MIT Dating Research by Matt McGann '00
A news story today looks at Media Lab research on dating.
There’s a story on the AP wire today (look for it in your local paper tomorrow) on a bit of Media Lab research called the Jerk-O-Meter (?!), which comes out of the Human Dynamics Group, run by Prof. Sandy Pentland (pictured at right).
Ever wonder if that spouse, friend or co-worker on the other end of the phone is really paying attention? The “Jerk-O-Meter” may hold the answer.
Researchers at the Massachusetts Institute of Technology are developing software for cell phones that would analyze speech patterns and voice tones to rate people — on a scale of 0 to 100 percent — on how engaged they are in a conversation.
The article also touches on a related project on speed dating:
The researchers also tested the technology at a bar in Cambridge where a group of singles were “speed-dating,” rotating through a series of five-minute conversations.
“Mathematically modeling” each person’s speaking style let the research team predict whether a speed-dater would agree to a real date. It was a good sign, Madan said, if the speed-daters engaged in “back and forth exchanges,” punctuated by “ahas” and “yups.”
This isn’t the first time Pentland’s dating research hit the press. Last year, the New Scientist‘s Kate Douglas wrote on the topic.
Having trouble attracting women? Here’s what I find attractive in a man: give me lots of short words of encouragement and let me talk. And I’m not the only one. This seems to be one of the key rules to winning women over, according to researchers at the Massachusetts Institute of Technology.
Sandy Pentland and colleagues analysed conversations during speed dating sessions and discovered a way to predict whether couples are attracted to each other. They then turned it into software that runs on a hand-held computer. Used during speed dating sessions, it will show how you’re doing in real-time. […]
Pentland has analysed nearly 60 speed-dating conversations. He can now predict the verdict for 80 per cent of couples and get it right three-quarters of the time; the remainder are too close to call.
Crucially, the researchers don’t study the content of the conversation. Instead they measure how long each person spends talking, how much the pitch and volume of each voice varies and how often one interrupts the other. Pentland thinks of these features as a form of “verbal body language”, the vocal equivalent of crossing your arms or fiddling with your hair. “We’re largely unconscious of it, but it matters hugely,” he says.
Analysing these features produces a set of numbers which Pentland combines with different weightings to give an “attraction factor” . The weight assigned to each feature depends on how much it influenced the outcome of the dates the team recorded: the greater its influence, the greater its weight. So, a real-time analysis could warn you the score was low and suggest what features you might change to up it.
One indicator of attraction was the number of interjections in a conversation that lasted less than a second, typically words or phrases such as “OK”, “I see” and “go on”. “Some people had 30 in 5 minutes, others used only five,” says Madan. If men used lots of these, women were more likely to report attraction at the end of the session.
And, somewhat unbelievably, the Human Dynamics Group isn’t even the only group at the Media Lab doing dating research, as they are joined by the Online Dating Research Center.
If you’re interested in doing research on the dynamics of dating — or the Kitchen of the Future, the Opera of the Future, or LEGO Mindstorms — it just so happens the Media Lab is the largest employer of UROP student researchers on campus. Also, they have a pretty cool freshman year program: “The goals of the program are to introduce students to university research, both how it’s carried out and how current research projects use the concepts presented in first-year subjects; to involve students in the Media Lab community, particularly through the learning-by-apprenticeship that is a hallmark of the MAS academic program; and to expose students to the intersection of technology and communication/expression that is the mission of the Media Lab.”
And that’s just the tip of the Media Lab iceberg…
Interesting. But of course, the speed-dating analyzation only works with women who like to talk. There’s too much generalization for it to ever be really practical. It’s a really cool idea though.
Suh-weet! This sounds like a really awesome idea.
Speed dating is blah, but the idea of predicting outcomes of relationships and/or conversations based on the analysis of vocal patterns and things like that is awesome.
I think I’ll be dropping by the Media Lab for a bit in the fall…see what else is cookin’.
Later,
Mike.
That’s very interesting indeed. I’m suprised that using short interjections often was effective, however.
Speaking of scientists analyzing the mysteries of love…
BBC has a technology news article today about Physicists at Oxford who model radioactive decay looking for insights that may explain how and why people seek certain relationships and/or partners.
http://news.bbc.co.uk/2/hi/technology/4144782.stm
what will they come up with next??? =)
hahaha… i like that!
hmm, it is reasonable because then women feel that they are being respected and that the men are interested in what they are saying. i guess?