Obesity as a cascade effect
There are many real life scenarios that can be modeled as cascading network effects. Obesity, for example, spreads in a very similar way to the network effects we have seen in class. There are even specific cascading quotients that occur.
You could look at groups of friends as clusters of nodes. As groups of friends have only cursory interaction with groups outside, then their are many different clusters, which have only a couple connections to outside clusters. In a study done by Nicholas A. Christakis, groups of obese people were modeled as clusters with 3 degrees of separation. They found that a person has a 57% chance of becoming obese during a given time interval if one of their friends becomes obese. The effects were slightly different depending on the relationship between the two people. If it was a pair of siblings, and one sibling became obese, the other had a 40% chance of becoming obese. If the pair examined were a married couple, then the result of one spouse becoming obese was that the other had a 37% chance of doing the same. Interestingly enough, physical distance did not have an effect on the results. What did, however, was the gender of the people becoming obese. Christakis found that members of the same sex had a greater influence on increased chances of obesity than members of opposite sexes.
This has huge ramifications for public health officials. If the CDC or the WHO used the type of graph theory we have seen in class to model the spread of obesity, then they could use their resources to best cut it off, in the same way we find bridges that contain great power as they connect different social clusters. The type of edges also affects the spread of obesity, so modeling those edges has a different quotient would yield more accurate results. For example, a friendship edge would have a lower quotient than a sibling edge, which would have a lower quotient than a spouse edge. This study is a great example of the application of graph and cascading effect theory to real life.
Source: http://www.nejm.org/doi/full/10.1056/NEJMsa066082