Accounting for variation in distrust of local police

Citation

Lytle, D. (2009), "Accounting for variation in distrust of local police", Policing: An International Journal, Vol. 32 No. 3. https://doi.org/10.1108/pijpsm.2009.18132cae.003

Accounting for variation in distrust of local police

Article Type: Perspectives on policing From: Policing: An International Journal of Police Strategies & Management, Volume 32, Issue 3

E. Sharp and P. Johnson Justice QuarterlyVol. 26 No. 12009pp. 157-82

Sharp and Johnson (2009) examined influences of citizens trust in the police. Specifically, the authors sought to explain why some individuals distrust the police and others do not. The authors hypothesized that race plays a role but so does community context. However, unlike previous research, which has focused on neighborhood variation in satisfaction with the police, Sharp and Johnson (2009) focus on variation between cities as a whole.

The data for this study comes from the Social Capital Community Benchmark Survey. The authors use data from 33 of the communities surveyed across the United States. Certain communities were excluded because they were a combination of communities and not a single city. The dependent variable for the study was the amount of distrust respondents had with the police. This variable was dichotomized from a four-point ordinal Likert scale. While there was variation between races, respondents were generally trustful of the police. A total of 60 percent of black respondents had at least some trust of the police and about 85 percent of white respondents had at least some trust in the police. This study used two-levels of predictors, individual and city level. Individual level variables included: race, age, marital status, political interest, education, gender, home owner status, distrust of neighbors, distrust of own race group, quality of life, and political efficacy. City level predictors included: black representation on police force, the amount of deadly force used against blacks, the presence of a black police chief, an interaction term between use of force against blacks and black police chief, total police force size, total crime rate, violent crime rate, responsiveness of policy scale, and whether or not black was the majority race.

For black respondents, higher quality of life, a greater interest in politics and those who feel the government can effectively deal with problems were significant predictors of police trust. Black respondents that scored higher on those variables were more trustworthy of the police. For white respondents, quality of life, being married, and females was associated with more trust of the police. For both racial groups, distrust of neighbors and distrust of own race indicated higher levels of distrust of the police. Blacks were influenced by use of force against blacks but white respondents were not significantly influenced by this variable. While black chief was not a significant variable for either racial group, the interaction term decreased distrust of police was a significant predictor for black respondents. Black respondents that lived in cities with larger police departments and cities with more total crime had more distrust of the police. Both races had more distrust of the police when there were more black officers on the police force. Overall, there were many more city level predictors for black than white respondents.

Early intervention program criteria: evaluating officer use of forceT. Bazley, T. Mieczkowski and K. LerschJustice QuarterlyVol. 26 No. 12009pp. 107-24

Bazley et al. (2009) examined two methods of early warning system officer identifiers. According to the authors, early intervention systems routinely use the number of use of force incidents as a major criterion in identifying potentially problematic officers. In this paper, the authors test the predictive validity of a simple count of use of force incidents and a weighted force factor system. By examining both methods, the authors hope to determine if one method is better at predicting problematic officers or if both methods are equally good predictors.

The paper used logistic regression as the primary analytical technique. The dependent variable was whether or not the officer was identified by the original early warning system as an at risk officer. Bazley et al. (2009) reported that 33 officers were identified by the police department’s current early warning system. The two independent variables were first, a count of the number of use of force incidents and second, each officer’s weighted force factor. The force factor was calculated by comparing the amount of suspect resistance in an encounter compared to the amount of force used by the officer. The amount of force was subtracted from the amount of resistance so that negative values indicated more force was used in an encounter. Thus, each interaction could score between a−5, where the officer used more force than was allowed and a+5, where the officer used less force than necessary. Each incident of use of force per officer was measured, and then each of these calculations was summed to create a total force factor for the officer.

The model was fairly predictive with a pseudo R² of 0.343 (Bazley et al., 2009). Both methods of prediction were statistically significant. The unexpected finding was that the force factor variable was in the positive direction indicating that officers tended to use less force in high resistance situations. The authors were expecting that this value would be negative. The number of use of force reports variable was in the predicted direction, positive, indicating that those that had a higher number of use of force reports were more likely to be flagged for early intervention. The authors expressed concern that the two methods had varying results. They suggested that police departments should not rely exclusively on a count of the number of use of force incidents.

Race, class or neighborhood context: which matters more in measuring satisfaction with the policeY. Wu, I. Sun and R. TriplettJustice QuarterlyVol. 26No. 12009pp. 124-56

There has been a great deal of research that has examined factors that influence satisfaction with the police. A lot of the research has focused on race and class. More recently, neighborhood context has become a more prominent factor in satisfaction research. Wu et al. (2009) attempt to determine which of these three factors is the most important predictor. Using hierarchical linear modeling (HLM) the authors assess the effects of each of these three main predictors.

The data for this study comes from the project: Informal Social Control of Crime in High Drug Use Neighborhoods in Louisville and Lexington, Kentucky, 2000. The study was conducted in two parts: first, face to face interviews, which accounted for about a quarter of the data and second, telephone interviews, which accounted for the remaining data. In total, the project interviewed 1,963 individuals from 66 neighborhoods. The dependent variable was satisfaction with the police, which was dichotomized from a four-point ordinal Likert scale. The predictor variables were in two sets, individual level and neighborhood level. The individual level predictors included: race (white and non-white), middle class (middle class and all else), age, gender, crime victimization (no and yes), and police harassment (no and yes). The measures at the neighborhood level included: the racial composition of the neighborhood (categories included predominantly black, racially mixed, and predominantly white, which was the reference category), the amount of concentrated disadvantage, the amount of residential mobility, and the violent crime rate. The authors produced five different HLM models. The first model was only individual level variables. Model 2, added the neighborhood racial composition variables. Model 3, added the concentrated disadvantage variable to the previous model. Model 4, added the residential mobility variable, and Model 5 added the violent crime rate variable.

When only individual level variables were in the model, race was a significant variable, as was class, age, crime victimization and police harassment. At the individual level, black respondents were less likely to be satisfied with the police, a finding that was consistent with previous research. Model 2 added the first two neighborhood level variables, and when these models are added the race variable at the individual level, became non-significant. When concentrated disadvantage was added to the model, class and racially mixed neighborhoods became non-significant. Additionally, predominantly black and concentrated disadvantage were significant predictors of satisfaction of police. Areas that were predominantly black were less satisfied with the police and areas with more concentrated disadvantage were less satisfied with the police. When residential mobility was added to the model concentrated disadvantage was no longer significant. Finally, when violent crime rate was added to the model all variables that had dropped out remained out of the model and violent crime rate was a significant predictor indicating that respondents that lived in areas with higher violent crime rates were less satisfied with the police. The study indicates that neighborhood context was a stronger and more important predictor of satisfaction than race or class.

Police officers’ perceptual distortions during lethal force situations: informing the reasonableness standardD. Klinger and R. BrunsonCriminology and Public PolicyVol. 8 No. 12009pp. 117-39

The criteria for judging an officers use of force is the reasonableness standard. However, reasonableness tends to be a vague term, open for interpretation, and there has been little research to determine the exact nature of reasonableness. Since use of force is a major concern for both the public and the police, the present study attempted to determine what perceptual distortions may occur during a lethal force situation. Specifically, Klinger and Brunson (2009) examined the extent of visual, auditory and temporal distortions both before and after a lethal force incident.

The data for the study was collected via a snowball sampling process that included face-to-face interviews with officers that had been in an officer involve shooting. In total, the sample was comprised of 80 officers from 19 different agencies and included 113 incidents of police shootings. A total of 74 officers were male and six were female, and 62 officers were white, 18 were non-white. The mean age of the officers was 32 and the mean amount of experience was about eight years of service. The majority of officers were at the officer or deputy sheriff rank.

The greatest amount of variation was in auditory perceptions. Visual distortions were fairly consistent before and after the shootings, approximately the same percentages of officers experienced tunnel vision, heightened visual detail or both before and after the incident. There was a lot of variation in auditory distortions. Of the officers, 42 percent experienced auditory blunting before the incident but 79 percent experienced it after the incident. Before the event, 10 percent reported auditory acuity and only 5 percent reported this after. No officers experienced both before the event and 8 percent reported both after the event. Time distortions were somewhat consistent. Of the officers, 43 percent experienced slow motion before the event and 40 percent reported slow motion after the event. Before the event, 12 percent reported fast motion and 17 percent reported fast motion after the event. Finally, 2 percent of officers experienced both temporal distortions after the event. The authors conclude that reasonableness may vary depending on the situation and is not necessarily consistent across encounters.

Daniel LytleUniversity of Cincinnati, Cincinnati, Ohio, USA