Spreading Memes
https://www.scientificamerican.com/article/how-fake-news-goes-viral-mdash-heres-the-math/
This Scientific American article tackles the question of how so many memes go viral. The author, Madhusree Mukerjee, sites different models for how memes may spread. One model claims that memes follows a power law, so “a meme is 4x less likely to be tweeted twice than once” (Mukerjee, 2017). However, if this were really true, then we would not expect so many memes to be so widely distributed. Mathematician James Gleeson provided another model, in which a meme going viral is analogous to the one grain of sand which, if added to a sand pile, causes an avalanche. Essentially, people can slowly start sharing the meme and its audience will gradually build, but there is some tipping point where one more share will lead to the meme going viral.
Mukerjee also discusses how the structure of the network affects the spread of memes. In a random network (where links between nodes are assigned randomly), memes do not go viral. However, in more realistic networks where nodes’ links tend to follow a power law, the hubs which have many links will facilitate memes going viral. This claim, however, was disputed because it assumes that the models used to show how epidemics spread applies to memes. Lerman argues that the nodes with lots of connections spread disease because they come in contact with so many other nodes, whereas those hub nodes actually share relatively few memes because they get so many.
We have discussed all of these different models in Networks throughout the semester. We saw that popularity does not follow a normal distribution, but rather exhibits extremes—the most popular books are very widespread, but others are barely read (i.e., there is a “long tail”). We also discussed why we need different models for the spread of disease vs. the spread of some behavior. Whereas people choose to do a certain behavior, e.g. share a meme, people do not choose to spread a disease; there is some probability that it will happen when they come in contact with others. This discussion helps to strengthen Lerman’s argument, because it further explains why spreading memes would not work in the same way as spreading disease.