One of my professors in Doctoral work in public administration mentioned this in a class on the history of public administration.He reported that the advent of the street light in the early 20th century resulted in a reduction of crime at night especially burglary. His name was Dr. Daniel Martin and he wrote a book on annotation of the major research of public administration. It maybe that this source would be included. I will look around I have a copy of his book...

Hi Jim, the positive correlation between ASB reports and road lights seem interesting indeed. I think that ASBs, like any other human activity, need light to happen. I mean: we might implicitly associate the danger with the bad guy waiting for us in the 'dark alley', but maybe the truth is that the bad guy needs the light too to be bad. If we look at the issue from a 'density' perspective, I think that lit streets are more attended on average than dark streets. Thus we might infer that higher concentrations of people lead, proportionally, to more ASB.

 Hi Francesco and Phillip, thanks for your helpful answers.

Philip - I will look up Dr. Martin, it sounds like his work will be an a useful read.

Francesco - I agree, these are very plausible explanations for the effect. It would be good to try and find some evidence to confirm these. I think there could be a few different explanations that may interact with each other. From a lighting researcher perspective, it's important we understand what impact adding more lighting may have on ASB, or other aspects of crime and disorder.

Jim

I can think of a couple of issues - first is that the density of street lights is probably also correlated with many other things as well, possibly including "pro-social" behavior such as cleaning stoops, picking up litter, and so on. Certainly lighting reduces the fear of crime (The 1970s review by Tien is still very relevant to the question: https://www.ncjrs.gov/pdffiles1/Digitization/47011NCJRS.pdf), but as mentioned it also sometimes is necessary to commit crime. (Criminals need task lighting, too!)

Because street lighting density is also linked to general population density, and probably other factors that may be linked to safety/crime-related environmental factors such as vegetation, sidewalks, road medians, bus shelters, trash cans, and so on, it is important to try to include as many of these factors in your regression modeling as possible to try to isolate what lighting may be doing versus the combination of lighting, high population density and other things. You still have the issue that correlation does not imply causality, but if you can account for as many of the other factors as possible, you can at least begin to make a case for causality.

A rather far-out parallel that also comes to mind is the fact that dementia patients in nursing homes sometimes become more agitated when they do not have strong patterns of light during the day followed by darkness at night. While it seems far fetched to guess that individuals who commit antisocial behaviors suffer from dementia, perhaps there is something about reduced circadian light-dark patterns that makes us more irritable or even more antisocial?

Thanks for the response John. I agree street light density is likely to be correlated with a number of other things that might also be linked to ASB reports. It would be nice to try and tease out some of these different possible factors and identify which may be causally related, if you know of any evidence or data that could be used to do this or encounter any in the future it would be great to look at this.

You make an interesting point about the possible link between lighting, circadian patterns and irritable / antisocial behaviour. Something to add to the list of possible explanations, and a very interesting idea worth exploring!

I think that there is a certain bias in favor of certain neighborhoods for infrastructure. This includes good sidewalks, street lights that work and so on. In way, it is almost like city officials write off certain neighborhoods. I worked for two different city administrations over a period of 8 years and the both favored the south side of town. Other mayors favored the North side. No one really spent anytime worrying about the center of the city which was the center of crime was this coincidentally? Other areas that had problems to a lesser degree were the east and west side of the city. This could be considered anecdotal evidence or I would to term it as "ethnographic observation and participation."

En Uruguay se ha estado trabajando sobre el tema, del cual no soy experta.

Un factor a considerar es la segregación socio espacial que se da en el territorio y los servicios de seguridad, iluminación, limpieza, educación, salud, etc. que existen en el territorio. Solo luego de aislar la incidencia de las otras variables sería posible justificar como significativo el papel de la iluminación de una zona. Barrios muy iluminados sin presencia policial visible son pasibles de comportamientos antisociales. Incluso distintos actos antisociales (no solo la delincuencia) pueden ubicarse en distintos momentos del día y de la semana.

dejo un link de evaluación preliminar de incidencia de la presencia de cámaras en las denuncias por hurto:

https://www.minterior.gub.uy/observatorio/images/stories/camarascv.pdf

My guess is that antisocial behavior and crime are not equivalents, and that the positive correlation between lighting and ASB reports may in fact explain the negative correlation between lighting and crime, reflecting greater police presence and greater perceived risk of detection. 

My guess is based on an assumption that "antisocial behavior" includes suspicious behavior that is not criminal, as well as behavior that is criminal but not the kinds of crime that people tend to worry about (theft, assault etc).

For example, people might call to report "suspicious looking" unknown individuals loitering on the street, underage youth drinking beer, people being loud and boisterous. This can result in a police response but no crime has been committed in examples 1 and 3, and the crime in example 2 is a relatively minor one. 

Laura, gracias por su respuesta! My Spanish is not great but I understand the points you are making and I agree. There are many other factors that could also be correlated with road lighting density that might also influence ASB. Further work definitely needs to try and pick apart the role of lighting from other factors in my view. Thanks also for the pdf, it looks interesting, if you had an English-language version that would be great but don't worry if not, I'll put Google Translate to use.

Michael, thanks also for your response. I think youre guess seems a very good one to me, it is the explanation I am currently leaning towards. Lighting helps so-called ASB be seen, perhaps because the people committing the ASB don't consider being able to be seen as important. For those committing more serious crimes, being seen may be a bigger consideration, and therefore more lighting leads to less crime. Well, that's one possible explanation anyway.

I thought this was an interesting question. I have heard of other anecdotal occurrences of this happening and the explanation is that the lighting allows more people to use/ loiter in the area, thereby heightening the risk for inappropriate behaviour. I think this supports Francesco's explanation (above) and would fit in with Routine Activity Theory. I also think its also a good example of a situational crime prevention initiative not working all the time, in all situations (which is often the case).

NI have no data to present, but it makes sense.  Anti-social people are as afraid of the dark as social people.  One test would be to see if there is a positive correlation between the number of street lights and the number of people present.  More people present suggest more antisocial people present with suggests more antisocial behaviors.

Density of population would be a factor for both street lights and certain common crimes per 100,000. I think this is easily provable. But, when we looking at some statistics like serial killing these defy common sense because I am seeing interesting data on that. Take Larry Hall of Wabash Indiana. He committed some 40 murders in very small towns with populations similar to his hometown. Wabash is billed as the "first electrically lighted City." It's population is 10,500.

The Interstate 65 and interstate 10 murderers killed in lonely motels and truck parking lots all across the US in very low population centers. Some crimes defy classification.

I think Phillip hit on the most important first step: controlling for population density. Number of ASB reports is going to increase with population, so the more useful comparison would be to first see whether road-distance normalized road lighting is positively associated with incidence rates of ASB (e.g., # ASB reports per annum/100k residents in geographic area). If the relationship holds for incidence rates, rather than just for number of reports, then this is a really intriguing finding.  In that case, I'd lean towards your [more lighting] --> [more likely to be seen, and thus, reported] explanation