orgtheory.net

network analysis vs. public choice

I just wrapped up my undergrad course in networks for seniors. Near the end, in the week on networks and crime, we discussed Papachristos’ work on homicide in Chicago. If you haven’t read it, he has a very rich data set on gangs and traces the back and forth of gang revenge homicides. Great stuff. So I asked my students: “You are the police and now you have read this research, what did you learn?”

Student 1: You should target the most central gangs. They seem to generate a lot of violence.

Me: Good, what else?

Student 1: Since a lot seems to focus on revenge, maybe police should focus on friends of homicide victims. Maybe counsel them so they won’t get revenge and keep the cycle going.

Student 2: That would never work.

Me: Why?

Student 2: The cops gets no credit for counseling. Only for arrests.

Bingo. Great insight. In other words, we have a lot of good data on homicides and we know that a lot of it has to do with gang/revenge cycles. And that implies a solution – go after survivors and do what you can to keep them from acting out. But it is very hard to see how anyone could ever be rewarded in the system where people get promoted for arrests rather than crime prevention. It’s sad that you need have someone murdered first before you can be praised for being a good cop.

If people buy $500 of my books by Christmas, I will leave David Graeber alone: From Black Power/Grad Skool Rulz

About these ads

Written by fabiorojas

December 17, 2013 at 12:10 am

Posted in fabio, networks, sociology

7 Responses

Subscribe to comments with RSS.

  1. When you look at Chicago’s homicide problem, one of the first things you can notice is that obvious neighborhood pattern. Crime ridden neighborhoods receive the fewest resources, have the lowest civic participation especially in terms of informants for crime prosecution and also have a majority of men who contribute low parenting investments. Neighborhoods in which the people pull together to address crime tend to lower the rate. This may be related to the incarceration return problem – individuals who have prior arrests and have served time return to the same neighborhoods they came from; this might also be called a gang problem but given the mis-perception that the term gang produces, other concepts are used. There is a relation between the ratio of ex-cons to crime rate. But, the problem does not seem to transform itself which implies that either no interventions are made or the ones which are implemented do not have any effect. Probably the first move should be to increase police presence at the hot spots but I am unaware that even this minimal intervention has occurred. So, one place to look for why the problem persists is at the leadership: at the members of the city council and the style of leadership in the city as a whole. As RJ Sampson notes in his analysis of Chicago, “Great American City,” cohesive leadership among the elite has a decisive effect on the homicide rate. This seems to be the one main thing missing. The elected leaders in the vulnerable neighborhoods are incredibly weak and either have no clue what to do or demand or the powers of the executive are merely containing the problem. There is a leadership vacuum and there is a very complex problem due to the nature of mobility in the Chicago area.

    Like

    Fred Welfare

    December 17, 2013 at 2:32 am

  2. There are programs that assign officers to what are essentially counseling roles. I had the opportunity to be involved in a fascinating meeting that involved mostly police officers and a few academics discussing concerns about “community noncooperation” around homicide investigations that ended up involving some interesting intra-police dynamics between neighborhood police officers who spend a lot of time building relationships and the higher-ranking detectives who drop into communities to investigate homicides. The discussion emphasized the importance of spending a lot of time in one place to build relationships and trust and how that ran counter to career issues since a) neighborhood officers are relatively low-ranking and b) getting promoted involves moving around a lot and getting a variety of experience.

    I also brought up Robert Vargas’s research that found that another reason victims did not want to cooperate with police is that they were, in fact, targets of revenge and that a lot of the problem was police ineptitude in acting in ways that publicly revealed who was talking to police. That also added to the dialog about neighborhood officers’ ability to speak privately with informants versus the detectives’ strategy of stopping people on the street and interrogating them in front of others.

    Like

    olderwoman

    December 17, 2013 at 3:31 am

  3. Thanks folks, interesting discussion. I want to pick up on Fabio’s point about the balance of incentives (“people get promoted for arrests rather than crime prevention”) because this doesn’t really square with the general reading I’ve done. Hasn’t there been a big shift in the US towards the use of trend statistics (e.g. trends in murder rates) to measure police performance? If that’s the case, then wouldn’t there be solid incentives for interrupting revenge cycles (e.g. via counselling) in order to improve these trend stats?

    What got me thinking about this was a Wired article that reports that the NYPD was using interactions on social media to interrupt revenge cycles (http://www.wired.com/underwire/2013/09/gangs-of-social-media/all/). It’s not a central point in the article, but it stuck because it had me thinking back to Gould’s ASR paper on social networks and revenge in Corsica!

    Like

    Sam

    December 17, 2013 at 8:24 am

  4. This is a good example of how “knowledge is power”, but I don’t mean that in the usual sense of knowledge providing us with a basis to solve problems.

    Imagine you live in a neighborhood with a high murder rate. This means you are statistically more at risk of being murdered than someone who lives a neighborhood with a low murder rate, and you rightly demand that your politicians think seriously about how to solve your problem. But now suppose it becomes known that the murder rate is driven up by “revenge cycles” (rather than, say, poverty or the criminalization of drugs). Well, now when your friend is killed you not only lose a friend, you also gain a police “counselor” (at least in the ideal policy environment Fabio imagines) who “goes after” you and prevents you from “acting out”. That is, you become a suspect in a pre-crime investigation. That’s the price you pay for the knowledge that “implies a solution” to your high risk of being killed on your own street.

    “It’s sad that you need to have someone murdered first before you can be praised for being a good cop,” sounds obvious and true and right. But perhaps we don’t actually want police to prevent crime. Perhaps we want them mainly to deter it, i.e., to increase the risk of being punished for a crime. After all, you still need for someone to be in a gang that sells drugs, and those drugs still have to be illegal, before anyone will praise you for being a good cop/counsellor. It was better when cops were praised for solving crimes that had actual victims, like the murders themselves.

    It would also be better to focus on the production of “crime” through laws that certain criminalize activities, or to focus on economic and social conditions that foster violence. It could be argued that this also requires knowledge. But against this I’d argue that it only requires the responsible use of power. Drug laws and poverty are not bad because they cause crime. They’re bad because they unjustly restrict freedom.

    Knowledge is power when it implies more police in your life.

    Like

    Thomas

    December 17, 2013 at 12:41 pm

  5. Fabio,

    I’m honored that you and your class read the paper! Thanks.

    But, I have to point out that police across the country are doing precisely the thing you say they aren’t: using SNA to focus on the friends/associates of victims to try and get ahead of shootings. And, not just one city, but dozens: Cincinnati, NOLA, New Haven, Bridgeport, Stockton, Newark, East Palo Alto, Chicago, among others. There is a strong movement (yes, movement) in policing towards integrating such technology and techniques into honest to goodness prevention efforts. The COPS office even put out a grant recently explicitly to develop community policing efforts along these lines. And, more importantly, evaluation evidence in some of those cities are related to impressive in short-term declines in gun violence.

    To see more about this is playing out in practice, check out the work of @NNSCommunities (or David Kennedy’s book Don’t’ Shoot). Happy to point towards evaluation of these types of initiatives.

    Or, if you want a decent journalistic account of how this stuff is being used in Chicago’s version, check out today’s FastCo.

    http://www.fastcoexist.com/3022154/using-social-networks-to-track-and-predict-gun-violence

    There’s more to it, but, really, it is using SNA to find the groups involved in current disputes and directing services of all sorts (including law enforcement, but not limited to them or even dominated by them) to said groups before another shooting occurs.

    Student 1 was more right than you might think!

    Like

  6. Well, this speaks to a lot of social problems that the US is confronting. For example, doctors are rewarded for providing treatments to patients regardless, but not preventing diseases. Same for pharmaceuticals that they get rewarded for coming up with drugs that only treat symptoms and probably cause drug dependency, but they are not that motivated to come up with inexpensive drugs that could cure disease….

    Like

    John

    December 17, 2013 at 10:17 pm

  7. I wonder what point has to be reached before the “gang” violence is assessed as subversion or terrorism.

    Like

    Fred Welfare

    December 18, 2013 at 4:31 am


Comments are closed.

Follow

Get every new post delivered to your Inbox.

Join 1,200 other followers

%d bloggers like this: