Offense, Defense, or Both: Why Do Some Poor Minority Inner-City Youth Carry Guns?

As much as I have enjoyed my recent foray into recreational gun culture (hunting, target/sport shooting, and collecting), I am returning from my hiatus from thinking about guns in connection with violence because we are discussing violence in my Sociology of Guns seminar now.

As I wrote previously, following James Wright, “most of the gun violence problem results from the wrong kinds of people carrying guns at the wrong time and place” (“Ten Essential Observations on Guns in America,” p. 66). Understanding why these people – let’s just call them “bad guys” for short — carry guns could help us to figure out ways to get them to stop and thereby mitigate the harmful effects of such violence.


Many people, Wright included, have found that bad guys carry guns for the same reasons that good guys carry guns: for protection. Insofar as the bad guys live in much more violent worlds than good guys, their defensive motivation for carrying guns is presumably that much stronger. If a suburban concealed carry “operator” or “warrior” dials .357 rather than 911, what must a poor, inner-city youth think about the value of relying on the police rather than himself as a first line of defense?

However, unlike good guys, bad guys also contribute to the violent milieus in which they live. They are not innocent victims of crime, but are also perpetrators of crime. They carry guns to protect themselves because they do things that require protection.

The idea that there are two distinct groups in areas of concentrated violence – victims and offenders – has been questioned since the 1980s. And many believe that most of the gun violence problem actually exists within this overlapping group of victim-offenders. Andrew Papachristos’s work on networks of offending and victimization risk supports this interpretation. Hence the sadly popular idea among some that we should just build a fence around high crime neighborhoods and let “them” shoot it out.

Unfortunately, just asking individuals why they carry guns doesn’t allow us to understand the complexity of offensive gun carrying vs. defensive gun carrying vs. offensive and defensive gun carrying.

Enter a very interesting study by Richard Spano and John Bolland published in 2013 in the journal Crime & Delinquency, “Disentangling the Effects of Violent Victimization, Violent Behavior, and Gun Carrying for Minority Inner-City Youth Living in Extreme Poverty.”

Spano and Bolland used data from the Mobile Youth Survey, a series of surveys of high poverty (mostly African American) youth living in Mobile, Alabama, to understand victimization and offending as predictors of initiating gun carry. Significantly, they have data on the same individuals at two points in time, so they can distinguish those who did not carry a gun at Time 1 (T1) but did carry a gun at Time 2 (T2). They also have data on whether the respondent engaged in violent behavior and/or was a victim of violent behavior at T1 so they can use those as independent variables to predict the likelihood of initiating gun carrying at T2 (statistically controlling for other factors such as family structure, gender, and gang membership).

From T1 to T2, 8% of the youth in the survey initiated gun carrying – 83 of 995 respondents.

Two figures highlight interesting descriptive findings. Figure 1 shows that a slight majority of respondents were NEITHER victims nor offenders at T1 (52.4%). The second largest group of respondents (31.2%) were only offenders at and the smallest group (4.9%) were only victims at T1. A relatively small group (11.6%) were in the overlapping group of those who are both victims and offenders.

These data suggest that a remarkable number of youth in this community engage in violent behavior (42.8%), but also that a small majority manage to avoid violence as victims and offenders. It certainly doesn’t suggest the propriety of building a wall and letting them fight it out.


Figure 2 then looks at these same groups of victims and/or offenders in terms of the proportion of each group that initiated gun carrying at T2. The highest proportion of new gun carriers come from the overlapping group (19%) and the second highest come from the victim only group (16%). These groups were 3 to 4 times more likely to initiate gun carrying than those who were neither victims nor perpetrators of violence (5% initiated gun carry). In the middle were those who only engaged in violent behavior at T1 (9%).


So, compared to those with no exposure to violence either as victims or perpetrators, victimization does seem to motivate gun carrying. This finding holds true in multivariate statistical anlayses conducted by Spano and Bolland (Table 3, Model 1 in their paper).

The point of comparison matters, however. Spano and Bolland also found that compared to those who only engaged in violent behavior at T1, those who were victimized (either only victims or in the overlapping group of victim-offenders) were no more likely to initiate gun carrying (Table 3, Model 2). That is, there was no statistically significant difference in initiating gun carrying for offensive versus defensive purposes once other factors were held constant statistically.

Note: Technique here is logistic regression, and the regression coefficients have been transformed into odds ratios for ease of interpretation. Odds ratios > 1 mean a positive relationship to the dependent variable, < 1 means a negative relationship. Only ratios with at least one * are statistically significant; the more * the stronger the statistical relationship.

Both of these findings suggest that, although retrospective accounts of why bad guys carry guns can tell us something, they don’t tell us the whole story of the complex motivations for gun carrying.

Also of note is the fact that in the multivariate analyses, being a gang member at T1 is the strongest predictor of initiating gun carry at T2. 3% of respondents were gang members, and they were 300%+ more likely than non-gang members to initiate gun carrying, ceteris paribus. Being male is also a consistent and strong predictor of gun carry compared to being female.

The only measured factor that reduces the likelihood of initiating gun carry at T2 is having a two parent (or parent equivalent) family structure at T1. This reduced the likelihood of gun carrying by 50% compared to not having such a family structure.


  1. This is very interesting. I do wish, however, we’d drop the language of “good guys” with guns vs. “bad guys” with guns. Pretty much all people think of themselves as “good guys” and violent crimes like domestic violence occur across ALL socioeconomic levels. Historically, women who used their own guns defensively to stop a batterer have been labeled as the “bad guys.” How else might we discuss whose gun ownership is legitimate or whose makes people around them safer?


  2. Re: “good guys” and “bad guys.” Crimes, including domestic violence, are not committed at a socioeconomic level, they are committed by individual people, and individual people are not all equally likely to commit violent crime. It is well known that some particular individuals are far more likely to commit violent crime than are others. How do we know this? Because this is the actual observed data. For example, I expect that both David and I will go our whole lives without committing a rape, an armed robbery, an aggravated assault, a burglary, or an act of domestic violence, yet other individuals commit these crimes with some frequency. I suggest that the former group consists of “good guys” and the latter of “bad guys.” This is so obvious I am mystified by the effort to pretend that it’s not so. Is that effort partly driven by a desire to obfuscate how much violent crime is committed by the “bad guys” who happen to fall into the ~5% of the population that is young black males (fully half the violent crime in America, per DOJ/FBI). Obfuscating that reality only makes it more difficult to identify and address the underlying causes of this catastrophically high crime rate, to wit an profoundly broken urban culture, and it does a terrible disservice to “good guy” people living in those urban centers (who also happen to be largely black) who are overwhelmingly the victims of this violent crime.


  3. It would also be interesting to have David discuss the prospective defects in any study based on a survey which asks people, many of whom are believed to be criminal predators, if they are criminal predators or victims of crime, and the biases that might be introduced into their responses both by human nature and their environment (and these might have opposite effects). For example, most people even if guilty of some criminal activity would deny such guilt to a stranger taking a survey (what’s the upside of admitting criminal conduct to such a stranger–science?). On the other hand, if one lives in a violent community it might be prudent for one’s own safety to foster a perception as someone capable of acts of violence. I would note that studies of gun ownership based on similar survey methodologies, for example, are patently and severely defective for similar reasons.


  4. So how do we work this science into gun violence prevention? I recently left a disappointed comment at Mike the Gun Guy’s site because so many in the gun control movement, seemingly Mike included, want to only look at gun violence vs. gun laws and numbers of guns. They know damn well that this is utterly myopic, but I see little change in the Everytown et al people. As “self” implies, the vast majority can live with guns and never use them wrongly. Education (or lack), jobs (or lack), war on drugs, social associations, etc provide motive. Instruments of violence only provide tools to act out.


    • It is a LOT easier to blame “the other guy” for our problems than it is to deal with our problems ourselves.
      I do it as much (or more) than everyone else, so I am equally to blame.
      However, Orestes didn’t get rid of the Furies by blaming Apollo for his actions, Apollo got rid of the Furies when Orestes took credit for the works done by his own hand.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.