Perspectives on policing

Citation

Lytle MS, D.J. (2008), "Perspectives on policing", Policing: An International Journal, Vol. 31 No. 1. https://doi.org/10.1108/pijpsm.2008.18131aae.003

Perspectives on policing

Article Type: Perspectives on policing From: Policing: An International Journal of Police Strategies & Management, Volume 31, Issue 1.

The impact of order-maintenance policing on New York City homicide and robbery rates: 1988-2001

Richard Rosenfeld, Robert Fornango and Andres F. RengifoCriminologyVol. 45 No. 22007pp. 355-384

Rosenfeld et al. (2007) analyzed the impact of order-maintenance policing (OMP) on New York City homicide and robbery rates. The authors studied data from the medical examiners office and the NYPD from 1988 to 2001. The authors sought to address six limitations of previous research on broken windows theory and order maintenance policing. First, Rosenfeld et al. (2007) argue that use of the violent crime index is a poor measure because it is an aggregate of homicide, felonious assault, rape, and robbery rates. By aggregating the measure variances across crimes cannot be examined. Rosenfeld et al. (2007) examine only robbery and homicide rates and looked for differences in the effects of OMP across the two crimes. Second, they assert that previous studies have omitted a measure of ordinance-violation arrests. Third, Rosenfeld et al. (2007) argue that previous research have not incorporated a measure of complaints of disorder. Fourth, the authors include a measure of felony arrests to account for the relationship between OMP and levels of serious crime. Fifth, they attempt to control for indirect effects of variables acting through violent crime trends, such as issues with manpower availability in a given precinct. Sixth, Rosenfeld et al. (2007) assert that previous studies have failed to control for spatial autocorrelation. They use Hierarchical Linear Modeling (HLM) to analyze their data.

The main dependent variables in this study were changes in the homicide and robbery rates. The main explanatory variable was a measure of OMP. Rosenfeld et al. (2007) measured OMP by the number of misdemeanor and ordinance violation arrests per 10,000 precinct residents. The authors measured disorder by the number of misdemeanor and ordinance violation complaints per 10,000 residents. The authors used the number of felony arrests per 10,000 felony complaint ratio to control for the amount of resources used to combat serious crime. They created a measure of the number of prison admissions per 1,000 felony arrests. Rosenfeld et al. (2007) also incorporated a manpower measure, the rate of officers per 10,000 residents, and a drug market measure, the amount of cocaine overdose deaths per 10,000 residents. The authors created a factor score control variable for SES disadvantage:

1. 1.

   percent Hispanic;

2. 2.

   female headed-households with children under the age of 18;

3. 3.

   percent living in Puerto Rico five years before;

4. 4.

   male unemployment rate, poverty rate;

5. 5.

   percent of households receiving public assistance;

6. 6.

   median family income; and

7. 7.

   percent of the population aged 15 to 24 years.

They measured residential instability with a factor score consisting of:

1. 1.

   the divorce rate;

2. 2.

   percent of the population living in the same house five years before;

3. 3.

   percent vacant housing units;

4. 4.

   percent owner-occupied housing; and

5. 5.

   the population density.

They also created a factor score for immigration consisting of:

1. 1.

   the percent of the population living outside the United States and Puerto Rico five years before; and

2. 2.

   the percent foreign born. Rosenfeld et al. (2007) controlled for percent Black.

Because previous research has argued that effects of OMP are merely mean regressions, the authors controlled for base rates of robbery and homicide as a single measure of violent crime. Finally, to control for possible spatial autocorrelation they created weights based on the ten closes precincts.

Rosenfeld et al. (2007) found that the rates of robbery and homicide fell at a faster rate after aggressive OMP policies were implemented by the NYPD. Areas that had the greatest increases in OMP experienced the largest decreases in homicide and robbery. Specifically, the authors found that a one percent increase in OMP was associated with a 0.09 to 0.14 percent decrease in robbery rates and a 0.24 to 0.33 percent decrease in homicide rates. They found that the base rates of robbery were associated with disorder measures but homicide base rates were not associated with disorder.

Rosenfeld et al. (2007) examined the influences on the use of OMP. They found that base levels of OMP were not significantly related to precincts with more manpower and drug markets. Yet, areas with high levels of police per capita and drug markets experienced the greatest growth in OMP. OMP was also higher at both base levels and growth rates in precincts with high felony arrest-complaint ratios. The authors also found base levels of OMP were lower in areas where there were high levels of disadvantage and minorities area and these same areas also experienced some of the largest growth in OMP.

While examining for direct and indirect effects, Rosenfeld et al. (2007) found that drug markets, police size and levels of disorder all had direct effects on robbery rates. Each of those variables had indirect effects through OMP as well. Drug markets and levels of disorder had direct effects on homicide rates, but police size only acted on homicide rates indirectly via levels of OMP. Drug markets and levels of disorder also had indirect effects through OMP. The authors concluded by noting that OMP contributed to about 4 percent of the reduction in robbery rates between 1988 and 2001, and about 10 percent in the homicide rates between 1988 and 2001.

Daniel J. Lytle MS University of Cincinnati, Cincinnati, USA