| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|---|---|---|---|---|---|
| 1 | 2 | |||||
| 3 | 4 |
5 |
6 |
7 | 8 |
9 |
| 10 | 11 | 12 | 13 | 14 | 15 | 16 |
| 17 | 18 | 19 |
20 | 21 | 22 | 23 |
24 |
25 |
26 |
27 | 28 |
29 |
« The dark side of social networks - Insights into a WoW goldfarmer's life | Main | Social-network sites give businesses ideas for new collaboration »
18 June 2007
Peter Gloor and Daniel Oster, some of my collaborators from MIT and the University of Cologne, respectively, recently showed that they were able to predict neuroticism and extroversion on a standard psychological test (NEO-FFI) using data from our Sociometric badges in an experiment at a German bank's marketing division (see this earlier post for a description).
They examined the disparities in the infra-red (IR) hits between individuals, using this to determine how "balanced" their communication was. One badge can receive an unreciprocated IR hit from another badge because IR requires a direct line of sight to transmit data. Therefore, if one individual is not directly facing another person, they may recieve IR transmissions. Their conversation partner, however, will not receive anything.
The authors then processed this sensor information into a higher level "contribution index". The contribution index was previously defined by Gloor for e-mail data in online standards communities, and it is essentially: (number of messages sent – number of messages received)/(number of messages sent + received). Here it is defined as numbers of infrared readings picked up by an actor. The higher the value, the more signals an actor picks up. A contribution index of 1 means that an actor looks at other people and is never looked at, while a contribution index of –1 describes an actor who is only looked at while never looking another actor squarely into the face.
The authors found a correlation of -0.74 with neuroticism and a 0.52 correlation with extroversion, which intuitively makes a lot of sense. This work shows the promise of using precise sensor data to map not only social relations on a large scale, but using it to quantify individual personality types as well.
This work is currently under review for submission.
Posted by Ben Waber at June 18, 2007 5:16 PM