26, 2001

Shackbar: if male/female relationship problems are caused by the differences between men and women then surely all homosexual relationships would be carefree love-ins without any of the same problems. Is that the case?
posted by davidgentle at 4:55 PM on February 26, 2001

Why is everyone getting all analytical about a test on TheSpark? Am I missing something here? You did read the names of their other tests and quizzes right? For instance oh I don't know...The Pregnancy Test?!
posted by crushed at 11:53 PM on February 26, 2001

Not to mention the fact that these are the wonderful people behind the Stinky Meat projects, the Fat Project and the Stinky Feet Project (their archive).

It would be interesting to revisit this in two months or so, after it's been reasonably thoroughly exhausted by the e-maillers and webloggers, and see how it rescores everyone.

I'm a mid-60% man, in case anyone cares.
posted by cCranium at 5:48 AM on February 27, 2001

86% BABY!!!!

Yes, yes, yes!!!!

Oh. 86% man, that is.

Yes, yes, yes!!!!
posted by sonofsamiam at 8:20 AM on February 27, 2001

I would like to know what it takes to get 100% man on this thing. Apparently I have a few womany things in my make up that I need to eradicate.

I have been comparing answers with my wife, who tested 86% woman, and from what we can remember, the only question we answered differently, was my desire to fall to my death vs. her preference to drown.

I just remembered I answered 2,4,6, which now seems very feminine compared to the manly 2,4,8 choice. I femmed on math :(

posted by thirteen at 8:44 AM on February 27, 2001

Gah! Who would want to drown to death? The falling would suck, but at least the dying would be done with quick. The last thing I want to spend my last 5 minutes doing is thinking "Oh crap, I need air, oh crap my lungs hurt oh crap it hurts it hurts it hurts!"

I'd much rather spend it crying "ohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgod Hey I'm flying! ohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgodohgod*SPLAT*"
posted by cCranium at 11:24 AM on February 27, 2001

I'll just wander over to the corner now. I must remember to allow for line breaks in my text.

sigh.
posted by cCranium at 11:25 AM on February 27, 2001