7 ms·

Gini coefficient

25 points by sbolt 7y ago

woadwarrior01 7y ago
Here's an online calculator[1] for Gini coefficient that I found some time ago. Tinkering around with some numbers there was what finally gave me an proper intuitive understanding of the Gini coefficient and the Lorentz curve. [1]: http://shlegeris.com/gini http://shlegeris.com/gini
wodenokoto 7y ago
One of the most interesting uses of gini coefficient I've heard of, was counting messages, winks and likes received on a dating site as money or wealth and measuring the gini coefficient for men and women. Apparently the distribution is pretty fair for women and extremely skewed for men.
- xiphias2 7y ago
  I think this is the article you are refering to. https://medium.com/@worstonlinedater/tinder-experiments-ii-guys-unless-you-are-really-hot-you-are-probably-better-off-not-wasting-your-2ddf370a6e9a https://medium.com/@worstonlinedater/tinder-experiments-ii-g...
  - claudiawerner 7y ago
    It's a little odd that the author almost consistently refers to men as "men" but uses the word "femlaes" consistently to refer to women. A take on the experiment is here: https://medium.com/@ericschulman/2-big-problems-with-the-tinder-experiments-meme-6dbfdfcee6a8 https://medium.com/@ericschulman/2-big-problems-with-the-tin... I feel as though thinking of dating as a kind of economy with transactions likely isn't the best way to find someone, especially when the "training" for such "jobs" in the economy is so one-sided - men are very rarely taught how to properly look after their appearance, but it's a core part of every girl's life, 12 and up (and usually sooner than that). It does not surprise me that those with the training (through receiving positive and negative signals on their appearance) are more picky. You'd find that a Javascript programmer asked to choose a JS framework would probably make a better decision (and of course, be much more demanding given the options) than a documentation writer or graphic designer on the same project.
    - xiphias2 7y ago
      I agree with you that the study was probably biased, but the take on experiment have even more faults than the experiment itself: ,,Problem #1: Personality Matters'' Of course it matters, it's just far not as important as fitness indicators of the other sex. The study was clear that it's about distribution of swipes, which is a prerequisite of even getting to know the personality of eachother. ,,men are very rarely taught how to properly look after their appearance'' I had this problem in my life, but nowdays the internet has all the information easily accessible and for free, and instagram gives people the tools for imitation/experimentation, so taking care of appearence got much easier for men in the last decade. ,,part of every girl's life, 12 and up (and usually sooner than that)'' My niece is 3 years old, and I see her already thinking about her appearance all the time (as opposed to my nephews who just want to play games). ,, It does not surprise me that those with the training (through receiving positive and negative signals on their appearance) are more picky.'' Evolutionary biology shows that it's much much deeper within the sexes than just training. It's a fact that we have twice as many female ancestors than male, and this fact is true in other species as well.
      - claudiawerner 7y ago
        >The study was clear that it's about distribution of swipes, which is a prerequisite of even getting to know the personality of eachother. The author notes that the content of the picture itself is highly relevant to judging personality from a picture, and how we perceive people in pictures. Painters, for instance, have known to use this fact for the past two thousand years when portraying people - often, the context the subject is placed in is just as important, if not moreso, than the appearance of the person themselves. >so taking care of appearence got much easier for men in the last decade. Being easy does not mean that it is something that they (in general) follow through with. It is now more easy than ever to watch films, but that does not mean I will dedicate considerable time to learning to appreciate films, or learn what differentiates a good film from a bad one. >Evolutionary biology shows that it's much much deeper within the sexes than just training. I agree, but this just reinforces my point that it is a fallacy to expect a "market" that is so heavily inbalanced in this skill to be anything near equal. Men have to put a lot more effort into the first picture they show, since there is a large body of research showing that women are less stimulated by visual representations than men. Many other variables were ignored in the experiment, all of which directly and importantly influence the charactersitics of the "Tinder economy"; profiles are assigned an ELO on Tinder, such that those who swipe more are pushed down the stack, using the boost/super like functionality dramatically increases the chance of getting a match, women and men both select on the contents of the bio, or whether it has an Instagram link, and how many followers are on the user's page, Tinder shows the mutual connections within social groups, etc. When I'm on Tinder, at least, I scroll through all the pictures, and I don't just look at the first one. It's unclear whether the author even used multiple pictures, forcing people to make decisions based on just the one - decisions they may not have made if there were multiple. Not to mention the extremely small sample size, assuming that women all rate men (and indeed they rate them along the same axes as men rate women, even given just a picture), the use of deception by the author, and the bizarre comparison to real-world economies whose inequality is shaped by a variety of factors, from ownership of capital, unemployment, gender, race (the author even put South Africa on the graph!), etc. "Attractiveness" is left undefined qualitatively (and therefore meaningless to assign any value to its quantiative measurement, as if such a thing were objective, even on Tinder), and it is unclear whether this includes the user's bio or not. A like, on its own, on Tinder doesn't get you anywhere. It opens the potential (and a low potential) to find something (usually sex, in which it is no surprise that people selected based on atractiveness and extrapolate on that atractiveness from personal biases). The currency of Tinder, if anything, sohuld be modelled as "successful" meets.
  - wodenokoto 7y ago
    I remember it as being done on OK cupid, but this is a really good link. Thank you!
  - 0max 7y ago
    is there a non-hetronormative analysis available as well?
RocketSyntax 7y ago
My mind went straight to gini impurity
marriedWpt 7y ago
I think this is a bad way to gauge an economy. If you want to look at why people are poor, I would look towards corruption rather than merely inequality. Or maybe people in a country are poor on paper but can afford a hundred dollars of luxuries a week. Using gini reminds me of GIGO. You used a poor metric for information, your decisions are going to be poor.
- dang 7y ago
  Could you please stop creating accounts for every few comments you post? We ban accounts that do that. This is in the site guidelines: https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html. HN is a community and we want it to remain one. For that, users need some identity for others to relate to. Otherwise we may as well have no usernames and no community, and that would be a different kind of forum. https://hn.algolia.com/?sort=byDate&dateRange=all&type=comment&storyText=false&prefix&page=0&query=by:dang%20community%20identity https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme... You needn't use your real name, of course.

25 points by sbolt 7y ago

Here's an online calculator[1] for Gini coefficient that I found some time ago. Tinkering around with some numbers there was what finally gave me an proper intuitive understanding of the Gini coefficient and the Lorentz curve. [1]: http://shlegeris.com/gini http://shlegeris.com/gini

wodenokoto 7y ago

One of the most interesting uses of gini coefficient I've heard of, was counting messages, winks and likes received on a dating site as money or wealth and measuring the gini coefficient for men and women. Apparently the distribution is pretty fair for women and extremely skewed for men.

xiphias2 7y ago
I think this is the article you are refering to. https://medium.com/@worstonlinedater/tinder-experiments-ii-guys-unless-you-are-really-hot-you-are-probably-better-off-not-wasting-your-2ddf370a6e9a https://medium.com/@worstonlinedater/tinder-experiments-ii-g...
- claudiawerner 7y ago
  It's a little odd that the author almost consistently refers to men as "men" but uses the word "femlaes" consistently to refer to women. A take on the experiment is here: https://medium.com/@ericschulman/2-big-problems-with-the-tinder-experiments-meme-6dbfdfcee6a8 https://medium.com/@ericschulman/2-big-problems-with-the-tin... I feel as though thinking of dating as a kind of economy with transactions likely isn't the best way to find someone, especially when the "training" for such "jobs" in the economy is so one-sided - men are very rarely taught how to properly look after their appearance, but it's a core part of every girl's life, 12 and up (and usually sooner than that). It does not surprise me that those with the training (through receiving positive and negative signals on their appearance) are more picky. You'd find that a Javascript programmer asked to choose a JS framework would probably make a better decision (and of course, be much more demanding given the options) than a documentation writer or graphic designer on the same project.
  - xiphias2 7y ago
    I agree with you that the study was probably biased, but the take on experiment have even more faults than the experiment itself: ,,Problem #1: Personality Matters'' Of course it matters, it's just far not as important as fitness indicators of the other sex. The study was clear that it's about distribution of swipes, which is a prerequisite of even getting to know the personality of eachother. ,,men are very rarely taught how to properly look after their appearance'' I had this problem in my life, but nowdays the internet has all the information easily accessible and for free, and instagram gives people the tools for imitation/experimentation, so taking care of appearence got much easier for men in the last decade. ,,part of every girl's life, 12 and up (and usually sooner than that)'' My niece is 3 years old, and I see her already thinking about her appearance all the time (as opposed to my nephews who just want to play games). ,, It does not surprise me that those with the training (through receiving positive and negative signals on their appearance) are more picky.'' Evolutionary biology shows that it's much much deeper within the sexes than just training. It's a fact that we have twice as many female ancestors than male, and this fact is true in other species as well.
    - claudiawerner 7y ago
      >The study was clear that it's about distribution of swipes, which is a prerequisite of even getting to know the personality of eachother. The author notes that the content of the picture itself is highly relevant to judging personality from a picture, and how we perceive people in pictures. Painters, for instance, have known to use this fact for the past two thousand years when portraying people - often, the context the subject is placed in is just as important, if not moreso, than the appearance of the person themselves. >so taking care of appearence got much easier for men in the last decade. Being easy does not mean that it is something that they (in general) follow through with. It is now more easy than ever to watch films, but that does not mean I will dedicate considerable time to learning to appreciate films, or learn what differentiates a good film from a bad one. >Evolutionary biology shows that it's much much deeper within the sexes than just training. I agree, but this just reinforces my point that it is a fallacy to expect a "market" that is so heavily inbalanced in this skill to be anything near equal. Men have to put a lot more effort into the first picture they show, since there is a large body of research showing that women are less stimulated by visual representations than men. Many other variables were ignored in the experiment, all of which directly and importantly influence the charactersitics of the "Tinder economy"; profiles are assigned an ELO on Tinder, such that those who swipe more are pushed down the stack, using the boost/super like functionality dramatically increases the chance of getting a match, women and men both select on the contents of the bio, or whether it has an Instagram link, and how many followers are on the user's page, Tinder shows the mutual connections within social groups, etc. When I'm on Tinder, at least, I scroll through all the pictures, and I don't just look at the first one. It's unclear whether the author even used multiple pictures, forcing people to make decisions based on just the one - decisions they may not have made if there were multiple. Not to mention the extremely small sample size, assuming that women all rate men (and indeed they rate them along the same axes as men rate women, even given just a picture), the use of deception by the author, and the bizarre comparison to real-world economies whose inequality is shaped by a variety of factors, from ownership of capital, unemployment, gender, race (the author even put South Africa on the graph!), etc. "Attractiveness" is left undefined qualitatively (and therefore meaningless to assign any value to its quantiative measurement, as if such a thing were objective, even on Tinder), and it is unclear whether this includes the user's bio or not. A like, on its own, on Tinder doesn't get you anywhere. It opens the potential (and a low potential) to find something (usually sex, in which it is no surprise that people selected based on atractiveness and extrapolate on that atractiveness from personal biases). The currency of Tinder, if anything, sohuld be modelled as "successful" meets.
- wodenokoto 7y ago
  I remember it as being done on OK cupid, but this is a really good link. Thank you!
- 0max 7y ago
  is there a non-hetronormative analysis available as well?

RocketSyntax 7y ago

My mind went straight to gini impurity

marriedWpt 7y ago

I think this is a bad way to gauge an economy. If you want to look at why people are poor, I would look towards corruption rather than merely inequality. Or maybe people in a country are poor on paper but can afford a hundred dollars of luxuries a week. Using gini reminds me of GIGO. You used a poor metric for information, your decisions are going to be poor.

dang 7y ago
Could you please stop creating accounts for every few comments you post? We ban accounts that do that. This is in the site guidelines: https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html. HN is a community and we want it to remain one. For that, users need some identity for others to relate to. Otherwise we may as well have no usernames and no community, and that would be a different kind of forum. https://hn.algolia.com/?sort=byDate&dateRange=all&type=comment&storyText=false&prefix&page=0&query=by:dang%20community%20identity https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme... You needn't use your real name, of course.