Suicide by Cop: Police Shooting as a Method of Self-harming

Citation

Deryol, R. (2011), "Suicide by Cop: Police Shooting as a Method of Self-harming", Policing: An International Journal, Vol. 34 No. 2. https://doi.org/10.1108/pijpsm.2011.18134aaa.003

Suicide by Cop: Police Shooting as a Method of Self-harming

Suicide by Cop: Police Shooting as a Method of Self-harming

Article Type: Perspectives on policing From: Policing: An International Journal of Police Strategies & Management, Volume 34, Issue 2

Vivian B. Lord and Michael W. Sloop,Justice Quarterly,Vol. 38,2010,pp. 889-95

Deciding police shootings as a mean of self-harming is a complex action to be understood by the people who do not have sufficient knowledge of attempting person’s individual life. Person’s interactions with others and difficulty of measuring his intention make self-harming action complicated. This study proposes a set of indicators that may help to determine police shootings as taxonomy of suicide by cop. The authors used the decision tree developed by Best et al. (2004) to assess observable acts implying suicidal intentions of police shootings. They classified 242 cases under two headings using Hostage Barricade Data System (developed by FBI’s Critical Incident Response Group): self inflicted suicide and suicide attempts as suicide by cop (SbC). This classification formed the binary dependent variable and decision tree indicators are used as independent variables predicting the cases’ log odds of classification under either heading. The Hostage Barricade Data System (HOBAS) stores historical data from participating law enforcement agencies across the nation on hostage/barricade, suicide, kidnapping, and attempted suicide incidents.

The study tries to answer two research questions:First research question evaluated the personal and history differences between suicides and SbC with regard to demographics, prior suicide attempts, mental illness, drug addiction, criminal history, crisis, criminal act, drug use during incident, weapon used, refuse to surrender, and outcome of the act.

Results of the comparison indicated no significant difference in terms of demographics. However, some historical and behavioral characteristics differed significantly. First, previous suicide attempts differed significantly in that individuals who took part in SbC cases have more multiple suicide attempts and less one-time suicide attempts than those who took part in self-inflected suicide cases. Second, self-inflicted subjects were more likely to have history of mental hospital commitments. Third, with regard to characteristics of individuals during the incident, a criminal act as part of the incident immediately before the incident, the weapon used, and surrender to police were significantly different (p < 0.01): SbC subjects were more likely to involve in criminal act just before the case, more likely to possess a weapon, and less likely to surrender to police than self-inflicted suicide subjects. Last, there was a significant difference in the outcome of the incident. The majority of the self-inflicted suicides remained as “attempt” since no injury occurred as a result of the event. On the other hand, SbC subjects’ action ended up with death or no injury with the odds of 40.4 to 44.7 percent respectively.

The second research question is answered by slightly modifying the Best et al.’s decision tree. The decision tree has the same format: it has 11 components answering a question subsumed within four headings. Table I shows the summary of the decision tree.

Table I The decision tree

The primary indicators of suicidal intent are combined and an ordinal scaled variable is created ranging from 0 to 3 with the higher scores indicated more intent to induce police officers to shoot them. For the other three headings, each component is entered into the model as a separate variable. Thus, nine variables are used from 11 decision tree questions. Apart from these nine variables, demographics are used as control variables in the analysis; however none of these demographics were significant. The results of the binary logistic regression indicated that the only significant predictor was primary indicators variable (coefficient: 4.541). In other words, the more likely the subjects communicated, showed, and/or planned their intentions to induce the officers to kill them, the more likely their action were classified as SbC. Explained variance was 85.5 percent (Nagelkerke R²).

Voluntary submission of cases by police agencies to the HOBAS database can be a limitation of the study; however, its fields contain rich source of incident and subject information. Furthermore, the authors argue that, to be able to measure intent more accurately, SbC criteria need to include observable verbal or non-verbal intent. Thus, research collaboration between mental health agencies and law enforcement might help to develop better measurement of SbC and help to identify these cases beforehand and save lives.

Abbreviations:RQ1.; Do individuals, whose attempts or actions are classified as either self-inflicted or SbC, differ from each other in terms of personal and individual characteristics?; RQ2.; Do the independent variables, used in binary logistic regression, significantly discriminate between self inflicted suicide and SbC cases?

Rustu DeryolUniversity of Cincinnati, Cincinnati, OH, USA