Tactical Social Network Analysis: Using Affiliation Networks to Aid Serial Homicide Investigation
Bichler suggested two main points: the stronger two people are tied to each other, the more likely their social relations overlap; those who are frequently share a social space/activity are more likely to form social relations with each other. These points are developed to discover and understand crime patterns by looking for overlaps among victims. I find it interesting that they do not start by mapping out the social relations of the victims, instead they first model the connections among people(victims and suspects) and the places they frequent. From there, they may map out the social relations among people and find how the victim crossed paths with their offender. For example, if someone and the victim are both seen at one place at the same time, they have a common affiliation or they may be considered weakly connected. If they are seen at the same place more than once, it is likely that they belong to the same social circle. This is because their bond strengthens as they frequently share the same social space or activity. In a serial homicide, if one suspect belongs to each victim’s social circle while other suspects do not, it may become suspicious. I often see detectives on a crime show attempt to find the connection between the victims through their social circle and then that doesn’t work, they focus on where the victims could have met the offender–the place where the connection was formed. The victims may never have been in the same place at the same time, but the victim and the offender most definitely were.
Bichler also mentions that the offender of a serial homicide would more likely be the only one from each victim’s social group who is connected to other victims from different social groups/clusters. In other words, imagine there are two clusters of nodes and their only connection is that they both connect to only one node. Cluster A has Node A(victim) connected to Node C(suspect) who is also connected to Node B(another victim) in cluster B. We may ask “who else but the offender in a crime series would link to victims from different social spaces?”. Indeed, in a serial crime, it is very suspicious when the victims(and their social worlds) only overlap through one person. This reminds me of the Strong Triadic Closure Property which is supported by the idea that if two nodes are both connected to one other node, then the two nodes are more likely to have a connection as well. However, if the two victims and their social worlds do not have any connection other than the suspect, then it is reasonable to be suspicious.
Source: https://journals.sagepub.com/doi/full/10.1177/1088767916671351