15 ms·

OkCupid on Trial - 4 Month Experiment

136 points by avlok 14y ago

untog 14y ago
For anyone not aware, OkCupid has a fantastic blog where they do all sorts of statistical analysis on their members: http://blog.okcupid.com/ http://blog.okcupid.com/ It hasn't been updated in over a year. Downer. Getting bought by match.com has clearly had an effect. I have the extremely dubious honour of having had my profile featured on there once as being of "average attractiveness". People still visit my profile to this day, despite me not used it in a few years. Now that's good long tail traffic. Anecdotal evidence: all the girls I know on OkCupid have said that they are harassed the second they turn on the IM system, so they have it turned off. Not that it makes a huge difference- they get flooded with messages, varying from the inane ("hey baby, what's up?") to the creepy. Given the sheer amount of crap they have to deal with, I don't think for a second that online dating is "easier" for women. Having to reject hundreds of suitors might sound preferable to having to approach tens, but I doubt that it really is.
- pie 14y ago
  And sadly, the article "Why You Should Never Pay For Online Dating" bashing match.com was taken down right after the acquisition. Very interesting use of statistics by these guys. There seems to be a mirror here http://interestingreads.posterous.com/why-you-should-never-pay-for-online-dating-ok http://interestingreads.posterous.com/why-you-should-never-p...
- DrJ 14y ago
  if anything, I wish they would update the blog. News from some inside sources says that they are so over extended (running too lean?) that it takes them weeks to get new stuff started (that aren't in the pipe line).
  - KingMob 14y ago
    I wonder if it's one of the consequences of writing it in C++. Seems like that would hurt developer velocity relative to other sites.
- AncientPC 14y ago
  There are a few relevant blog posts: 1. Race and gender data: http://blog.okcupid.com/index.php/your-race-affects-whether-people-write-you-back/ http://blog.okcupid.com/index.php/your-race-affects-whether-... 2. Attractiveness data: http://blog.okcupid.com/index.php/your-looks-and-online-dating/ http://blog.okcupid.com/index.php/your-looks-and-online-dati...
- ig1 14y ago
  Had. They've stopped updating it post-Match.com acquisition.
- personlurking 14y ago
  There's a very interesting behavior that happens on the streets of Brazilian cities which your comment on anecdotal evidence made me remember. Due to a seemingly high percentage of men here acting like they've never seen a woman before in their life (ie, doing a quadruple-take when one walks by) and doing lots of cat-calling (generally in groups), women have learned to not make eye contact. Being that I'm not like these kinds of men, I still get treated like them thus women don't make eye contact with me, in passing.
jsnk 14y ago
>The fact that the first stage of online dating is so heavily stacked in women’s favour doesn’t necessarily mean that it’s any easier for them, compared to men, to reach the end goal of pure love or perfect sex. Of course, men have it harder than women in the mating game. This conclusion seems like a white lie the author created for women. No one is looking for perfect mate. Both men and women choose their partners from whoever is available. If women got 20 times more messages than men do, then women's pool of mates to select is 20 times larger than that of men. And even if some people happen to hold the notion of perfect mate, women still has a better chance of finding the perfect mate than men do. With exception of special cases (maybe alpha males or older women), women on the average have is so much easier when it comes to finding their partner.
- bishnu 14y ago
  Not necessarily. You're assuming the only difference between the women and the men is relative numbers. Let's say 15 women are on a dating website and 30 men. Taken in and of itself, this sounds great for women. But the distribution of objectives matter. Let's say 13 of the women are looking long-term committed relationships and 2 are looking for something casual. This is a gross oversimplification but bear with me. If 25 men are looking for something more casual and 5 are in it for something more long-term, all of the sudden you have a system with 45 people where only 7 of them "have it easy" if you go by numbers - 2 of whom are women. Now these aren't real figures but anecdotally I would say I'm in the right ballpark. At any rate, it sufficiently illustrates that "easiness" has nothing to do with relative size of the gender pools.
  - jsnk 14y ago
    I understand where you are coming from. In a game as complex as dating with so many variables, you can easily come up with many situations where the general rule doesn't follow. However, on average, there are some obvious conclusions one can draw.
- nostrademons 14y ago
  You're familiar with the Stable Marriage Algorithm, right? The math doesn't bear you out: http://en.wikipedia.org/wiki/Stable_marriage_problem http://en.wikipedia.org/wiki/Stable_marriage_problem It's been proven that the algorithm is optimal for the initiators of proposals, i.e. men in contemporary Western culture, although women can hack this by asking guys out themselves. The confusion is because you've artificially restricted the universe to a subset of its true size. Women's pool of mates to select from isn't 20 times larger than men, because men have already done a round of selection in deciding who to message. You don't actually message every single girl you see on OKCupid, do you? If you're fat or ugly as a woman, you're basically screwed (or rather, not screwed) in the mating game, because you become invisible to most men. That's why I tell girls on the OKCupid Reddit that messaging guys can only help them. Once in a blue moon (as in once every month or two), I get an unsolicited message on OKC. A majority of the time, I don't reply. The thing is, I would never have messaged these girls to begin with, while occasionally I'll take a chance on a first date if their personality seems interesting, so they've only helped themselves. It's also worth remembering that there are subsequent rounds of selection too. If women receive (on average) 20x more messages than men, it must mean that men send (on average) 20x more messages than women, since each message has exactly one sender and recipient. They're only going to end up with one girlfriend. So if those girls respond to their favorite guys, many of them will be culled out after the first date, third date, whenever. I've been on I think 16 first dates off OKCupid. I asked only about 3 of them out for a second date. Some (several?) of the remaining 13 probably were not interested in me anyway, but that's still a sizeable number of girls who did not get who they want.
Breefield 14y ago
I was totally with this up until: I then herded our collection of fake people...to five different US cities, where they would be allocated in pairs. The best looking man and woman in one city, second best boy and girl in another, and so on. So the control is lost, even though care was taken to give them similar usernames. In my experience OkCupid is drastically different in different cities. I feel that the attractiveness scale is completely useless when the profile gender pairs were seeded into different cities. I actually stopped reading at that point though, so perhaps there was more to learn, but that was kind of an ender for me.
- Terretta 14y ago
  Because of the matching system, two profiles with identical answers would have shown up identically ranked, therefore obvious to any searchers. So they would have been flagged. And any variation in answers or questions answered would have matched different personalities, throwing off stats. Had to be done.
richardv 14y ago
Curious as to why you tried to control the experiment as much as you could, but then decided to move all of the profiles into different cities? They should of been located in the same city. How do we not know that LA girls and guys are more loose, and Boston women don't even use online dating.
- dalore 14y ago
  Problem is that the profiles in the same city would all show up as similar users with similar names and with the profiles matching exactly. People would easily tell that they are fake and not send any message.
continuations 14y ago
Interestingly the SECOND BEST looking woman got far more messages than the best looking woman, both in US and UK. I wondered why.
- jacques_chester 14y ago
  Learned helplessness. Most men stop contacting the most attractive female profiles because they never hear back.
  - BadassFractal 14y ago
    It's also because most men believe that they have no chance with a 9 or a 10. Why bother messaging someone who they believe is vastly out of their league? Some men are great at hacking the system and ignoring a woman's physical attractiveness, thus scoring really desirable mates. Others instead implicitly follow the rules of thumb that psychology discovered a while ago, according to which an attractive person is much more likely to be with another attractive person.
  - SideburnsOfDoom 14y ago
    Most men stop contacting the most attractive female profiles because they never hear back, and the effort required to make a message that stands out - i.e. is original and interesting - does not justify the very low chance that there will be any response.
  - planetguy 14y ago
    This might be the case sometimes; in this case the best-looking one isn't in the unattainably-beautiful class though. Her nose is kinda big. In this case I think it just indicates that the initial judges did a poor job of judging who the most attractive one was. Or perhaps that the black and white picture used by the most attractive woman is a turn-off. As for the "most attractive" man, he has a nice smile and nice eyes but his baldness and undefinable ethnicity might be a turnoff for the average woman.
- leeny 14y ago
  higher perceived odds of success by the men, perhaps. there's an okcupid blog post about this very observation: http://blog.okcupid.com/index.php/the-mathematics-of-beauty/ http://blog.okcupid.com/index.php/the-mathematics-of-beauty/
jacques_chester 14y ago
I don't know if any of this strikes me as particularly new. Every new female account on any dating site gets absolutely hammered with messages in the first few hours. What's missing from a lot of discussions is simple supply and demand. I was on OKCupid for a few weeks and paid the $10 for the creepy stalker upgrade. Out of curiosity I started comparing male and female profiles in my city. In Perth, Australia, for profiles above a 90% match to mine, there were 8 male profiles for every female profile. For profiles averaging five stars, there were 8.5 male profiles for every female profile. Supply and demand means that the females on the site can really hold out for more. I suspect that this means that the distribution of actual dating activity will be wildly uneven -- the 1% or 2% of males who are really very attractive will clean up almost the entire female pool's attention. And the 50% of females who are above average will have their pick of the top 25% of the male pool. Perth is not necessarily a good sample, as this town already has a lopsided male:female ratio. But every city I looked at had the same phenomenon. If I turned attractiveness up to maximum and only went to high matches, males always greatly outnumbered females. I've given consideration to setting up a dating website in a niche market, but that particular niche market will have an even more lopsided male-centric demographic. I've been seriously wondering about just letting men and women know up front what the current ratios are. The theory being that it will deter some men and encourage some women. Alternatively I might try bar tactics: men pay at the door, women drink for free. Again to deter male users and encourage female users. Basically it is men who are using these sites, not women. And until someone cracks the code on that, it's going to be a fairly one-sided experience.
- BadassFractal 14y ago
  For ccomparison, it would have been extremely interesting to host this experiment in Eastern European countries such as Russia. Due to alcoholism and many other issues, quality women there outnumber quality men and have to deal with competition that is much more alike that for men in the West.
  - mootothemax 14y ago
    It would have been extremely interesting to host this experiment in Eastern European countries such as Russia. Due to alcoholism... I can't speak for alcoholism here in Poland, but after moving to Warsaw from London, I was quite surprised at the number of young, single ladies. I've wondered if a seeming small difference in the age 15-64 sex ratio[1] of 0.99 vs 1.03 could be responsible, but really have no idea. [1] http://en.wikipedia.org/wiki/List_of_countries_by_sex_ratio http://en.wikipedia.org/wiki/List_of_countries_by_sex_ratio
eternalban 14y ago
"[...] the largely undifferentiated onslaught of male attention [...]" The algorithm.
dude_anon123 14y ago
In other news, gay people don't even exist. Seriously, if you are going to attempt this type of "study" gay people are a perfect group, as you can mostly control for gender bias.
- rmc 14y ago
  Exactly. I was reading this, and had the inbult assumption that all men like women and all women like men. Way to erase a minority. It also slightly feeds into the idea that "women have it easy!". I'd be interested if a straight female were to conduct a similar experiement and write it up. What would her perspectives be?
- teuobk 14y ago
  It's an interesting question: what determines how many messages a gay guy receives? What type of guys are most likely to send the initial message? In the first four months I was on OKCupid, I got unsolicited messages from about 115 guys. In the same period, I sent unsolicited messages to about 10 guys, of which 7 elicited responses. My sense is that my numbers are higher than if I were a straight guy, but I have no idea how they compare to "typical" for gay guys.
  - tibbon 14y ago
    I'm a straight guy, but I've had a lot of experiences observing how some of my gay male friends are able to date online and offline. From what I've casually observed (which lines up with your guess) is that it is relatively easier for gay men to find others to meet/hookup with on almost any website (OkCupid, Grinder, Manhunt, etc) and also in real life. One of my friends, who was not exceptionally attractive, witty or any other specific characteristic (let's say he's a 5/10) was able to get guys over to his apartment using Manhunt faster than I was able to get a pizza delivered. Not exaggerating a bit. I have to wonder however, if the primary thing here is that its more equal 2-way attention. Men are actively looking for men, and men are actively looking for me. Whereas for heterosexual relationships it appears to be men looking for women, and women not looking for men. If I remember right, OkCupid did some stuff on non-heterosexual relationships and it was interesting. I'll have to dig for that.
goodside 14y ago
Hi. I work at OkCupid. I coud have saved you a lot of trouble. Your results confirm what all dating web sites have always known, and which we've made no secret about: Nearly all initial messages are sent by men. This does not mean that the site is "better" for women than for men, relative to any other form of heterosexual courtship. It's a deeply ingrained cultural, and perhaps biological, norm that men make proposals and the women evaluate them. If you went to a singles bar, you'd find that all drinks bought for strangers were purchased by men for women. As others have pointed out, the choice of different cities interferes with what you were trying to demonstrate, though the effect size was so huge it was difficult to not confirm it despite the confounding factor.
- 3pt14159 14y ago
  I think he had a good time doing the experiment, so it wasn't really "trouble". Trouble is when you spend a year and a half performing experiments for data that you need as the foundation of your PhD, only to find them available for free in German off of the University of Hannover's website.
- tghw 14y ago
  Is there any chance of OkTrends coming back any time soon, if ever?
  - goodside 14y ago
    There are no plans to permanently stop doing posts as far as I know. OkCupid's user base has been growing rapidly over the past year, to put it mildly, but we're still only ~30 people so we haven't had much time for the more peripheral aspects of the site. I know there are some things in the works, but I'm not directly involved so I can't really give a time estimate.
    - alttab 14y ago
      I glad we have someone from OKCupid on here. Just want to let you know personally that I tried everything when I moved to a new city. Match, PlentyOfFish, and a billion others I can't remember. I got good at writing profiles. OKCupid did an auto-match for someone I didn't find during my own searches. I thought I looked at everything. We apparently thought each other were attractive enough, and thats what spawned the initial prod from OKCupid. The similarities didn't stop there. This was 4 years ago and now I have an amazing wife, the perfect wife. She makes me a better person. It is fairly easy for me to say I would have never met this woman without OKCupid (the site, and the science that drives it). Your service has literally changed my life permanently, for the better. And you did it without charging me a thing. So thank you, and hopefully you guys keep your spirit and innovation as part of a larger match.com portfolio!
erehweb 14y ago
Whenever there's an experiment, we have to weigh the costs against the benefits. Benefits here seem slight - we relearn that women, and pretty women in particular, get more messages than men. Costs are non-trivial - about 1000 people deceived. Admittedly a minor deception, but experimental subjects normally get some protection, and a minute of time wasted by each translates to 16 hours overall. I'm presuming he got consent to use all the photos in his experiment and on the site, but he's silent on that. At least we have a thousand small disappointments. Experiments involving deceiving people can be useful (e.g. ones investigating whether employers / landlords discriminate), but this one doesn't seem to rise to that level.
personlurking 14y ago
"If they are hot, the girls can pick and choose which men they interact with." There was a video on Youtube on dating and game theory where 5 women and 5 men were given a number which corresponded to their attractiveness. Basically, those who were 5's, when accepted into the 8's group were made more attractive in the eyes of the opposite sex, even if they were still 5's physically. Conversely, those who were 10's got the pick of the litter and had to do the least work to find a mate.
- pmiller2 14y ago
  Do you happen to have the link?
  - personlurking 14y ago
    I searched for it but couldn't recall the search terms. I tried a few to see if I got a bite but nada. I originally saw it a year or two ago.
sophacles 14y ago
It may be an artifact of the different city problem in data collection, but assuming it wasn't, I find it interesting that the woman who was 'second most attractive' got significantly more messages than the 'most attractive'. To me this suggests interesting game theoretical stuff going on here. Does anyone know any good research/math on strategies that would optimize for this phenomenon? e.g. I can see a thought process happening along the lines of "she is too good looking for me, I won't try, rejection sucks. I'll try this slightly less attractive woman who probably isn't hit up as often because my chances will be better, because everyone is hitting up the really good looking woman". Of course if most men follow this, you get the discrepancy I noted.
- planetguy 14y ago
  Or else men just thought the "second best looking" was better looking than the "best looking" one. Or else men don't like it when the only photo is a black and white close-up from an odd angle -- makes it look like you're hiding something.
- allenwlee 14y ago
  I was not surprised due to this study that I read recently where You see similar nonlinearities http://dienekes.blogspot.com/2009/09/physical-attractiveness-and.html http://dienekes.blogspot.com/2009/09/physical-attractiveness...
eqdw 14y ago
THIS JUST IN: man spends 4 months confirming what everyone already knows

136 points by avlok 14y ago

untog 14y ago

For anyone not aware, OkCupid has a fantastic blog where they do all sorts of statistical analysis on their members: http://blog.okcupid.com/ http://blog.okcupid.com/ It hasn't been updated in over a year. Downer. Getting bought by match.com has clearly had an effect. I have the extremely dubious honour of having had my profile featured on there once as being of "average attractiveness". People still visit my profile to this day, despite me not used it in a few years. Now that's good long tail traffic. Anecdotal evidence: all the girls I know on OkCupid have said that they are harassed the second they turn on the IM system, so they have it turned off. Not that it makes a huge difference- they get flooded with messages, varying from the inane ("hey baby, what's up?") to the creepy. Given the sheer amount of crap they have to deal with, I don't think for a second that online dating is "easier" for women. Having to reject hundreds of suitors might sound preferable to having to approach tens, but I doubt that it really is.

pie 14y ago
And sadly, the article "Why You Should Never Pay For Online Dating" bashing match.com was taken down right after the acquisition. Very interesting use of statistics by these guys. There seems to be a mirror here http://interestingreads.posterous.com/why-you-should-never-pay-for-online-dating-ok http://interestingreads.posterous.com/why-you-should-never-p...
DrJ 14y ago
if anything, I wish they would update the blog. News from some inside sources says that they are so over extended (running too lean?) that it takes them weeks to get new stuff started (that aren't in the pipe line).
- KingMob 14y ago
  I wonder if it's one of the consequences of writing it in C++. Seems like that would hurt developer velocity relative to other sites.
AncientPC 14y ago
There are a few relevant blog posts: 1. Race and gender data: http://blog.okcupid.com/index.php/your-race-affects-whether-people-write-you-back/ http://blog.okcupid.com/index.php/your-race-affects-whether-... 2. Attractiveness data: http://blog.okcupid.com/index.php/your-looks-and-online-dating/ http://blog.okcupid.com/index.php/your-looks-and-online-dati...
ig1 14y ago
Had. They've stopped updating it post-Match.com acquisition.
personlurking 14y ago
There's a very interesting behavior that happens on the streets of Brazilian cities which your comment on anecdotal evidence made me remember. Due to a seemingly high percentage of men here acting like they've never seen a woman before in their life (ie, doing a quadruple-take when one walks by) and doing lots of cat-calling (generally in groups), women have learned to not make eye contact. Being that I'm not like these kinds of men, I still get treated like them thus women don't make eye contact with me, in passing.

jsnk 14y ago

>The fact that the first stage of online dating is so heavily stacked in women’s favour doesn’t necessarily mean that it’s any easier for them, compared to men, to reach the end goal of pure love or perfect sex. Of course, men have it harder than women in the mating game. This conclusion seems like a white lie the author created for women. No one is looking for perfect mate. Both men and women choose their partners from whoever is available. If women got 20 times more messages than men do, then women's pool of mates to select is 20 times larger than that of men. And even if some people happen to hold the notion of perfect mate, women still has a better chance of finding the perfect mate than men do. With exception of special cases (maybe alpha males or older women), women on the average have is so much easier when it comes to finding their partner.

bishnu 14y ago
Not necessarily. You're assuming the only difference between the women and the men is relative numbers. Let's say 15 women are on a dating website and 30 men. Taken in and of itself, this sounds great for women. But the distribution of objectives matter. Let's say 13 of the women are looking long-term committed relationships and 2 are looking for something casual. This is a gross oversimplification but bear with me. If 25 men are looking for something more casual and 5 are in it for something more long-term, all of the sudden you have a system with 45 people where only 7 of them "have it easy" if you go by numbers - 2 of whom are women. Now these aren't real figures but anecdotally I would say I'm in the right ballpark. At any rate, it sufficiently illustrates that "easiness" has nothing to do with relative size of the gender pools.
- jsnk 14y ago
  I understand where you are coming from. In a game as complex as dating with so many variables, you can easily come up with many situations where the general rule doesn't follow. However, on average, there are some obvious conclusions one can draw.
nostrademons 14y ago
You're familiar with the Stable Marriage Algorithm, right? The math doesn't bear you out: http://en.wikipedia.org/wiki/Stable_marriage_problem http://en.wikipedia.org/wiki/Stable_marriage_problem It's been proven that the algorithm is optimal for the initiators of proposals, i.e. men in contemporary Western culture, although women can hack this by asking guys out themselves. The confusion is because you've artificially restricted the universe to a subset of its true size. Women's pool of mates to select from isn't 20 times larger than men, because men have already done a round of selection in deciding who to message. You don't actually message every single girl you see on OKCupid, do you? If you're fat or ugly as a woman, you're basically screwed (or rather, not screwed) in the mating game, because you become invisible to most men. That's why I tell girls on the OKCupid Reddit that messaging guys can only help them. Once in a blue moon (as in once every month or two), I get an unsolicited message on OKC. A majority of the time, I don't reply. The thing is, I would never have messaged these girls to begin with, while occasionally I'll take a chance on a first date if their personality seems interesting, so they've only helped themselves. It's also worth remembering that there are subsequent rounds of selection too. If women receive (on average) 20x more messages than men, it must mean that men send (on average) 20x more messages than women, since each message has exactly one sender and recipient. They're only going to end up with one girlfriend. So if those girls respond to their favorite guys, many of them will be culled out after the first date, third date, whenever. I've been on I think 16 first dates off OKCupid. I asked only about 3 of them out for a second date. Some (several?) of the remaining 13 probably were not interested in me anyway, but that's still a sizeable number of girls who did not get who they want.

Breefield 14y ago

I was totally with this up until: I then herded our collection of fake people...to five different US cities, where they would be allocated in pairs. The best looking man and woman in one city, second best boy and girl in another, and so on. So the control is lost, even though care was taken to give them similar usernames. In my experience OkCupid is drastically different in different cities. I feel that the attractiveness scale is completely useless when the profile gender pairs were seeded into different cities. I actually stopped reading at that point though, so perhaps there was more to learn, but that was kind of an ender for me.

Terretta 14y ago
Because of the matching system, two profiles with identical answers would have shown up identically ranked, therefore obvious to any searchers. So they would have been flagged. And any variation in answers or questions answered would have matched different personalities, throwing off stats. Had to be done.

richardv 14y ago

Curious as to why you tried to control the experiment as much as you could, but then decided to move all of the profiles into different cities? They should of been located in the same city. How do we not know that LA girls and guys are more loose, and Boston women don't even use online dating.

dalore 14y ago
Problem is that the profiles in the same city would all show up as similar users with similar names and with the profiles matching exactly. People would easily tell that they are fake and not send any message.

continuations 14y ago

Interestingly the SECOND BEST looking woman got far more messages than the best looking woman, both in US and UK. I wondered why.

jacques_chester 14y ago
Learned helplessness. Most men stop contacting the most attractive female profiles because they never hear back.
- BadassFractal 14y ago
  It's also because most men believe that they have no chance with a 9 or a 10. Why bother messaging someone who they believe is vastly out of their league? Some men are great at hacking the system and ignoring a woman's physical attractiveness, thus scoring really desirable mates. Others instead implicitly follow the rules of thumb that psychology discovered a while ago, according to which an attractive person is much more likely to be with another attractive person.
- SideburnsOfDoom 14y ago
  Most men stop contacting the most attractive female profiles because they never hear back, and the effort required to make a message that stands out - i.e. is original and interesting - does not justify the very low chance that there will be any response.
- planetguy 14y ago
  This might be the case sometimes; in this case the best-looking one isn't in the unattainably-beautiful class though. Her nose is kinda big. In this case I think it just indicates that the initial judges did a poor job of judging who the most attractive one was. Or perhaps that the black and white picture used by the most attractive woman is a turn-off. As for the "most attractive" man, he has a nice smile and nice eyes but his baldness and undefinable ethnicity might be a turnoff for the average woman.
leeny 14y ago
higher perceived odds of success by the men, perhaps. there's an okcupid blog post about this very observation: http://blog.okcupid.com/index.php/the-mathematics-of-beauty/ http://blog.okcupid.com/index.php/the-mathematics-of-beauty/

jacques_chester 14y ago

I don't know if any of this strikes me as particularly new. Every new female account on any dating site gets absolutely hammered with messages in the first few hours. What's missing from a lot of discussions is simple supply and demand. I was on OKCupid for a few weeks and paid the $10 for the creepy stalker upgrade. Out of curiosity I started comparing male and female profiles in my city. In Perth, Australia, for profiles above a 90% match to mine, there were 8 male profiles for every female profile. For profiles averaging five stars, there were 8.5 male profiles for every female profile. Supply and demand means that the females on the site can really hold out for more. I suspect that this means that the distribution of actual dating activity will be wildly uneven -- the 1% or 2% of males who are really very attractive will clean up almost the entire female pool's attention. And the 50% of females who are above average will have their pick of the top 25% of the male pool. Perth is not necessarily a good sample, as this town already has a lopsided male:female ratio. But every city I looked at had the same phenomenon. If I turned attractiveness up to maximum and only went to high matches, males always greatly outnumbered females. I've given consideration to setting up a dating website in a niche market, but that particular niche market will have an even more lopsided male-centric demographic. I've been seriously wondering about just letting men and women know up front what the current ratios are. The theory being that it will deter some men and encourage some women. Alternatively I might try bar tactics: men pay at the door, women drink for free. Again to deter male users and encourage female users. Basically it is men who are using these sites, not women. And until someone cracks the code on that, it's going to be a fairly one-sided experience.

BadassFractal 14y ago
For ccomparison, it would have been extremely interesting to host this experiment in Eastern European countries such as Russia. Due to alcoholism and many other issues, quality women there outnumber quality men and have to deal with competition that is much more alike that for men in the West.
- mootothemax 14y ago
  It would have been extremely interesting to host this experiment in Eastern European countries such as Russia. Due to alcoholism... I can't speak for alcoholism here in Poland, but after moving to Warsaw from London, I was quite surprised at the number of young, single ladies. I've wondered if a seeming small difference in the age 15-64 sex ratio[1] of 0.99 vs 1.03 could be responsible, but really have no idea. [1] http://en.wikipedia.org/wiki/List_of_countries_by_sex_ratio http://en.wikipedia.org/wiki/List_of_countries_by_sex_ratio

eternalban 14y ago

"[...] the largely undifferentiated onslaught of male attention [...]" The algorithm.

dude_anon123 14y ago

In other news, gay people don't even exist. Seriously, if you are going to attempt this type of "study" gay people are a perfect group, as you can mostly control for gender bias.

rmc 14y ago
Exactly. I was reading this, and had the inbult assumption that all men like women and all women like men. Way to erase a minority. It also slightly feeds into the idea that "women have it easy!". I'd be interested if a straight female were to conduct a similar experiement and write it up. What would her perspectives be?
teuobk 14y ago
It's an interesting question: what determines how many messages a gay guy receives? What type of guys are most likely to send the initial message? In the first four months I was on OKCupid, I got unsolicited messages from about 115 guys. In the same period, I sent unsolicited messages to about 10 guys, of which 7 elicited responses. My sense is that my numbers are higher than if I were a straight guy, but I have no idea how they compare to "typical" for gay guys.
- tibbon 14y ago
  I'm a straight guy, but I've had a lot of experiences observing how some of my gay male friends are able to date online and offline. From what I've casually observed (which lines up with your guess) is that it is relatively easier for gay men to find others to meet/hookup with on almost any website (OkCupid, Grinder, Manhunt, etc) and also in real life. One of my friends, who was not exceptionally attractive, witty or any other specific characteristic (let's say he's a 5/10) was able to get guys over to his apartment using Manhunt faster than I was able to get a pizza delivered. Not exaggerating a bit. I have to wonder however, if the primary thing here is that its more equal 2-way attention. Men are actively looking for men, and men are actively looking for me. Whereas for heterosexual relationships it appears to be men looking for women, and women not looking for men. If I remember right, OkCupid did some stuff on non-heterosexual relationships and it was interesting. I'll have to dig for that.

goodside 14y ago

Hi. I work at OkCupid. I coud have saved you a lot of trouble. Your results confirm what all dating web sites have always known, and which we've made no secret about: Nearly all initial messages are sent by men. This does not mean that the site is "better" for women than for men, relative to any other form of heterosexual courtship. It's a deeply ingrained cultural, and perhaps biological, norm that men make proposals and the women evaluate them. If you went to a singles bar, you'd find that all drinks bought for strangers were purchased by men for women. As others have pointed out, the choice of different cities interferes with what you were trying to demonstrate, though the effect size was so huge it was difficult to not confirm it despite the confounding factor.

3pt14159 14y ago
I think he had a good time doing the experiment, so it wasn't really "trouble". Trouble is when you spend a year and a half performing experiments for data that you need as the foundation of your PhD, only to find them available for free in German off of the University of Hannover's website.
tghw 14y ago
Is there any chance of OkTrends coming back any time soon, if ever?
- goodside 14y ago
  There are no plans to permanently stop doing posts as far as I know. OkCupid's user base has been growing rapidly over the past year, to put it mildly, but we're still only ~30 people so we haven't had much time for the more peripheral aspects of the site. I know there are some things in the works, but I'm not directly involved so I can't really give a time estimate.
  - alttab 14y ago
    I glad we have someone from OKCupid on here. Just want to let you know personally that I tried everything when I moved to a new city. Match, PlentyOfFish, and a billion others I can't remember. I got good at writing profiles. OKCupid did an auto-match for someone I didn't find during my own searches. I thought I looked at everything. We apparently thought each other were attractive enough, and thats what spawned the initial prod from OKCupid. The similarities didn't stop there. This was 4 years ago and now I have an amazing wife, the perfect wife. She makes me a better person. It is fairly easy for me to say I would have never met this woman without OKCupid (the site, and the science that drives it). Your service has literally changed my life permanently, for the better. And you did it without charging me a thing. So thank you, and hopefully you guys keep your spirit and innovation as part of a larger match.com portfolio!

erehweb 14y ago

Whenever there's an experiment, we have to weigh the costs against the benefits. Benefits here seem slight - we relearn that women, and pretty women in particular, get more messages than men. Costs are non-trivial - about 1000 people deceived. Admittedly a minor deception, but experimental subjects normally get some protection, and a minute of time wasted by each translates to 16 hours overall. I'm presuming he got consent to use all the photos in his experiment and on the site, but he's silent on that. At least we have a thousand small disappointments. Experiments involving deceiving people can be useful (e.g. ones investigating whether employers / landlords discriminate), but this one doesn't seem to rise to that level.

personlurking 14y ago

"If they are hot, the girls can pick and choose which men they interact with." There was a video on Youtube on dating and game theory where 5 women and 5 men were given a number which corresponded to their attractiveness. Basically, those who were 5's, when accepted into the 8's group were made more attractive in the eyes of the opposite sex, even if they were still 5's physically. Conversely, those who were 10's got the pick of the litter and had to do the least work to find a mate.

pmiller2 14y ago
Do you happen to have the link?
- personlurking 14y ago
  I searched for it but couldn't recall the search terms. I tried a few to see if I got a bite but nada. I originally saw it a year or two ago.

sophacles 14y ago

It may be an artifact of the different city problem in data collection, but assuming it wasn't, I find it interesting that the woman who was 'second most attractive' got significantly more messages than the 'most attractive'. To me this suggests interesting game theoretical stuff going on here. Does anyone know any good research/math on strategies that would optimize for this phenomenon? e.g. I can see a thought process happening along the lines of "she is too good looking for me, I won't try, rejection sucks. I'll try this slightly less attractive woman who probably isn't hit up as often because my chances will be better, because everyone is hitting up the really good looking woman". Of course if most men follow this, you get the discrepancy I noted.

planetguy 14y ago
Or else men just thought the "second best looking" was better looking than the "best looking" one. Or else men don't like it when the only photo is a black and white close-up from an odd angle -- makes it look like you're hiding something.
allenwlee 14y ago
I was not surprised due to this study that I read recently where You see similar nonlinearities http://dienekes.blogspot.com/2009/09/physical-attractiveness-and.html http://dienekes.blogspot.com/2009/09/physical-attractiveness...

eqdw 14y ago

THIS JUST IN: man spends 4 months confirming what everyone already knows