In a couple of previous posts (here and here), I promised to say more about a recent article on nonfatal gunshot injuries that was treated dismissively on another blog: “Tragic, but not random: The social contagion of nonfatal gunshot injuries,” by Andrew V. Papachristos, Christopher Wildeman, and Elizabeth Roberto, published in Social Science and Medicine, Vol. 125 (2015), pp. 139-150.
In the original post and comments there were two major reasons given for dismissing (or as OP Miguel subsequently called it “making fun of”) the article:
(1) It tells us what we already know, but in more convoluted language, because “the Intelligentsia.” Examples:
- “So basically the investigators found out that if you hang out with stupid people at stupid places during stupid hours, you win stupid prizes. We never heard this before, thank God we have the Intelligentsia to protect us and save us.”
- “Actually what pisses you off is that colleagues are being mocked because they discovered, after extensive analysis, that water is wet.”
- “It’s damned obvious that people with a low regard for the welfare and feelings of others are more likely to be involved in criminal activity. We all know that most people who get shot in this country are criminals getting shot by other criminals. That’s not really in doubt by anyone who bothers to read the relevant research.”
- “Let me help the intelligentsia: Birds of a feather get shot together. . . . Highly polysyllabic words and obtuse sentence structures only serve to obfuscate the facts.”
- “I think this study, while it’s put in pretentious language, says what we all know, or at least believe; That a small group of idiots makes a lot of trouble for everyone.”
- “Another way to characterize this study is that it validates what your Mom always told you: you are going to get in trouble if you hang out with the wrong people.”
- “Sink a hundred grand into regression analysis to show that what old people have known for hundreds of years may be plausable… maybe you can buy wisdom… No, probably not.”
This line of thinking is what caught my attention and motivated me to comment on the original blog. It smacks of the anti-intellectualism that has been attributed to those in gun culture. And yet the gun culture has its own intellectuals that it holds in high esteem, despite their “highly polysyllabic words,” “pretentious language,” and validation of “what your Mom always told you.” People like Massad Ayoob and John Lott. In fact, another gun blog, The Truth About Guns, calls its readers “The Armed Intelligentsia” and even sells “AI” gear at cafepress.com.
So, something else is surely going on, specifically point #2.
(2) It is part of a gun control agenda. OP Miguel, says: “Why do they keep making the studies? Because they are desperately seeking a different answer that conforms to the political narrative. It is never crime, it is never drug trafficking, it is never poverty, it has to be guns so we can ban them.”
Note here that Miguel is not only saying that the study can be used by people who have a political agenda, but more specifically that “the makers” (authors) of the study actually have a political agenda. Now, to be sure, this second objection has some merit for some studies of “gun violence,” as I will write about soon (I hope!). But reading the article in question, or even the newspaper article about it, would make clear that this criticism does not apply to this study, which is thoroughly – and for some, excessively – academic in nature (see criticism #1). Damned if you do, damned if you don’t.
(By the way, if anyone finds the authors, politicians, or advocacy groups using this study to promote gun control, please let me know! I would like to pursue that, too.)
In fact, at the end of 10 pages of two-column text (and tables, formulas, graphs), the authors devote one paragraph (185 words) to “gun violence reduction and prevention strategies” (p. 148). Even here, the recommendations are given with appropriate caution: “if…,” “suggest…,” “may…,” “might…”
The strongest statement in the conclusion was: “An approach such as this would argue against sweeping policies and practices based purely on categorical distinctions such as race and ethnicity and, instead, opt for interventions and policies that consider the observable risky behavior of individuals.” That is, no stop and frisk, for example. Or, the conclusion I drew from it, no “sweeping policies and practices based purely on categorical distinctions” such as gun ownership. What needs to be controlled is “the observable risky behavior of individuals.”
So, the objection then goes back to the first: a lot of money spent to fund “the Intelligentsia” to tell us what we already know. Here I would like to consider the value of this study and how it goes beyond the “painful elaboration of the obvious” often attributed to social scientists.
Beyond the anti-intellectualism and suspicion of a political agenda, the crux of the critique of this research is that it doesn’t tell us anything we don’t already know. Is that true?
I begin by noting some slippery and imprecise language about “what we already know” in the critiques:
- Certain people are “more likely” to be involved in criminal activity and “most people” who get shot are criminals, but no specification of how much more or what constitutes most.
- Uses pretentious language to tell us “what we all know, or at least believe,” as if knowing and believing are the same thing.
- Study tells us “what old people have known for hundreds of year may be plausible [sic.],” here conflating knowing and what may be plausible.
Although this may seem nit-picky, I hope to show its relevance because a major contribution of this line of research is its precision.
More Precise Specification of Risk Factors Associated with Violent Victimization
Those who study violent victimization tend to examine three sets of predictors: (1) individual demographic risk factors like being poor, young, and black; (2) neighborhood risk factors like living in an area with high poverty and crime rates, and (3) behavioral risk factors like involvement in gangs, delinquency, or criminal behavior.
This is, in a sense, “what we already know.” We know that some people in certain places that do certain things are more likely to end up shot or killed. What we don’t already know is why some with these risk factors end up injured or dead, while some with the same risk factors do not. So, our models – both common sensical and sociological — are incomplete.
Adding social networks to the picture makes these models more complete. For example, on gang involvement. “Everyone knows” being in a gang is dangerous. This is true, as far as it goes, but it doesn’t go very far. How dangerous is gang involvement and for whom? Papachristos’s study shows that the risk of victimization from being in a gang is mediated by race. Being white and in a gang is more dangerous than being white and not in a gang, but being white and in a gang is no more dangerous than being black or Hispanic and not being in a gang (by specific amounts I will not cite here). This has to do with the social structure of gangs and the distribution of violence within those structures. (Papachristos and colleagues have also written separately about the issue of gang violence to better explain these racial differences in terms of geography and social networks.)
So, conventional wisdom suggests “Birds of a feather get shot together.” But not all birds are equally likely to get shot. What we should really say is, Birds of a feather are more likely to get shot than non-birds of a feather, but some birds of a feather are more likely to get shot together than other birds of a feather. Not quite as pithy a saying, but more true to empirical reality, as this study documents.
In addition, and to my next point, using this research we can also say with much more precision how much more likely people are to get shot or killed based on their exposure to these social networks.
More Precise Measurement of Social Networks and Distribution/Concentration of Violence Therein
The network approach employed in this research is also more much precise in measuring the things most people only have a vague “sense” of. According to the dictum attributed to William Thomson (Lord Kelvin), “When you cannot measure your knowledge is meager and unsatisfactory.”
This research begins to help us know more precisely how expansive the social networks are that carry criminal violence and how concentrated violence is in those networks. I am quite certain this falls outside the realm of “what we already know.”
For example, how expansive are the social networks of criminal offenders? The study in question covers the entire city of Chicago between January 2006 and September 2012 and identifies 169,725 individuals who were arrested with another person during this period. 107,740 of these individuals (63%) were part of the same social network, connected to each other directly or indirectly. Some people were very tightly connected in this one large network, but the majority of individuals (57%) were only connected to 1 or 2 other people (pp. 142-43). Consequently, most individuals in the network are connected loosely to a very large network that is probably unknown to them, but which can nonetheless dramatically affect their lives.
How highly concentrated are gun injuries in these social networks? 70% of all gunshot injuries in Chicago during this time period took place within these networks of co-offenders, which comprised less than 6% of the city’s population. Furthermore, 62% of the gunshot injuries took place within the largest component of the network.
The article that occasioned this series of posts is just one part of a larger project that Papachristos is involved in. Other papers that have already been published highlight the concentration of violent victimization in other settings.
In a study of gun homicides in one high-crime neighborhood in Chicago (82,000 people living in a 6-mile area), Papachristos and Wildeman found that “41% of all gun homicides occurred in a network component consisting of approximately 4% of the population of the community” (p. 145).
Papachristos, Anthony Braga, and David Hureau have also studied the risk of gunshot injury in Boston’s Cape Verdean community. This study focused on networks of individuals who were observed in each other’s presence and recorded on Field Intelligence Observation cards by the Boston Police Department. Again, the risk of gunshot injury is highly concentrated in the largest component of these networks (which comprises 76% [579 of 763] individuals observed). 85% of gunshot victims are found in this network component, as depicted in the graphic below.
This line of research can also help to specify more precisely how much the odds of getting shot/killed increase and decrease as one moves from the core to the periphery of these social networks. In the Boston study, each “degree of separation” a person has from someone who has been shot reduces that person’s odds of getting shot by 25%. In the Chicago high crime community study, each social tie a person is removed from a homicide victim results in a 57% decrease in the person’s odds of being a homicide victim.
To be sure, this level of precision may not be of interest to people who are mostly concerned with gun rights and personal self-defense. Modeling, quantifying, and visually representing the underlying network structure of violence may not be everyone’s cup of tea. After all, I do not need to know how to read or write music to appreciate Mozart. Or how electricity works to turn a light on. Or the molecular structure of water to know it is wet (but see coda below). But just because I don’t know (or even care to know) these things does not mean they are unimportant. And it certainly doesn’t mean that the people who do know how these things work are only telling me what I already know.
So, I suspect that the issue here is not that these studies tell people “what they already know.” It is that it tells people about something they simply don’t care about, because many in the legal gun community – especially those active in the echo chamber which is Web 2.0 — are demographically, geographically, socially, and behaviorally distant from these crime “victims.”
Note that I use victimization in this post, but put “victims” in quotation marks here in recognition of the fact that most “victims” of gunshot injuries and death in inner-cities are also perpetrators. This makes these studies of even less interest to law-abiding gun owners (except insofar as the studies could be used to restrict gun rights). But even though most gun violence is concentrated in small pockets of various communities, the effects of street violence go beyond the immediate victims. So, there are people who through no fault of their own become collateral damage. And that is something that more people – including law-abiding gun owners — ought to be concerned about. Not as gun owners, but as people.
And those who do care about the effects of street violence on individuals and communities do well to heed the lesson of this study: emphasize targeted intervention on those who are most likely to be victims and perpetrators of such violence, not blanket regulation of guns and gun owners.
Coda: Is Water Actually Wet?
I wasn’t sure whether I should say this for fear of being called pedantic, but contrary to conventional wisdom/whatever everyone knows/what our mothers always told us, some would argue that water isn’t wet. Water is fluid owing to its chemical structure. The “wetness” of water may be said to be the subjective sensation of coming into contact with water owing to its fluid chemical structure, the process of evaporation, and other factors.