The OKCupid data gets thrown around quite a bit. Weirdly enough, both sides use it to make opposite points. The way the data is formatted makes it difficult to interpret, which is the main reason for the confusion. So I took a close look at it. What I discovered is that most people misinterpret the data to some degree. Even including Christian Rudder, the guy at OKCupid who compiled the data, seems to get it wrong. ( The blog post from OKCupid is here. )

First, women's judgements of men's attractiveness looks terrible.

https://i.imgur.com/L9Vu4Zo.png

But if we look at messaging patterns, things look a little better. Here's what that data looks like:

https://i.imgur.com/GSudEHM.png

It shows that:

  • The top 6% of men received 18% of all initial messages.
  • The top 6% of women received 18% of all initial messages.
  • The top 20% of men received 40% of all initial messages.
  • The top 20% of women received 44% of all initial messages.

From that initial data, it looks like men and women are equally interested in the top 6%. But, for the tier right below that, it looks like men are trying to "date up" more often than women, but there complications to this data which might make that statement false.

To get a better understanding of the data, I wanted to look at it on a "percentile" basis. For example, I wanted to compare how well a man or woman in the 20th, 50th, or 90th percentile do. Here's what the data looks like when I split it out by percentiles. (Note: Because the top two tiers of men are so incredibly small, I was worried about rounding errors, so I combined the top three categories together, so that it represents the top 6% of men.) The percentile chart looks like this:

https://i.imgur.com/kewvVqT.png

What this chart is showing is the ratio of messages received by men and women at different percentiles. The average is "1" for men and women - as in: if men send 500 messages and there are 100 women on the site, then a "1" indicates that a woman woman receives 5 messages (i.e. 500/100 = 5). A value of "3" means she gets 3x as many messages - i.e. 15 messages. For example, on the right side, we see that the top men and women receive 3x messages. For both men and women, this corresponds to people who are in the 94th-100th percentiles (the dot on the chart is shown at 97, which is the mid-point between 94 and 100).

We can see on this chart that top-tier (i.e. the top 6%) men and women receive 18% of all messages - which is 3x "their fair share" of messages. It's kind of amazing that these percentages are identical. Men aren't more or less likely than women to send messages to the very hottest members of the opposite sex. It does show that men are slightly more likely than women to send messages to the 60th-90th percentiles of women. And women are more likely than men to send messages to men who are in the 0th-50th percentiles of men.

This directly contradicts what Christian Rudder says in his blog post: "When it comes down to actually choosing targets, men choose the modelesque...So basically, guys are fighting each other 2-for-1 for the absolute best-rated females, while plenty of potentially charming, even cute, girls go unwritten. The medical term for this is male pattern madness." Obviously, Christian Rudder doesn't know what he's talking about here. Maybe he's confused himself by his poorly formatted data. Men aren't going for the very hottest women anymore than women are going for the very hottest men. However, men are slightly more likely than women to message above-average (60th-90th percentiles) members of the opposite sex. More specifically, women in the 87th percentile receive a 15% higher ratio of messages than men in the 87th percentile. And women at the 70th percentile receive a 12% higher ratio of messages than men in the 70th percentile. On the flip side, men in the 14th percentile receive about a 70% higher ratio of messages than women in the 14th percentile.

But, wait - there's more complications in the data. We're assuming that all men (regardless of attractiveness) and all women (regardless of attractiveness) are sending the same number of messages. If unattractive women and/or attractive men are sending more messages, then it would explain the discrepancy. Afterall, if hot guys are sending more messages than ugly guys, then why wouldn't he preferentially send messages to above-average women? And if unattractive women are sending more messages than other women, shouldn't most of her messages go to below-average men - who are in her own league? They're just messaging people who are near their own league. As it turns out - this is exactly what's happening. Good looking guys send the most messages (compared to other guys), and unattractive women send the most messages (compared to other women).

https://i.imgur.com/jyf4QUv.png

For unattractive women, this pattern makes a lot of sense. As one PPD commenter said: "I don't message men first because I don't have to". Well, that system probably works great unless you're an unattractive woman. Since women at the bottom of attractiveness can't rely as much on people messaging them, they take more initiative. To quote a comedian I heard once: "If you're a man or an ugly woman, you're going to have to make an effort". As for why the bottom 60% of men send fewer messages than the top 40% of men? My only guess is that attractive men find online dating more rewarding and less demoralizing than less attractive men. I certainly have male friends who have deleted Tinder based on feeling demoralized at the lack of response they'd get from women.

Regarding the chart above: I think this chart is a complete mess. First, the numbers on the left don't line-up with the horizontal lines on the chart. And does the bottom of the chart represent 1.25 or 0.0 messages sent? And second, do the dots represent actual data-points and the curve is just the result of a poor curve-fitting algorithm? Other sources say that men send 3.5 initial messages for every initial message women send, but this chart makes it look much more extreme - based on this chart it appears that men send 10+ messages for every message a woman sends. Taking into account the "3.5x" number, here's what I *think* the chart is trying to show:

https://i.imgur.com/rFPWfbw.png

The effect of this is that it increases the ratio of messages sent to attractive women, and increases the ratio of messages sent to unattractive men. Like this:

https://i.imgur.com/pgZO87D.png

It's hard to say for certain, but this would make the lines rather similar, and *might* cause the women's line to skew slightly towards a more hypergamous line (i.e. skewed more towards the most attractive men, relative to men's line). Still, it's hard to say, and it's probably not much more skewed than men's line is.

What about the claim that "the most attractive guys get 11x the messages the lowest-rated do. The medium-rated get about 4x." and "[The most attractive women] gets nearly 5 times as many messages as a typical woman and 28 times as many messages as a woman at the low end of our curve." This suggests that men, much more than women, are sending all their messages to the hottest members of the opposite sex. I'm unclear how he came up with these numbers, but I can see two potential problems with this claim:

First, if he's comparing the Tier 1 men (the top 1%) against the Tier 7 men (literally the bottom 26% of men) and then comparing the Tier 1 women (the top 6%) against the Tier 7 women (the bottom 6%), then that whole calculation is a bad one because you can't assume that guys in the the bottom 26% of men are an equivalent group to compare to women in the bottom 6% of women.

Second, the fact that attractive men send more messages and unattractive women send more messages throws off his whole calculation - because his graph only makes sense if he assumes that all people, regardless of attractiveness, send equal numbers of messages.

As a result, this graph from OKCupid is bunk: https://i.imgur.com/3QVMUoV.png

Overall, it looks like men and women have rather similar messaging patterns. In other words: Christian Rudder is wrong when he claims that men (and not women) are being unrealistic and only chasing the hottest members of the opposite sex. It also contradicts claims by women that men's dating problems are simply the result of men chasing the hottest women and not realizing that they're unattractive losers. The charts also undermine the (oft repeated) claim that women are virtuously less interested in physical attraction than men are. But, the flipside also seems true: there isn't a lot of evidence for rampant female hypergamy in these charts, and it doesn't look like the 80/20 rule is correct. Based on the charts, the top 20% of men are receiving 40% of the initial messages from women.

Still, I think I have explanations for why men find dating difficult:

First, men send more messages than women. From OKCupid: "Straight men are 3.5 times more likely to send the first message compared to straight women." This can result in men feeling like they're taking action and not getting a lot of results or validation. Meanwhile, women can avoid taking action, but still get results. And they are largely shielded from the pain of rejection since they can simply pick and choose from the men who have approached them.

Second, there are more men than women on dating apps and websites. I've seen some data from OKCupid showing that there were about 1.5 men for every woman on OKCupid, and other data showing 1.8 men for every woman on OKCupid.

The combined effect of men sending 3.5x as many messages and if there are 1.8x as many men means that women receive 6.3 messages for every message they send. This means the actual number of messages received by both genders looks something like this:

https://i.imgur.com/powehHB.png

This chart is fairly close to the chart released by OKCupid:

https://i.imgur.com/54jNjCA.png

This chart also undermines the claim by Christian Rudder than unattractive women are being ignored by men: "So basically, guys are fighting each other 2-for-1 for the absolute best-rated females, while plenty of potentially charming, even cute, girls go unwritten." I also thought it was interesting that a guy in the 99th percentile received about 30% fewer messages than a woman in the 50th percentile.

An additional factor in men's dating difficulty is that these charts don't examine what happens after the first message or first response is sent. I've been in plenty of conversations where women have suddenly ghosted. While I'm sure that happens to women, too, I think there is evidence that women ghost men more often than vice-versa. I'm reminded of that Tinder experiment where a woman ran a man's Tinder and she complained about how she'd get no responses and get ghosted far more often when she was running a man's Tinder profile than she did when she was on her own Tinder profile. She said:

"I struggled. Even in the conversations [that happened] I had to lead. Some of them put zero effort. In the last [few days of the experiment], I was like "I hate this. I don't want to do this again." ... I didn't understand what was the problem. It's weird to me. This whole thing is weird because guys don't do this on dating apps. They just don't stop replying. They don't do that. They don't ghost. And it's weird that women do that so often... I just feel like Tinder is unfair as hell. This is all a very weird reality. And maybe I was ignorant. I didn't know this was like this [for men]. I just feel sorry for guys. Like, no, I don't feel like this is good for anyone."

Based on data from the attractiveness chart, what could be going on is that - even when men and women at the same percentile start talking to each other - men are already attracted to the women they're talking to, while women are only somewhat attracted to the men, and they expect men to compensate for her lack of attraction by being extra interesting and engaging. This makes the conversation stage much more unstable for men because they have to bring a lot more to the table than women do. An additional explanation is that women have so many more options based on the number of men sending them messages and the fact that there are twice as many men as women on the website, and that results in women become much more flakey.

(To illustrate the point about the attractiveness chart: if a man in the 90th percentile is talking to a woman in the 90th percentile, then, based on the attractiveness chart, she sees him as a 4 out of 7 in attractiveness. Whereas, a woman who's in the 90th percentile is seen as a 6 out of 7 in attractiveness. For men and women at the 50th percentile, the man is seen as a 2 out of 7, whereas the woman is seen as a 4 out of 7. When women are talking to men at the same percentile of attractiveness, she sees him as quite a bit less attractive than he sees her. Thus the reason women expect more in the conversation to win her over, and the reason for the higher flake-rate.)

I should add that some other data has suggested that women are slightly more hypergamous than men. For example, this chart from the "Gendered Interactions in Online Dating" paper showed that women were slightly more likely than men to message the opposite sex who were in a "higher" attractiveness tier than they were. Data from Hinge shows a similar pattern: "The top 1% of guys get more than 16% of all likes on the app, compared to just over 11% for the top 1% of women." The pattern is similar for the top 5% and top 10% of men and women on Hinge.

The end result being that men have a variety of factors stacked against them in dating - and some of these difficulties might end up being attributed simply to hypergamy when it's actually a combination of things:

  • Too many men and not enough women results in lots of competition, and women picking between many options.
  • Even when conversations happen, it seems like women are less attracted to their equivalent male counterpart, so they seemingly want men to "make up the difference" by being extra interesting, funny, and engaging. This results in conversations where a disproportionate number of women will ghost or unmatch.
  • Some level of hypergamy by everyone, but it seems like women do it slightly more. (It's unclear from the OKCupid data if that's true, but other sources seem to confirm it.) Of course, some of this might be driven by the fact that, when there are more men than women on a dating website or app, women can more easily "date up".