Home » 2013 Categories » 2013 – Literature Review (Page 2)

Category Archives: 2013 – Literature Review

The Time-Space Connection Among Urban Burglaries, and its Possible Effect on Localized Urban Flight

by Chris Bowman Bowman_Chris_LitReview



When a household becomes the victim of burglary, Rountree (1996) suggests that the victimized household and surrounding neighborhood experience an increase in perceived crime risk, as well as the increased use of precautionary measures such as locked doors, alarm systems, light timer devices, etc.  However, the findings demonstrate that perceived risk and the utilization of security measures vary among neighborhoods with different racial and income compositions.  What Rountree (1996) fails to address is a more extreme precautionary measure: household relocation.


There exist a number of studies analyzing the spacio-temporal link between urban burglaries, each suggesting that burglaries tend to be correlated in both location and timing.  Therefore, the perception of elevated risk and use of precautionary measures following a burglary are justified.  Also, analysis of the linkage between urban crime, urban flight, and housing costs suggest that further study may be undertaken to utilize data on recent trends in localized property crime to predict the future effects on home sales and changes in property values.


Crime Risk

Predicting when and where crime is likely to occur can assist urban police forces in allocating their resources more efficiently.  Johnson and Bowers (2004) assert that criminology research can contribute to three things:

  1. The understanding of demographic and other neighborhood attributes that are associated with high incidences of burglary, so that crime prevention can be undertaken before victimization.
  2. The understanding of the linkage between the distance in space and time of burglaries, taking into account prior victimization.
  3. The understanding of the interaction between prior burglary victimization, sociodemographic data, and increased burglary risk.


In addressing the first point, Budd (1999) determines that the risk of domestic burglary is increased in households with a lack of security measures, low levels of occupancy, detached houses, inner city locations, and in households in which there is a single adult with children.  Budd also acknowledged that one-fifth of burgled homes have repeat burglaries within the year.


Johnson and Bowers (2004) investigate the spacio-temporal link in burglaries, suggesting that burglaries cluster in time and space.  Using crime data from Merseyside, England, Johnson and Bowers accumulate 1,692 records of burglary including the dates, locations, and times of the crimes.  They identify homes that have been burglarized more than once in the past year.  They first use the Mantel test, initially designed by Mantel (1967) to measure levels of disease transmission, to compare the expected and observed event distributions that take the form of:


d = distance between i and j in meters

t = time between i and j in days

The observed value incorporates actual times and distances, while the expected value assumes a random pairing between all times and distances between burglary events, and is assigned the mean value of this random distribution.  Johnson and Bowers then perform a z-test using the Mantel sum to check the variance between these values for significance. If there is evidence of space–time clustering, the z-score will be statistically significant, having a value greater than or equal to 1.96.  The actual Mantel z-score obtained in their research is 5.49, revealing that burglaries are more likely to occur near each other in both space and time.


Johnson and Bowers (2004) also test their data using the Knox (1964) technique, by creating a contingency table to produce a standard residual by comparing the number of observed and expected pairs that fall in each cell.  The Knox residual is found to be +6.5, further confirming the clustering of burglaries in space and time.  In the following figure, residuals greater than +1.96 indicate more burglaries than would be expected.  Residuals between +1.96 and -1.96 are in the expected range, and residuals lower than -1.96 demonstrate fewer burglaries than expected.  Figure 1 demonstrates that residuals are highest within one month and 400 meters of a burgled home.


Figure 1


In a companion paper, Bowers and Johnson (2005) build upon the spacio-temporal relationship of burglaries to include the sociodemographic characteristics of wards in Merseyside, England.  Using 75,101 records of domestic burglary over a period of 6.5 years, they first repeat the Mantel test as described above, and prove the presence of clustering.  They then use data on affluence, home type (detached, terraced, flat, etc…), and heterogeneity of neighborhood housing to identify which characteristics lend themselves to a greater likelihood of a space and time link between burglaries.  Bowers and Johnson (2005) conclude that homes in poor wards are more likely to have repeat burglaries at the same house, while more affluent neighborhoods are more likely to exhibit a spacio-temporal link.  Time-space burglary clusters are most likely to occur when homes are on the same side of the street, affluent, similar in structure, and immediate neighbors.


Crime and Residential Choice

Possibly one of the costliest measures in response to increased property crime in an area is moving the household to a new location.  If victims are more likely to move following an increase in crime, it is useful to know who is most likely to move, and how much they are able to sell their home for.  Cullen and Levitt (1997) find as urban crime rates increase, high-income households and those with children are more likely to move out of the city.  They claim that this outward crime-related mobility leaves the city with generally lower property values, a decreased tax base, and greater concentrations of poverty; these conditions put an increased strain on limited public services and may contribute to further crime, perpetuating the cycle.



As Rountree (1996) demonstrates, burglarized households and close neighbors are more likely to increase preventative measures following the crime, and other findings are consistent with the notion that the likelihood of a burglary occurring in a particular area is increased if a prior burglary occurred nearby in both location and time (Johnson and Bowers 2004; Bowers and Johnson 2005).  Because increased crime rates lead to urban flight (Cullen and Levitt 1997), it would be worthwhile investigating the effect of localized spates of crime on the rate of neighborhood flight, and the prices commanded by the homes sold.  I expect that outbreaks of crime in previously low-crime neighborhoods may prompt some households to relocate, and lower the price of the house being sold.  Owing to the time-space connection of burglaries, we may be able to predict a decrease in property value following a single burglary.

Works Cited



Bowers, Kate J., and Shane D. Johnson. “Domestic Burglary Repeats and Space-Time Clusters The Dimensions of Risk.” European Journal of Criminology 2.1 (2005): 67-92.


Budd, Tracey. “Burglary of domestic dwellings: Findings from the British Crime Survey.” (1999): 99.


Cullen, Julie Berry and Steven D. Levitt (1997). Crime, urban flight, and the consequences for cities. National Bureau of Economic Research, Inc. Working Paper No. 5737.

Johnson, Shane D., and Kate J. Bowers. “The Burglary as Clue to the Future The Beginnings of Prospective Hot-Spotting.” European Journal of Criminology1.2 (2004): 237-255.


Knox, G. (1964). Epidemiology of childhood leukemia in Northumberland and

Durham. British Journal of Preventative and Social Medicine 18, 7–24.


Mantel, N. (1967). The detection of disease clustering and a generalized regression approach. Cancer Research 27, 209–20


Rountree, Pamela Wilcox, and Kenneth C. Land. “Burglary victimization, perceptions of crime risk, and routine activities: A multilevel analysis across Seattle neighborhoods and census tracts.” Journal of Research in Crime and Delinquency 33.2 (1996): 147-180.