Some things we would like to show:
-
When we look at the global fraction (event magnitude vs. agent importance) graph, we see that, for a given interaction rule and decay, regardless of your importance, the event succeeds or fails. For example with slow decay (which is equivalent to fast interaction) it really doesn’t matter who you are, if your event is important then it will prevail. However in societies where interaction is slow/weak, the node’s importance will matter more. Real life parallel: in societies, such as ours, where communication is fast and easy, content is more important than how important you are (now assuming that important people don’t hush away content that’s not theirs). Average people have a chance to create viral funny videos, and people who experience and document injustice can bring awareness to a cause, or people can organize strikes from one day to the next. However in societies, where communication between people is slow, rare, difficult (decay is high), the events of important people matter much more (this still needs to be verified by our simulations). In such societies the events of the plebs don’t become viral so easily, and it is intrinsically difficult to oppose a regime. E.g. if the state controls most of the media and maybe blocks some forms of social media, or if the people are poor in the country and don’t spend so much time communicating over phone/internet, then injustice happening to people will not get conveyed to each other, and they won’t unite. To strengthen this argument, we have to look at different decay, and different interaction rules maybe.
-
We should make a graph that disregards the network structure completely: let’s only look at event magnitudes. Compare events with the same magnitudes and see what the distribution of their global awareness is. I.e. make a plat with event magnitude as y axis, max. awareness as x axis, and use colors to plot how many events are at each point. This is very similar actually to the other colored plot we have. It basically shows the effect of the graph for us. If the distributions are sharp, then we can infer that the social network’s structure isn’t important. But if the distributions are wide or skewed, then the social network really makes a difference. Since in the previous graph we observed that the social relations don’t matter that much I would expect sharp distributions.