Anyone reading this blog post who is enrolled in this class or in any other large lecture has almost certainly experienced the phenomenon of students starting to pack up shortly before class is actually over. A few students will start packing up as soon as the professor (or the clock) intimates that the lecture will be ending soon, and in the following 15-20 seconds more and more students jump on the packing up bandwagon until the professor is forced to end class because pretty much everyone has packed up and is ready to leave.
I had never given this kind of event too much thought in the past as it seems to be a ubiquitous and organic part of the conclusion of a large lecture. However, in reading The Tipping Point and thinking about the “power of the few,” mentioned therein I realized that this could be easily modeled in social network type of situation. Below is a link to a small Java Applet I created using NetLogo
that does just this.
When the simulation is run with the default settings, five students decide to start packing up for some reason or another. When their “neighbors” (i.e. the cells in the simulation to which the packers have edges) see that this is happening, they might decide to pack up as well. The likelihood of their deciding to pack is determined by their awareness of the situation and their susceptibility to peer pressure. Also, the more total students start to pack up, the more likely other students are to start packing up when they see their neighbors doing the same. This is meant to account for the sort of whole classroom rustling effect as the packing up stampede gains momentum.
For reference, the equation that underlies this model is:
A student will pack up when one of his or her neighbors packs up and: random(1,100) < Sensitivity % * Peer Pressure % * # of People Currently Packing Up
Since this is just meant to be a quick and dirty model of something I noticed, I won’t try to draw any conclusions from it. However, with many of the settings (play around with them!) a noticeable tipping point is reached once some critical mass of people start packing up, and soon after that everyone has packed. Hopefully this model will inspire people to think more about how commonplace social phenomena can be modeled using a (in this case, a very rough) sort of social network model.
[…] Fred Billups provides a very interesting simulation he created for a familiar experience — the way in which a few people packing up at the end of a large lecture class can trigger a large cascade of further students deciding to pack up. We’ll be talking a lot more about such cascading behavior later in the semester. Interestingly, in one of the papers we’ll be discussing — Mark Granovetter’s Threshold Models of Collective Behavior — the case of people leaving a lecture is actually explicitly invoked as an example. This type of packing-up phenomenon can also be viewed as an instance of synchronization, the subject of Steve Strogatz’s recent book Sync. […]