Of Science and Choice In Online Dating

Must be summertime, as online publications turn to the contemplation of Internet dating. The NY Times's piece (registration may be required) takes a not particularly deep look at the reality behind the "science" claims of chemistry.com, eHarmony.com, and others. "The question is how much it really matters to users if the methods have any scientific basis. A friend of mine... said she looked at several dating sites and chose the ones that looked like they had 'the least riffraff.'" Technology Review focuses on studies showing that the overwhelming number of choices presented by many dating sites can be counterproductive: "...more search options lead to less selective processing by reducing users' cognitive resources, distracting them with irrelevant information, and reducing their ability to screen out inferior options." The article concludes with a look at the startup Omnidate, which offers technology for 3D virtual dating. The site has had twice as many women (by percentage) sign up as the other dating sites typically see.

Comments

There's also okcupid

[Colin Smith]

Run by a couple of maths grads. Last time I looked they were using a regression analysis to match people.

The site's also free.

 

Re:Science, lol?

[radtea]

Do people even know what they want from a partner?

Yeah, they do. 99.9% of women want "a good man who loves to laugh and is fun and just an ordinary guy."

I'm a divorced man in a small (~100,000) town and have used online dating sites off-and-on for about five years--mostly Plenty Of Fish, but also LavaLife and OkCupid. I've met two absolutely wonderful women this way--both of whom were so wonderful that after a year or three with me their careers took them off to bigger, far-distant centres, although in both cases we're still friends.

I've also met the biggest collection of flakes, losers, liars, bores and nutjobs you could possibly imagine, and I am currently ready to slap anyone whose entire self-description is, "I love to laugh, like long walks on the beach and am just looking for an ordinary guy."

Seriously, have you ever met anyone anywhere who doesn't like to laugh? It's what we laugh at that's interesting, and hardly anyone ever says what that is.

The trick for all these sites is to weed out the common things that everyone has, and to reduce people who have zero self-awareness to abject silence until they come up with sufficient self-knowledge to say something about themselves that isn't woefully banal. OkCupid's system of questions does that, although I can think of some simple improvements that would make it better.

The key thing is to focus on the concrete. There should be very nearly zero abstraction in any of the information gathered from users, and the site should then generate the abstract categories the user is assigned to based on that information.

For example, don't ask people what their "body type" is (abstract category) but what their height and weight are, how fast they can run or walk a mile, how many miles they run or walk each week, when was the last time they walked more than a mile, or biked more than a five miles, or swam more than 500 m, and so on. Then generate the abstract category for them: "couch potato", "morbidly obese", etc, rather than letting users define "athletic" or "slim" or "average" any way they want to (I've seen morbidly obese people, who have posted pictures of themselves, categorize themselves as "average".)

Mostly, these sites are selling fantasies to liars (women) and idiots (men), so doing anything that would provide more accurate information about what differentiates one person from another is counter-productive relative to their business model. The few honest, intelligent people out there have to wade through a huge amount of dross to find each other. Fortunately, that is still possible, and despite their flaws these sites remain a sensible component of anyone's search for companionship. Just be prepared to do a lot of filtering by hand.