A common misconception in cities is that crime is isolated to certain areas of the city.  Typically, crime areas are thought to be located in areas of high density like our downtowns or areas “planned” for the greater dwelling units per acre. Today, we find that criminals have become incredibly mobile and criminal events occur often where we would not otherwise expect.
When we consider people per square mile, or potential criminals per square mile, the outcome should yield a greater number of crimes in the more dense neighborhood.  There certainly appears to be a correlation to density and crime, however I do not believe that such a correlation can be made.  Five minutes worth of research was all that I needed.  The following is what I found when I compared murders per capita in New York City, Denver and Colorado Springs.  The murders per capita are very comparable for the three cities (between 6.3 and 6.7 per 100,000 people each year).  The following are densities of New York City, Denver and Colorado Springs from www.CityData.com:

  1. New York City: 27,668 people per square mile
  2. Denver: 3,980 people per square mile
  3. Colorado Springs: 2,153 people per square mile

Crime Data Per 100,000 people from 2008. Source: www.citydata.com

As you can see, New York City has over ten times the density of Colorado Springs, yet the murder rate per person is equal.  Neighboring city, Denver, has nearly twice the density of Colorado Springs and again a murder rate that is nearly equal.  The cities shown above with the largest crime rates are Detroit and Kansas City.  Kansas City has 1,538 people per square mile (less than Colorado Springs) and Detroit has 6,564 people per square mile (three times as many people as Colorado Springs).

I believe that there are other factors that need to be considered that can make a profound difference.  One factor is the land use diversity of a place.  It is not only important to have a place that is alive 24-hours of the day, but also a place that does not have an expiration date, as referenced in the prior post Euclidean Effects on the Built Environment.
An area of town that lacks diversity in land use, segregates residential uses from commercial uses and industrial uses. It typically also segregates densities and types of construction.  Segregation creates homogeneous residential subdivisions with all of the residents in a narrow range of socioeconomic status and typically age.  This is perceived to have sociological advantages, but it certainly has a distinct disadvantage of a subdivision that has periods of time in the day when it is essentially closed for the day.
Crime in the form of burglary and break-ins often happen in subdivisions of working families during the day when the criminal believes that they have the best chance of not being seen.  When we mix the uses and densities together in a cohesive fabric, we begin to create places that have “eyes on the street” for 24 hours of the day.  Eyes on the street are an intangible security measure that was written about in the timeless book Death and Life of Great American Cities by Jane Jacobs.
I believe that the largest factor is one that is difficult to quantify, that is pride and a sense of ownership for the surrounding neighborhood.  This is accomplished many times by financial ownership of a home or parcel of land.  That is not always the case. In particular, within major cities where high density residential is common, the pride factor does not always come from financial ownership. It comes from the sense of ownership in the neighborhood.