The purpose of this study was to examine the impacts of broken windows policing at crime hot spots on fear of crime, ratings of police legitimacy and reports of collective efficacy among residents of targeted hot spots.
The project tests the impact of a broken windows policing approach upon 108 street segments, divided equally and randomly into treatment and control locations, that evidenced relatively high levels of crime and disorder for three cities in the San Bernardino Valley area of California (Redlands, Colton, and Ontario), over a time period of six months. The study is twofold, collecting both administrative police data on crime as well as survey data from residents and business owners in the area.
The survey portion of the study was collected by phone from 371 residents and business owners in each street segment and were administered in two waves. These surveys were taken by the same respondents both before and after the police intervention in order to gauge their perceptions of crime and disorder on an individual level as well as the success of the intervention program. Questions to the respondents inquired about their perceptions of disorder and crime on their street, as well as fear and collective efficacy. Phone surveys were given in either Spanish or English, based off of the preference of the interviewee.
The broken windows policing intervention portion of the study had three central principles. First, no discovered physical or social disorders should go ignored by the police in the target segments. Second, social disorder was to be dealt with in an escalating fashion with citations and arrests as the last resort options. Third, the key element of dealing with physical disorder was rapid repair. Police were to notify the relevant agencies for cleanup of graffiti, trash and other physical disorder issues, and then follow up with them if needed to make sure the problems were dealt with as quickly as possible. Arrests were to be the last choice of action. In order to best serve the project, and their community, it was decided that three hours of police presence per week would be implemented on each target segment throughout the intervention period. Officers involved in the project were put through a training program about broken windows policing theory and the study and intervention design. This included training protocol on how to address specific types of disorder in hopes of ensuring consistency throughout the police intervention. Police were also trained on how to fill out the required log sheets of the event in order to track the type of crime and how it was dealt with.
For the telephone interview portion of the survey, the initial sample of telephone numbers to call was pulled from the PowerFinder software provided by InfoUSA and additional phone numbers were pulled from the city water departments (as needed) in order to reach the goal of a sample size of 30 phone numbers per street segment. This sample included both residential and commercial numbers. For the pre-intervention survey for residential addresses, the first person over the age of 18 in a household willing to participate in the survey was interviewed; businesses interviews asked for the owner/manager, or the person in charge of daily operations. For the post-intervention survey, the person who had taken the pre-interview survey was asked to take it.
Street segments are defined as the two block faces on both sides of a street and were selected based off of the following criteria:
- The street segment had enough emergency calls for disorder in order to warrant an intervention. A threshold of 10 or more disorder calls and three or more Uniform Crime Report (UCR) Part I crime calls in a year was used to make this determination.
- The street segment showed some degree of serious crime as illustrated by reports of Part I (UCR) offenses.
- The study sites be distant enough from one another to reduce the risk of possible contamination of treatment into control sites (study segments to be separated by at least one full street segment in all directions).
- A segment must have at least seven phone numbers present in the PowerFinder database to be eligible for inclusion as a study site.
Residents and business owners in Colton, Ontario, and Redlands Califoria between March 2008 and April 2009 (pre- and post- phone surveys).
Calls for service, crime incidents and arrest data for street segments in three cities in California (Colton, Ontario and Redlands) between March 2008 and April 2009.
Police activity logs.
Pre-intervention phone surveys of residents and business owners in the study area.
Post-intervention phone surveys of residents and business owners in the study area who took the pre-intervention survey.
Calls for service data (CFS).
Police arrest incident reports.
administrative records data,
This study includes five datasets:
1) police_arrest_data.sav: This dataset includes 59 variables on the arrests made by police. It includes data on the number of different types of arrests on each street segment (n=108) in the pre-intervention period, intervention period, and post-intervention period including the number of arson, assault, burglary, theft, motor vehicle theft, homicide, robbery, carjacking, rape, drugs and alcohol related incidents, disturbances, vandalism, and prostitution arrests.
2) police_call_for_service_data.sav: This dataset includes 71 variables on the number of calls for service (CFS) on each street segment (n=108) during the pre, post, and intervention periods. It includes number of calls related to arson, assault, burglary, theft, motor vehicle theft, robbery, carjacking, rape, drugs and alcohol related incidents, disturbances, vandalism and prostitution.
3) police_incident_data.sav: This dataset includes 63 variables on the number of incidents that occurred on each street segment (n=108) and that were verified by police during the three phases of the study (pre, post, and intervention). It includes variables on arson, assault, burglary, theft, motor vehicle theft, homicide, robbery, carjacking, rape, drugs and alcohol related incidents, disturbances, vandalism, prostitution and vagrancy.
4) police_activity_logsheet.sav: This dataset includes 106 variables on the log sheet file that was filled out by police after an incident (n=5272). It includes variables about the logistical aspects of the incident including: officer id, date, time (arrived and departed), minutes of officer time on segment, catalyst for action, description of social disorder and description of physical disorder. Additionally, a number of variables illustrate a physical or social disorder and include: number of suspects, number of people talked to, number of suspects involved and talked to, number of stop and frisks, number of warnings, number of arrests, and number of actions taken by officer.
5) resident_interview_data.sav: This dataset includes 191 variables representing interview questions of residents and business owners who resided in the examined areas (n=371). The variables include:
- Demographic information: Variables identifying gender, age, education, income, marital status, employment, children, and race/ethnic group, whether residents had their interview in English or Spanish, years of residence (or years working at the business), whether resident was an owner or renter, and the type of home.
- Perceptions of neighborhood: Variables describing perceptions that the interviewees had about their neighborhood including whether they felt it was a close knit community, were people willing to help one another, can people be trusted, do people get along and do people share the same values, how often do they or people in their community get together with one another, with local politicians, or religious leaders to talk about problems. Additionally, perception variables described answers residents had about their neighborhood when given a scenario such as "If a group of kids were skipping school and hanging out on a street, how likely is it that one of your neighbors would do something about it?"
- Perceptions of crime: Variables were created about people's perception of the amount of crime that is happening on their street and how likely they feel they will be a victim of a crime.
- Perceptions of police: Variables were created about interviewees opinions on their dealings with the police as well as their satisfaction with the department and the city in the way they handle crime.
Overall a total of 836 responses were obtained on the final 108 street segments which accounted for a response rate of 38.4 percent for the pre-intervention surveys. For the post-intervention survey, 496 completed post-intervention surveys were collected from the 836 household/business addresses that completed the pre-intervention surveys, but only 371 were completed with the same person who took the pre-intervention survey, which was the number used for analysis.