TLDR #32 - An Imperfect Match

Monday, August 04, 2014 - 09:58 AM

Transcript

Last week, dating site OK Cupid put up a blog post describing experiments it conducted on its users. In one experiment, the site told users who were bad matches for one another that they were actually good matches, and vice versa. Alex and PJ talk to OK Cupid President and co-founder Christian Rudder about the ubiquity of online user experimentation and his defense of potentially sending OK Cupid's users on bad dates.

Thanks for listening. If you like our show, please subscribe to us on iTunes. Or you can follow PJ and Alex and TLDR on Twitter. 

Tags:

More in:

Comments [4]

Kevin Stevens from Buffalo

To say that he is employing the scientific method is disingenuous at best.

Oct. 08 2014 01:43 PM
Harold Goldner from Southeastern PA

I think you got this one dead wrong, and I found the representative from OK Cupid amazingly patient with your naivite. He was right, PJ's restaurant analogy was off-base.

This *isn't* eating at a restaurant. Persons engaging with a website are accepting terms and conditions offered at the website. There is an ancient common law concept of "assumption of the risk." Persons who are aware of certain risks and voluntarily encounter them "assume the risk" that something might happen because of their confrontation of that risk.

This isn't brain surgery; you don't get "informed consent." You sign up and participate in a dating website (or any other website for that mattter) or you don't. Everybody "experiments" with their product. If I had to tell my client every time I chose a different set of words in a brief for stylistic reasons or strategic reasons, I'd be completely incapable of practicing law.

You don't like what they do? Don't buy the product. Vote with your feet. But protesting because "we didn't know we were being used" is just plain disingenuous.

I love your stuff --- really I do, and listen to every podcast. I think you just got this one completely wrong, and tell Bob I said high (he's an old classmate).

Aug. 07 2014 11:19 AM
matt h

What an arrogant creep and a terrible spokesman for that company. Total jerk.

Aug. 06 2014 02:27 PM
Adam from Oregon

This episode reminded me of an episode of Planet Money which contrasts giving to developing countries:

http://www.npr.org/blogs/money/2013/08/16/212645252/episode-480-the-charity-that-just-gives-people-money

On Planet money, there was a very emotional reaction by the person defending traditional methods of philanthropy toward developting nations to the idea of direct giving. This, I think, stems from that person's ability to directly observe some instances where their kind of giving works: people are fed rather than hungry, etc. The argument of the direct giving proponent is basically: "How do you know unless you test?" to which the traditional giving proponent says: "You can't test, these are real peoples lives."

This seems incredibly similar to the argument in this episode of TLDR, although the people arguing the different sides is mixed around. PJ is saying, basically, that OkCupid works: "If I were single, I'd be back on your site", and the OkCupid rep is saying "How do you know unless we test?" and then the guys are saying "You can't test, these are real people."

To me, this is an issue of priority: If OkCupid didn't exist, nor any dating site, and we did the study, we would obviously want the OkCupid that is better. But the site exists before the study, and the fact that it seems to work undermines our ability to think rationally about whether it ought to be different in order to work better.

Listening to these two episodes, it also seems like outcome has a lot to do with public attitude. The Planet Money guys are very pro-direct giving. Direct giving has also been show to be effective in some small studies. Here on TLDR, the guys seem to be against the kind of study that OkCupid had done, but I have to wonder if they would have been if OkCupid had found that its algorithm is very ineffective.

Aug. 04 2014 12:48 PM

Leave a Comment

Email addresses are required but never displayed.

Supported by

Subscribe to Podcast iTunes RSS

Feeds