Dating apps are mostly a waste of time for guys unless you are really hot and the rest of us will all die sad and alone — a quantitative socio-economic study: Hinge Edition
Abstract (TL;DR)
The purpose of this article was to review and augment the analysis of Hinge Junior Growth Engineer Aviv Goldgeier concerning the inequal distribution of likes across the dating app Hinge. The analysis was carried out both as a function of ranked level of attractiveness and gender. These data were also compared and contrasted to a similar set of data previously reported by the author about the dating app Tinder. The Hinge ecosystem was treated as an economy with “likes received” being the currency. The distribution in wealth in the Hinge ecosystem was tabulated and analyzed to determine the Lorenz curve, Gini coefficient, and distribution of “likes” by ranked attractiveness for straight male and female Hinge users.
The Gini coefficient for straight men on Hinge as calculated by Goldgeier is 0.542. This would place Hinge as having more wealth inequality than 93.5% of countries in the world. This value is similar to the Gini coefficient of 0.58 calculated for straight men on Tinder. The Gini coefficient for women is 0.376. This puts women on Hinge roughly in the middle of the world country rankings (slightly more equal than the world average). From this analysis it was found that the most attractive men receive approximately as many likes as the average woman and that the bottom 50% of all men combined collect 1% of the total likes received.
In conclusion, if you are in the bottom 50% of men in terms of attractiveness you won’t get many matches (if any) on Hinge and will likely die sad and alone… or maybe we are looking at this all incorrectly and there is hope after all.
Introduction
A few weeks ago, I discovered that a random blog article I wrote seven years ago had somehow become recognized enough to be mentioned on Real Time with Bill Maher. The article was about how “likes” on the dating app Tinder are unequally distributed in the same way that wealth is unequally distributed in a national economy. More specifically, the article was about how most likes go to the most attractive men on Tinder leaving everybody else fighting for scraps. I discovered many other blog articles, websites, and reddit boards that cited and discussed my original blog article while tracking down the spread of my article across the internet. One of these citing articles deserves an individual examination due to both its relevance to the original article and the craziness that it exists at all.
It turns out that a data scientist working for the dating app Hinge named Aviv Goldgeier had read my article and used his access to Hinge user data to recreate the analysis I had performed. Apparently, this was part of his job (unlike me who did the original analysis to justify my inability to score hot make out sessions with cool babes). The Hinge sponsored blog that reported on his analysis has long been taken down, but I was able to find a copy of the original article using the magic of the internet wayback machine.
Goldgeier’s analysis solves two of the biggest problems with my original work. My original analysis was limited to 27 data points I got from interviewing women that I had catfished on Tinder. It had a small sample size that was susceptible to self-reporting bias (not to mention ethically dubious at best). Aviv had access to data from thousands of accounts and it was real actual data, not self-reported data. So how did Aviv’s analysis of Hinge compare to my study of Tinder? Do dating apps work or will we all die sad and alone? I already put the answer in the article title but stick around anyway and you just might learn something.
Don’t look back in anger
The main premise of my initial analysis was to consider Tinder as an economy. The wealth of an economy is quantified in terms of its currency. In most of the world the currency is money (or goats). The basic aim of the article was to see what would happen if you treated “likes” on Tinder as a sort of currency. The more “likes” you get the more wealth you have in the Tinder ecosystem. Wealth in an economy is not distributed equally. Attractive guys have more wealth in the Tinder economy (get more “likes”) than unattractive guys do. This concept isn’t surprising but begs for further quantification.
I need to pause here to add a disclaimer. In this case we are defining “the most attractive guys” as the ones that get the most likes. This assertion has a bit of circular reasoning built into it and therefore it can’t tell us what it is about these men that make them so attractive. It can only attest that for some reason they get the most likes. It can be assumed that physical appearance has something to do with attractiveness (since this is Tinder we are talking about), but a man could also be attractive because he has a really cute dog (or really knows how to hold a fish well). In this study the men with the most likes ARE the most attractive, by definition, and therefore also have the most Tinder wealth.
Tinder doesn’t supply any statistics or analytics about member usage, so I had to collect this data myself. I was able to collect data about how incoming “likes” were distributed across straight male Tinder users and compare the “likes” inequality across Tinder users to the wealth inequality in other traditional national economies. It turns out that the bottom 80% of men are fighting over the bottom 22% of women and the top 78% of women are fighting over the top 20% of men. This was summed up by the title of the study, “Guys, unless you are really hot you are probably better off not wasting your time on Tinder — a quantitative socio-economic study.”
Gini in a bottle
One metric often used to quantify and compare the inequality of an economy is the Gini coefficient. The Gini coefficient (Wikipedia link) is a number between 0 and 1, where 0 corresponds with perfect equality where everyone has the same income (damn commies) and 1 corresponds with perfect inequality where one person has all the income and everyone else has zero income (let them eat cake).
The world average Gini coefficient is 0.39. The US is a little higher at 0.41 which is in the 63rd percentile. When I calculated the Tinder Gini coefficient I found it to be 0.58. If Tinder were an economy it would have more wealth inequality than 95.1% of the countries in the world! So how does the Hinge economy compare?
Figure 2 compares the income Gini coefficient distribution for 162 nations and adds the Tinder economy and the Hinge economy to the list. The Gini coefficient for straight males on Hinge as calculated by Goldgeier is 0.542. This would place Hinge as having more wealth inequality than 93.5% of countries in the world. This is surprisingly close to the value I calculated for Tinder with my limited data set. Goldgeier also mentioned that half of all likes sent to men go to the top 15% of men. This is also very close to the Tinder data. I previously calculated that half of all likes sent to men go to the top 14% of men. This both validates my original analysis and suggests that the inequalities are similar in the Tinder and Hinge economies.
There can only be one
The Gini coefficient is calculated from the Lorenz curve. The Lorenz curve (Wikipedia link) is a graph showing the proportion of overall income or wealth held by the bottom x% of the people. If the wealth was equally distributed the graph would show a straight line between (0%, 0%) and (100%, 100%). The amount the curve bends below the line of equality shows the extent of wealth inequality.
A Gini coefficient can be calculated for every Lorenz curve, but there are an infinite number of Lorenz curves for every Gini coefficient. Therefore, the Lorenz curve can’t be determined from the Gini coefficient alone. One great example of this comes from a paper by Christian Damgaard and Jacob Weiner entitled “Describing inequality in plant size or fecundity.” I will paraphrase their example in terms of income (the original was based around variations in plant size):
Imagine two different countries: Country A has nine citizens that make $55,560 dollars a year and one person that makes $500,000 a year. Country B has five citizens that make $20,000 a year and five that make $180,000 a year. Look at Figure 3 to visualize these two economies. Which economy has more inequality? According to the Gini coefficient they both have the same amount of inequality. The Gini coefficient is 0.444 for both countries.
Knowing the shape of the Lorenz curve for the Hinge economy would be very useful for further analysis, but unfortunately it is impossible to know the exact shape of the curve from just the Gini coefficient. Luckily, we also have one more piece of information to help us determine the shape of the curve. Goldgeier mentioned that half of all likes go to the top 15% of men. This allows us to pin down one point on the curve. Using these two data points I was able to estimate what the Lorenz curve for the Hinge data might look like. Although the curve meets the two conditions of having the correct Gini coefficient and the correct percentage of men that receive half of all likes, I can’t be certain this is the exact shape. It is just a guess, but probably a decent guess.
Figure 4 shows how the estimated Lorenz curve for the Hinge economy compares to the United States economy and also to the Lorenz curve I calculated for Tinder from the limited data set I collected for my previous article:
The Lorenz curves for Tinder and Hinge are very similar, especially at high levels of wealth, with Hinge having slightly less inequality on average. Both are more unequal than the United States economy.
What a girl wants
In my original study I only collected information on the distribution of male likes by interviewing women about their swiping patterns. Goldgeier had information available to him on both men and women. How does the Hinge economy compare for men and women? The economy of likes for women is much more equally distributed than for men. The woman Gini coefficient is 0.376. This puts women on Hinge roughly in the middle of the world rankings (slightly more equal than the world average). Additionally, half of all likes sent to women go to the top 25% as opposed to the top 15% for men. Figure 5 shows the women ranking added to the previously shown Gini coefficient graph (Hinge — F).
Once again, we can’t exactly determine the Lorenz curve from these data, but we can estimate what the shape of the curve looks like. Figure 6 illustrates the estimated Lorenz curve for female Hinge users and compares it to the curve for male users. We can see that the inequality for women is significantly less than the inequality for men. It is difficult to comprehend how these different levels of inequality affect the Hinge economy just by looking at Gini coefficients and Lorenz curves. Therefore, a deeper analysis is needed.
Unhinged
Deconstructing the Lorenz curve allows us to determine the relative number of likes Hinge users receive at different levels of attractiveness. Again, in this case, ranked attractiveness only refers to the number of likes received and says nothing about whether the person is physically attractive, has a good personality, or just has an interesting profile. Figure 7 shows this trend for female and male Hinge users. In this graph, a ranked attractiveness of 100% represents the most attractive people in that group. The difference between these two curves illustrates the difference in the Hinge experience between women and men. The best way to look at this graph is to start from the 100% ranked attractiveness condition (the most attractive people on Hinge) and work down the attractiveness scale. The increased inequality in the distribution of likes for men as compared to women is already evident by the time we drop to 80% ranked attractiveness. Women ranked in the top 80% of attractiveness still receive 64% of the likes of the most attractive women. Men ranking in the top 80% of attractiveness only receive 37% of the likes of the most attractive men. This disparity increases as you work your way down the ranked attractiveness scale. The average women on Hinge (50% ranked attractiveness) still receives 28% of the likes of the highest ranked women, but average men only receive 8.7% of the likes of the highest ranked men.
This divergence is amplified when you take into account the difference in total number of users and average number of likes given between women and men. Goldgeier pointed out in his Hinge blog that men on average give over three times as many likes as women. Other sources indicate that men outnumber women on Hinge by approxmiately 2 to 1. Using these data we begin to see a more complete picture of the distribution of Hinge likes. Figure 8 shows distribution of likes by ranked attractiveness again, but this time the likes for both men and women are normalized to the most attractive women. The most attractive men receive 28% of the likes of the most attractive women. In fact, the most attractive men only receive as many likes as the average women. The men ranked at 80% attractiveness receive less likes than the women ranked in the bottom 10% of attractiveness. So, unless you are in the top 20% of men you are not going to get very many likes on hinge.
Figure 9 shows the ratio of likes received by women to the likes received by men for the same ranked attractiveness percentage. The ratio decreases as ranked attractiveness increases. The most attractive women receive 3.6 times the likes of the most attractive men. The average women receive 11.6 times the likes of average men. At the low end of attractiveness, the women in the bottom 10% of attractiveness receive 42.6 times as many likes as the men in the same range. So Hinge can still be effective for less attractive women, especially compared to less attractive men.
I think we’re alone now
So obviously it is bad to be a man of lower than average attractiveness on Hinge, but how bad is it really? Pretty bad. Figure 10 shows a break out of the percent of total likes for men and women for the groups of 80% — 100% attractiveness, 50% to 80% attractiveness, and 0% to 50% attractiveness. Women between 80% and 100% attractiveness receive 37% of all of the likes on Hinge. Women in the bottom 50% of attractiveness still receive 18% of all likes. This is more than all men combined, which receive 14% of the total Hinge likes. Most of these likes (9 out of 14) go to the men in the top 20% of attractiveness. So how bad is it to be a less than average attravtiveness man? The bottom 50% of men combined, which represents 1/3 of the total Hinge users, only receive 1% of the total likes. In conclusion, especially for men, datings apps are a waste of time unless you are among the most attractive, and the rest of us will likely die sad and alone.
So Aviv, was I close?
Epilogue — Can I get a connection?
It isn’t surprising that Hinge would take down an article about how hard it is to be a less than uber attractive man on Hinge, even if they didn’t actually show the full analysis to demonstrate just how bad it really is. There is some hope though. It may be nice to get a lot of likes in your dating app inbox, but the actual goal is to meet people, fall in love, and have lots of hot make out sessions. This requires two people to like each other and connect. Goldgeier discussed the differences between likes and connections. It turns out that connections are much more equally distributed than likes. The Gini coefficient for male connections was calculated to be 0.324. This is even more equally distributed than female likes. Figure 11 shows how likes and connections are distributed as a function of male ranked attractiveness (based on estimated Lorenz curves). The lowest 10% of men in terms of attractiveness receive about 1% of the likes of the most attractive men, but 18% of the connections.
Even though low attractiveness men don’t receive many likes, they tend to connect with the likes they do receive much more than attractive men. Luckily for less attractive men, attractive men seem to reject most of their likes leaving plenty of connections for the rest of us. So maybe we won’t die sad and alone after all. Maybe you just need to keep on hoping and keep on swiping.
— Worst Online Dater
