Informal Social Control of Crime in High Drug Use Neighborhoods in Louisville and Lexington, Kentucky, 2000 (ICPSR 3412)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Hide

Study Purpose View help for Study Purpose

Understanding the community context in which behaviors occur is viewed as critical in predicting a wide variety of social problems for an increasing number of disciplines. While sociology and public health disciplines have a long tradition of understanding the ecological context of social problems, other disciplines are just beginning to move toward more ecological approaches. Within criminology, this trend of context-sensitivity has produced a literature that has become increasingly concerned with understanding the processes or mechanisms through which community characteristics affect crime rates. The most common community-level approach to understanding crime has been within a social disorganization framework. However, empirical examinations of this model have generally only examined aspects of structural disorganization, ignoring what has been referred to as cultural disorganization. The idea of cultural disorganization focuses on the extent to which culture has become weakened and can no longer provide the basis for social control. Social ties among community members are viewed as an essential mediating characteristic between the community structure and the informal social control of crime. Social ties are the foundation for informal social control as they provide the mechanism through which articulation of shared values occurs and support for enforcing those values is generated within communities. Communities with wider friendship and associational ties have been argued to have a greater potential for informal social control. This neighborhood-level study sought to explore the effect of cultural disorganization, in terms of both weakened conventional culture and value heterogeneity, on informal social control, and the extent to which these effects may be conditioned by the level of drug use in the neighborhood.

Study Design View help for Study Design

Because many of the issues of values and norms seemed particularly relevant to neighborhoods in which drugs were heavily used and in which a drug culture may have developed, researchers wanted to intentionally include neighborhoods known to be home to high levels of drug activity. While there has been much discussion on how to best define neighborhoods geographically, most community-level crime studies have used relatively large boundaries such as census tracts, political wards, municipally defined neighborhoods, or police beats. These "neighborhoods," however, are generally quite large, comprised of several thousand residents. The census block group was chosen to define neighborhoods for this study. Approximately 60 neighborhoods were used in this study. Data for Part 1 were collected from face-to-face and telephone interviews with households in the targeted sample (on average, 35 households per census block group/neighborhood). Part 2 is comprised of data collected from the United States Census 1990 Summary Tape File 3A (STF3A) as well as the United States Census 2000 Population counts, Lexington and Louisville police crime incident reports, and police data on drug arrests. The responses gleaned from the survey used in Part 1 were aggregated to the census block group level, which are included in Part 2. Each sampled household was mailed a letter explaining the purpose of the study and stating that the household might be contacted to participate in the study. The letter further informed recipients that participation in the study was voluntary and completely confidential, and that survey respondents would be paid 15 dollars for participating in the survey, which would last approximately 20-25 minutes. Cover letters were mailed to sampled households in Louisville on February 10, February 16, February 22, and February 25, 2000, and to households in Lexington on April 1, April 5, and April 10, 2000. The mailing of the letters was staggered in order to assure that each household would receive the letter only days before interviewers would contact them. At this point households with phones were separated from the households without phones. The "with phones" sample was then turned over to the University of Kentucky Survey Research Center, which conducted the telephone interviews. Households without phones were interviewed face-to-face by members of the research team. Telephone interviews were conducted between February 16, 2000, and June 11, 2000. All telephone interviews began with a confirmation of the respondent's address, so that researchers were certain to interview only persons at the sampled addresses. Interviewers were instructed to speak with the person in the household who had most recently had a birthday and who was at least 18 years of age. Disconnected numbers were tried again two weeks later in order to attempt to capture temporary disconnects. If those numbers were still disconnected, they became eligible to be included in the "no-phone" list. At least 20 attempts, and in some neighborhoods as many as 30 attempts, were made to contact a household member at the listed phone number. Even though researchers used the most recent city-wide directories, many of the sampled households were vacant or researchers had a telephone number that was no longer in service or no longer associated with the sampled address. Therefore, in some neighborhoods it was necessary to draw a second random sample in order to achieve the desired 35 respondents/households per neighborhood. Households without telephones, or with disconnected numbers, were interviewed using face-to-face interviews. Interviewers were trained by the research team and included both males and females, and African Americans and Caucasians. Face-to-face interviewers attempted to make contact at each address five times. Approximately 75 percent of the completed surveys were conducted over the phone and 25 percent were conducted in person. Two neighborhoods were eventually dropped from the study. One neighborhood in the medical area of Louisville had only a small number of respondents and another neighborhood consisting of elderly housing had a low response rate. For Part 2, police data for each of the two study cities were also collected. Researchers collected police incident reports for 1997, 1998, and 1999, as well as drug arrest data for 1999. All police data were geocoded using ArcView. After geocoding, lists of incidents were printed for each neighborhood. These lists were verified by a research assistant, who checked addresses on all boundary streets to be sure only those addresses on the included side of the street (odd versus even) were counted. Further, incidents that occurred in the intersection of boundary streets were treated in two ways. Because these incidents could be counted in as many as four different neighborhoods (for example, if the intersection was a boundary on the southeast corner of neighborhood A, the southwest corner of neighborhood B, the northeast corner of neighborhood C, and the northwest corner of neighborhood D), researchers counted them in each neighborhood in which they were included (total counts) and randomly assigned them to one of the neighborhoods in which they could have been included (random counts).

Sample View help for Sample

The sample for this study consists of 66 block groups in two urban communities in Kentucky. The sampling plan was developed to assure a sufficient number of high drug use neighborhoods and to assure an adequate distribution of predominantly white, predominantly racially mixed, and predominantly minority neighborhoods. To achieve these goals, a non-proportional stratified sampling of block groups was used. Once block groups were sampled, all street segments within those block groups were identified. Using the "street section" of city-wide directories, all addresses on these street segments were then identified, and a sample of approximately 60 households from each block group was selected using systematic random sampling. Since the cross-reference directory distinguished between residences and businesses, only residences were included in the sampling frames.

Universe View help for Universe

Census block groups in two urban cities located in Kentucky and individuals, representing households, from within the block groups.

Unit(s) of Observation View help for Unit(s) of Observation

Data Source View help for Data Source

Data Type(s) View help for Data Type(s)

Description of Variables View help for Description of Variables

Data for Part 1 include respondents' answers to questions about their interaction with their neighbors such as how often they borrowed or exchanged things with their neighbors, how often they invited neighbors over for a meal, to play cards, or to talk, how often they discussed their children, neighborhood children, or church with their neighbors, how often they went out with their neighbors to dinner or to a sporting event, and how often they disagreed with their neighbors. Respondents were also asked about the number of times within the last six months certain crimes took place in their neighborhood such as fights involving or not involving a weapon, sexual assault or rape, robbery, or muggings, and whether any member of their household was a victim of any of these crimes. Whether any of the following events ever occurred in the neighborhood was reported by the respondents, as well as if the respondent called the police to report these events: people using drugs, noisy parties, loud arguments, broken street lights, abandoned cars, and people drinking or buying and selling drugs. Respondents also provided their opinion about the importance of education, drug use, trusting neighbors, living by a value system, spousal abuse, and the police and their interaction with the neighborhood. Demographic variables include the gender, age, race, and marital status for each respondent. For Part 2, variables that were used in Part 1 were aggregated up so that they represented a "neighborhood" or census block group as opposed to an individual household. Additional variables for Part 2 include the 1990 Census population totals for specific age groups (e.g., total persons aged 22 to 24), ratio of income to poverty taken from the 1990 Census, and homicides, rapes, assaults, robberies, burglaries, and auto thefts from 1997, 1998, and 1999, taken from police reports.

Response Rates View help for Response Rates

For Part 1, the overall cooperation rate was 60 percent.

Presence of Common Scales View help for Presence of Common Scales

Several Likert-type scales were used.

Hide

Original Release Date View help for Original Release Date

2003-12-19

Version History View help for Version History

Hide