Vol. 78, No. 1External submissions

Man argues with two police officers.

Taking matters into your own hands

Public support for vigilantism and confidence in police

The Dutch study sought to determine whether support for vigilantism is linked to a lack of confidence in police. Credit: Robert Hoetink


Beating up shoplifters, stabbing sex offenders and shooting robbers: these types of violence do not only result in outrage, but also in public support. In Canada in 2009, there was quite a controversy when storeowner David Chen was arrested for chasing a repeat shoplifter and locking him up in his van. Chen was charged with assault and forcible confinement. The criminal charges and trial sparked a heated public debate. Was Chen a hero or a criminal?

Chen was not convicted for the use of excessive violence against the suspect; he was acquitted. There are numerous cases in which citizens have used considerable violence against alleged criminals and were subsequently praised for doing so by the public. These cases sparked heated debates about private action and the boundaries and shortcomings of law enforcement. The public showed admiration for the vigilantes and labelled their behaviour as righteous justice.

When citizens express support for crime and criminals, this begs the question what causes such reactions. Public support for vigilantism is frequently interpreted as a sign that citizens have lost confidence in the police.

After all, when citizens applaud those who take the law into their own hands, does that not imply that they believe that they cannot count on the police to properly deal with the situation?

Citizens who take the law into their own hands defy the state monopoly on the legitimate use of force. The police are commonly seen as the embodiment of this monopoly, as they tend to be much more visible to the public than other criminal justice professionals. It is thus understandable that when citizens support illegitimate acts in response to crime, this is considered a sign that police confidence is at stake. But is this really the case?

To find out, we conducted a study to assess whether support for vigilantism is attributable to a lack or low level of confidence in police. Importantly, we distinguished between confidence in police on both a general and specific level.


We presented a vignette, or brief case study, about an act of vigilantism to our respondents, and asked them to answer some questions. The vignette describes two criminal acts: a precipitating shoplifting crime and a subsequent violent act of vigilantism. The story concerns Ann, a storeowner who suspects a particular customer has been shoplifting. Her suspicion is corroborated by evidence from surveillance tapes. She sends the video footage to the police, hoping that they will undertake some action. A few days later, that same customer enters her store again. This time the woman steals a t-shirt, but upon realizing that she has been seen, manages to escape. The next day, the storeowner is downtown on a day off when she happens to catch sight of the shoplifter. She forcefully grabs the woman's arm and physically assaults her.

After reading the vignette, respondents completed a survey. Respondents indicated their agreement with various statements using a response scale that varied between one (fully disagree) to five (fully agree). The support for vigilantism measure was comprised of 16 items, addressing various aspects such as approval of vigilantism, empathy with the offender and the victim, punishment, deservingness and blame.

These included items such as "What Ann did is justified" and "Thanks to people like Ann at least something is done against crime." To measure general confidence in police, eight items about the police were presented. These include "The police do their job well" and "The police are there when you need them."

To study whether support for vigilantism is related to specific police responsivess, we varied police responsiveness between conditions. In the high responsiveness condition, a police officer shows up at the store to pose questions about the shoplifting. He promises more frequent surveillance of the store, and gives the storeowner a phone number so he can be reached at all times. In the low responsiveness condition, the storeowner does not hear back from the police after sending the tapes. Upon contacting the police again herself, she is told that they do not have time to deal with the shoplifting.

We were furthermore interested in finding out whether support for vigilantism also depends on situational factors. Specifically we focused on variations in the amount of vigilante violence used. In the low violence condition, the vigilante hits the shoplifter, resulting in a black eye and a headache. In the high violence condition, the shoplifter falls to the ground after being hit by the vigilante. The storeowner then goes on to kick her in the head, leaving her with a broken jaw and a heavy concussion.

Data were collected by handing out questionnaires to train passengers in the Netherlands. This method allowed us to reach a mixed sample of Dutch citizens relatively easily. The response level was 70 per cent, resulting in a final sample of 385 people. Mean age was 35 years; 55 per cent was male.


The average level of support for the presented case of vigilantism is 2.73 on a five-point scale. Respondents are overall thus not very positive about the act of vigilantism in the vignette. The item that resulted in most agreement is "Ann should have looked for another solution" indicating that vigilantism is not the preferred reaction to the shoplifting. Nonetheless, a large majority — 74 per cent — expressed an understanding for the vigilante's behaviour. Another noteworthy observation is that less than 12 per cent of respondents felt pity with the victim of vigilantism.

The mean rating of confidence in police is 3.13, which lies slightly above the neutral midpoint of the scale (i.e. 2.5). Respondents expressed least agreement with the item "The police are there when you need them."

Interestingly, this matches the topic of concern for police responsiveness to the precipitating crime. Most agreement was found with an item expressing respect for police. This implies that dissatisfaction with certain aspects of police does not necessarily cause an overall lack of respect for them.

We analysed the role of various determinants of support (i.e. general confidence in the police, police responsiveness to the precipitating incident, and level of vigilante violence). The findings indicate that more general confidence in police results in less support for vigilantism. Importantly, however, the additional impact of situational characteristics reveals that support for vigilantism is not only attributable to a low level of confidence. Police responsiveness also played a role: the more actively the police responded after receiving the report and footage of the shoplifting, the less the subsequent vigilantism act was supported. Likewise, the violence used by the vigilante affected support: a more violent vigilantism act led to less support.


Our study provides empirical evidence for the often-assumed relation between support for vigilantism and confidence in police. However, confidence in police did not only play a role on the general level: police responsiveness on a situational level also affected support. When police were less responsive to the report of shoplifting, people expressed more support for the subsequent act of vigilantism.

Importantly, this suggests that the role of police in the event leading up to vigilantism can have a considerable impact on public opinion about a specific case. In our study, when police took the victim of shoplifting seriously and showed genuine interest in the case, this already affected public response to a subsequent act of vigilantism, despite the fact that arrests were not yet made.

Our findings suggest that the impact of the actions of criminal justice agencies on a concrete, situational level should not be underestimated as they may play an important role in the events leading up to an act of vigilantism.

Even if someone has a high level of general confidence in the police, he may support vigilantism when he perceives them to have failed in their response to a specific crime situation. As such, police responsiveness may be a factor leading to or preventing an act of vigilantism itself.

To prevent vigilantism, police may therefore be advised to put substantial effort into explaining their response to a crime, or lack thereof, to those involved and to society. Although this is speculative, our findings so far do stress the importance of the response of police to a crime on a situational level.

Furthermore, as some acts of vigilantism may be impossible to prevent, proper police responsiveness and subsequent popular education about that responsiveness, may keep widespread public support for vigilantism to a minimum.

Dr. Nicole Haas is assistant professor at the Erasmus School of Law, Criminology Department, Erasmus University Rotterdam, the Netherlands.

Prof. Dr. Jan de Keijser is a professor at the Institute for Criminal Law and Criminology, Leiden University, the Netherlands.

A version of this article by Nicole Haas, Jan de Keijser and Gerben Bruinsma originally appeared in the journal Policing and Society (vol. 24, 2014, issue 2, pp. 224-241).

Date modified: