Last night, I traveled to the beautiful Bell House in the Gowanus Section of Brooklyn for this month's Secret Science Club lecture with Mathematical Sociologist Duncan Watts, author of the book Everything is Obvious* *Once You Know the Answer.
After a brief discussion of his career trajectory, Dr Watts made a wry observation about the sciences, riffing off the cliche "It's not rocket science." Considering the success rate of rocket science, rocket science is easy compared to the social sciences- it's a lot easier to launch a satellite into orbit than it is to predict human behavior. One pitfall in describing human behavior is a tendency to make appeals to "common sense". The utility of common sense is limited to simple day-to-day activities. Much of our perception of the utility of common sense comes from hindsight- Dr Watts cited Paul Lazarfeld's "American Soldier" study, in which assertions made about men serving in the armed forces were justified post hoc- for example, if the adaptability of soldiers from rural backgrounds was posited, it could be justified by stating that they were more comfortable living outdoors for long durations, while an assertion that urbanites adapted more readily to military service could be justified by asserting that it was due to their living in close proximity to others.
Because it is so difficult to predict trends (unless, as Dr Watts observed, you're Nouriel Roubini), it's easier for a company to roll out a variety of products on a limited basis, and discontinue the unsuccessful products while ramping up production of the popular product line.
Another cognitive bias that can cause problems when evaluating something is the halo effect, often we base our judgments about an individual's intelligence or moral character on their physical attractiveness or likeability.
Since I'm a bit pressed for time, here's an article about common sense by Dr Watts. I should mention, however, that he spoke extensively about the role of the internet, and how the sheer vastness of the data that is now available about internet users will revolutionize the social sciences. Some bastard asked a question about the possibility of epistemic closure if consumers choose to seek out "information" that adheres to their personal biases. Dr Watts did indicate that filter bias could occur, but he indicated that the increased amount of information could help to counteract such bias.
On the whole, the lecture was great, but I'm more comfortable with writing recaps in which I can regurgitate a bunch of factoids, so this recap seems perfunctory to me. Dr Watts was an extremely engaging speaker, and I imagine that his new book is an entertaining, informative read with subject matter that is very accessible to the layperson. Getting your hands on it would be a common sense move.
I majored in a Social Science, Economics.
ReplyDeleteI wonder how many of its esteemed leaders in the past few decades have declared their findings based on the "wingnut welfare" effect?
Magick of the Markets, and such as. I'm looking at you, Chicago.
~
My assumption is that 'Rocket Science' is and will always be more straighforward than the social sciences primarily because in the hard sciences everything can be quantified and expressed mathematically. While there have always been attempts to do so in the 'softer' sciences, the plain fact is human behavior may not be amenable to such efforts.
ReplyDeleteHowever, it is worth noting that the very large internet sites are providing input data for unprecedented analyses of human behavior. This is a pretty interesting article:
http://www.technologyreview.com/featured-story/428150/what-facebook-knows/
unless, as Dr Watts observed, you're Nouriel Roubini
ReplyDeleteNow there is someone with experience at predicting trends after they've happened.
I majored in a Social Science, Economics.
ReplyDeleteSocial Science, or Dismal Science?
My assumption is that 'Rocket Science' is and will always be more straighforward than the social sciences primarily because in the hard sciences everything can be quantified and expressed mathematically. While there have always been attempts to do so in the 'softer' sciences, the plain fact is human behavior may not be amenable to such efforts.
Also, those components don't deceive experimenters.
Now there is someone with experience at predicting trends after they've happened.
He's my favorite prophet of doom. I used to have lunch with him about once a week when I had a big blond afro.
Because it is so difficult to predict trends (unless, as Dr Watts observed, you're Nouriel Roubini), it's easier for a company to roll out a variety of products on a limited basis, and discontinue the unsuccessful products while ramping up production of the popular product line.
ReplyDeleteThis kind of thing is easy with easy products. Otherwise, bye-bye Nortel and Research in Motion.
Thanks to the social sciences, I understand that the two strangest environments I've ever lived in were textbook cases. It's oddly reassuring to find out how normal oddness can be. For black and white thinkers, however, this can't be right.
ReplyDeleteHe's my favorite prophet of doom.
ReplyDeleteI do recall a chapter in Black Swan about the pointlessness of imposing government oversight or restrictions upon the finance sector... apparently bankers know the full range of risks better than anyone else, and have self-interest to motivate them -- and the knowledge that any wrong decisions will impact upon them first of all.