In recent years, a significant amount of research regarding procedural justice, which is theorised to contain four sub-constructs: respect (or dignity), trustworthiness, neutrality, and voice, has been undertaken (e.g., Antrobus et al., 2019; Madon et al., 2017; Madon & Murphy, 2021; Mazerolle, Bennett, et al., 2013; Solomon, 2019). This research has occurred in a wide variety of criminal justice contexts including the courts (e.g., Gover et al., 2007; Wales et al., 2010), police (Bennett et al., 2019; Mazerolle, Antrobus, et al., 2013; Murphy et al., 2019; Sargeant et al., 2017), and corrections (e.g., Barkworth & Murphy, 2021; Maguire et al., 2021; Ryan & Bergin, 2022).

Research suggests that there are benefits associated with using procedural justice processes in policing interactions. For instance, procedural justice improves perceptions of police legitimacy which predicts cooperation with law enforcement (Bolger & Walters, 2019). Additionally, procedural justice appears to increase legitimacy which then improves compliance (Walters & Bolger, 2019). However, the direct relationship between perceptions of procedural justice and behavioural compliance is relatively weak and not seen when only considering at longitudinal studies (Walters & Bolger, 2019). Nagin and Telep (2020) argue there is a need for further research in this area to provide evidence that a causal relationship exists.

Despite the amount of procedural justice research that has been undertaken, there are research gaps. For instance, much of the research has focused on face-to-face interactions and been survey based (e.g., Bennett et al., 2016; Radburn & Stott, 2019; Reisig et al., 2014; Wells et al., in press) and therefore focused on perceptions and intentions as opposed to the impact on actual behaviour. This paper will address the research gaps by reporting on a randomised field experiment that identified if a camera detected traffic infringement notice that was accompanied by a letter that incorporated the principles of procedural justice influenced subsequent speeding behaviour.

Procedural justice and road policing

Road policing is a significant investment by policing agencies around the world to help reduce deaths on the road and provide a visible police presence to the public (Castillo-Manzano et al., 2019). To date, most road policing initiatives are based on deterrence theory (Bates et al., 2012; Truelove et al., 2017), although procedural justice is increasingly being considered as a framework to help improve road policing outcomes (Bates, 2014; Tudor-Owen, 2021). This is because earlier research has indicated that the use of procedural justice reduces subsequent offending. For instance, Paternoster et al. (1997) identified using a sample of male domestic violence perpetrators that a procedurally just interaction with a police officer affects subsequent offending.

While some research studies in this area have been conducted using video interactions (Lowrey et al., 2016) or vignette studies (Barkworth & Murphy, 2015; Madon et al., in press), randomised field experiments have sought to operationalise procedural justice in road policing contexts. For instance, a key study in this field was the Queensland Community Engagement Trial (QCET), a field experiment that randomly assigned 60 roadside random breath testing operations, each consisting of 300–400 police-public encounters. In the experimental condition, police interacted with the drivers in a more procedurally just manner by communicating elements of neutral decision making (randomly selected for the breath test), trustworthy motives (the current road toll and impact of alcohol related road injuries/fatalities on family and police), treating drivers with respect (acknowledging and thanking drivers for their time), and providing voice (asking drivers for questions or comments). Whilst the interaction took significantly longer than the standard random breath test (Mazerolle et al., 2015), drivers in the experimental condition were more likely to report that their views on drinking and driving had changed as well as hold higher levels of self-reported compliance and satisfaction when compared with drivers in the control group (Mazerolle, Bennett, et al., 2013). The interaction also positively affected the views of the police officers (Bates et al., 2015).

Sahin et al. (2017) conducted an experiment with a sample of 458 drivers stopped for speeding in Turkey. Drivers in the experimental group received procedurally just treatment from a police officer which included providing information regarding the number of speed-related deaths in Turkey and how these police operations were designed to save lives. Interviews with the drivers following the police-driver interactions identified that the perceptions of the specific encounter were positive. However, there was not an improvement in general perceptions of traffic police officers. A further study in Turkey identified that individuals interacting with police at traffic stops perceived that an interaction with a police officer is more procedurally just when the officer has a body worn camera (Demir et al., 2020; Demir & Kule, 2022). Generally speaking, very few experiments occur within a road safety context despite the results of studies that use this methodology being considered trustworthy (Elvik, 2021).

In contrast to the field experiments, Bates, Allen, and Watson (2016) conducted a survey of 237 young novice drivers in Queensland, Australia, to identify if procedural justice was perceived to have an influence in more automated policing encounters such as camera detected speed offences. They considered how the four elements of procedural justice were perceived in relation to two types of speed cameras: point-to-point speed cameras and mobile speed cameras. The only procedural justice element related to these types of cameras and self-reported speeding by young drivers was neutrality.

While much of the existing research in this space is quantitative, there is a small amount of qualitative work that has explored procedural justice in a road policing context. Bates et al. (2021) undertook focus groups with young drivers in metropolitan and regional areas in two Australian states. They identified that young drivers perceived their interactions with police (which generally occurred in face-to-face contexts) to be respectful, gave them a voice and police officers and agencies could be trusted. However, they did not perceive that they were treated neutrally.

Wells (2008) had earlier undertaken research regarding procedural justice and camera based enforcement within the UK. She analysed discussion forums, held focus groups, and observed sessions of a Speed Awareness Course. The results indicated that while cameras used to detect speed were considered consistent, they were also considered unfair in many ways including by denying the driver a ‘voice’ in the interaction (Wells, 2008).

Increasingly, the use of procedural justice in road policing is being undertaken in non-western contexts. For instance, Soo Kim et al. (2019) used a South Korean adult driver convenience sample (n = 1241) to examine the effect of procedural justice on police legitimacy and then the effect of legitimacy on the intention to speed. While they found a relationship between procedural justice and police legitimacy, they did not find one between legitimacy and intention to speed. Tankebe et al. (2020) used cross-sectional survey data from 415 commercial vehicle drivers from Accra and Kumasi in Ghana to look at the effect of fairness on traffic violations. In contrast to research outlined above, they did not find a relationship between perceptions of police fairness and self-reported violations of traffic laws. Thus, it appears that the cultural context in which procedural justice interventions take place appears important. However, studies where the outcome is attitude-related or based on self-report data may not be directly comparable to studies that measure actual behavioural change. It is possible that cultural context may have a more limited impact on interventions that influence behavioural, as opposed to attitude, change.

It is important to note that not all road policing interventions based on procedural justice are successful even in western nations. MacQueen and Bradford et al. (2015) used the Festive Road Safety Campaign by Police Scotland to run an experiment. The experiment involved police officers providing key messages and a leaflet based on procedural justice to drivers. However, this experiment did not improve trust in police or legitimacy. This was possibly because the experiment was not implemented as planned due to breakdowns in communication leading to misunderstandings (MacQueen & Bradford, 2017).

As outlined above, most procedural justice research to date, particularly in the road safety field, has focussed on individual’s perceptions of the interaction. Additionally, with the exception of one or two studies that examine procedural justice within a speed camera context (Bates, Allen, & Watson, 2016; Wells, 2008), most research focuses on face-to-face interactions despite calls for investigations into how technologically mediated contacts with police can occur in a procedurally just manner (Wells et al., in press). The current study therefore fills an important gap by identifying if (a) a procedurally just intervention has an effect if it is delivered in a written format and (b) if there is an impact on subsequently detected speeding behaviour as opposed to perceptions.

Research context

Speeding is one of the ‘fatal five’ major contributing factors to fatalities on Queensland roads. In 2017, the year of our experiment, 52 or 21.1% of fatalities, and 292 or 4.5% of hospitalised casualties in Queensland involved speeding drivers or riders (Department of Transport and Main Roads, 2018). Research consistently has identified that age is related to speeding with younger drivers more likely to engage in speeding behaviours (Hoye, 2020; Norton et al., 2021; Truelove et al., 2021).

Speed cameras are used to detect speed offences committed by drivers. They are used in a diverse range of jurisdictions including Australia (Freeman et al., 2017), Britain (Wells & Wills, 2009), France (Blais & Carnis, 2015), Israel (Schechtman et al., 2016), Qatar (Shaaban, 2017), Sweden (Belin et al., 2010), the United States (Retting et al., 2008), and New Zealand (Keall et al., 2001). There are three main types of speed cameras. Mobile cameras require an operator to deploy them either from inside a vehicle or on the side of the road (Freeman et al., 2017). In contrast, fixed cameras are permanently installed in a particular location (Freeman et al., 2017). Mobile and fixed cameras can be operated in either an overt manner where they are visible to drivers or a covert manner where they are hidden (Bates et al., 2012). The third type of camera is a point-to-point camera which measures the average speed of travel between two particular points on the road (Soole et al., 2013). Evidence indicates that speed camera programs decrease injuries and fatalities associated with crashes as well as reducing travelling speeds (e.g., Blais & Dupont, 2005; Carnis & Blais, 2013; Pilkington & Kinra, 2005; Quistberg et al., 2019; Tilahun, 2023; Wilson et al., 2010). Queensland operates mobile, fixed and point-to-point speed camera programs (Freeman et al., 2017). When a speed camera detects a vehicle exceeding the speed limit, a traffic infringement notice (TIN) including details of the alleged offence is mailed to the registered owner of the vehicle. When the owner receives the TIN, they can either pay the fine, transfer the fine to the responsible person, or dispute the fine.

Drivers do not report positive views of speed cameras and speeding enforcement. For instance, research conducted in the UK indicates that most drivers engage in travelling above the posted speed limit (Corbett, 2000) and that many drivers perceive speed cameras to be unjust and unfair (Wells, 2008; Wells & Wills, 2009). Australian research suggests that drivers hold paradoxical views regarding speed enforcement considering that it both encourages road safety and is a tool for revenue raising (Fleiter & Watson, 2006; Soole et al., 2008).

Methods

Intervention and randomised field trial

The trial involved a full sample of 16,406 drivers who received a camera-detected speed infringement between late August and early September 2017. Before the traffic infringement was issued to them, batches of traffic infringement notices were randomly allocated to either the control (business as usual) or the experimental condition. While both groups received the standard infringement notice, the experimental condition involved participants also receiving a letter constructed using the principles of procedural justice (see Appendix A).

The letter incorporated the principles of procedural justice and key road safety messages drawing on substantive foundational research and intervention design alongside direct consultation with road safety practice experts from the policing and transport agencies in Queensland and internationally. Specifically, the letter adapted and replicated messaging from (a) the QCET road police-driver dialogue (Mazerolle et al., 2012) to denote the number of people killed and injured because of speed-related crashes on Queensland roads, and the role of road policing to help reduce this toll, and (b) a letter trial ‘No driver means to kill. They were just going too fast’ conducted in a collaboration between the West Midlands Police and the Behavioural Insights Team UK that aimed to reduce repeat speeding and improve compliance with fine payments (Behavioural Insights Team, 2016). As camera detected traffic infringement notices do not allow for police-driver dialogue, these initial statements sought to activate elements of trust and respect as core foundations for procedural justice and legitimacy. The letter also sought to communicate trust, neutrality, and respect by including information on how speed limits are set to reduce serious collisions and how monies collected from speed camera fines are re-invested to improve road safety (e.g., road improvements, safety education, and Queensland Health blood products). The resulting letter aimed to reduce speeding infringements through drivers’ improved perceptions of procedural justice and awareness of the role of speeding and speed cameras on road safety. If a diffusion effect was present, the intervention may also reduce other traffic infringements.

Data

The Department of Transport and Main Roads provided the research team with the de-identified traffic offense data of 16,406 Queensland drivers that received a TIN during the experimental trial period. Drivers with incomplete or missing license or demographic information were excluded (approximately 1% of the total sample). The final sample, after data cleaning, was 16,155 Queensland drivers (see Table 1 for the sample description). Data was collected for 12 months before and after the intervention.

Table 1 Demographics and tests of difference across the control (n = 7946) and experimental groups (n = 8209)
Full size table

Drivers who had two or more general traffic offenses or with at least one high-speed traffic offence (i.e., greater than 40km over the speed limit) in the 12 months preceding the trial were classified as high-risk drivers. Drivers with a maximum of one traffic offence in the 12 months preceding the trial were classified as low-risk drivers. Additionally, across the same time period, the number of months before the driver had a second traffic offence following the intervention was calculated.

Data analysis

Analysis employed four difference-in-difference tests with a negative binomial approach to evaluate the effectiveness of the intervention by identifying if drivers had received subsequent speeding infringements. The reason for this analysis methodology is due to a strong positive skew found in the number of traffic offences recorded by drivers in this sample. This is unsurprising, as many participants would receive between 0 and 2 additional traffic offences across the 24-month timeframe (12-months pre- and post-intervention). Difference-in-difference tests with a negative binomial approach allow for the use of data with a strong positive skew, while also testing the efficacy both the within (time) and between groups (condition) components of a randomised control trial, as well as balancing characteristics, both observed and unobserved, across the two conditions (McEwan, 2010). This approach has been used in previous studies into evaluations of speed camera-based traffic offences and incidences to reduce the impact of such data and samples on the results (Chainey et al., 2021; Li et al., 2012; Martínez-Ruíz et al., 2019; Yau et al., 2003).

The two dependant variables were number of speeding offenses, and number of other traffic offences (e.g., failing to stop at a stop sign, failing to stop at a traffic light). Given the relationship between age, risk taking, and traffic offences (Boufous et al., 2010; Hoye, 2020; Rodwell et al., 2023; Rolison et al., 2014), two additional analyses were proposed, one examining the effects of the program on young people (under 25 year-olds), and the other examining the effect on adults (25 years and older). The independent variables included time (12-months pre- and post-intervention), condition (0 = control, 1 = intervention), difference-in-difference (time × condition effect) age, gender (0 = female, 1 = male), driving profile (0 = low risk drivers, 1 = high risk drivers), and number of months until re-offence.

Results

The difference-in-difference tests found that individuals over the age of 25 years who received a procedural justice letter were 11% less likely to get an additional speeding fine in the 12 months following the initial speeding fine than individuals who received a standard speeding fine. No other significant interactions were found for other traffic offenses or speeding offenses for individuals under the age of 25. It is worth noting that being male was associated with higher numbers of all traffic offenses (6–33%), as was having a high-risk driver profile (193–570%). The results of the four difference-in-difference tests are shown in Table 2.

Table 2 Results for the difference-in-difference tests for under 25 years and 25-years and older across the control (n = 971 and n = 6975) and experimental groups (n = 915 and n = 7294)
Full size table

Discussion

The randomised field experiment reported here suggests that the inclusion of a letter incorporating the principles of procedural justice with a camera detected traffic infringement notice reduces subsequent speeding offending behaviour for drivers 25 years or older in the next 12 months by 11%. These findings are consistent with earlier research suggesting that a short procedurally just interaction with police officers can change perceptions of drunk driving and police (Mazerolle et al., 2012) and that the procedural justice element of respect is associated with fewer subsequent arrests for intimate partner violence (Maxwell et al., 2020). However, this study builds upon this earlier work by demonstrating that the effects of procedural justice (a) can be delivered in a written format and (b) have an effect on subsequent offending behaviour. These are key findings indicating that a relatively small intervention can have a demonstrable impact on offending behaviour.

Longitudinal research conducted in Maryland in the USA suggests that a single speeding infringement notice has a limited effect on reducing the likelihood that a person will receive a subsequent infringement notice (Lawpoolsri et al., 2007). Thus, it appears that for those 25 years or older, the addition of a letter based on procedural justice principles to a camera detected infringement notice strengthens the likelihood that the driver will reduce their speeding behaviour in the future. One possible reason for demonstrating the success of the intervention may be because speed cameras, while considered consistent, are not always considered fair (Wells, 2008). Thus, the inclusion of the letter may help enhance the perceptions of speeding offenders that speed cameras are fair.

This finding is consistent with the work of Tyler et al. (2021) who examined whether the use of procedurally just processes in a social media context assist individuals to self-regulate their behaviour and prevent the posting of inappropriate content (nudity). Although the sample size in the study was small, if people felt that they had been treated fairly by the social media platform, they were less likely to post nudity content in the future (Tyler et al., 2021).

It is possible that the mere presence of the letter in the current study may have created a Hawthorne effect (where participants change their behaviour because they believe they are being observed). This would mean that it was not the content of the letter that affected subsequent speeding behaviour. However, as there were no significant differences for all traffic offences or younger drivers, the differences that were found are unlikely to be related to a Hawthorne effect.

There was no reduction in speeding offences for young drivers. There are a range of social and situational factors (e.g., social group and peers), exposure factors (e.g., the times of day that they drive), core attributes (e.g., gender and personality), modifiable attributes (e.g., skill and experience), and situational assessment, decision making and hazard perception that mean young drivers crash at much higher rates than other drivers (Bates et al., 2014). There is also some emerging evidence that road policing interventions based on deterrence theory are not as effective for young drivers (Allen et al., 2017; Bates et al., 2017; Bates, Scott-Parker, et al., 2016) and that there may be a need to use more innovative approaches such as incorporating parents (Bates et al., 2019). The lack of reduction in speeding offences as a result of the intervention in this study is consistent with the idea that it is more difficult to reduce offending for younger drivers when compared to those who are 25 or older.

In this study, the procedural justice letter was focused on speed. This was because it was sent to individuals who had committed a camera detected speeding offence. However, we did consider if the letter had either a diffusion or a displacement effect on other road offending behaviour. The lack of a significant difference between the intervention and the control group indicates that there is no diffusion or displacement effect present.

Not all procedurally just interventions demonstrate the same positive effect found in this study. For instance, as noted in the literature review, the Festive Road Safety Campaign by Police Scotland did not result in improved trust in police or perceptions of legitimacy (MacQueen & Bradford, 2015, 2017). Similarly, Tankebe et al. (2020) did not find a direct relationship between perceptions of police fairness and self-reported violations of traffic laws in Ghana. This suggests that the type of intervention, the way procedural justice is included in the interaction, the outcome measure (whether it is attitudinal or behavioural), and the cultural context all affect whether or not a procedurally just intervention is successful.

This study identified that women are less likely than men to have a follow-up traffic offence or speeding offence after being detected speeding by an enforcement camera. This is consistent with other research that identifies that male drivers are more likely than female drivers to receive a speeding infringement notice (e.g., Factor, 2018; Lawpoolsri et al., 2007). Consistent with the work of Watson et al. (2014), high range speed offenders were more likely to re-offend within 12 months of receiving a camera detected infringement notice when compared with low range speed offenders. High-range speeding offenders, and other high risk traffic offenders, are also more likely to engage in other types of crime (Bates et al., 2022). Thus, the use of alternative punishments for high-range speed offenders, such as vehicle impoundment, may reduce re-offending rates for both traffic and criminal offences (Watson et al., 2020).

One limitation of this study is the assumption that it was the letter in the experimental condition that increased perceptions of procedural justice, and thus influenced subsequent behaviour. However, this was not explicitly tested. Thus, it is possible that something other than improved perceptions of procedural justice such as increased understanding of the association between speeding, speed cameras, and road safety affected participants subsequent speeding behaviour.

A key strength of this study was the randomised field experiment design. As noted by J. Li et al. (2011), more randomised controlled evaluations are needed to inform traffic safety policies. However, due to the nature of the administrative data collected, it was not possible to include psycho-social variables that have been associated with speeding behaviours such as personality (Cestac et al., 2011; Delhomme et al., 2012; Machin & Sankey, 2008), or attitudes (Hatfield et al., 2008; Newnam et al., 2004). Future research could take this into account by combining survey data with administrative infringement data. Further research could also consider the impact of vicarious experiences of procedural justice on traffic offending as it appears this also has an effect on compliance with the law (Kaiser & Reisig, 2019).

Additionally, camera detected infringement notices are issued to the registered owner of the vehicle in Queensland. If this person was not driving, they can then transfer this infringement to the person who was driving at the time of the infringement. However, we were not able to identify if infringement notices were transferred to other drivers. Therefore, our follow-up administrative data relates to the registered owner of the vehicle at the time of the infringement. If tickets are transferred more frequently from men to women, or from younger to older persons, these differences may affect the results.

This study joins the growing body of procedural justice research that has occurred in western contexts such as the USA (e.g., Weisburd et al., in press), UK (Bradford et al., 2015), Europe (e.g. Sytsma et al., 2021), and Australia (e.g. Madon et al., in press; Zahnow et al., in press). However, there is a need for more research in non-western cultures to identify if the effects of a procedurally just non face-to-face interaction are as effective. There is also a need for more research into road policing within developing communities (Sam, 2022).

Despite these limitations, this study demonstrates that the inclusion of a letter drafted using the principles of procedural justice reduces subsequent speeding behaviour for those who are 25 years or older. Thus, it is possible for police agencies to engage in procedurally just interactions with individuals even when they are not interacting with them face-to-face and that these interactions will increase compliance. At a practical level, police agencies should test and incorporate procedurally just interactions even when technology facilitates their interactions with members of the public.