Complexity and Social Networks Blog

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 5:16 PM | Permalink | Comments (0)

Some days ago I attended a talk on human information processing by Thomas Mussweiler from the University of Cologne who spoke at the Columbia Business School. Mussweiler and colleagues conducted an impressive number of experiments on the mechanisms and influences of individual information processing. A simple example would be to ask you to determine your best athletic performance. You have two basic options: 1) You think of every single athletic moment in your life, i.e. you engage in absolute information processing, or 2) you compare what you recollect as some of your best performances to a given standard, e.g. a famous athlete’s performance (or a famous couch potato’s performance). Not surprisingly it turns out that comparison allows to process information in a more efficient manner.

Mussweiler went on to talk about various factors that influence the comparisons we make, most importantly the standards we employ for comparing information. His experiments used a technique called “priming” to activate certain standards – for example, subjects were asked to judge a trait in a person. The result shows that priming a trait concept (such as aggressiveness) will induce the subject to judge the target person according to that trait. In other words, once activated, standards are spontaneously compared to the target person.

While I was listening to the talk, I kept asking myself how the way we process information relates to how we search for it. Some possible bridges might be that the search itself is the result of some form of information comparison (my search is triggered by a comparison of the information I have to a “standard”, which is the knowledge I believe I need to possess), and/or that we subconsciously use standards to determine the source to turn to when searching for information. I don’t know if there’s literature out there that links cognitive psychology to advice networks, but it's definitely something useful to look into.

Posted by Maria Binz-Scharf at 12:48 PM | Permalink | Comments (4)