Dollars and Sex (10 page)

Before online dating sites started making this data available to researchers, reliable data on how relationships form was extremely difficult to obtain. You might think we could just observe existing relationships, but that information doesn't come close to telling us what we need to know because what we observe to be a “couple” is really the outcome of a dating market that has already closed. The reason why this is a problem is that in this particular market the “equilibrium” (i.e., the couple we can observe) is as much a function of the preferences of the other players on the market (the people we don't end up with) as it is a function of the preferences of the two people who have formed the couple.

Let me give an example of what it means to say that this market is in equilibrium. Suppose (hypothetically) that I looked at data collected from married couples only and found that women with small breasts are often married to men who are balding. If I were very naive I might jump to the conclusion that this is proof that balding men prefer small-breasted women. Of course, it doesn't prove that at all.

It might be the case, for example, that large-breasted women have a particular preference for men with hair, leaving only balding men available to marry small-breasted women. Or perhaps women don't care whether men have hair or not, and the reason we see this particular matching is that men who are not balding prefer large-breasted women, leaving only small-breasted women for the balding men to marry.

Either way, knowing that small-breasted women are often married to balding men tells us nothing about those individuals' preferences for breast size or hair thickness because at the end of the market for spouses, men and women do not necessarily marry the type of person their initial preferences dictate: they marry the person who is both available and willing to be married to them.

I am not suggesting that they don't prefer their partners to other men/ women, only that they prefer them to the subset of all the men/women who are willing to be with them. Who is available in that particular subset of willing partners is determined largely by the decisions made by everyone else on the market.

The example above may seem silly, but let me give you an example from the real world. According to the 2006 U.S. census, there are far more black men married to white women (6.6 percent) than there are white men married to black women (0.2 percent). We know that this outcome is not the result of a shortage of single black women (black women are significantly more likely to be single than women of any other race; in 2007, only 33 percent of black women were married). So what does this data tell us about the racial preference of men and women who are searching for a spouse?

Absolutely nothing.

To understand racial preferences in dating, we need to look at data collected while markets are in action, and some very clever ways have been found to do just that. For example, economists Raymond Fisman, Sheena Iyengar, Emir Kamenica, and Itamar Simonson arranged speed-dating trials in which men and women (Columbia University graduate students) met
with a series of potential partners for short three- to five-minute “dates.” Following the trial, participants were given an opportunity to share their contact information (through the organizer) with those whom they would like to meet again for a real date. The point of this experiment was to untangle whose preferences were driving the racial segregation we observe in marriage data.

Based on the evidence in these trials, it appears that the racially segregated marriage market is driven almost exclusively by the same-race preferences of women; men in this trial exhibited far less same-race preference in dating. White women, on average, preferred to date white men, but the preferences of black women to date black men were even stronger.

This evidence tells us that the reason that white women are married to black men more frequently than black women are married to white men is not because white men are not attracted to black women, but because black women prefer to date men of the same race as themselves.

Another dating market study that looks at partner preferences used data generated from an online dating service that included information on member-to-member first contact e-mails (i.e., “Hi, I read your profile and it looks like we have a lot in common,” etc.) and found similar results to the Columbia speed-dating trial.

Dan Ariely, Günter Hitsch, and Ali Hortacsu find that even after controlling for all other factors that might contribute to an individual's decision to send a first contact e-mail (age, marital status, income, education, children, etc.), every ethnic group “discriminated” based on race and the same-race preferences of women exceeded those of men.

In fact, to measure just how much ethnicity matters to white women, these particular researchers decided to measure ethnic preference against another male characteristic that women seems to care quite a bit about: a man's income.

To do this, they essentially posed a question that can be answered using the data: If a hypothetical woman cares about her potential mate's
income and whether or not he is the same race as she is, what would a specific man's income have to be in order to make a woman initiate contact with him even if he belongs to a different racial group?

In this experiment a hypothetical man earns $62,500 a year and is the same race as the woman. The other hypothetical man earns $X per year and is one of three races differing from the woman's. All other observable characteristics about these two men are identical. The metric that measures how much a woman cares about race over income is how large X has to be in order to encourage the woman to contact the man who is not in her ethnic group.

The results, using the online data, are informative. For a white woman, the experiment predicts that a black man would have to earn $154,000 more than the white man in order for a white woman to prefer to contact him instead of the otherwise identical white man. A Hispanic man would need to earn $77,000 more than the white man, and an Asian man would need a remarkable $247,000 in additional annual income to make her want to contact him.

The results are even stronger for black women. In this case, the experiment predicts that a white man would have to earn $220,000 more than a black man in order for the black woman to prefer to contact him, and a Hispanic man would have to earn $184,000 more.

Alternatively, Asian women appear to have stronger preferences for white men than they do for Asian men—a white man could earn $24,000 less than an otherwise identical Asian man, and an Asian woman would still prefer to contact him.

These income differentials are large, but don't let that fool you into thinking that all women care about is income. In fact, if that were true and income really mattered to women, these numbers would be quite small; it would take very little additional income to entice a woman to date a man of a different race. The fact that the numbers are so large suggests that a man's race is very important relative to his income.

If you are wondering what this experiment tells us about the choices of men, it shows that men care so little about a woman's income that it is completely meaningless as a measure of how men value different female traits. Even though men's online behavior suggests they care less about their partner's race than do women, the income needed to encourage men to make the trade-off between a same- and different-race partner is incalculably large. This doesn't indicate a strong same-race preference though; it just shows that in men's mating preferences a woman's income is irrelevant. To really estimate how much men care about race, you would have to find a different measure that is important to men—like physical beauty.

AN ECONOMIC APPROACH TO THE “MEET” MARKET

This exercise I just described above is a good example of how online dating operates like a market, with individual players trading off characteristics they value in a mate in an attempt to find the best possible match given their own value on the market.

The dating market is like any other market in that it is made up of buyers and sellers, and, just like any other market, it is in equilibrium only after all the prices have adjusted so that buyers and sellers are both willing to trade. In most dating markets, there is no explicit exchange of money; that is not the price that I am talking about. Prices on this market are determined by the forgone alternative opportunities—or as economists like to call them, “opportunity costs”—that people accept when they finally meet a person with whom they choose to have a relationship.

Consider the following as an example: You are searching an online dating site and see the profile of a man/woman who is extremely physically attractive. In fact, if you ranked all people who are currently on your market, he/she would be in the top 10 percent of the distribution in terms of looks. Now, someone this attractive is going to be very “expensive” because everyone on the market will be competing for his/her attention and competition for a limited product always drives up its price. Whether or not
you in particular will be successful at attracting this (incredibly hot) person will depend on your own price on the market because, in this market, everyone is both a buyer and a seller.

When the market clears, I can guarantee that the expensive people will end up with other expensive people, the midrange people end up with the other midrange people, and so on through the market until inexpensive people are left to date the other equally inexpensive people.

When men and women end up in relationships with others who are similar to themselves, economists call their behavior
assortative mating
. There is plenty of evidence that, on average, people end up with a partner who is very similar to themselves in terms of education, income, and even physical attributes such as height, weight, and beauty—that says that assortative mating is very common.

BEAUTY IS NOT REALLY IN THE EYES OF THE BEHOLDER

Just to give you an idea of how competitive the online market is in terms of physical beauty, let me introduce a study that was conducted using data from a website that allows users to rate others' level of attractiveness and, if they choose, send them a message in the hopes of initiating a relationship.

Individual users on the site Hot or Not (
www.hotornot.com
) upload a picture of themselves along with a few lines of text, if they wish to say something. Visitors to the site can rate the level of hotness of members on a 10-point scale based on members' pictures that appear on their screen in a random order. If the visitor comes across someone whom they would like to meet, they can log on and hit a “Meet Me” link indicating their interest in meeting the person in the photo.

Using just ten days of data, the authors of this study (Leonard Lee, George Loewenstein, Dan Ariely, James Hong, and Jim Young) were able to observe the behavior of 16,550 members—75.3 percent male and 24.7 percent female. Each member viewed an average of 144 pictures over the ten-day period, and each of the more than two million observations in the data set is an individual decision to hit the “Meet Me” link. By observing the choices made by users as to whom they would like to meet, researchers were given the opportunity to measure the preference for a mate based on physical attractiveness alone.