Social Capital accounts for 82% of variation in homicide

[phantom power]

You can learn more about social capital here: http://en.wikipedia.org/wiki/Social_capital

The basic result seems to be that lack of community (social relations and cooperation) is most correlated with violence and homicide. I would add that in real life you have to consider feedback loops. A decaying sense of community is a very plausible root cause for homicide and high gun ownership rates, but once you have a society with high violent crime rates and high gun ownership rates, that also becomes an enabling factor with a life of its own.

In 1999 Ichiro Kawachi at the Harvard School of Public Health led a study investigating the factors in American homicide for the journal Social Science and Medicine (pdf here). His diagnosis was dire.

“If the level of crime is an indicator of the health of society,” Kawachi wrote, “then the US provides an illustrative case study as one of the most unhealthy of modern industrialized nations.” The paper outlined what the most significant causal factors were for this exaggerated level of violence by developing what was called “an ecological theory of crime.” Whereas many other analyses of homicide take a criminal justice approach to the problem–such as the number of cops on the beat, harshness of prison sentences, or adoption of the death penalty–Kawachi used a public health perspective that emphasized social relations.

In all 50 states and the District of Columbia data were collected using the General Social Survey that measured social capital (defined as interpersonal trust that promotes cooperation between citizens for mutual benefit), along with measures of poverty and relative income inequality, homicide rates, incidence of other crimes–rape, robbery, aggravated assault, burglary, larceny, and motor vehicle theft–unemployment, percentage of high school graduates, and average alcohol consumption. By using a statistical method known as principal component analysis Kawachi was then able to identify which ecologic variables were most associated with particular types of crime.

The results were unambiguous: when income inequality was higher, so was the rate of homicide. Income inequality alone explained 74% of the variance in murder rates and half of the aggravated assaults. However, social capital had an even stronger association and, by itself, accounted for 82% of homicides and 61% of assaults. Other factors such as unemployment, poverty, or number of high school graduates were only weakly associated and alcohol consumption had no connection to violent crime at all. A World Bank sponsored study subsequently confirmed these results on income inequality concluding that, worldwide, homicide and the unequal distribution of resources are inextricably tied. (see Figure 2). However, the World Bank study didn’t measure social capital. According to Kawachi it is this factor that should be considered primary; when the ties that bind a community together are severed inequality is allowed to run free, and with deadly consequences.

But what about guns? Multiple studies have shown a direct correlation between the number of guns and the number of homicides. The United States is the most heavily armed country in the world with 90 guns for every 100 citizens. Doesn’t this over-saturation of American firepower explain our exaggerated homicide rate? Maybe not. In a follow-up study in 2001 Kawachi looked specifically at firearm prevalence and social capital among U.S. states. The results showed that when social capital and community involvement declined, gun ownership increased (see Figure 3). Kawachi points out that it is impossible to prove whether one factor caused the other, but the most reasonable interpretation is that people who don’t trust their neighbors are more likely to think guns will provide security. In this way the number of guns and the number of homicides both stem from the same root, suggesting that guns don’t cause murders anymore than cars cause fatal accidents.

http://blogs.scientificamerican.com/primate-diaries/2012/07/26/the-jokers-wild/

[xchrom]

1. Du rec. Nt

[HiPointDem]

2. and 'social capital' correlates with capital, period. it doesn't develop in an economic vacuum.

[phantom power]

3. you mean, poverty decreases social capital?

[gejohnston]

4. I disagree with him one one thing

Kawachi points out that it is impossible to prove whether one factor caused the other, but the most reasonable interpretation is that people who don’t trust their neighbors are more likely to think guns will provide security. In this way the number of guns and the number of homicides both stem from the same root, suggesting that guns don’t cause murders anymore than cars cause fatal accidents.

Most Americans buy guns for the same reasons most Canadians and Norwegians do, target shooting and hunting. Also, if you look at percentage of households with guns, we are slightly behind Finland and not that different than Canada and Norway. Where the large number per person comes in is simply, we have too much shit. It applies to TVs, floor space in McMansions, as it does guns. The thinking is the same.
I grew up where everyone had guns, and left the doors unlocked. Guns were unloaded and properly stored.

Another point he misses is that the vast majority of our murders are gang related. People with criminal records killing other people with criminal records. Most of these are concentrated in a few areas in a few cities. That is why Minnesota has the same if not slightly lower murder rate than Manitoba and El Paso has murder rate that is half to Vancouver's.

[phantom power]

6. I think that is consistent, because...

areas with a lot of dense gang activity would be areas with low social capital, where people did not feel secure or trust each other.

Although that makes me think about sub-populations and social capital. The sub population of "people involved with gangs and/or drug dealing" will have extremely low social capital. The neighborhoods they live in might have higher social capital, I assume still lower than other areas.

[DanTex]

5. Interesting.

One thing to point out about the guns is that if you look at the studies comparing gun prevalence and homicide, you find that gun prevalence specifically correlates with gun homicide, with no significant effect on non-gun homicide. So this strongly indicates that there is an instrumentality effect: that guns actually cause, or at least facilitate murders. If it were actually the case that the guns/homicide correlation were due to a common correlation with low social capital, you'd expect to find a correlation between guns and all kinds of homicides, not just gun homicides.

In fact, a likely interpretation of this data is that gun prevalence is one of the ways that the social capital causes more homicide (at least partially so). In other words, when social capital is low, people go out and buy more guns, and when people have more guns, they end up committing more homicide.

This is also consistent with what social scientists have observed in other studies of gun crimes. Most homicides result from either escalating conflicts or other crimes (e.g. robbery, assault), but since gun assaults are much more lethal than assaults with other weapons, the presence of a gun makes it far more likely that these other situations result in an actual homicide rather than just an injury. And this explains, for example, why the US has comparable overall violent crime statistics to other industrialized nations, but we have far more homicides -- it's because criminals in the US have a much easier time getting guns.