I think particularly in the United States status is an under acknowledged variable in predicting social behaviors. Somehow it is considered rude or inegalitarian to admit that you are ignoring or seeking people out because of their social status. However when we are designing social systems to enable social networking, I believe it would be unwise for us to ignore its importance.
Back at Microsoft Research I did a project called Point to Point, exploring knowledge management behaviors in the context of social networks. That is, how do differences in social network variables between two people (number of overlapping people, distance in org chart, relative organizational status) impact information seeking and sharing.
The system bootstrapped the Microsoft network based on communication group memberships. In the study I then had people explore a visualization of the relatedness between them and a series of pairs of individuals, and for each pair they indicated whom would they prefer to meet to seek out or share information.
I found, as expected, that the more social connections you have in common, the more you were likely to seek out or share information with that person. (See Figure below.) What surprised me however was the extent to which this effect was completely shadowed by the impact of status: people were much more likely to accept or seek out information from a high status person—in this case status operationalized as the number of people reporting under you.
People are much more likely to chose to share information with higher status person.
(N = 17, repeated measures design with 16 choices from randomly selected pairs)
Right now, with Pathable we are exploring social matching at events. The question for us is to what extent do we take into account implict goals around social status to determine our social recommendations.
In the field of Social Psychology there’s an entire line of related research around the “Matching Hypothesis”. The basic idea is while everyone might desire a match with as attractive a person as possible, in the negotiation of meeting and getting to know people, they generally end up with someone who is at the same level of attractiveness. (There’s a great class demo of this: have students put cards with numbers on their foreheads facing out. Their task in twenty minutes is to mill around and find a match with as high a number as possible. Amazingly, most people end up with someone a number or two off from theirs even though they didn’t know their own number.)
I see a similar pattern at events: people are hoping to have a meaningful connection with someone who is clearly of higher status. However the higher status person does not want to spend a lot of time talking to lower status people. Consequently over time people are more likely to develop collaborative relationships and friendships with their peers.
Conferences are unusual because while there are great disparities in status, attendees are not always aware of their status level relative to the people with whom they are talking – the only clue is if their badges indicate their job title or whether they are a speaker. Some people prefer to remain in the dark, because it facilitates random pollination across social strata (inspiring more creativity, senior people find good interns, lower status person does not get nervous, etc.) However others are annoyed at wasting their time on people with whom they are unlikely to develop any meaningful connection.
Social status is something we can determine from our Pathable data: the question remains, would people prefer to be aware of relative status in their badges? Should status differentials impact whom we recommend as similar? Interestingly, at FOO camp, an invitation only event with a lot of higher status people in the industry, we were asked to remove status indicators from the Pathable badge. I wonder however how effective this would be at events without any restraint on who’s attending. Wouldn’t Steve Jobs be very annoyed if he had several interns listed on his badge as whom we recommended he meet based only on similar interests?
