Richard Callahan

From Santa Fe Institute Events Wiki

Hi, my name is Richard Callahan and I'm a fourth-year graduate student in Sociology at the University of Washington. Back in 2007 at the SFI summer school I was interested in studying the spread of epidemics through social networks, and in particular researching the spread of HIV/AIDS in China. I also have a background in criminology, and my adviser is Dr. Ross Matsueda. Mark Feldman's and Li Shuzhuo's presentations on China's gender imbalance changed my direction a bit, though, as I became interested in understanding how the marriage crunch induced by the gender imbalance could affect the crime rate in China. Taiwan has a similar cultural phenomenon and some excellent data, and I'm fortunate enough to have the opportunity to go to Taipei in September to research this question on a Fulbright scholarship. Such a great opportunity! I'll be able to travel back to the Mainland a couple of times next year as well, so I'll be around.

A couple of things on the research plate right now:

1. How does employment affect crime? An important question, especially considering that any changes in population policy won't have an affect on crime in China for years, while changes in employment policies could constitute a stopgap measure.

2. I'm working with Dan Hruschka on understanding social learning processes using criminological data in a network survey.

3. I'm interested in working with a high-tech firm in America to conduct a network analysis survey and model two dimensions of guanxi, replicating some work by a professor at Qinghua on a US sample. I've been too busy with the other work to start on this one.

4. This is one that could be a great SFI collaboration: using Bayesian networks and possibly some of Aaron Clauset's work on hierarchical network structures to model the emergence of the "code of the street," a culture of street crime. The idea is that in many cities in the US, especially where there may be a perception that the police are not always capable of providing adequate protection, a set of norms has emerged in which people have to act tough and be willing to get into fights in order to not be perceived as weak (and therefore be more likely to be victimized). I'd like to improve upon a recent simulation by an economist that explores under what conditions a code of the street emerges. It would be interesting to explicitly model individual perceptions about how "tough" others are when evaluating whether to initiate or consent to a fight, try to talk their way out, or run. Such perceptions about the probability of winning a fight and the cost of losing would be updated after each encounter based on the flow of information through social networks. A "code of the street" emerges when individuals who would not otherwise engage in fights begin to feel compelled to do so to minimize their future risk. Both the network structure and individual attributes would affect the emergence of the code. I hope we could get a chance to work on this! It's the sort of project that lends itself well to collaboration, as you have the Bayesian component, the network structure, the substantive literature and a rational choice framework in which people are trying to optimize some utility related to the outcome of the fight.

5. While my current research involves causal analysis and panel surveys, earlier I was involved in some research involving global trade networks so I know something about the literature and data there. If anyone shares a similar interest, a collaboration would be fun.

Just a couple of ideas. Talk to you guys soon!


P.S. Here's a link to my Web site: <>. I haven't updated since last fall but there it is.