Marie Claire has a fascinating short interview with psychologist Mark Thompson who was apparently hired by a big name internet dating website to work on ‘scientific matchmaking’ – but recently jumped ship when he became disillusioned with the industry.
Buyer beware: the guy has just written his own book on sex and relationships, although his comments on dating sites don’t seem to directly bear on his book promotion efforts.
Regardless, it’s actually quite refreshing to hear someone give a sensible take on the limit of ‘scientific matchmaking’ as, since it has become popular, science news has regularly been bogged down by lots of poorly disguised PR fluff based on exaggerated findings or dodgy unpublished ‘statistics’.
MC: What made you leave e-dating?
MT: I hated the way we overpromised and underdelivered. Our studies showed that the odds of meeting someone online and dating him more than a month are roughly one in 10. So it’s great that all those people on the TV commercials met their spouses, but they are the exceptions, not the rule. No computer can accurately predict whom you should be with. The function of the math will make vastly more false predictions than accurate ones.
MC: But isn’t blind dating always hit or miss?
MT: Yes, but you don’t have to pay $30 a month to be set up by your friend. And you don’t go in believing that science is behind the match. There’s a different set of expectations. When diet companies show someone who lost a bunch of weight in six weeks, they have to say, “Results not typical.” I think eHarmony and other sites should do the same.
MC: Do you think online dating can be fixed?
MT: It really depends on people’s willingness to come back and tell us why each date didn’t work out so the system could get smarter. It would be like Netflix, which learns from your preferences to make better predictions for you.
Netflix for dates. Actually, it’s not such a bad idea. “If you liked this date, you might also like…” could actually come in useful you had the hots for the other person, but they weren’t so keen on you. Or just even if you’re not in it for the long-term thing perhaps.
Link to Marie Clarie interview with Mark Thompson (via @DrPetra).
4 thoughts on “Falling out of love with e-dating”
Actually, OKCupid works in a very similar way- when you view a profile, it shows you alternative matches in the sidebar, tagged according to their personalities.
So you’ll view X, and in the sidebar it says “Y- like X, but less geeky” or “Z- even geekier”.
That already exists, and it’s called okcupid.com. On top of that, the people who run it are crazy-awesome nerds, and blog about their mathematical analyses of various dating-related notions — including an analysis about why online dating as done by for-pay matchmaking companies is doomed; see http://blog.okcupid.com/index.php/why-you-should-never-pay-for-online-dating/
Another problem, which no one ever seems to notice, or mention, is that dating sites are companies with no incentive to effectively provide the advertised service. If you find your perfect match you’ll cancel your subscription and go and enjoy your wonderful new relationship. If you occasionally find a date but it never works out you might continue paying that subscription fee indefinitely. These companies are much better off only half working, and that’s the way they are likely to continue. There’s a kind of natural selection going on, any dating site that actually solved the problems mentioned above would fail before it properly established itself.
Also noteworthy is Dan Ariely’s interview about online dating and why it is “so unsatisfying”.
“Online dating sites assume that people are easy to describe on searchable attributes.”