le 25 février 2014
The main argument: The sciences that focus on human behavior, meaning the social sciences, have traditionally relied mainly on surveys and lab experiments in their investigations. While valuable to a degree, these sources of evidence do have their shortcomings. Most significantly, surveys offer but indirect evidence of human behavior (and can also be compromised by deception and self-deception); while lab experiments tend to be somewhat artificial, and fail to capture the complexities of real life.
Recently, however, new digital technology has opened up a whole new way to study human behavior. This proves to be the case since mobile devices and sensors of all kinds are now able to record a dizzying array of human activity—everything from where we go, to what we buy, to whom we interact with and for how long, to our body language, and even our moods etc. When placed in the hands of social scientists these new sources of information can prove very valuable (and are far preferable than either surveys or lab experiments); for they allow scientists to study us in our natural environments—out in the real world—and they also allow scientists to study what we actually do, rather than what we say (which are sometimes quite different).
The method of investigating human behavior in our natural environments using digital technology has come to be called reality mining, and it is revolutionizing the social sciences.
One of the pioneers and leaders in the field of reality mining is Alex Pentland, a researcher out of MIT. Pentland’s main field of interest is using reality mining to explore the properties and patterns of interactions between people—what he calls social physics. Specifically, Pentland uses reality mining to investigate the social physics in a wide range of groups and situations, from social and peer groups; to social media platforms; to institutional settings such as schools and businesses; to even whole cities. And in his new book Social Physics: How Good Ideas Spread—The Lessons from a New Science Pentland takes time out to catch us up on his findings.
One of Pentlands’s main findings thus far has to do with the importance of social interaction in influencing our behavior. Indeed, Pentland has found that much of our behavior is dominated by the influence of our close relations and the peer groups we are embedded in—everything from our diet and body weight to our political opinions and all things in between.
The influence of our social world is so great, in fact, that Pentland argues it is much more appropriate to think of ourselves as group-oriented than self-directed. This is important because Western society as a whole tends to take the opposite view. The result is that many of our policies and institutions are ill-fitted to our true nature—which leads to less than desirable outcomes. Thankfully, Pentland does offer some advice with regards to how we can re-design our policies and institutions in a way that better accommodates our nature.
A second of Pentland’s main findings has to do with how ideas and behavior spread through human interactions and groups—and also, and even more important, what kinds of interactions produce the best results in terms of generating the most creative and productive ideas.
Specifically, Pentland has found that the most creative and productive groups tend to have something very important in common: the group members have numerous interactions with highly diverse people outside of the group, and the group members are also highly connected to one another.
In terms of explaining why this pattern works best Pentland argues that the interactions outside of the group are important in becoming familiar with many different types of ideas, while the interactions within the group function to winnow out what are the best ideas, and also help build common norms of behavior and trust that allow the group to work well and cooperatively together.
I was happy to get the opportunity to learn about a very new and promising science from one of its leading practitioners. Many of the ‘living lab’ experiments outlined in the book are very interesting and I certainly learned a lot. My only complaints are that the book does have a fair bit of repetition and jumps around some, so I question the writing and organization a bit. All in all, though, a very good and interesting read about a new field that we are sure to hear more from moving forward.
le 12 mars 2014
Le Professeur Sandy Pentland vit dans le futur au MIT Media Lab, et ce livre est son témoignage.
Écrit pour le grand public, Social Physics nous emmène aux frontières de la recherche; où la technologie devient l'instrument de mesure des sciences sociales et du comportement et où les mathématiques deviennent leur nouveau formalisme.
Social Physics décrit la recherche de Pentland et de son groupe de recherche "Human Dynamics": des badges intelligents appelés à remplacer nos cartes d'accès, le "reality mining" et les techniques d'analyse des meta-données de téléphones portables ou encore openPDS, une approche open-source de la protection de la vie privée.
Autant d'applications concrètes qui donnent de bons exemples de ce qui nous attend au 21ème siècle!
A l'aide de senseurs, l'auteur a capté les interactions de différentes équipes (équipe projet, centre de services téléphonique ou groupe d'étudiants) : sur plusieurs jours, il a pu suivre les conversations, les gestuelles, le ton de la voix, etc. En analysant ces milliards de données, il a pu démontrer que plus nos interactions sont variées et nombreuses, plus nous sommes productifs et créatifs. Pas d'idées réellement neuves, mais la confirmation d'intuitions sur l'importance de favoriser les échanges au sein des équipes et entre elles.