Science desk
< February 20	<< Jan \| February \| Mar >>	February 22 >

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

February 21

Green cones in the perception of colour

Hey, I was looking at this picture:

and it occured to me that we never really see green without quite a lot of red, and so we must be really sensitive to the difference between what our red and green cones are picking up. The purest green we'll ever see still looks like it'll maybe be 55% green to 45% red.

So I was wondering what it would be like to artificially stimulate only your green cones. This would essentially mean receiving RGB data in a ratio that you'd never experienced before, greener than the green. Would the experience of this be basically the same as seeing a normal bright green, or would one be able to tell that this is a colour they had never seen?

Thanks, Luke —Preceding unsigned comment added by 82.35.84.214 (talk) 01:00, 21 February 2010 (UTC)[reply]

I don't know what the result of that would be (or any way to do the experiment, actually), but it isn't easy to predict, because we don't "see" the raw output of the cones -- the neural activity is quite substantially transformed and recombined before it reaches the cerebral cortex, and then transformed even more inside the cortex. Along the way, the three-color cone activity pattern gets turned into an opponent process representation for which the primary hues are red, green, yellow, and blue. Looie496 (talk) 02:18, 21 February 2010 (UTC)[reply]

It's an interesting thought - what would happen if you got a green stimulus without any red stimulus at all? Certainly it's a bit of a guess because our brains are 'wired' to see things that way. Some colorblind people have no red sensors at all - so they are seeing what you're proposing. But the trouble is that their brains have formed since babyhood without that red sensor - so they don't know any different. The best answer is "We can't possibly guess without doing the experiment". If I had to guess, I think the result would be exactly what we see when we've been staring at a pure 625nm-ish red surface for a LONG time - then quickly glance at a green surface. Staring at the red surface will cause Neural adaptation which will make the brain ignore the "red" signal for a while. Then, when you look at a yellow or green surface, the red sensors are dumbed down and all you're registering is the green sensors. This is the basis of various Afterimage illusions. All I see when I do that is a really brilliant lime-green. But we have to be careful - we can't be 100% sure about the comparability of the result of this practical test with the impractical one you're thinking about. SteveBaker (talk) 02:46, 21 February 2010 (UTC)[reply]

Right. The color stimuli that correspond to impossible combinations of the three cone type activations are called Imaginary colors. Indeed, you can use the fact that human color adaptation is not instant, and have the L-cone gain reduced by strong input at 650-700 nm wavelength. Then, prompt preferential stimulation of the M-cones at around 550 nm should produce a "greener than green" color percept. However, I doubt the effect is going to be too strong, as V1 and higher visual areas have multiple negative feedback loops, making a highly unusual response next-to-impossible. Reducing this inhibition pharmacologically may actually produce an "impossible color" sensations via this mechanism; I think this was indeed reported as one of the LSD effects. Needless to say, any unauthorized use of controlled substances is VERY strongly discouraged. --Dr Dima (talk) 03:32, 21 February 2010 (UTC)[reply]

To be entirely accurate, LSD does not directly reduce inhibition. Still, it increases the excitability of the sensory areas of the cortex. --Dr Dima (talk) 03:40, 21 February 2010 (UTC)[reply]

The color you would see would not be green. Green is the color that is "55% green to 45% red". If you did only green cones you would have some other color, but not green. You have to remember our color is calibrated by the light we actually see, not by the labels we give the color sensors. Ariel. (talk) 04:33, 21 February 2010 (UTC)[reply]

Indeed. It is best not to refer to the different types of cones by colour, but by "short", "medium" and "long" (as they are labelled in that picture) to avoid confusion. --Tango (talk) 06:03, 21 February 2010 (UTC)[reply]

I think this has a straightforward answear. A situation where only the green cones are stimulated is closest to a situation where the green cones are stimulated the most and the blue and red cones are stimulated only a little and with equal intensity. From the figure it would seem that percentagewise, green is stimulated the most around 500nm, which corresponds to sea-green or blue-green according to a picture in the color vision article. EverGreg (talk) 13:15, 21 February 2010 (UTC)[reply]

Unfortunately "equal intensity" on that diagram is meaningless because the curves are arbitrarily normalized to have the same peak value. You would need to instead normalize them to have the same inner product with a particular white, such as D65, then look for equal heights on that graph. I think the easiest way to guess the effect of stimulating a particular cone alone is to convert (1,0,0) or (0,1,0) or (0,0,1) in an LMS color space to XYZ using one of the matrices in that article, then convert to linear sRGB, then add white to make everything positive, then convert to nonlinear sRGB. Using the Hunt-Pointer-Estevez matrix, I get these hues: L alone, M alone, S alone. Using the Bradford matrix, I get these hues: L alone, M alone, S alone. I think that Hunt-Pointer-Estevez is more accurate. Stimulating the S cone alone should produce violet or magenta, not blue, unless the L cone has a second sensitivity peak in the violet, which as far as I know it does not. -- BenRG (talk) 00:55, 23 February 2010 (UTC)[reply]

Maybe I'm just looking at this too late at night, but isn't the answer 'a less bright green'? The curves in this diagram don't represent colors, but the sensitivity of the cones to certain wavelengths. A pure 'green' wavelength (say 550nm) is easily created these days. Light of this wavelength entering a normal eye (both S, M and L cones functioning) gives the owner the impression of "wow, that's really a bright green light". When all the L-cones are suddenly turned of in some science-fiction manner and only the M-cones working (S-cones being insensitive to this light), this means that only half of the normally receiving cones will be receiving this specific wavelength, thus the visual input of this -and at that moment only- wavelength (to the not yet adapted eye) will be reduced by 50% (in case of an equal distribution of M- and L-cones), giving the perception of a "hey, who turned down that bright green light by 50%". But then again, maybe it IS to late right now...190.98.50.179 (talk) 03:45, 23 February 2010 (UTC)[reply]

No, that's not right. The cones can't tell what wavelength the light hitting them has, it just gives the brain a single value that depends on the wavelength and the intensity. You can only separate those two factors by comparing the responses from different cones. --Tango (talk) 18:22, 24 February 2010 (UTC)[reply]

Spectroscopy spectrum order

Okay, so I'm doing a lab right now on spectroscopy, and I'm given this equation: dsinθ=mλ. m is defined in the lab manual as "a positive integer equal to the order of the spectrum", but I am having trouble finding anything having to do with the order or how to know what value it may be.--十八 04:40, 21 February 2010 (UTC)[reply]

A light bulb of a flashlight seen through a transmissive grating, showing three diffracted orders. The order m = 0 corresponds to a direct transmission of light through the grating. In the first positive order (m = +1), colors with increasing wavelengths (from blue to red) are diffracted at increasing angles. Cuddlyable3 (talk) 23:14, 21 February 2010 (UTC)[reply]

We're missing a lot of context about what kind of spectroscopy you're doing, so I'll answer in general. The idea is that as you go along, the the paths taken by two waves vary in length, and therefore the phase difference between them changes. Say they start off in-phase; as the path-length difference increases, they get progressively more out-of-phase and then they wind up coming back into phase again, and then back out-of-phase, and then back into phase, and so on. The "order" talks about how many times that happens: at the lowest m they are very close (m=0 is the starting in-phase phase); at higher m, they have gone through many complete cycles and beyond. DMacks (talk) 04:53, 21 February 2010 (UTC)[reply]

I'm doing visible spectroscopy.--十八 06:16, 21 February 2010 (UTC)[reply]

Are you using a diffraction grating, and is our article helpful? TenOfAllTrades(talk) 15:00, 21 February 2010 (UTC)[reply]

how does magnet works in space?????????

hi,i am one of your member and i want to know , how permanent magnet works in space ? I will be waiting for your answer. —Preceding unsigned comment added by Debiprasadnayak (talk • contribs) 06:14, 21 February 2010 (UTC)[reply]

A magnetism works through the vacuum of outer space just like it works in air. Magnetism is like light; it is not transmitted via real particles or pressure the way sound is. DMacks (talk) 06:17, 21 February 2010 (UTC)[reply]

Light IS transmitted via real "particles" though... --antilived^{T | C | G} 11:02, 21 February 2010 (UTC)[reply]

No it's transmitted by waves. No, wait, it's particles after all! Actually it's neither, it's something else that sometimes appears particle-like, and sometimes wave-like, depending on how you measure it. 87.81.230.195 (talk) 20:26, 21 February 2010 (UTC)[reply]

Magnetism, the subject of the actual question, however, is transmitted by virtual particles, so DMack's is correct where it matters. --Tango (talk) 00:17, 22 February 2010 (UTC)[reply]

Coed Olympic Events

I just read the informative article on Olympic Sports but it does not cover the origin of the gender division, nor whether there have ever been any challenges to it. It occurred to me last night that there are more than a few Olympic events that could be fairly contested by both men & women together. In most shooting events strength plays only a very minor role, if not none at all (air pistols). I also suspect that gender would not imbalance a curling tournament, though perhaps the men could grind out a little extra bend to their throws?

Q1: Has there ever been a movement to unify some of the Olympic events?
Q2: In which events would gender impart very little or no advantage?

This question could have gone on Entertainment, but I find Q2 the more interesting and so I post it here. Thank you. 218.25.32.210 (talk) 07:15, 21 February 2010 (UTC)[reply]

Check out the equestrian events. B00P (talk) 08:25, 21 February 2010 (UTC)[reply]

Rightly or wrongly, the commentator on the US vs UK Women's Curling competition last night said something like: If this were the men's competition then this shot wouldn't have been a subject of debate - but it's a measure of how far Women's curling has come that they are able to even contemplate taking it" (after which the shot was taken and messed up). This suggests that the women's teams are getting closer to the men's game - but that they aren't there yet. I agree that there is nothing obvious in the biology of men and women to explain why women and men need to have separate games (unlike...say...weight-lifting). But the effect of merging the two events would undoubtedly to simply exclude women from Olympic curling. That would probably have a devastating effect on the number of young girls who see a female olympic curler and decide to take up the sport - and might well (effectively) keep women out of the competition forever. So I guess the Olympic committees prefer to keep male and female events separate until both genders are at about the same level. Is this "right"? Is it "fair"? Is it "sexist"? I don't know - it's like positive racial descrimination in college admissions - no matter what you do, there will be winners and losers. There are good arguments on both sides and no clear moral high-ground. If there were big protests amongst women curlers about being excluded from the men's game - then perhaps we'd be doing this differently. SteveBaker (talk) 16:12, 21 February 2010 (UTC)[reply]

Probably not what the OP means but it was a good play. Cuddlyable3 (talk) 17:00, 21 February 2010 (UTC)[reply]

I will note that there may be a preference for maintaining separate men's and women's events even where men and women are playing at comparable skill levels. This would be purely because it doubles the number of athletes who can participate, the number of events that can be televised, the number of medals to be won, and (potentially) the number of sponsorships and endorsement contracts available.

On the subject of curling, I will note that men (on average) are bigger and stronger than women. While this should make no difference for the majority of shots (which depend heavily on touch and finesse, rather than raw power), it will have an effect on a few of the high-power, high-risk shots (double peels, multiple raise takeouts, etc.) where a large amount of power needs to be finely controlled. I suspect that the biggest difference probably comes in sweeping, however. Bigger, heavier, stronger men are going to be able to put more weight on the brooms, and are going to be able to affect the course of a moving stone more effectively. TenOfAllTrades(talk) 18:07, 21 February 2010 (UTC)[reply]

There are still sex-related differences in areas like spatial cognition Sex-related differences in spatial cognition. Therefore even if a sport doesn't favour one sex on physical grounds, it may favour one sex through mental differences. Curling and shooting would both seem to require the kind of spatial cognition that would favour males. 82.132.248.88 (talk) 00:01, 22 February 2010 (UTC)[reply]

Litmus paper

how does Litmus paper or Universal indicator work. what chemicals are on the paper? are these harmful? and where can i buy the paper? —Preceding unsigned comment added by 67.246.254.35 (talk) 08:16, 21 February 2010 (UTC)[reply]

Have you tried reading the article on litmus paper? It is harmless (but obviously you wouldn't want to eat it). It is easy to buy on line.--Shantavira|^{feed me} 08:46, 21 February 2010 (UTC)[reply]

whats in it thou? —Preceding unsigned comment added by 67.246.254.35 (talk) 08:58, 21 February 2010 (UTC)[reply]

litmus is extracted from lichens(a symbiosis of algae and fungi). its really a solution of purple color and changes to red when in contact with acid . and turn to blue when in contact with base. if u just want a acid base indicator you can use turmeric it turns red with base( try it on soap water) and shows no difference for acid

most probably litmus when in contact with H⁺ ions give red color and when in contact with OH^- it gives blue.--Myownid420 (talk) 09:26, 21 February 2010 (UTC)[reply]

As Shantavira hinted at, what exactly from the article requires further clarification? Nil Einne (talk) 11:49, 21 February 2010 (UTC)[reply]

Here's a picture of Roccella tinctoria used to make Litmus paper. The article São Jorge Island under "Economy" mentions a possible source of this lichen. Cuddlyable3 (talk) 16:55, 21 February 2010 (UTC)[reply]

As the litmus test article indicates, the compound that changes color with a pH change is 7-hydroxyphenoxazone (pictured in article). It works like all other pH indicators - that is, it's a compound where the conjugate acid and the conjugate base have different colors (so it changes color around the pKa when the molecule gains or loses a proton). And the two have different colors because the presence or absence of a proton (hydrogen ion) changes the electronics of the conjugated system, perturbing the energy levels of the molecular orbitals, altering the energy difference between them, and thus the wavelength (color) of the photons the compound can absorb. -- 174.21.254.47 (talk) 18:37, 21 February 2010 (UTC)[reply]

field lines of magnet

lets assume a straight conductor studded in a square cardboard,which is parallel to the ground, If xA current flows through conductor than the concentric field lines formed are at distance of m cm from each other. now current yA flows though the conductor, Such that yA > xA, and dist between circles is n cm. so will n = m or n>m or n<m? in simple words if current is increased in a conductor the field lines get denser or distance between them remains same?--Myownid420 (talk) 09:41, 21 February 2010 (UTC)[reply]

First you need to know that the magnetic field lines fill the whole space so it is not meaningfull to talk about the distance between them. But conventionally when drawing them they're spaced inversely proportional to the magnetic field intensity. The magnetic field intensity around a wire drops with the distance from the wire so you really shouldn't draw equaly space concentric field lines. Dauto (talk) 15:00, 21 February 2010 (UTC)[reply]

Those "field lines" are not things that really exist. Think of them as the contour lines on a map of a mountain - when you visit the actual mountain, there are no lines to be seen! The spacing between the lines is something that the map-maker (or the magnetic-field-illustrator) chooses to make the illustration be informative. As Dauto points out, if the field lines were intended to represent fixed intervals of magnetic intensity (like the 100' foot contours on a map that are equally spaced in height) then they wouldn't be uniformly spaced in the first place. If they are uniformly spaced then they must represent varying 'step' in magnetic field intensity. SteveBaker (talk) 15:55, 21 February 2010 (UTC)[reply]

The title is about a magnet but the question is about the magnetic field of a straight conductor. The cardboard and the ground seem to be unnecessary parts of the question. The magnetic field lines are indeed concentric with the wire but as Dauto pointed out the lines are just a way of showing the direction of the field which actually fills the whole space (it is a Vector field). As one moves away from the wire the distance between the lines should increase because the field gets weaker. To the question: Increasing the current by the factor y/a increases the magnetic field at a given point by the same factor so the line spacing gets less, n<m. You may like the article Field line but note that its top diagram shows electric field lines not magnetic ones. Cuddlyable3 (talk) 16:38, 21 February 2010 (UTC)[reply]

Michael Faraday was a 19th century genius and an early experimenter in electricity and magnetism. He had little or no mathematical ability. He used "lines of magnetic force" to create theories of electromagnetism. James Clerk Maxwell was another 19th century genius who theorized about electricity and magnetism, and who was a gifted mathematician. Maxwell considered at length Faraday's writings and then said that their views of electromagnetic fields agreed 100%. "Field lines" do not exist. The field exists everywhere, and not just in discrete "lines." But they are a useful aid to gaining a physical understanding of magnetism and electromagnetism. In the crude model of "field lines," greater current would result in more densely spaced field lines, which would correspond to a stronger magnetic field. If you consider a permanent magnet in the form of a horseshoe or bar, with a north and south pole, the lines of force leave the north pole and end at the south pole. They are densely clustered at the poles, and some take a short path between the poles of the horseshoe, while others branch out to cover the entire plane, getting farther apart as you move away from the poles. The strength of the magnetic field, as seen in its effect on a current carrying wire or one pole of a long bar magnet, varies with the density of the fictitious "field lines." When iron filings are sprinkled over a pane of glass or a sheet of cardboard over a bar or horseshoe magnet or an electromagnet, the filings tend to align like the fictitious "lines of force," but there is a bit of hand waving in the demo. Each iron filing becomes by induction a little magnet itself, and they attract each other. Tapping on the glass or cardboard causes the "visible lines" to shift closer to the magnet's poles, so the "lines" are not discrete in space at all. "Electric field lines" can be modelled in a shallow tray of water between a DC + and - terminal just like the magnetic "field lines", with a voltmeter probe tracing out equal potential contours, but the electric field exists everywhere in the tray, and not just on the "field lines." More voltage between the + and - terminals in the tray would create more "one volt contours" just as more current in a wire creates denser "magnetic field lines." Edison (talk) 05:16, 22 February 2010 (UTC)[reply]

No Edison, you are describing equipotential contours not electric field lines. Electric field lines such as these show the direction of force on an isolated positive charged Test particle. Mathematically electric field lines are the Gradient of the scalar electric potential field and they lie at right angles to your measured contours. ~~BTW you won't measure anything in pure water~~. Cuddlyable3 (talk) 19:24, 22 February 2010 (UTC)[reply]

Thanks for the correction. This lab writeup describes the procedure, and in fact calls for "deionized water", which would not work in your understanding. Tap water should also work fine. They use low voltage AC. Edison (talk) 20:20, 22 February 2010 (UTC)[reply]

Thank you for the Lab writeup link. Electrical_conductivity quotes the conductivity of deionised water as 5.5E-6 S/m so I grant that a "water potentiometer" should work with a very high impedance probe. Interestingly, deionised water acts as an insulator in this X-ray generator. I agree that the probing experiment should work as well with tap water or even better with an electrolyte. However the writer seems very concerned about corrosion of the copper electrodes. I criticise the writeup for implying in its introduction that there is a conducting medium between the plates of a capacitor. It is not clear to me how the "dual probe assembly" that is not shown is able to determine the direction of the field of an AC source. Cuddlyable3 (talk) 21:40, 23 February 2010 (UTC)[reply]

Stars ejected from the Milky Way

Are there any known stars that are no longer part of the Milky Way but are just floating away in extragalactic space? TheFutureAwaits (talk) 17:54, 21 February 2010 (UTC)[reply]

This section of our Milky Way article describes some globular clusters that are a long way outside of the galactic disk which may be leaving the galaxy. Also, it is thought that the Canis Major Dwarf Galaxy is close enough to the Milky Way to be ripped apart by it - and it would be surprising if some of the Milky Way's stars were not being pulled out in the opposite direction. But at these great distances, we aren't really looking so much at individual stars as larger groups like globular clusters. So I doubt we could point to a particular "known star" and say "Ooops! We lost that one!". Also, this stuff happens on truly spectacular time-scales - so it's not like we can see it happening. SteveBaker (talk) 18:17, 21 February 2010 (UTC)[reply]

See hypervelocity star. According to that article (or rather section), 10 are currently known. Looie496 (talk) 18:23, 21 February 2010 (UTC)[reply]

There are nine stars in Category:Extragalactic stars. TenOfAllTrades(talk) 18:39, 21 February 2010 (UTC) ...though most of those aren't free-flyers, but are in the Large Magellanic Cloud.... TenOfAllTrades(talk) 18:43, 21 February 2010 (UTC) [reply]

nutrition labels

How would a nutrition label for the average human body (male and female) be broken down and for the normal distribution of (male and female) humans? No, I am not planning to join a cannibal cabal but just want to know what nutrients the human body already contains so I can print a human nutrition label on a t-shirt to give people who look at me in that sort of way a printed answer to their question. 71.100.3.6 (talk) 19:44, 21 February 2010 (UTC)[reply]

The nutritional content of the edible parts of a human would not be significantly different than what you would find on the nutritional label for other types of meat (say pork, for example). So you can just find one of those and copy it - though some things (like fat content) are going to vary from one individual to the next. There would be no significant difference between a male and female body, though serving size may vary depending on the consumer's predilections. —Preceding unsigned comment added by 72.94.164.21 (talk) 21:25, 21 February 2010 (UTC)[reply]

Females have a higher fraction of fat on average (some tables), but you could personalize the label by bioelectrical impedance analysis. Icek (talk) 21:43, 21 February 2010 (UTC)[reply]

Wikipedia has an article on Cannibalism which does not give a nutritional breakdown. It would not be a good idea to wear the OP's T-shirt in North Korea where some people reportedly do look at one in that sort of way. Cuddlyable3 (talk) 23:06, 21 February 2010 (UTC)[reply]

Actually I do not sympathize with cannibals since it is very easy to have a complete and balanced diet with all of the calories you need from items you might not think of. All that is required is that you know the nutrients the items have, such as a nutrient label provides. While such things as snails and slugs might not be on your normal breakfast menu, as with the information provided by any nutrition label, if you had one for them a meal could be prepared to fulfill your every nutritional need simply through using the right proportion. Instructables: Meal in a Cup 71.100.3.6 (talk) 23:31, 21 February 2010 (UTC)[reply]

Here is an example of data from one such source... Mollusks and Snails as Shellfish 71.100.5.197 (talk) 03:41, 22 February 2010 (UTC)[reply]

research methods

A society has a very high gini coefficient, a high level of social inequality. The same society has a very high level of HIV infection. The hypothesis could be that social inequality 'causes' or exacerbates HIV infection, or conversely that high HIV infection leads to social inequality. We're trying to avoid a fundamental methodological mistake...is it possible that both can be true at the same time, or does scientific method assume that only one of these hypothesis can actually be 'true'? —Preceding unsigned comment added by 83.98.238.113 (talk) 20:02, 21 February 2010 (UTC)[reply]

What of the possibility that neither are true? 87.81.230.195 (talk) 20:17, 21 February 2010 (UTC)[reply]

I have not considered this specific case in detail, but our article on Positive feedback may be relevant. You may also wish to consider that the causes and effects in this situation (or indeed any situation) may not be constant over time.131.111.185.68 (talk) 20:28, 21 February 2010 (UTC)[reply]

They can certainly both be true—a feedback situation would mean that each exacerbates the other dynamically. Or, put another way, HIV exacerbates social inequality, and social inequality exacerbates HIV transmission. There is no reason to expect a complex socio/epidemiological issue to be monocausal, and nothing about the scientific method suggests it has to just be one way or the other a priori. It's also possible that neither are true, that the correlations you see do not imply any causation. (Classic example: in the summer, crime rates and ice cream sales both increase. This is not because ice cream causes crime or the other way around, it is because both have some causation due to outside temperatures.) --Mr.98 (talk) 20:31, 21 February 2010 (UTC)[reply]

A lot depends on how good the correlation is. In the case of crime versus ice cream - I doubt the correlation is as exact as (for example) lung cancer and smoking. Probably the crime wave starts a month before the ice cream sales spike - and outside of summer there may be other crime and ice-cream 'spikes' that are completely uncorrelated. But when the correlation is exceedingly close, it's more likely that there is a causative link than when it's merely a sloppy fit. So it is perhaps better to say that "correlation cannot prove causation" - but it can certainly provide a sufficiently strong suspicion as to justify a presumption of causation until such time as better information may be found. SteveBaker (talk) 20:51, 21 February 2010 (UTC)[reply]

Even in cases of very tight fits there are always the possibilities of more complicated dynamics, though. Anyway, I wasn't trying to say that correlation and causation have nothing to do with each other, just that without really studying something closely, seeing a correlation does not tell you what the causation is. In the case given here relating to social inequality and HIV transmission, both could be by-products of corrupt government, for example, one that lacks both services in public health/education and supports a regime of inequality. In such a case you might expect extremely tight correlation between public health issues and economic ones, but it doesn't necessarily mean that the public health situation is causing the economic situation or vice versa (or, at least, that such an answer is an adequate understanding). (I'm not saying that is the case here, but one would, I assume, want to make sure one is not missing out on a larger causative factor when trying to draw conclusions about variables that might otherwise be linked.) --Mr.98 (talk) 21:44, 21 February 2010 (UTC)[reply]

The measurements being compared may have different time scales. HIV was first clinically reported in 1981 and has spread rapidly to reach the latest reliable figures that the OP has, perhaps for a year like 2007. I don't know what the timeline of the data being used to calculate Gini coefficient of incomes dispersion but only a change in consistently calculated gini since 1981 can support the "HIV infection leads to social inequality" hypothesis. The converse hypothesis "social inequality exacerbates HIV infection" (I think it is obvious that it does not cause HIV) is best tested by comparing the rate of HIV infections in two (preferably more) different societies of different long-term ginis. Cuddlyable3 (talk) 22:41, 21 February 2010 (UTC)[reply]

Agree 100% that a question like this must have historical data incorporated into a convincingly causal conclusion. HIV is not transhistorical (and neither probably is social inequality, but the latter is probably generally more constant). --Mr.98 (talk) 04:32, 22 February 2010 (UTC)[reply]

Our article on correlation does not imply causation may provide some insight into the risks of drawing these sorts of parallels. (Is there a third factor – perhaps poor access to antiretroviral therapy – which is associated with both HIV transmission and social inequality?) TenOfAllTrades(talk) 23:04, 21 February 2010 (UTC)[reply]

Well, the fundamental methodological mistake is to think that it is possible to conclude anything at all from post-hoc analysis of a single data point.