Field of Science

Why are there more Christian congregations where there is more crime?

Take a middle-American US city – a fairly typical city with the usual mix of rich and poor, down-town and suburban, black and white. Indianapolis, let’s say. Which areas do you think would have the highest levels of crime?

Well, the poor areas of course. No surprises there. Down-town areas and those area with low population density are also at risk – probably a result of increased opportunities. Racially mixed areas have higher level of theft and burglary, although not violent crime. And there also seems to be a strong ‘cultural’ effect. There are pockets of high crime that persist even after taking into account all the other factors.

And, last but not least, you also get more crime in neighbourhoods that have more Christian congregations.

Now, the effect isn’t across the board. Catholic and non-Protestant congregations are not related, either positively or negatively, to crime levels. And although Black Protestant and mainline Protestant congregations tend to be located in areas of high commercial burglary (and larceny, in the case of mainline Protestants), they aren’t associated with other types of criminal behaviour.

It’s evangelical Protestants (aka ‘fundamentalists’) that show the strongest connection. Those areas of Indianapolis with more evangelical Protestant churches also have more robbery, aggravated assault, vehicle theft, commercial burglary and larceny. Blimey.

This is just a statistical association. So it could simply be that these churches set up shop in those areas with highest need – with the highest crime rates. But remember that this association remains even after controlling for all those factors I mentioned above. These churches are located in areas that have more crime than you would expect, given the level of deprivation and other factors that predispose to crime.

It could be that these evangelical churches actually increase crime rates as a direct result of their teachings. Often these churches extol the virtues of defensive and punitive violence, and Manichean (i.e. dualistic, ‘heaven and hell’ religious concepts) have been linked to more violent societies. But evangelical churches were not associated with more homicide in this study (although they were linked to more aggravated assault).
The researchers (led by Scott Desmond at Purdue University in Indiana) think that it has something to do with the relative newness of Conservative congregations. In particular, it might be that many people commute to these congregations, rather than living locally. The normal social networks that help forge society are undermined when people travel to meet people over long distances, rather than getting to know their neighbours.

There is, however, one final possibility suggested by the researchers. In the US, churches are a fundamental of social fabric. But the evangelical churches are highly polarising. Could it be that having one arrive in the middle of your neighbourhood actually leads to suspicion, resentment and even hostility?

If that were so, then more evangelical churches could actually have a destructive effect on local society.

PS. I'm currently away, but will catch up on comments and emails when I get back!

ResearchBlogging.orgDesmond, S., Kikuchi, G., & Morgan, K. (2010). Congregations and Crime: Is the Spatial Distribution of Congregations Associated with Neighborhood Crime Rates? Journal for the Scientific Study of Religion, 49 (1), 37-55 DOI: 10.1111/j.1468-5906.2009.01491.x

Creative Commons License This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.


  1. There's only one way to check that: a controlled experiment. Take a bunch of sets of equal neighborhoods and start a congregation in one, not in the other. Perhaps start something else there.

    If that's not an option, look at crime rates following the foundation (or bankruptcy) of congregations.

    In both cases, if your prediction is correct, crime rates should change over time. One could get an indication of the outcome by assessing inhabitants' attitudes before doing the experiment.

    At the moment, I tend to think there shouldn't be a causal relationship. Being religious may change someone's attitudes, but I feel it hardly ever changes behaviour.

  2. What's the association between Christian congregations and average income levels of the neighborhoods they are in?

    Both # of churches and amount of crime may be mediated by income.

  3. @Marius 't Hart: I would take exception with your last assertion. Religious indoctrination most certianly can change behaviour. As many have said before me;

    "A good man will do good on their own. An evil man will do evil on their own. But it takes relgion to make a good man do evil."

    With irrational justifications come attrocious behaviours. We see that time and time again in soceity and experiments. This interesting statistical anommaly may just be one more datapoint.

    I'm sure that Epiphenom has shown this old study: (Make sure to actually click for additinal figures at the end, and see how screwed up overall our society is when religion gets factored in.

  4. I wonder if this might relate to the "cultures of honor" concept written about by Richard Nisbett and others. A very short description:
    A review:
    (And of course there's a wiki entry.)

  5. This is from Gregory Paul's oft quoted study mentioned by Larian:

    "There is evidence that within the U.S. strong disparities in religious belief versus acceptance of evolution are correlated with similarly varying rates of societal dysfunction, the strongly theistic, anti-evolution south and mid-west having markedly worse homicide, mortality, STD, youth pregnancy, marital and related problems than the northeast where societal conditions, secularization, and acceptance of evolution approach European norms"

    I suspect that both religiosity and criminal activity have a stronger relationship with a third factor like I.Q. or education, but, in any case, religion doesn't seem to curtail such behavior. Heck, threats of hell can't stop clergymen from molesting kids, so why should it stop other crimes.

  6. There's another paper by Gary Jensen in the same journal which is actually the one I was thinking of:

    He shows that nations with a high level of dualism (i.e. believe in both god and the devil, as many in the USA do) also have higher levels of homicide.

  7. I accidentally deleted a comment by T.A. Lewis, so here it is:

    I just wonder if this has anything to do with diagnostic causation?

    I have just been introduced to this concept in cognitive science by Steve Sloman’s book “Causal Models”.

    The relevant quote reads: “We all have all had the experience of eating something or listening to something or smoking something not because we enjoyed it, but because we felt that if we’re a certain kind of person (sophisticated or sensitive or cool), then that’s how we would behave. The decision was made not because of the beneficial causal consequences of the action but instead because the action was diagnostic of a cause that we wanted to be true.” (2005, p. 99).

    I would extend his example to church-going. Many people might attend because they think this is how members of a low-crime neighborhood ought to behave.

    After all, in the popular consciousness, piety = morality. Could it be that the higher levels of crime make it easier for people to decide to go to church because they think this is diagnostic of what makes for a good neighborhood?

  8. This is fascinating - could someone here help me out in reading Table 2? I've got no experience in reading sociological charts. From what I gather, the higher the number in the chart, the greater the association between the type of crime and the potential factor; for instance, the association between Density and Assault is 0, but between Disadvantage and Assault it's 9.535, meaning that density has no effect on assault but that disadvantage is very strongly associated with it. If that's the way to read the chart, Catholic congregations have a much stronger effect on robbery and aggravated assault than evangelical congregations (.324 vs. .758 for robbery, and .498 vs. .717 for aggravated assault), which doesn't seem to match the conclusion. How am I misreading the chart?


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS