- Email: [email protected]
Is there an observational effect? An exploratory study into speed cameras and self-reported offending behaviour
Accident Analysis and Prevention 108 (2017) 201–208
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Full length article
Is there an observational effect? An exploratory study into speed cameras and self-reported offending behaviour
MARK
⁎
J. Freeman , S-A. Kaye, V. Truelove, J. Davey Queensland University of Technology (QUT), Centre for Accident Research and Road Safety − Queensland(CARRS-Q), 130 Victoria Park Road, Kelvin Grove, 4059, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: Speed cameras Speeding Exposure Deterrence and age
Fixed and mobile speed cameras are an important element of enforcement initiatives designed to create a strong deterrent effect and improve road safety. Despite the widespread use of the technology and the need to create a strong deterrent effect, research has yet to determine if there is a relationship between levels of exposure to the devices and subsequent self-reported deterrent effects. As a result, licensed motorists (N = 536; 51% female) in Queensland (Australia) were recruited to complete a questionnaire that measured exposure to speed cameras and associated offending behaviours. Data were analyzed utilising descriptive, bivariate and multivariate statistics. The key findings that emerged were: the sample reported a higher level of exposure to fixed cameras (even though there are more operational mobile cameras), younger males were most likely to speed and be observant of speed cameras and that perceived certainty of apprehension was the largest reported deterrent force. However, a positive (rather than negative) relationship was found between perceived camera exposure levels and speeding behaviours, which indicates a range of additional factors (both legal and non-legal factors as well as driving exposure levels) influence speed limit non-compliance. Furthermore, multivariate analysis revealed that higher levels of perceptual certainty were associated with general speed compliance and perceptions of the severity and swiftness of sanctions, rather than levels of self-reported camera exposure. This paper is the first to reveal that while motorists prone to speed may be more cognisant of speed camera operations, this in itself does not ensure appropriate behaviour modification.
1. Speeding behaviour Violating speed limits has been consistently demonstrated to increase crash risk as well as the severity of injuries associated with crashes (Fleiter and Watson, 2006; Petridou and Moustaki, 2000). In essence, it remains one of the largest contributors to the Australian road toll (BITRE, 2016). Despite this, speeding behaviour remains socially acceptable among some subgroups, and as such, a sizeable proportion of drivers still continue to speed (Fleiter and Watson, 2006; Job et al., 2013). For example, the 2013 Australian community attitudes to road safety survey, revealed that while 89% of the 1500 respondents reported that a crash at 80 km/h was more severe than a crash at 70 km/ h, 5%1 reported that they always, nearly always, or mostly drive at 10 km/h over the speed limit and 65% reported that they sometimes or occasionally drive at 10 km/h above the posted speed limit (Petroulias, 2014). Similar findings were reported 10 years earlier in the 2003 Australian community attitudes to road safety survey (Pennay, 2004). These findings highlight the importance of implementing effective ⁎
1
countermeasures to reduce speeding behaviour and reduce the significant road toll associated with such violations. Speed cameras are one such approach that have been increasingly implemented. 2. History of speed camera Speed cameras have been adopted in many counties worldwide in a coordinated effort to identify, apprehend and deter offenders as well as promote general rule compliance. In Australia, for instance, mobile speed cameras were first trialled in Victoria in 1985, and were operational throughout the rest of Australia from the early-to-late 1990s (Delaney et al., 2005a,b; Newstead and Cameron, 2003). In Britain, mobile speed cameras were first operational in 1991 (Delaney et al., 2005a,b) and in New Zealand, mobile speed cameras were introduced in 1993 (Tay, 2000). Since this period of time, the approach has evolved to now incorporate both mobile and fixed approaches. Mobile cameras are implemented by traffic camera operators, either from inside a stationary police van or inside a stationary vehicle which is not
Corresponding author. E-mail addresses: [email protected] (J. Freeman), [email protected] (S.-A. Kaye), [email protected] (V. Truelove), [email protected] (J. Davey). These figures were based on drivers who had driven within in past two years (n = 1365).
http://dx.doi.org/10.1016/j.aap.2017.08.020 Received 10 March 2017; Received in revised form 3 July 2017; Accepted 15 August 2017 Available online 12 September 2017 0001-4575/ © 2017 Elsevier Ltd. All rights reserved.
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
should be noted that these evaluations do not take into account specific biases (such as regression to the mean) and the broader background trend in safety improvements (e.g., road infrastructure), particularly in regards to reductions in fatalities. Research has also yet to determine if fixed or mobile camera operations create the strongest general deterrent effect (briefly outlined below) nor how much exposure is required to create a lasting effect.
identified as a police van, or from a hand-held device on the side of the road. As such, mobile cameras provide an opportunity to randomise and/or tailor police enforcement initiatives to high risk locations (Cameron and Delaney, 2008; Delaney et al., 2005a,b). In contrast, fixed speed cameras are permanently installed at specific locations and were adopted by all Australian States and Territories by late 2000. Fixed cameras can be signed or unsigned, however and due to media attention and location information provided on speeding tickets, the locations for unsigned fixed cameras do not remain concealed for long. Fixed speed cameras are often located on high-risk roads such as blackspots and tunnels; (Queensland Government, 2014) and information on the specific camera sites are displayed on Government websites (e.g., see Queensland Government, 2016) in order to promote a general deterrent effect (see Section 4). Mobile and fixed speeding cameras can be operated in either an overt (e.g., visual cameras) or covert (e.g., hidden cameras) manner (Bates et al., 2012). It is noted that hidden cameras (e.g., overt) can also promote a general deterrent effect (even if unseen) simply by increasing motorists’ perceptions regarding the likelihood of being apprehended (if motorists are aware that overt policing operations are undertaken in the area). In regards to speed camera operations in Queensland, speed camera enforcement operations usually involve a variety of vehicles (e.g., sedan, van and 4WD), overt and covert operations are not usually accompanied by direct signage (although general warnings about speed camera operations may be installed in an area) and portable cameras are usually installed in unmarked vehicles. It should be noted that substantial variances in operations exist between jurisdictions in Australia. In addition to mobile and fixed cameras, point-to-point cameras were introduced in Australia in 2007 and these cameras measure a driver’s average speed between two specific points on the road (Soole et al., 2013). If the driver’s average speed surpasses that of the legal speed limit between the camera’s start and end points, then the information of the vehicle and offence are recorded (Soole et al., 2013). Point-to-point technology is a relatively new concept and only operates at a small number of locations in Australia.2 In the current study, pointto-point enforcement techniques are included in the “fixed camera” aspect of the current study.
4. Promoting speed limit compliance with deterrence Deterrence theory remains the foundation of many road safety countermeasures designed to improve road safety. Deterrence theory consists of two types of processes: general deterrence and specific deterrence. General deterrence proposes that offending behaviours will be reduced if the population consider associated penalties to be certain, severe, and swift (Davey and Freeman, 2011; Freeman and Watson, 2009; Homel, 1988; Taxman and Piquero, 1998). Mass media campaigns warning drivers’ of the penalties associated with illegal behaviour (e.g., speeding behaviour) and visible law enforcement approaches are vital to maximise the effectiveness of general deterrence (Elvik and Christensen, 2007; Taxman and Piquero, 1998; Vingilis and Salutin, 1980). Specific deterrence, however, proposes that individuals who have previously been apprehended for illegal behaviour will avoid further reoffending due to previously experiencing direct punishment associated with their conviction, such as fines or licence loss (Homel, 1988). While deterrence-based initiatives have proven extremely successful in reducing the prevalence of road rule violations, particularly within the drink driving domain (Homel, 1988; Watson et al., 2005), a number of outstanding questions remain regarding methods to enhance deterrent effects in order to maximise rule compliance. In the current case, this endeavour is particularly important given the above reviewed research that indicates: (i) a sizeable proportion of motorists continue to speed and (ii) speeding remains one of the largest (if not the largest) contributor to the road toll. Arguably of most importance is the question of how much exposure to deterrent-based enforcement (e.g., speed cameras) is needed to create a strong deterrent effect. It has long been proposed that drivers need to be constantly exposed to deterrencebased messages in order for a strong deterrent effect to be sustained (Homel, 1988). However, it remains unknown how much exposure to speed cameras actually influences: (a) perceptions of apprehension certainty and (b) subsequent compliance with posted speed limits. This may be considered a significant oversight given the tremendous amount of police resources required to maintain speed camera operations as well as the need to implement targeted and effective deterrence-based strategies to maximise the impact of speed cameras. More broadly, the sizeable body of research that has focused on models of learning and experimental psychology has demonstrated the importance of frequent exposure to stimuli in order to create behavioural change (Nagin and Pogarsky, 2001). However, this knowledge has not been transferred to the road safety domain. As a result, the current study aims to:
3. Effectiveness of speed cameras A number of reviews of speed camera operations have provided support for the effectiveness of the approach to reduce crashes and associated injuries and fatalities (Pilkington and Kinra, 2005; Wilson et al., 2010). For example, Pilkington and Kinra (2005) reviewed 14 studies which examined the effectiveness of fixed, mobile, and a combined of fixed/mobile speed camera. The findings from the review highlighted that speed cameras resulted in crash reductions (5–69%), reductions in injuries (12–65%), and reductions in fatalities (17–71%). In addition, Wilson et al. (2010) reviewed 35 studies which had utilised repeated measure designs to assess the potential effectiveness of mobile and fixed speed cameras. The review found that of the studies which had examined speed outcomes, all had reported a reduction in speed post-camera installation. Further, all studies had reported reductions in all types of crashes, with an average reduction of 11% (500 m from the camera site) and 13% (1 km from the camera site) for more serious crashes (e.g., those resulting in serious injuries or fatalities). In regards to the Queensland context, Newstead and Cameron (2003), examined the effectiveness of mobile speed cameras from May 1997 to June 2001 and reported fatal crash reductions of 45% and a reduction in all crashes by28% within 2 km of the mobile speed camera sites. However it
1. Determine the frequency of a group of urban motorists perceived exposure to fixed and mobile speed cameras; 2. Explore what level of exposure (and what personal characteristics) influences compliance with speed limits; and 3. Examine the effect such exposure has on levels of perceptual certainty of apprehension. 5. Method 5.1. Participants
2 In the State of Queensland, Australia, where the current study was undertaken, there are 3 point-to-point speed camera locations, 16 fixed speed camera locations (excluding combined red light and speed camera locations), and an active deployment of mobile speed cameras (Queensland Government, 2016).
South Eastern Queensland motorists (N = 536) were recruited via an online advertisement and a snowballing technique to take part in this study. The participants (51% female and 49% male) were aged 202
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
5.3. Analysis
between 18 and 73 years (Mage = 32.45, SD = 12.30). The median range participants drove was between 6–10 h per week. On average, participants had been licensed for 14.58 years (SD = 12.10). Participants received a AUS$10 gift voucher as an incentive for study participation.
Frequency and descriptive statistics examined self-reported speeding behaviour,4 perceived camera exposure, and perceived deterrence. To compare the relationships between age, perceptual certainty, self-report speeding behaviour, and perceived camera exposure, bivariate correlations were undertaken. Further, a series of independent group t-tests were implemented to examine the effects of gender and previous speeding offences on self-reported speeding behaviour. A linear regression was then conducted to examine the predictors of perceived certainty of apprehension. Bonferroni adjustments were utilised for all multiple comparisons. Data screening revealed that assumptions of linearity and normality were not extensively breached, and thus, parametric tests were subsequently utilised. Where homogeneity of variance was breached and consistent with recommendations by Field (2009), the equal variances not assumed statistic is interpreted and reported.
5.2. Measures and procedure3 Participants completed either an online version or a pen-paper copy of the questionnaire and corresponding between-group analyses did not reveal any differences on key independent/dependent variables in regards to response type. Participants were required to respond to questions relating to demographics (e.g., age and gender), licence status, speeding behaviour and conviction history. The average of the following two questions was used to measure speeding behaviour: “How often do you exceed speed limits by more than 10 km/h on a highway?” and “How often do you exceed speed limits by more than 10 km/h in a town?” Participants responded to these two questions on a 7-point Likert Scale (1 = Never, 7 = Always), and a higher speeding threshold was utilised (e.g., 10 km/hr) to: (a) ensure there was no ambiguity regarding what actually constituted speeding and (b) previous research has found that a sizable proportion of motorists drive more than 10 km/ hr over the speed limit (Fleiter and Watson, 2006). Furthermore, participants were asked “on average, how many hours of driving would you do per a week?” Driving exposure was measured via a 5-point scale (1 = less than 5 hours per a week, 5 = more than 30 hours per a week). The remaining questions focused on exposure to fixed and mobile speed cameras (i.e., whether participants had seen any speed enforcement operations lately) and deterrence factors (outlined below).
6. Results 6.1. Self-report measure of speeding The majority of participants reported that they ‘never’ or ‘rarely’ exceeded the speed limit by more than 10 km/h on highways or in town settings (see Table 1). However a closer examination revealed that 12.5% of participants reported that they ‘often’, ‘nearly always’ or ‘always’, exceeded the speed limit by more than 10 km/h on highways. A smaller proportion of participants (6.2%) reported that they ‘often’, ‘nearly always’ or ‘always’ exceeded the speed limit by more than 10 km/h in a town. A related samples t-test revealed that participants were significantly more likely to report exceeding the speed limit by 10 km/h on highways compared with towns, (M = 2.53, SD = 1.31 vs. M = 2.04, SD = 1.11), t[538] = 10.16, p < 0.001, 95% CI [0.39, 0.58].
5.2.1. Exposure to fixed and mobile speed cameras Prior to responding to the questions on fixed and mobile speed cameras, participants were provided with a short description of each camera type. The short description for mobile speed cameras was: “Mobile speed cameras include marked and unmarked police vehicles fitted with speed camera equipment parked on the side of the road, and portable devices that are hand-held or set up on a tripod on the side of the road by police officers”. For fixed speed cameras, the description provided to participants was: “Fixed speed cameras are cameras that are permanently installed on roads, tunnels, or intersections”. For each camera type, participants were asked, “In the last 3 months how often have you seen a [fixed/mobile] speed camera?” Participants responded from 1 = not at all, 6 = more than once a day on most days. Higher scores reflect greater camera exposure.
6.2. Self-report exposure to fixed and mobile cameras Table 2 presents the rate of exposure to fixed and mobile speed cameras. Participants were asked to report if they had seen a fixed and mobile speed camera over the past 3 months. Over a third of the sample (41.2%) reported that they had seen a fixed speed camera ‘almost every day’ and ‘more than once a day on most days’ in the past 3 months. A smaller proportion of participants (10.5%) reported seeing a mobile speed camera ‘almost every day’ and ‘more than once a day on most days’ in the past 3 months. A related samples t-test revealed that participants were significantly more likely to report exposure to fixed speed cameras compared with mobile speed cameras (M = 4.17, SD = 1.21 vs. M = 3.23, SD = 1.07), t[514] = 17.05, p < 0.001, 95% CI [0.84, 1.05].
5.2.2. Deterrence A 7-point Likert Scale (1 = strongly disagree, 7 = strongly agree) was used to assess deterrence factors of interest and were modelled on previous psychometrically sound drink driving deterrence items (Freeman et al., 2006; Freeman and Watson, 2009). Five items were used to assess the three Classical Deterrence factors. Two items assessed certainty (i.e., “The chances of getting caught for speeding are high” and “If I were to speed, I’d worry that I would get caught”), two items assessed severity (i.e., “The penalty I would receive for speeding would cause a considerable impact on my life” and “A penalty for speeding would be severe to me”), and one item assessed swiftness (i.e., “The time between getting caught for speeding and receiving a penalty would be very short”). On average, the questionnaire took 20 min to complete.
6.3. Influence of age, gender, and road exposure on self-reported speeding behaviour and perceived camera exposure Pearson correlations were conducted to examine the relationship between perceived camera exposure and: (a) age, (b) exposure to the road (e.g., number of driving hours per a week) and (c) speeding behaviour. As shown in Table 3, there were small significant positive relationships between the number of driving hours, age, exceeding the speed limits by more than 10 km/h on a highway, and perceived camera exposure (both fixed and mobile speed cameras). These finding suggests that as the number of driving hours increases, so too does age, exceeding the posted speed limit by 10 km/h on a highway, and
3 Only those measures relevant to this paper are reported. The larger research project focused more generally on road rule compliance for both speeding and drink driving.
4 Self-reported speeding behaviour included both automated speed enforcement and speed tickets issued by a law enforcement officer.
203
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
exposure to fixed and mobile speed cameras. The effect of exposure was further illuminated through examination of partial correlations, which revealed that deliberate speeding on highways (e.g., more than 10 km/ hr) and fixed camera exposure is essentially halved after consideration of exposure (e.g., r = 1.00* to r = 0.050). Further, there was a small negative relationship between age and exceeding the speed limits by more than 10 km/h in a town. This finding suggests that as age decreases, speeding behaviour increases in built up areas. There were no significant relationships between age and exceeding the speed limits by more than 10 km/h on a highway (see Table 3). There were small positive correlations between speeding by more than 10 km/h on a highway and exposure to fixed speed cameras in the past three months and between speeding more than 10 km/h on a highway and exposure to mobile speed cameras in the past three months. These findings may suggest that as self-reported speeding increases so too does: (a) perceived camera exposure and/or (b) sensitivity to observing speed cameras. However, there were no significant relationships between those who reported exceeding the speed limit by more than 10 km/h in a town and exposure to fixed or mobile speed cameras. Additional Spearman correlations were conducted to assess the relationship between average speeding behaviour (e.g., in both towns and on highways) and exposure to fixed and mobile speed cameras. There were small significant positive correlations between drivers who reported on average exceeding the speed limit by 10 km/h and observing fixed speed cameras, r = 0.091, p = 0.037, and exceeding the speed limit by 10 km/h and observing mobile speed cameras, r = 0.103, p = 0.019. Therefore, as self-report speeding behaviour increases so too does perceived camera exposure.
Table 1 Self-reported Measures of Speeding Behaviour. Frequency
n
%
Frequency
n
%
190 210 93 9 29 3 1 535
35.5 39.3 17.4 1.7 5.4 0.6 0.2
n
%
30 94 198 145 46 9
5.7 18.0 37.9 27.8 8.8 1.7
How often do you exceed the speed limit by more than 10 km/h in a town?
How often do you exceed speed limits by more than 10 km/h on a highway? Never Rarely Sometimes Uncertain Often Nearly always Always Total
101 216 144 9 48 12 7 536
18.8 40.3 26.9 1.5 9.0 2.2 1.3
Never Rarely Sometimes Uncertain Often Nearly always Always Total
Table 2 Self-reported Exposure to Fixed and Mobile Speed Cameras. Frequency
n
In the last 3 months, how often have you seen a fixed speed camera? Not at all Once or twice overall 1–3 times per month Once or twice weekly Almost everyday More than once a day on most days Total
10 38 97 162 141 74 522
%
1.9 7.3 18.6 31.0 27.0 14.2
Frequency In the last 3 months, how often have you seen a mobile camera? Not at all Once or twice overall 1–3 times per month Once or twice weekly Almost everyday More than once a day on most days Total
522
Table 3 The Relationships between Age, Driving Exposure, Speeding Behaviour, and Exposure to Fixed and Mobile Speed Cameras.
1. 2. 3. 4. 5. 6.
Age Driving exposure (hours driver per a week) Speeding highway (more than 10 km/h) Speeding town (more than 10 km/h) Exposure to fixed speed cameras (past 3 months) Exposure to mobile speed cameras (past 3 months)
1
2
3
4
5
6
–
0.243** –
−0.065 0.190** –
−0.135** 0.066 0.572** –
0.115** 0.321** 0.100* 0.055 –
0.089* 0.335** 0.126** 0.056 0.381** –
Note: * p < 0.05; ** p < 0.01.
A series of independent groups t-tests were undertaken to assess the effect of gender on speeding behaviour, number of driving hours per a week, and perceived camera exposure. Compared with females, males were significantly more likely to report exceeding the speed limit by 10 km/h on highways and in towns. Further, males were significantly more likely to report driving more hours per a week and, as a result, observing a fixed and mobile speed cameras within the last three months (see Table 4).
Table 4 The effect of Gender on Self-reported Speeding Behaviour, Number of Driving Hours per a week, and Exposure to Speed Cameras. Variable
n
M (SD)
t
p
Exceeding speed limit by more than 10 km/h on highway Male 264 2.72 (1.42) Female 272 2.35 (1.17) 3.321 0.001 Exceeding speed limit by more than 10 km/h in town Male 263 2.15 (1.16) Female 272 1.94 (1.04) 2.18 0.030 Road exposure (driving h per a week) Male 261 2.51 (1.09) Female 271 1.95 (0.90) 6.411 Exposure to fixed speed camera (past 3 months) Male 258 4.45 (1.12) Female 264 3.89 (1.23) 5.39 Exposure to mobile speed camera (past 3 months) Male 258 3.52 (1.04) Female 264 2.91 (1.03) 6.67
95% CI
0.15, 0.60
0.02, 0.39
< 0.001
0.39, 0.73
> 0.001
0.35, 0.76
> 0.001
0.43, 0.78
6.4. Self-report measures of deterrence Table 5 presents the descriptive statistics for the deterrence variables. The threat of perceptual certainty of apprehension (M = 5.29) was the highest mean, followed by severity and then swiftness of sanctions. To further explore if there were significant differences between the deterrence variables, a one-way repeated measures ANOVA was performed. The mean for certainty was significantly higher than the mean for severity, F[1, 539] = 17.03, p < 0.001, and significantly higher than the mean of swiftness, F[1, 537] = 39.44, p < 0.001. This is in contrast to previous research that has reported perceptual severity to be higher (than certainty) for drug driving (Freeman et al., 2010) as well as for drink driving (Freeman et al., 2006). The mean for severity was significantly higher than the mean for swiftness, F[1, 537]
Note: 1 = Equal variances not assumed statistic is reported here as homogeneity was breached; CI = Confidence Interval.
204
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Table 5 Self-reported Measures of Deterrence. Deterrence
Mean (SD)
Classical deterrence Certainty 5.29 (1.19) Severity 5.06 (1.43) Swiftness 4.85 (1.52)
Strongly disagree
Disagree
Somewhat disagree
Neither agree nor disagree
Somewhat agree
Agree
Strongly agree
1 (0.2%) 11 (2.0%) 9 (1.7%)
22 (4.1%) 31 (5.8%) 39 (7.3%)
36 (6.7%) 53 (9.9%) 59 (11.1%)
62 (18.3%) 101 (18.9%) 98 (18.4%)
154 (28.7%) 137 (25.6%) 100 (18.8%)
177 (33.1) 135 (25.2%) 167 (31.3%)
48 (9.0%) 67 (12.5%) 61 (11.4%)
In terms of road exposure, there was a significant difference between those who had received a speeding ticket (M = 2.45, SD = 1.01) compared with those who had not received a speeding ticket (M = 1.84, SD = 0.97) on number of driving hours per a week, t[529] = 6.90, p < 0.001, 95% CI [0.44, 0.79], suggesting that drivers who received a speeding ticket reported higher on-road exposure. In terms of deterrence, there was no significant difference between those who had received a speeding ticket (M = 5.36, SD = 1.25) compared with those who had never received a speeding ticket (M = 5.24, SD = 1.16) on perceived certainty, t[538] = −1.15, p = 0.251, 95% CI [-0.33, 0.09]. However, participants who had reported receiving a speeding ticket were significantly more likely to report that a penalty would be severe (M = 4.88, SD = 1.49) compared with those participants who had not received a speeding ticket (M = 5.34, SD = 1.28), t (484.62)5 = −3.81, p < 0.001, 95% CI [−0.70, −0.22]. The following analyses only included those individuals who had reported receiving a speeding ticket. Similar to above, those who had reported receiving a speed ticket within the last three years (M = 2.92, SD = 1.38) were significantly more likely to report exceeding the speed limit by more than 10 km/h on a highway compared with those who had received a speeding ticket more than three years ago (M = 2.48, SD = 1.19), t[339] = 3.13, p = 0.002, 95% CI [0.16, 0.71]. Further, those who had received a speeding ticket within the past three years (M = 2.35, SD = 1.17) were significantly more likely to report exceeding the speed limit by more than 10 km/h in a town compared with those who had received a speeding ticket more than three years ago (M = 1.96, SD = 0.99), t[338] = 3.32, p = 0.001, 95% CI [0.16, 0.63]. In terms of road exposure, there was no significant difference between those who had received a speeding ticket within the past three years (M = 2.54, SD = 0.96) and those who had received a speeding ticket over three years ago (M = 2.35, SD = 1.05) on number of driving hours per a week, t[335] = 1.72, p = 0.087, 95% CI [-0.03, 0.40]. In terms of deterrence, there were no significant difference between those who had received a speeding ticket within the past three years (M = 5.24, SD = 1.10) than those who had received a speeding ticket over three years ago on perceived levels of certainty (M = 5.29, SD = 1.20), t[339] = −0.40, p = 0.686, 95% CI [−0.29, 0.19] nor were there any significant difference between those who had received a speeding ticket within the past three years (M = 4.85, SD = 1.57) and those who had received a speeding ticket over three years ago on severity (M = 4.97, SD = 1.40), t[338] = −0.76, p = 0.450, 95% CI [-0.44, 0.20].
Table 6 The Relationships between Age, Driving Exposure, and Deterrence Factors.
1. Age 2. Driving exposure (hours driven per a week) 3. Perceived Certainty 4. Severity 5. Swiftness
1
2
3
4
5
–
0.241** –
0.040 0.038
0.009 0.063
0.130** 0.095*
–
0.498** –
0.309** 0.174** –
Note: *p < 0.05; ** p < 0.01.
= 10.94, p = 0.015. Spearman’s correlations were conducted to assess the relationships between age, exposure to road (number of driving hours per a week), and the deterrence variables (see Table 6). The findings revealed a significant positive relationship between age and swiftness indicating that as age increases so too do perceptions of swiftness of sanctions. There were significant positive relationships between exposure to the road and perceived swiftness, suggesting that as the number of driving hours increase so too does one’s perceptions of swiftness of sanctions. A series of independent groups t-tests were undertaken to examine the effect of gender on self-reported deterrence. As presented in Table 7, males were significantly more likely to report higher levels of swiftness of apprehension, although there were no other significant gender effects on self-reported deterrence.
6.5. Effect of demerit points A series of independent groups t-tests were also undertaken to examine the effect of ever receiving a speeding ticket on current driving behaviour, road exposure, and perceptual deterrence and severity. The findings revealed that those who reported receiving a speeding ticket (M = 2.73, SD = 1.30) were significantly more likely to report exceeding the speed limit by more than 10 km/h on a highway compared with those who had never received a speeding ticket (M = 2.20, SD = 1.27), t[533] = 4.56, p < 0.001, 95% CI [0.31, 0.75]. Similarly, compared with those who had never received a speeding ticket (M = 1.84, SD = 1.11), those who reported receiving a speeding ticket were significantly more likely to report exceeding the speed limit by more than 10 km/h in a town (M = 2.17, SD = 1.08), t[532] = 3.41, p < 0.001, 95% CI [0.14, 0.52]. Table 7 The effect of Gender on Deterrence. Variable
n
Classical Deterrence Perceived certainty Male 264 Female 277 Severity Male 264 Female 276 Swiftness Male 263 Female 275
M (SD)
6.6. Effect of demerit points on perceived camera exposure (controlling for road exposure) t
p
95% CI
5.19 (1.24) 5.38 (1.15)
−1.89
0.060
−0.36, 0.01
5.02 (1.41) 5.09 (1.45)
−0.61
0.541
−0.32, 0.17
5.00 (1.53) 4.71 (1.49)
2.21
0.028
0.03, 0.54
A Multivariate Analysis of Covariance (MANCOVA) was undertaken to examine the effect of ever receiving a speeding ticket (Independent Variable) on perceived camera exposure (mobile and fixed cameras) (Dependent Variables) after controlling for road exposure (number of driving hours per a week) (Covariate). The findings revealed that there was no significant effect of ever receiving a speeding ticket on perceived speed camera exposure when controlling for number of driving 5
205
Equal variances not assumed statistic is reported here as homogeneity was breached.
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Table 8 Beta Weights with their Significance, Unstandardised Co-efficients (B) and their Confidence Intervals and the Squared Semi-Partial Correlations of the Predictor Variables for Perceptual Certainty. Variables
B
SE
β
p
95% CI for B Lower
Age Gender Road exposure Self-reported speeding Exposure to fixed cameras Exposure to mobile cameras Severity Swiftness
−0.001 0.166 −0.018 −0.214 0.035 0.047 0.382 0.180
0.004 0.091 0.047 0.042 0.039 0.046 0.031 0.029
−0.009 0.069 −0.015 −0.194 0.036 0.042 0.457 0.227
0.809 0.068 0.700 > 0.001 0.367 0.302 > 0.001 > 0.001
−0.01 −0.01 −0.11 −0.30 −0.04 −0.04 0.32 0.12
Upper
sr 2
0.01 0.35 0.07 −0.13 0.11 0.14 0.44 0.24
−0.011 0.082 −0.017 −0.225 0.042 0.046 0.486 0.272
Note: CI = Confidence Intervals, sr2= semi partial correlation squared.
hours per a week, F[2, 503] = 1.36, p = 0.258, Wilk’s ʌ = 0.995, ηp2 = 0.005.
compared with the covert operations which are often employed with mobile speed cameras, and thus, motorists may not be aware of a mobile camera’s presence. Finally, many of these fixed cameras are located in high-traffic density areas (e.g., highways and busy intersection) and operate 24 h per a day (compared with several hours for mobile cameras). Taken together, it is not surprising that drivers were more likely to report greater exposure towards the fixed speed cameras over the past three months compared with the mobile speed cameras. What was also expected was that greater exposure to the road (e.g., more kms driven) would be associated with higher exposure to speed cameras, which was confirmed. While this may be considered of little significance, it is encouraging that drivers who probably were exposed to greater levels of speed cameras actually reported this outcome, which is a foundational component of deterrence principles e.g., drivers need to be constantly exposed to deterrent factors (Homel, 1988). This also strengthens the veracity of the current self-report data, which is mentioned in a proceeding section. In regards to the second aim, motorists who reported speeding more often actually reported a higher frequency of exposure to speed cameras, which was further supported by the positive association between younger males and perceived camera exposure. This finding is consistent with scant research in the area (Wundersitz et al., 2002), and suggests that drivers who consciously exceed the speed limit may be more aware (or sensitive) to the threat of speed enforcement approaches. Although, it should also be noted that the relationship between speeding and camera exposure was inflated by levels of driving. This finding needs to be explored through further research, particularly in regards to the origins of the lack of a deterrent effect, although it consistent with Australian research that has indicated higher levels of perceived certainty is actually associated with an increase in self-reported speeding frequency (Fleiter and Watson, 2006; Fleiter et al., 2009). In regards to perceptual deterrent factors, perceptions of apprehension certainty were found to be higher than severity or swiftness. This finding is in contrast to previous deterrent-based research on other offending behaviours (Freeman et al., 2006, 2010), but may be reflective of the increased likelihood of apprehension for speeding versus other rule violations e.g., drink driving and drug driving. The result is nonetheless encouraging as scholars have historically postulated that the most powerful deterrent effects on offending behaviour are produced by the perceived threat of the certainty of apprehension (Homel, 1988; Nagin and Pogarsky, 2001; Von Hirsch et al., 1999). A complementary analysis was undertaken to determine the link between age and perceptual deterrence, which revealed younger drivers reported lower levels of perceptions regarding the threat of sanctions. Similar to above, scant research has considered this issue, although preliminary evidence exists that suggests older drivers are more likely to respond appropriately to deterrent threats (Bushway et al., 2013; Sampson and Cohen, 1988). This may be associated with
6.7. Predictors of perceptual certainty Finally, a linear regression analysis was undertaken to examine the contribution of age, gender, exposure to the road, self-reported speeding behaviour, perceived camera exposure (both fixed and mobile), and deterrence factors (i.e., severity, swiftness) in predicting higher levels of perceptual certainty of apprehension for speeding behaviour. These factors accounted for 40.4% of the variance in perceptual certainty, F[8, 503] = 39.16, p < 0.001. Table 8 shows that speeding behaviour, severity, and swiftness were all unique significant predictors of perceptual certainty. 7. Discussion This study aimed to conduct one of the first exploratory investigations into the deterrent effect of observing fixed and mobile speed cameras on self-reported speeding behaviours. More specifically, the study aimed to: (a) examine the frequency of urban motorists’ exposure to fixed and mobile speed cameras, (b) investigate what level of exposure (and what personal characteristics) influences compliance with speed limits and (c) explore the differential effect such exposure has on levels of perceptual certainty of apprehension. In regards to self-reported speeding behaviours, the sample were more likely to report speeding on highways compared to towns, which is consistent with scant research in this area (Fleiter and Watson, 2006). While there is limited published research on the effects of road infrastructure (Fleiter, 2010), it may be suggested that speeding on highways is linked to: (a) increased traffic flow and associated opportunities to speed e.g., lower congestion and traffic lights (Wundersitz et al., 2002) and (b) greater opportunities to observe speed cameras in advance and engage in corrective behaviours. What was more predictable in the current sample was that males and younger drivers were at an increased likelihood to report speeding compared with females as well as older drivers (e.g., Fleiter and Watson, 2006; Shinar et al., 2001; Watson et al., 2009; Wundersitz et al., 2002). In terms of the first aim, the findings revealed that drivers reported greater exposure to fixed speed cameras compared with mobile speed cameras within the last three months. While there are no complementary studies for comparison, this result may be linked to motorists’ driving patterns being somewhat routine (or habituated) such as driving for the purposes of going to and from work. Additionally, while there are over 3500 active mobile camera sites in Queensland (Queensland Police, 2015) compared with 16 fixed speed camera sites (Queensland Government, 2016), all the fixed speed cameras are located in South-East Queensland, where the study was undertaken. In addition, fixed speed cameras tend to involve more overt operations 206
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
emerging findings that indicate younger individuals experience difficulties recognizing (and responding appropriately) to risk (Albert and Steinberg, 2011), which may in part be due to the underdeveloped prefrontal cortex (Lebel and Beaulieu, 2011). Consistent with the positive association between self-reported speeding and perceived camera exposure (outlined above), those who had incurred a speeding ticket were also more likely to report: (a) engaging in further speeding behaviours and (b) greater exposure to the road. While the latter finding is intuitive in nature, the former is consistent with research that indicates applied speeding sanctions are not sufficient to create a deterrent effect (Fleiter, 2010) and/or the apprehension process creates a “resetting effect” e.g., an individual believes that the chances of being re-apprehended is much lower according to the belief that their initial experience of getting caught was inevitable. More broadly, previous research has demonstrated that drivers who are apprehended for speeding have a greater chance of receiving further tickets, (e.g., Lawpoolsri et al., 2007), which is supportive of the assumption that past behaviour is an efficient predictor of future behaviour (Danner et al., 2008). Overall, the research suggests that speed tickets in isolation may not deter drivers from speeding (Fleiter, 2010) and that additional countermeasures (including non-legal sanctions) are required to enhance drivers’ compliance with the road rules. Similarly, there was no relationship between receiving a speeding ticket and perceptual certainty, which is again consistent with previous counterintuitive findings that indicate higher levels of perceived certainty and severity is actually associated with an increase in self-reported speeding frequency (Fleiter and Watson, 2006; Fleiter et al., 2009). Furthermore, no relationship was found between perceived camera exposure and receiving a speeding ticket, which might be expected given that a large proportion of the sample reported abiding by the speed limits. Finally, a positive relationship was found between receiving a speeding ticket and perceived severity, which is evidence of the specific deterrent effect of speeding sanctions. Finally, a linear regression analysis to determine the predictors of higher levels of self-reported perceptual certainty revealed it was linked with: (a) general speed compliance and (b) higher sanction severity and swiftness perceptions. The finding is consistent with research that has demonstrated those who comply with rules report higher levels of apprehension certainty (Homel, 1988; Paternoster and Piquero, 1995; Piquero and Pogarsky, 2002) and that a positive relationship can exist between the three constructs of the classical deterrence model (Piquero et al., 2011). That is, sanctions need to not only be certain, but severe (Homel 1988). In contrast, the analysis confirmed that a direct (and substantial) relationship was not evident between perceived camera exposure and perceptual certainty for the current sample. There were a number of limitations with this research that should be considered when interpreting the results. For example, due to the crosssectional design of this study, the current research methodology does not permit an examination of possible causal effects, nor was there any comparison with actual speed data e.g., in-vehicle measurement devices. Additionally, self-report measures were used that may be susceptible to participant biases (e.g., social desirability) (Wåhlberg et al., 2010) as well as inaccuracies with memory recall (McCarroll, 2017). As a result, future research should employee more objective measures of speeding behaviour, such as offence data or in-vehicle monitoring devices. Nevertheless, previous research has reported significant moderate positive correlations exist between objective and subjective measures of speeding behaviour (e.g., Åberg and Wallén Warner, 2008; Helman and Reed, 2015; Kaye et al., 2016), suggesting that self-report measures are a suitable measure of actual speeding behaviour. The sample consisted of a higher proportion of younger compared with older drivers. The majority of the sample did not regularly engage in speeding behaviours, which limits variance and the ability to identify trends in the data. The deterrent threat of a sanction (particularly the severity of a sanction) may be dependent upon a motorist’s demerit point accumulation at the time, and thus, future research would benefit
from considering this issue. Future research needs to disentangle the difference between “sensitivity to cameras” and actual “exposure to cameras”, as there may be natural variations in these concepts among the motoring population. Further attempts to extract direct effects also need to be directed towards the relationship between speeding, observing cameras and driving exposure, as the latter factor (e.g., amount of driving) may influence both the frequency of speeding and subsequent camera observations. Finally, the study focused on the Queensland speed camera program and, therefore, caution should be used when interpreting these findings outside of Queensland, Australia. Future research is required to not only extend upon the current study’s aims but also include a more representative sample of the driving population, which may be considered more random in nature. 8. Conclusion This study was one of the first to explore the deterrent effect of observing fixed and mobile speed cameras on self-reported speeding behaviours. Taken together, the investigation did not reveal clear linear links between exposure to cameras and corresponding behaviours, although this may be in part because of the above mentioned limitations. Despite this, the study did illuminate key preliminary findings associated with: (a) age, gender and deterrability, (b) those most likely subgroup to observe speed cameras, (c) the self-reported effect of speeding sanctions and (d) the factors that promote perceptual certainty. Given that deterrence-based mechanisms will continue to remain a cornerstone of road safety initiatives into the foreseeable future, it is warranted to conduct further research to determine the general deterrent and specific deterrent effect of speed cameras on overall rule compliance. Acknowledgement This research was funded by the Australian Research Council Discovery Scheme. References Åberg, L., Wallén Warner, H., 2008. Speeding–deliberate violation or involuntary mistake? revue europÅenne de psychologie appliquÅe/. Eur. Rev. Appl. Psychol. 58 (1), 23–30. http://dx.doi.org/10.1016/j.erap.2005.09.014. Albert, D., Steinberg, L., 2011. Judgement and decision making in adolescence. J. Res. Adolesc. 21 (1), 211–224. http://dx.doi.org/10.1111/j.1532-7795.2010.00724.x. Bates, L., Soole, D., Watson, B., 2012. The Effectiveness of Traffic Policing in Reducing Traffic Crashes. Palgrave MacMillian, New York,U.S. Bushway, S., DeAngelo, G., Hansen, B., 2013. Deterability by age. Int. Rev. Law Econ. 36, 70–81. http://dx.doi.org/10.1016/j.irle.2013.04.006. Bureau of Infrastructure, Transport, and Regional Economics [BITRE], 2016. Road Deaths Australia. Retrieved from https://bitre.gov.au/publications/ongoing/rda/files/RDA_ Dec_2015.pdf. Cameron, M.H., Delaney, A.K., 2008. Speed enforcement ? effects, mechanisms, intensity and economic benefits of each mode of operation. In: Paper Presented at the Joint Australiasian College of Road Safety and Queensland Parliamentary Travelsafe Committee Conference: High Risk Road Users – Motivating Behaviour Change: What Works and What Doesn't Work? Queensland, Australia. Danner, U., Aarts, H., de Vries, N., 2008. Habit vs intention in the prediction of future behaviour. The role of frequency, context stability and mental acessibility of past behaviour. Br. J. Soc. Psychol. 47 (2), 245–265. Davey, J.D., Freeman, J.E., 2011. Improving road safety through deterrence-based initaitives: a review of research. Sultan Qaboos Univ. Med. J. 11 (1), 312–320. Delaney, A., Ward, H., Cameron, M., 2005a. The History and Development of Speed Cameras. Retrieved from Melbourne, Australia: https://www.monash.edu/muarc/ research/our-publications/muarc242. Delaney, A., Ward, H., Cameron, M., Williams, A.F., 2005b. Controversies and speed cameras: lessons learnt internationally. J. Public Health Policy 26 (4), 404–415. Retrieved from: http://www.jstor.org/stable/4125166. Elvik, R., Christensen, P., 2007. The deterrent effect of increasing fixed penalties for traffic offences: the Norwegian experience. J. Safety Res. 38 (6), 689–695 1016/ j.jsr.2007.09.007. Field, A., 2009. Discovering Statistics Using SPSS, 3rd. ed. SAGE, London. Fleiter, J., Watson, B., 2006. The speed paradox: the misalignment between driver attitudes and speeding behaviour. J. Aust. Coll. Road Saf. 17 (2), 23–30. Fleiter, J., Watson, B., Lennon, A., King, M., Shi, K., 2009. Speeding in Australia and China: a comparison of the influence of legal sanctions and enforcement practices on
207
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Retrieved from: https://infrastructure.gov.au/roads/safety/publications/2013/pdf/ community_att_13.pdf. Pilkington, P., Kinra, S., 2005. Effectiveness of speed cameras in preventing road traffic collisions and related casualties: systematic review. BMJ: Br. Med. J. 330 (7487), 331–334. Retrieved from: http://www.jstor.org/stable/25458895. Piquero, A.R., Pogarsky, G., 2002. Beyond Stafford and Warr’s reconceptualization of deterrence: personal and vicarious experiences, impulsivity, and offending behavior. J. Res. Crime Delinquency 39 (2), 153–186. Piquero, A., Paternoster, R., Pogarsky, G., Loughran, T., 2011. Elaborating the individual difference component in deterrence theory. Ann. Rev. Law Soc. Sci. 7, 335–360. Queensland Government, 2014. Speed and Red Light Cameras. Retrieved from: http:// www.qld.gov.au/transport/safety/fines/speed/cameras/#fixed. Queensland Government, 2016. Location of Speed and Red Light Cameras. Retrieved from: https://www.qld.gov.au/transport/safety/fines/speed/location/. Queensland Police, 2015. Active Mobile Speed Camera Sites. Retrieved from: https:// www.police.qld.gov.au/programs/roadSafety/activespeedcamerasites.htm. Sampson, R.J., Cohen, J., 1988. Deterrent effects of the police on crime: a replication and theoretical extension. Law Soc. Rev. 22 (1), 163–189. Shinar, D., Schechtman, E., Compton, R., 2001. Self-reports of safe driving behaviors in relationship to sex, age, education and income in the US adult driving population. Accid. Anal. Prev. 33 (1), 111–116. http://dx.doi.org/10.1016/S0001-4575(00) 00021-X. Soole, D.W., Watson, B.C., Fleiter, J.J., 2013. Effects of average speed enforcement on speed compliance and crashes: a review of the literature. Accid. Anal. Prev. 54, 46–56. http://dx.doi.org/10.1016/j.aap.2013.01.018. Taxman, F.S., Piquero, A., 1998. On preventing drunk driving recidivism: an examination of rehabilitation and punishment approaches. J. Criminal Justice 26 (2), 129–143. http://dx.doi.org/10.1016/S0047-2352(97)00075-5. Tay, R., 2000. Do speed cameras improve road safety? Traffic Transp. Stud. 44–51. Vingilis, E., Salutin, L., 1980. A prevention programme for drinking driving. Accid. Anal. Prev. 12 (4), 267–274. http://dx.doi.org/10.1016/0001-4575(80)90003-2. Von Hirsch, A.V., Bottoms, A.E., Burney, E., Wikstrom, P., 1999. Criminal Deterrence and Sentence Severity: an Analysis of Recent Research. Hart Publishing, Oregon. Wåhlberg, A.E., Dorn, L., Kline, T., 2010. The effect of social desirability on self reported and recorded road traffic accidents. Transp. Res. F Traffic Psychol. Behav. 13 (2), 106–114. http://dx.doi.org/10.1016/j.trf.2009.11.004. Watson, B., Hart, S., Freeman, J., Tay, R., Davey, J., Madden, J., Goff, A., 2005. Review of Random Breath Testing (RBT) in Queensland. Queensland University Technology. Watson, B., Watson, A., Siskind, C., Fleiter, J., 2009. Characteristics and predictors of high-range speeding offenders. In: Paper Presented at the Australasian Road Safety Research, Policing and Education Conference. Sydney, Australia. Wilson, C., Willis, C., Hendrikz, J.K., Le Brocque, R., Bellamy, N., 2010. Speed cameras for the prevention of road traffic injuries and deaths. Cochrane Database Syst. Rev. 10. http://dx.doi.org/10.1002/14651858.CD004607.pub3. Wundersitz, L., Kloeden, C., McColl, R., Baldock, M., McLean, J., 2002. Speeding in metropolitan Adelaide: a situation analysis. In: Paper Presented at the Road Safety Research, Policing and Education Conference. Adelaide.
car drivers. In: Proceedings of the Australasian Road Safety Research, Policing and Education. Sydney. Fleiter, J., 2010. Examining Psychosocial Influences on Speeding in Australian and Chinese Contexts: a Social Learning Approach. Queensland University of Technology PhD thesis. Freeman, J., Watson, B., 2009. Drink driving deterrents and self-reported offending behaviours among a sample of Queensland motorists. J. Safety Res. 40 (2), 113–120. http://dx.doi.org/10.1016/j.jsr.2008.12.009. Freeman, J., Liossis, P., Schonfeld, C., Sheehan, M., Siskind, V., Watson, B., 2006. The self-reported impact of legal and non-legal sanctions on a group of recidivist drink drivers. Transp. Res. F: Traffic Psychol. Behav. 9 (10), 53–64. Helman, S., Reed, N., 2015. Validation of the driver behaviour questionnaire using behaviour data from an instrumented vehicle and high-fidlity driving simulator. Accid. Anal. Prev. 75, 245–251. http://dx.doi.org/10.1016/j.aap.2014.12.008. Homel, R.J., 1988. Policing and Punishing the Drinking Driver: A Study of Specfic and General Deterrence. Springer-Verlag, New York. Job, S., Sakashita, C., Mooren, L., Grezbieta, R., 2013. Community perceptions and beliefs regarding low-level speeding and suggested solutions. In: Paper Presented at the 2013 Annual Meeting of the Transportation Research Board. Washington, U.S.. Kaye, S.-A., Lewis, I., Algie, J., White, M.J., 2016. Young drivers' responses to antispeeding advertisements: comparison of self-report and objective measures of persuasive processing and outcomes. Traffic Inj. Prev. 17 (4), 352–358. http://dx.doi. org/10.1080/15389588.2015.1084419. Lawpoolsri, S., Li, J., Braver, E.R., 2007. Do speeding tickets reduce the likelihood of receiving subsequent speeding tickets? a longitudinal study of speeding violators in Maryland. Traffic Inj. Prev. 8 (1), 26–34. http://dx.doi.org/10.1080/ 15389580601009764. Lebel, C., Beaulieu, C., 2011. Longitudinal development of human brain wiring continues from childhood into adulthood. J. Neurosci. 31 (30), 10937–10947. http://dx.doi. org/10.1523/JNEUROSCI.5302-10.2011. McCarroll, C., 2017. Looking the past in the eye: distortion in memory and the costs and benefits of recalling from an observer perspective. Conscious. Cogn. 49, 322–332. Nagin, S., Pogarsky, G., 2001. Integrating celerity, impulsivity: and extralegal sanction threats into a model of general deterrence: theory and evidence. Criminology 39, 865–889. Newstead, S., Cameron, M., 2003. Evaluation of the Crash Effects of the Queensland Speed Camera Program. Retrieved from Monash University Accident Research Centre: http://www.monash.edu/__data/assets/pdf_file/0008/216368/muarc204. pdf. Paternoster, R., Piquero, A., 1995. Reconceptualizing deterrence: an empirical test of personal and vicarious experiences. J. Res. Crime Delinquency 32 (3), 251–286. Pennay, D., 2004. Community Attitudes to Road Safety: Community Attitudes Survey Wave 16. Retrieved from: https://infrastructure.gov.au/roads/safety/publications/ 2004/pdf/Comm_Att_16.pdf. Petridou, E., Moustaki, M., 2000. Human factors in the causation of road traffic crashes. Eur. J. Epidemiol. 16 (9), 819–826. http://dx.doi.org/10.1023/a:1007649804201. Petroulias, T., 2014. Community Attitudes to Road Safety – 2013 Survey Report.
208
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Full length article
Is there an observational effect? An exploratory study into speed cameras and self-reported offending behaviour
MARK
⁎
J. Freeman , S-A. Kaye, V. Truelove, J. Davey Queensland University of Technology (QUT), Centre for Accident Research and Road Safety − Queensland(CARRS-Q), 130 Victoria Park Road, Kelvin Grove, 4059, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: Speed cameras Speeding Exposure Deterrence and age
Fixed and mobile speed cameras are an important element of enforcement initiatives designed to create a strong deterrent effect and improve road safety. Despite the widespread use of the technology and the need to create a strong deterrent effect, research has yet to determine if there is a relationship between levels of exposure to the devices and subsequent self-reported deterrent effects. As a result, licensed motorists (N = 536; 51% female) in Queensland (Australia) were recruited to complete a questionnaire that measured exposure to speed cameras and associated offending behaviours. Data were analyzed utilising descriptive, bivariate and multivariate statistics. The key findings that emerged were: the sample reported a higher level of exposure to fixed cameras (even though there are more operational mobile cameras), younger males were most likely to speed and be observant of speed cameras and that perceived certainty of apprehension was the largest reported deterrent force. However, a positive (rather than negative) relationship was found between perceived camera exposure levels and speeding behaviours, which indicates a range of additional factors (both legal and non-legal factors as well as driving exposure levels) influence speed limit non-compliance. Furthermore, multivariate analysis revealed that higher levels of perceptual certainty were associated with general speed compliance and perceptions of the severity and swiftness of sanctions, rather than levels of self-reported camera exposure. This paper is the first to reveal that while motorists prone to speed may be more cognisant of speed camera operations, this in itself does not ensure appropriate behaviour modification.
1. Speeding behaviour Violating speed limits has been consistently demonstrated to increase crash risk as well as the severity of injuries associated with crashes (Fleiter and Watson, 2006; Petridou and Moustaki, 2000). In essence, it remains one of the largest contributors to the Australian road toll (BITRE, 2016). Despite this, speeding behaviour remains socially acceptable among some subgroups, and as such, a sizeable proportion of drivers still continue to speed (Fleiter and Watson, 2006; Job et al., 2013). For example, the 2013 Australian community attitudes to road safety survey, revealed that while 89% of the 1500 respondents reported that a crash at 80 km/h was more severe than a crash at 70 km/ h, 5%1 reported that they always, nearly always, or mostly drive at 10 km/h over the speed limit and 65% reported that they sometimes or occasionally drive at 10 km/h above the posted speed limit (Petroulias, 2014). Similar findings were reported 10 years earlier in the 2003 Australian community attitudes to road safety survey (Pennay, 2004). These findings highlight the importance of implementing effective ⁎
1
countermeasures to reduce speeding behaviour and reduce the significant road toll associated with such violations. Speed cameras are one such approach that have been increasingly implemented. 2. History of speed camera Speed cameras have been adopted in many counties worldwide in a coordinated effort to identify, apprehend and deter offenders as well as promote general rule compliance. In Australia, for instance, mobile speed cameras were first trialled in Victoria in 1985, and were operational throughout the rest of Australia from the early-to-late 1990s (Delaney et al., 2005a,b; Newstead and Cameron, 2003). In Britain, mobile speed cameras were first operational in 1991 (Delaney et al., 2005a,b) and in New Zealand, mobile speed cameras were introduced in 1993 (Tay, 2000). Since this period of time, the approach has evolved to now incorporate both mobile and fixed approaches. Mobile cameras are implemented by traffic camera operators, either from inside a stationary police van or inside a stationary vehicle which is not
Corresponding author. E-mail addresses: [email protected] (J. Freeman), [email protected] (S.-A. Kaye), [email protected] (V. Truelove), [email protected] (J. Davey). These figures were based on drivers who had driven within in past two years (n = 1365).
http://dx.doi.org/10.1016/j.aap.2017.08.020 Received 10 March 2017; Received in revised form 3 July 2017; Accepted 15 August 2017 Available online 12 September 2017 0001-4575/ © 2017 Elsevier Ltd. All rights reserved.
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
should be noted that these evaluations do not take into account specific biases (such as regression to the mean) and the broader background trend in safety improvements (e.g., road infrastructure), particularly in regards to reductions in fatalities. Research has also yet to determine if fixed or mobile camera operations create the strongest general deterrent effect (briefly outlined below) nor how much exposure is required to create a lasting effect.
identified as a police van, or from a hand-held device on the side of the road. As such, mobile cameras provide an opportunity to randomise and/or tailor police enforcement initiatives to high risk locations (Cameron and Delaney, 2008; Delaney et al., 2005a,b). In contrast, fixed speed cameras are permanently installed at specific locations and were adopted by all Australian States and Territories by late 2000. Fixed cameras can be signed or unsigned, however and due to media attention and location information provided on speeding tickets, the locations for unsigned fixed cameras do not remain concealed for long. Fixed speed cameras are often located on high-risk roads such as blackspots and tunnels; (Queensland Government, 2014) and information on the specific camera sites are displayed on Government websites (e.g., see Queensland Government, 2016) in order to promote a general deterrent effect (see Section 4). Mobile and fixed speeding cameras can be operated in either an overt (e.g., visual cameras) or covert (e.g., hidden cameras) manner (Bates et al., 2012). It is noted that hidden cameras (e.g., overt) can also promote a general deterrent effect (even if unseen) simply by increasing motorists’ perceptions regarding the likelihood of being apprehended (if motorists are aware that overt policing operations are undertaken in the area). In regards to speed camera operations in Queensland, speed camera enforcement operations usually involve a variety of vehicles (e.g., sedan, van and 4WD), overt and covert operations are not usually accompanied by direct signage (although general warnings about speed camera operations may be installed in an area) and portable cameras are usually installed in unmarked vehicles. It should be noted that substantial variances in operations exist between jurisdictions in Australia. In addition to mobile and fixed cameras, point-to-point cameras were introduced in Australia in 2007 and these cameras measure a driver’s average speed between two specific points on the road (Soole et al., 2013). If the driver’s average speed surpasses that of the legal speed limit between the camera’s start and end points, then the information of the vehicle and offence are recorded (Soole et al., 2013). Point-to-point technology is a relatively new concept and only operates at a small number of locations in Australia.2 In the current study, pointto-point enforcement techniques are included in the “fixed camera” aspect of the current study.
4. Promoting speed limit compliance with deterrence Deterrence theory remains the foundation of many road safety countermeasures designed to improve road safety. Deterrence theory consists of two types of processes: general deterrence and specific deterrence. General deterrence proposes that offending behaviours will be reduced if the population consider associated penalties to be certain, severe, and swift (Davey and Freeman, 2011; Freeman and Watson, 2009; Homel, 1988; Taxman and Piquero, 1998). Mass media campaigns warning drivers’ of the penalties associated with illegal behaviour (e.g., speeding behaviour) and visible law enforcement approaches are vital to maximise the effectiveness of general deterrence (Elvik and Christensen, 2007; Taxman and Piquero, 1998; Vingilis and Salutin, 1980). Specific deterrence, however, proposes that individuals who have previously been apprehended for illegal behaviour will avoid further reoffending due to previously experiencing direct punishment associated with their conviction, such as fines or licence loss (Homel, 1988). While deterrence-based initiatives have proven extremely successful in reducing the prevalence of road rule violations, particularly within the drink driving domain (Homel, 1988; Watson et al., 2005), a number of outstanding questions remain regarding methods to enhance deterrent effects in order to maximise rule compliance. In the current case, this endeavour is particularly important given the above reviewed research that indicates: (i) a sizeable proportion of motorists continue to speed and (ii) speeding remains one of the largest (if not the largest) contributor to the road toll. Arguably of most importance is the question of how much exposure to deterrent-based enforcement (e.g., speed cameras) is needed to create a strong deterrent effect. It has long been proposed that drivers need to be constantly exposed to deterrencebased messages in order for a strong deterrent effect to be sustained (Homel, 1988). However, it remains unknown how much exposure to speed cameras actually influences: (a) perceptions of apprehension certainty and (b) subsequent compliance with posted speed limits. This may be considered a significant oversight given the tremendous amount of police resources required to maintain speed camera operations as well as the need to implement targeted and effective deterrence-based strategies to maximise the impact of speed cameras. More broadly, the sizeable body of research that has focused on models of learning and experimental psychology has demonstrated the importance of frequent exposure to stimuli in order to create behavioural change (Nagin and Pogarsky, 2001). However, this knowledge has not been transferred to the road safety domain. As a result, the current study aims to:
3. Effectiveness of speed cameras A number of reviews of speed camera operations have provided support for the effectiveness of the approach to reduce crashes and associated injuries and fatalities (Pilkington and Kinra, 2005; Wilson et al., 2010). For example, Pilkington and Kinra (2005) reviewed 14 studies which examined the effectiveness of fixed, mobile, and a combined of fixed/mobile speed camera. The findings from the review highlighted that speed cameras resulted in crash reductions (5–69%), reductions in injuries (12–65%), and reductions in fatalities (17–71%). In addition, Wilson et al. (2010) reviewed 35 studies which had utilised repeated measure designs to assess the potential effectiveness of mobile and fixed speed cameras. The review found that of the studies which had examined speed outcomes, all had reported a reduction in speed post-camera installation. Further, all studies had reported reductions in all types of crashes, with an average reduction of 11% (500 m from the camera site) and 13% (1 km from the camera site) for more serious crashes (e.g., those resulting in serious injuries or fatalities). In regards to the Queensland context, Newstead and Cameron (2003), examined the effectiveness of mobile speed cameras from May 1997 to June 2001 and reported fatal crash reductions of 45% and a reduction in all crashes by28% within 2 km of the mobile speed camera sites. However it
1. Determine the frequency of a group of urban motorists perceived exposure to fixed and mobile speed cameras; 2. Explore what level of exposure (and what personal characteristics) influences compliance with speed limits; and 3. Examine the effect such exposure has on levels of perceptual certainty of apprehension. 5. Method 5.1. Participants
2 In the State of Queensland, Australia, where the current study was undertaken, there are 3 point-to-point speed camera locations, 16 fixed speed camera locations (excluding combined red light and speed camera locations), and an active deployment of mobile speed cameras (Queensland Government, 2016).
South Eastern Queensland motorists (N = 536) were recruited via an online advertisement and a snowballing technique to take part in this study. The participants (51% female and 49% male) were aged 202
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
5.3. Analysis
between 18 and 73 years (Mage = 32.45, SD = 12.30). The median range participants drove was between 6–10 h per week. On average, participants had been licensed for 14.58 years (SD = 12.10). Participants received a AUS$10 gift voucher as an incentive for study participation.
Frequency and descriptive statistics examined self-reported speeding behaviour,4 perceived camera exposure, and perceived deterrence. To compare the relationships between age, perceptual certainty, self-report speeding behaviour, and perceived camera exposure, bivariate correlations were undertaken. Further, a series of independent group t-tests were implemented to examine the effects of gender and previous speeding offences on self-reported speeding behaviour. A linear regression was then conducted to examine the predictors of perceived certainty of apprehension. Bonferroni adjustments were utilised for all multiple comparisons. Data screening revealed that assumptions of linearity and normality were not extensively breached, and thus, parametric tests were subsequently utilised. Where homogeneity of variance was breached and consistent with recommendations by Field (2009), the equal variances not assumed statistic is interpreted and reported.
5.2. Measures and procedure3 Participants completed either an online version or a pen-paper copy of the questionnaire and corresponding between-group analyses did not reveal any differences on key independent/dependent variables in regards to response type. Participants were required to respond to questions relating to demographics (e.g., age and gender), licence status, speeding behaviour and conviction history. The average of the following two questions was used to measure speeding behaviour: “How often do you exceed speed limits by more than 10 km/h on a highway?” and “How often do you exceed speed limits by more than 10 km/h in a town?” Participants responded to these two questions on a 7-point Likert Scale (1 = Never, 7 = Always), and a higher speeding threshold was utilised (e.g., 10 km/hr) to: (a) ensure there was no ambiguity regarding what actually constituted speeding and (b) previous research has found that a sizable proportion of motorists drive more than 10 km/ hr over the speed limit (Fleiter and Watson, 2006). Furthermore, participants were asked “on average, how many hours of driving would you do per a week?” Driving exposure was measured via a 5-point scale (1 = less than 5 hours per a week, 5 = more than 30 hours per a week). The remaining questions focused on exposure to fixed and mobile speed cameras (i.e., whether participants had seen any speed enforcement operations lately) and deterrence factors (outlined below).
6. Results 6.1. Self-report measure of speeding The majority of participants reported that they ‘never’ or ‘rarely’ exceeded the speed limit by more than 10 km/h on highways or in town settings (see Table 1). However a closer examination revealed that 12.5% of participants reported that they ‘often’, ‘nearly always’ or ‘always’, exceeded the speed limit by more than 10 km/h on highways. A smaller proportion of participants (6.2%) reported that they ‘often’, ‘nearly always’ or ‘always’ exceeded the speed limit by more than 10 km/h in a town. A related samples t-test revealed that participants were significantly more likely to report exceeding the speed limit by 10 km/h on highways compared with towns, (M = 2.53, SD = 1.31 vs. M = 2.04, SD = 1.11), t[538] = 10.16, p < 0.001, 95% CI [0.39, 0.58].
5.2.1. Exposure to fixed and mobile speed cameras Prior to responding to the questions on fixed and mobile speed cameras, participants were provided with a short description of each camera type. The short description for mobile speed cameras was: “Mobile speed cameras include marked and unmarked police vehicles fitted with speed camera equipment parked on the side of the road, and portable devices that are hand-held or set up on a tripod on the side of the road by police officers”. For fixed speed cameras, the description provided to participants was: “Fixed speed cameras are cameras that are permanently installed on roads, tunnels, or intersections”. For each camera type, participants were asked, “In the last 3 months how often have you seen a [fixed/mobile] speed camera?” Participants responded from 1 = not at all, 6 = more than once a day on most days. Higher scores reflect greater camera exposure.
6.2. Self-report exposure to fixed and mobile cameras Table 2 presents the rate of exposure to fixed and mobile speed cameras. Participants were asked to report if they had seen a fixed and mobile speed camera over the past 3 months. Over a third of the sample (41.2%) reported that they had seen a fixed speed camera ‘almost every day’ and ‘more than once a day on most days’ in the past 3 months. A smaller proportion of participants (10.5%) reported seeing a mobile speed camera ‘almost every day’ and ‘more than once a day on most days’ in the past 3 months. A related samples t-test revealed that participants were significantly more likely to report exposure to fixed speed cameras compared with mobile speed cameras (M = 4.17, SD = 1.21 vs. M = 3.23, SD = 1.07), t[514] = 17.05, p < 0.001, 95% CI [0.84, 1.05].
5.2.2. Deterrence A 7-point Likert Scale (1 = strongly disagree, 7 = strongly agree) was used to assess deterrence factors of interest and were modelled on previous psychometrically sound drink driving deterrence items (Freeman et al., 2006; Freeman and Watson, 2009). Five items were used to assess the three Classical Deterrence factors. Two items assessed certainty (i.e., “The chances of getting caught for speeding are high” and “If I were to speed, I’d worry that I would get caught”), two items assessed severity (i.e., “The penalty I would receive for speeding would cause a considerable impact on my life” and “A penalty for speeding would be severe to me”), and one item assessed swiftness (i.e., “The time between getting caught for speeding and receiving a penalty would be very short”). On average, the questionnaire took 20 min to complete.
6.3. Influence of age, gender, and road exposure on self-reported speeding behaviour and perceived camera exposure Pearson correlations were conducted to examine the relationship between perceived camera exposure and: (a) age, (b) exposure to the road (e.g., number of driving hours per a week) and (c) speeding behaviour. As shown in Table 3, there were small significant positive relationships between the number of driving hours, age, exceeding the speed limits by more than 10 km/h on a highway, and perceived camera exposure (both fixed and mobile speed cameras). These finding suggests that as the number of driving hours increases, so too does age, exceeding the posted speed limit by 10 km/h on a highway, and
3 Only those measures relevant to this paper are reported. The larger research project focused more generally on road rule compliance for both speeding and drink driving.
4 Self-reported speeding behaviour included both automated speed enforcement and speed tickets issued by a law enforcement officer.
203
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
exposure to fixed and mobile speed cameras. The effect of exposure was further illuminated through examination of partial correlations, which revealed that deliberate speeding on highways (e.g., more than 10 km/ hr) and fixed camera exposure is essentially halved after consideration of exposure (e.g., r = 1.00* to r = 0.050). Further, there was a small negative relationship between age and exceeding the speed limits by more than 10 km/h in a town. This finding suggests that as age decreases, speeding behaviour increases in built up areas. There were no significant relationships between age and exceeding the speed limits by more than 10 km/h on a highway (see Table 3). There were small positive correlations between speeding by more than 10 km/h on a highway and exposure to fixed speed cameras in the past three months and between speeding more than 10 km/h on a highway and exposure to mobile speed cameras in the past three months. These findings may suggest that as self-reported speeding increases so too does: (a) perceived camera exposure and/or (b) sensitivity to observing speed cameras. However, there were no significant relationships between those who reported exceeding the speed limit by more than 10 km/h in a town and exposure to fixed or mobile speed cameras. Additional Spearman correlations were conducted to assess the relationship between average speeding behaviour (e.g., in both towns and on highways) and exposure to fixed and mobile speed cameras. There were small significant positive correlations between drivers who reported on average exceeding the speed limit by 10 km/h and observing fixed speed cameras, r = 0.091, p = 0.037, and exceeding the speed limit by 10 km/h and observing mobile speed cameras, r = 0.103, p = 0.019. Therefore, as self-report speeding behaviour increases so too does perceived camera exposure.
Table 1 Self-reported Measures of Speeding Behaviour. Frequency
n
%
Frequency
n
%
190 210 93 9 29 3 1 535
35.5 39.3 17.4 1.7 5.4 0.6 0.2
n
%
30 94 198 145 46 9
5.7 18.0 37.9 27.8 8.8 1.7
How often do you exceed the speed limit by more than 10 km/h in a town?
How often do you exceed speed limits by more than 10 km/h on a highway? Never Rarely Sometimes Uncertain Often Nearly always Always Total
101 216 144 9 48 12 7 536
18.8 40.3 26.9 1.5 9.0 2.2 1.3
Never Rarely Sometimes Uncertain Often Nearly always Always Total
Table 2 Self-reported Exposure to Fixed and Mobile Speed Cameras. Frequency
n
In the last 3 months, how often have you seen a fixed speed camera? Not at all Once or twice overall 1–3 times per month Once or twice weekly Almost everyday More than once a day on most days Total
10 38 97 162 141 74 522
%
1.9 7.3 18.6 31.0 27.0 14.2
Frequency In the last 3 months, how often have you seen a mobile camera? Not at all Once or twice overall 1–3 times per month Once or twice weekly Almost everyday More than once a day on most days Total
522
Table 3 The Relationships between Age, Driving Exposure, Speeding Behaviour, and Exposure to Fixed and Mobile Speed Cameras.
1. 2. 3. 4. 5. 6.
Age Driving exposure (hours driver per a week) Speeding highway (more than 10 km/h) Speeding town (more than 10 km/h) Exposure to fixed speed cameras (past 3 months) Exposure to mobile speed cameras (past 3 months)
1
2
3
4
5
6
–
0.243** –
−0.065 0.190** –
−0.135** 0.066 0.572** –
0.115** 0.321** 0.100* 0.055 –
0.089* 0.335** 0.126** 0.056 0.381** –
Note: * p < 0.05; ** p < 0.01.
A series of independent groups t-tests were undertaken to assess the effect of gender on speeding behaviour, number of driving hours per a week, and perceived camera exposure. Compared with females, males were significantly more likely to report exceeding the speed limit by 10 km/h on highways and in towns. Further, males were significantly more likely to report driving more hours per a week and, as a result, observing a fixed and mobile speed cameras within the last three months (see Table 4).
Table 4 The effect of Gender on Self-reported Speeding Behaviour, Number of Driving Hours per a week, and Exposure to Speed Cameras. Variable
n
M (SD)
t
p
Exceeding speed limit by more than 10 km/h on highway Male 264 2.72 (1.42) Female 272 2.35 (1.17) 3.321 0.001 Exceeding speed limit by more than 10 km/h in town Male 263 2.15 (1.16) Female 272 1.94 (1.04) 2.18 0.030 Road exposure (driving h per a week) Male 261 2.51 (1.09) Female 271 1.95 (0.90) 6.411 Exposure to fixed speed camera (past 3 months) Male 258 4.45 (1.12) Female 264 3.89 (1.23) 5.39 Exposure to mobile speed camera (past 3 months) Male 258 3.52 (1.04) Female 264 2.91 (1.03) 6.67
95% CI
0.15, 0.60
0.02, 0.39
< 0.001
0.39, 0.73
> 0.001
0.35, 0.76
> 0.001
0.43, 0.78
6.4. Self-report measures of deterrence Table 5 presents the descriptive statistics for the deterrence variables. The threat of perceptual certainty of apprehension (M = 5.29) was the highest mean, followed by severity and then swiftness of sanctions. To further explore if there were significant differences between the deterrence variables, a one-way repeated measures ANOVA was performed. The mean for certainty was significantly higher than the mean for severity, F[1, 539] = 17.03, p < 0.001, and significantly higher than the mean of swiftness, F[1, 537] = 39.44, p < 0.001. This is in contrast to previous research that has reported perceptual severity to be higher (than certainty) for drug driving (Freeman et al., 2010) as well as for drink driving (Freeman et al., 2006). The mean for severity was significantly higher than the mean for swiftness, F[1, 537]
Note: 1 = Equal variances not assumed statistic is reported here as homogeneity was breached; CI = Confidence Interval.
204
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Table 5 Self-reported Measures of Deterrence. Deterrence
Mean (SD)
Classical deterrence Certainty 5.29 (1.19) Severity 5.06 (1.43) Swiftness 4.85 (1.52)
Strongly disagree
Disagree
Somewhat disagree
Neither agree nor disagree
Somewhat agree
Agree
Strongly agree
1 (0.2%) 11 (2.0%) 9 (1.7%)
22 (4.1%) 31 (5.8%) 39 (7.3%)
36 (6.7%) 53 (9.9%) 59 (11.1%)
62 (18.3%) 101 (18.9%) 98 (18.4%)
154 (28.7%) 137 (25.6%) 100 (18.8%)
177 (33.1) 135 (25.2%) 167 (31.3%)
48 (9.0%) 67 (12.5%) 61 (11.4%)
In terms of road exposure, there was a significant difference between those who had received a speeding ticket (M = 2.45, SD = 1.01) compared with those who had not received a speeding ticket (M = 1.84, SD = 0.97) on number of driving hours per a week, t[529] = 6.90, p < 0.001, 95% CI [0.44, 0.79], suggesting that drivers who received a speeding ticket reported higher on-road exposure. In terms of deterrence, there was no significant difference between those who had received a speeding ticket (M = 5.36, SD = 1.25) compared with those who had never received a speeding ticket (M = 5.24, SD = 1.16) on perceived certainty, t[538] = −1.15, p = 0.251, 95% CI [-0.33, 0.09]. However, participants who had reported receiving a speeding ticket were significantly more likely to report that a penalty would be severe (M = 4.88, SD = 1.49) compared with those participants who had not received a speeding ticket (M = 5.34, SD = 1.28), t (484.62)5 = −3.81, p < 0.001, 95% CI [−0.70, −0.22]. The following analyses only included those individuals who had reported receiving a speeding ticket. Similar to above, those who had reported receiving a speed ticket within the last three years (M = 2.92, SD = 1.38) were significantly more likely to report exceeding the speed limit by more than 10 km/h on a highway compared with those who had received a speeding ticket more than three years ago (M = 2.48, SD = 1.19), t[339] = 3.13, p = 0.002, 95% CI [0.16, 0.71]. Further, those who had received a speeding ticket within the past three years (M = 2.35, SD = 1.17) were significantly more likely to report exceeding the speed limit by more than 10 km/h in a town compared with those who had received a speeding ticket more than three years ago (M = 1.96, SD = 0.99), t[338] = 3.32, p = 0.001, 95% CI [0.16, 0.63]. In terms of road exposure, there was no significant difference between those who had received a speeding ticket within the past three years (M = 2.54, SD = 0.96) and those who had received a speeding ticket over three years ago (M = 2.35, SD = 1.05) on number of driving hours per a week, t[335] = 1.72, p = 0.087, 95% CI [-0.03, 0.40]. In terms of deterrence, there were no significant difference between those who had received a speeding ticket within the past three years (M = 5.24, SD = 1.10) than those who had received a speeding ticket over three years ago on perceived levels of certainty (M = 5.29, SD = 1.20), t[339] = −0.40, p = 0.686, 95% CI [−0.29, 0.19] nor were there any significant difference between those who had received a speeding ticket within the past three years (M = 4.85, SD = 1.57) and those who had received a speeding ticket over three years ago on severity (M = 4.97, SD = 1.40), t[338] = −0.76, p = 0.450, 95% CI [-0.44, 0.20].
Table 6 The Relationships between Age, Driving Exposure, and Deterrence Factors.
1. Age 2. Driving exposure (hours driven per a week) 3. Perceived Certainty 4. Severity 5. Swiftness
1
2
3
4
5
–
0.241** –
0.040 0.038
0.009 0.063
0.130** 0.095*
–
0.498** –
0.309** 0.174** –
Note: *p < 0.05; ** p < 0.01.
= 10.94, p = 0.015. Spearman’s correlations were conducted to assess the relationships between age, exposure to road (number of driving hours per a week), and the deterrence variables (see Table 6). The findings revealed a significant positive relationship between age and swiftness indicating that as age increases so too do perceptions of swiftness of sanctions. There were significant positive relationships between exposure to the road and perceived swiftness, suggesting that as the number of driving hours increase so too does one’s perceptions of swiftness of sanctions. A series of independent groups t-tests were undertaken to examine the effect of gender on self-reported deterrence. As presented in Table 7, males were significantly more likely to report higher levels of swiftness of apprehension, although there were no other significant gender effects on self-reported deterrence.
6.5. Effect of demerit points A series of independent groups t-tests were also undertaken to examine the effect of ever receiving a speeding ticket on current driving behaviour, road exposure, and perceptual deterrence and severity. The findings revealed that those who reported receiving a speeding ticket (M = 2.73, SD = 1.30) were significantly more likely to report exceeding the speed limit by more than 10 km/h on a highway compared with those who had never received a speeding ticket (M = 2.20, SD = 1.27), t[533] = 4.56, p < 0.001, 95% CI [0.31, 0.75]. Similarly, compared with those who had never received a speeding ticket (M = 1.84, SD = 1.11), those who reported receiving a speeding ticket were significantly more likely to report exceeding the speed limit by more than 10 km/h in a town (M = 2.17, SD = 1.08), t[532] = 3.41, p < 0.001, 95% CI [0.14, 0.52]. Table 7 The effect of Gender on Deterrence. Variable
n
Classical Deterrence Perceived certainty Male 264 Female 277 Severity Male 264 Female 276 Swiftness Male 263 Female 275
M (SD)
6.6. Effect of demerit points on perceived camera exposure (controlling for road exposure) t
p
95% CI
5.19 (1.24) 5.38 (1.15)
−1.89
0.060
−0.36, 0.01
5.02 (1.41) 5.09 (1.45)
−0.61
0.541
−0.32, 0.17
5.00 (1.53) 4.71 (1.49)
2.21
0.028
0.03, 0.54
A Multivariate Analysis of Covariance (MANCOVA) was undertaken to examine the effect of ever receiving a speeding ticket (Independent Variable) on perceived camera exposure (mobile and fixed cameras) (Dependent Variables) after controlling for road exposure (number of driving hours per a week) (Covariate). The findings revealed that there was no significant effect of ever receiving a speeding ticket on perceived speed camera exposure when controlling for number of driving 5
205
Equal variances not assumed statistic is reported here as homogeneity was breached.
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Table 8 Beta Weights with their Significance, Unstandardised Co-efficients (B) and their Confidence Intervals and the Squared Semi-Partial Correlations of the Predictor Variables for Perceptual Certainty. Variables
B
SE
β
p
95% CI for B Lower
Age Gender Road exposure Self-reported speeding Exposure to fixed cameras Exposure to mobile cameras Severity Swiftness
−0.001 0.166 −0.018 −0.214 0.035 0.047 0.382 0.180
0.004 0.091 0.047 0.042 0.039 0.046 0.031 0.029
−0.009 0.069 −0.015 −0.194 0.036 0.042 0.457 0.227
0.809 0.068 0.700 > 0.001 0.367 0.302 > 0.001 > 0.001
−0.01 −0.01 −0.11 −0.30 −0.04 −0.04 0.32 0.12
Upper
sr 2
0.01 0.35 0.07 −0.13 0.11 0.14 0.44 0.24
−0.011 0.082 −0.017 −0.225 0.042 0.046 0.486 0.272
Note: CI = Confidence Intervals, sr2= semi partial correlation squared.
hours per a week, F[2, 503] = 1.36, p = 0.258, Wilk’s ʌ = 0.995, ηp2 = 0.005.
compared with the covert operations which are often employed with mobile speed cameras, and thus, motorists may not be aware of a mobile camera’s presence. Finally, many of these fixed cameras are located in high-traffic density areas (e.g., highways and busy intersection) and operate 24 h per a day (compared with several hours for mobile cameras). Taken together, it is not surprising that drivers were more likely to report greater exposure towards the fixed speed cameras over the past three months compared with the mobile speed cameras. What was also expected was that greater exposure to the road (e.g., more kms driven) would be associated with higher exposure to speed cameras, which was confirmed. While this may be considered of little significance, it is encouraging that drivers who probably were exposed to greater levels of speed cameras actually reported this outcome, which is a foundational component of deterrence principles e.g., drivers need to be constantly exposed to deterrent factors (Homel, 1988). This also strengthens the veracity of the current self-report data, which is mentioned in a proceeding section. In regards to the second aim, motorists who reported speeding more often actually reported a higher frequency of exposure to speed cameras, which was further supported by the positive association between younger males and perceived camera exposure. This finding is consistent with scant research in the area (Wundersitz et al., 2002), and suggests that drivers who consciously exceed the speed limit may be more aware (or sensitive) to the threat of speed enforcement approaches. Although, it should also be noted that the relationship between speeding and camera exposure was inflated by levels of driving. This finding needs to be explored through further research, particularly in regards to the origins of the lack of a deterrent effect, although it consistent with Australian research that has indicated higher levels of perceived certainty is actually associated with an increase in self-reported speeding frequency (Fleiter and Watson, 2006; Fleiter et al., 2009). In regards to perceptual deterrent factors, perceptions of apprehension certainty were found to be higher than severity or swiftness. This finding is in contrast to previous deterrent-based research on other offending behaviours (Freeman et al., 2006, 2010), but may be reflective of the increased likelihood of apprehension for speeding versus other rule violations e.g., drink driving and drug driving. The result is nonetheless encouraging as scholars have historically postulated that the most powerful deterrent effects on offending behaviour are produced by the perceived threat of the certainty of apprehension (Homel, 1988; Nagin and Pogarsky, 2001; Von Hirsch et al., 1999). A complementary analysis was undertaken to determine the link between age and perceptual deterrence, which revealed younger drivers reported lower levels of perceptions regarding the threat of sanctions. Similar to above, scant research has considered this issue, although preliminary evidence exists that suggests older drivers are more likely to respond appropriately to deterrent threats (Bushway et al., 2013; Sampson and Cohen, 1988). This may be associated with
6.7. Predictors of perceptual certainty Finally, a linear regression analysis was undertaken to examine the contribution of age, gender, exposure to the road, self-reported speeding behaviour, perceived camera exposure (both fixed and mobile), and deterrence factors (i.e., severity, swiftness) in predicting higher levels of perceptual certainty of apprehension for speeding behaviour. These factors accounted for 40.4% of the variance in perceptual certainty, F[8, 503] = 39.16, p < 0.001. Table 8 shows that speeding behaviour, severity, and swiftness were all unique significant predictors of perceptual certainty. 7. Discussion This study aimed to conduct one of the first exploratory investigations into the deterrent effect of observing fixed and mobile speed cameras on self-reported speeding behaviours. More specifically, the study aimed to: (a) examine the frequency of urban motorists’ exposure to fixed and mobile speed cameras, (b) investigate what level of exposure (and what personal characteristics) influences compliance with speed limits and (c) explore the differential effect such exposure has on levels of perceptual certainty of apprehension. In regards to self-reported speeding behaviours, the sample were more likely to report speeding on highways compared to towns, which is consistent with scant research in this area (Fleiter and Watson, 2006). While there is limited published research on the effects of road infrastructure (Fleiter, 2010), it may be suggested that speeding on highways is linked to: (a) increased traffic flow and associated opportunities to speed e.g., lower congestion and traffic lights (Wundersitz et al., 2002) and (b) greater opportunities to observe speed cameras in advance and engage in corrective behaviours. What was more predictable in the current sample was that males and younger drivers were at an increased likelihood to report speeding compared with females as well as older drivers (e.g., Fleiter and Watson, 2006; Shinar et al., 2001; Watson et al., 2009; Wundersitz et al., 2002). In terms of the first aim, the findings revealed that drivers reported greater exposure to fixed speed cameras compared with mobile speed cameras within the last three months. While there are no complementary studies for comparison, this result may be linked to motorists’ driving patterns being somewhat routine (or habituated) such as driving for the purposes of going to and from work. Additionally, while there are over 3500 active mobile camera sites in Queensland (Queensland Police, 2015) compared with 16 fixed speed camera sites (Queensland Government, 2016), all the fixed speed cameras are located in South-East Queensland, where the study was undertaken. In addition, fixed speed cameras tend to involve more overt operations 206
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
emerging findings that indicate younger individuals experience difficulties recognizing (and responding appropriately) to risk (Albert and Steinberg, 2011), which may in part be due to the underdeveloped prefrontal cortex (Lebel and Beaulieu, 2011). Consistent with the positive association between self-reported speeding and perceived camera exposure (outlined above), those who had incurred a speeding ticket were also more likely to report: (a) engaging in further speeding behaviours and (b) greater exposure to the road. While the latter finding is intuitive in nature, the former is consistent with research that indicates applied speeding sanctions are not sufficient to create a deterrent effect (Fleiter, 2010) and/or the apprehension process creates a “resetting effect” e.g., an individual believes that the chances of being re-apprehended is much lower according to the belief that their initial experience of getting caught was inevitable. More broadly, previous research has demonstrated that drivers who are apprehended for speeding have a greater chance of receiving further tickets, (e.g., Lawpoolsri et al., 2007), which is supportive of the assumption that past behaviour is an efficient predictor of future behaviour (Danner et al., 2008). Overall, the research suggests that speed tickets in isolation may not deter drivers from speeding (Fleiter, 2010) and that additional countermeasures (including non-legal sanctions) are required to enhance drivers’ compliance with the road rules. Similarly, there was no relationship between receiving a speeding ticket and perceptual certainty, which is again consistent with previous counterintuitive findings that indicate higher levels of perceived certainty and severity is actually associated with an increase in self-reported speeding frequency (Fleiter and Watson, 2006; Fleiter et al., 2009). Furthermore, no relationship was found between perceived camera exposure and receiving a speeding ticket, which might be expected given that a large proportion of the sample reported abiding by the speed limits. Finally, a positive relationship was found between receiving a speeding ticket and perceived severity, which is evidence of the specific deterrent effect of speeding sanctions. Finally, a linear regression analysis to determine the predictors of higher levels of self-reported perceptual certainty revealed it was linked with: (a) general speed compliance and (b) higher sanction severity and swiftness perceptions. The finding is consistent with research that has demonstrated those who comply with rules report higher levels of apprehension certainty (Homel, 1988; Paternoster and Piquero, 1995; Piquero and Pogarsky, 2002) and that a positive relationship can exist between the three constructs of the classical deterrence model (Piquero et al., 2011). That is, sanctions need to not only be certain, but severe (Homel 1988). In contrast, the analysis confirmed that a direct (and substantial) relationship was not evident between perceived camera exposure and perceptual certainty for the current sample. There were a number of limitations with this research that should be considered when interpreting the results. For example, due to the crosssectional design of this study, the current research methodology does not permit an examination of possible causal effects, nor was there any comparison with actual speed data e.g., in-vehicle measurement devices. Additionally, self-report measures were used that may be susceptible to participant biases (e.g., social desirability) (Wåhlberg et al., 2010) as well as inaccuracies with memory recall (McCarroll, 2017). As a result, future research should employee more objective measures of speeding behaviour, such as offence data or in-vehicle monitoring devices. Nevertheless, previous research has reported significant moderate positive correlations exist between objective and subjective measures of speeding behaviour (e.g., Åberg and Wallén Warner, 2008; Helman and Reed, 2015; Kaye et al., 2016), suggesting that self-report measures are a suitable measure of actual speeding behaviour. The sample consisted of a higher proportion of younger compared with older drivers. The majority of the sample did not regularly engage in speeding behaviours, which limits variance and the ability to identify trends in the data. The deterrent threat of a sanction (particularly the severity of a sanction) may be dependent upon a motorist’s demerit point accumulation at the time, and thus, future research would benefit
from considering this issue. Future research needs to disentangle the difference between “sensitivity to cameras” and actual “exposure to cameras”, as there may be natural variations in these concepts among the motoring population. Further attempts to extract direct effects also need to be directed towards the relationship between speeding, observing cameras and driving exposure, as the latter factor (e.g., amount of driving) may influence both the frequency of speeding and subsequent camera observations. Finally, the study focused on the Queensland speed camera program and, therefore, caution should be used when interpreting these findings outside of Queensland, Australia. Future research is required to not only extend upon the current study’s aims but also include a more representative sample of the driving population, which may be considered more random in nature. 8. Conclusion This study was one of the first to explore the deterrent effect of observing fixed and mobile speed cameras on self-reported speeding behaviours. Taken together, the investigation did not reveal clear linear links between exposure to cameras and corresponding behaviours, although this may be in part because of the above mentioned limitations. Despite this, the study did illuminate key preliminary findings associated with: (a) age, gender and deterrability, (b) those most likely subgroup to observe speed cameras, (c) the self-reported effect of speeding sanctions and (d) the factors that promote perceptual certainty. Given that deterrence-based mechanisms will continue to remain a cornerstone of road safety initiatives into the foreseeable future, it is warranted to conduct further research to determine the general deterrent and specific deterrent effect of speed cameras on overall rule compliance. Acknowledgement This research was funded by the Australian Research Council Discovery Scheme. References Åberg, L., Wallén Warner, H., 2008. Speeding–deliberate violation or involuntary mistake? revue europÅenne de psychologie appliquÅe/. Eur. Rev. Appl. Psychol. 58 (1), 23–30. http://dx.doi.org/10.1016/j.erap.2005.09.014. Albert, D., Steinberg, L., 2011. Judgement and decision making in adolescence. J. Res. Adolesc. 21 (1), 211–224. http://dx.doi.org/10.1111/j.1532-7795.2010.00724.x. Bates, L., Soole, D., Watson, B., 2012. The Effectiveness of Traffic Policing in Reducing Traffic Crashes. Palgrave MacMillian, New York,U.S. Bushway, S., DeAngelo, G., Hansen, B., 2013. Deterability by age. Int. Rev. Law Econ. 36, 70–81. http://dx.doi.org/10.1016/j.irle.2013.04.006. Bureau of Infrastructure, Transport, and Regional Economics [BITRE], 2016. Road Deaths Australia. Retrieved from https://bitre.gov.au/publications/ongoing/rda/files/RDA_ Dec_2015.pdf. Cameron, M.H., Delaney, A.K., 2008. Speed enforcement ? effects, mechanisms, intensity and economic benefits of each mode of operation. In: Paper Presented at the Joint Australiasian College of Road Safety and Queensland Parliamentary Travelsafe Committee Conference: High Risk Road Users – Motivating Behaviour Change: What Works and What Doesn't Work? Queensland, Australia. Danner, U., Aarts, H., de Vries, N., 2008. Habit vs intention in the prediction of future behaviour. The role of frequency, context stability and mental acessibility of past behaviour. Br. J. Soc. Psychol. 47 (2), 245–265. Davey, J.D., Freeman, J.E., 2011. Improving road safety through deterrence-based initaitives: a review of research. Sultan Qaboos Univ. Med. J. 11 (1), 312–320. Delaney, A., Ward, H., Cameron, M., 2005a. The History and Development of Speed Cameras. Retrieved from Melbourne, Australia: https://www.monash.edu/muarc/ research/our-publications/muarc242. Delaney, A., Ward, H., Cameron, M., Williams, A.F., 2005b. Controversies and speed cameras: lessons learnt internationally. J. Public Health Policy 26 (4), 404–415. Retrieved from: http://www.jstor.org/stable/4125166. Elvik, R., Christensen, P., 2007. The deterrent effect of increasing fixed penalties for traffic offences: the Norwegian experience. J. Safety Res. 38 (6), 689–695 1016/ j.jsr.2007.09.007. Field, A., 2009. Discovering Statistics Using SPSS, 3rd. ed. SAGE, London. Fleiter, J., Watson, B., 2006. The speed paradox: the misalignment between driver attitudes and speeding behaviour. J. Aust. Coll. Road Saf. 17 (2), 23–30. Fleiter, J., Watson, B., Lennon, A., King, M., Shi, K., 2009. Speeding in Australia and China: a comparison of the influence of legal sanctions and enforcement practices on
207
Accident Analysis and Prevention 108 (2017) 201–208
J. Freeman et al.
Retrieved from: https://infrastructure.gov.au/roads/safety/publications/2013/pdf/ community_att_13.pdf. Pilkington, P., Kinra, S., 2005. Effectiveness of speed cameras in preventing road traffic collisions and related casualties: systematic review. BMJ: Br. Med. J. 330 (7487), 331–334. Retrieved from: http://www.jstor.org/stable/25458895. Piquero, A.R., Pogarsky, G., 2002. Beyond Stafford and Warr’s reconceptualization of deterrence: personal and vicarious experiences, impulsivity, and offending behavior. J. Res. Crime Delinquency 39 (2), 153–186. Piquero, A., Paternoster, R., Pogarsky, G., Loughran, T., 2011. Elaborating the individual difference component in deterrence theory. Ann. Rev. Law Soc. Sci. 7, 335–360. Queensland Government, 2014. Speed and Red Light Cameras. Retrieved from: http:// www.qld.gov.au/transport/safety/fines/speed/cameras/#fixed. Queensland Government, 2016. Location of Speed and Red Light Cameras. Retrieved from: https://www.qld.gov.au/transport/safety/fines/speed/location/. Queensland Police, 2015. Active Mobile Speed Camera Sites. Retrieved from: https:// www.police.qld.gov.au/programs/roadSafety/activespeedcamerasites.htm. Sampson, R.J., Cohen, J., 1988. Deterrent effects of the police on crime: a replication and theoretical extension. Law Soc. Rev. 22 (1), 163–189. Shinar, D., Schechtman, E., Compton, R., 2001. Self-reports of safe driving behaviors in relationship to sex, age, education and income in the US adult driving population. Accid. Anal. Prev. 33 (1), 111–116. http://dx.doi.org/10.1016/S0001-4575(00) 00021-X. Soole, D.W., Watson, B.C., Fleiter, J.J., 2013. Effects of average speed enforcement on speed compliance and crashes: a review of the literature. Accid. Anal. Prev. 54, 46–56. http://dx.doi.org/10.1016/j.aap.2013.01.018. Taxman, F.S., Piquero, A., 1998. On preventing drunk driving recidivism: an examination of rehabilitation and punishment approaches. J. Criminal Justice 26 (2), 129–143. http://dx.doi.org/10.1016/S0047-2352(97)00075-5. Tay, R., 2000. Do speed cameras improve road safety? Traffic Transp. Stud. 44–51. Vingilis, E., Salutin, L., 1980. A prevention programme for drinking driving. Accid. Anal. Prev. 12 (4), 267–274. http://dx.doi.org/10.1016/0001-4575(80)90003-2. Von Hirsch, A.V., Bottoms, A.E., Burney, E., Wikstrom, P., 1999. Criminal Deterrence and Sentence Severity: an Analysis of Recent Research. Hart Publishing, Oregon. Wåhlberg, A.E., Dorn, L., Kline, T., 2010. The effect of social desirability on self reported and recorded road traffic accidents. Transp. Res. F Traffic Psychol. Behav. 13 (2), 106–114. http://dx.doi.org/10.1016/j.trf.2009.11.004. Watson, B., Hart, S., Freeman, J., Tay, R., Davey, J., Madden, J., Goff, A., 2005. Review of Random Breath Testing (RBT) in Queensland. Queensland University Technology. Watson, B., Watson, A., Siskind, C., Fleiter, J., 2009. Characteristics and predictors of high-range speeding offenders. In: Paper Presented at the Australasian Road Safety Research, Policing and Education Conference. Sydney, Australia. Wilson, C., Willis, C., Hendrikz, J.K., Le Brocque, R., Bellamy, N., 2010. Speed cameras for the prevention of road traffic injuries and deaths. Cochrane Database Syst. Rev. 10. http://dx.doi.org/10.1002/14651858.CD004607.pub3. Wundersitz, L., Kloeden, C., McColl, R., Baldock, M., McLean, J., 2002. Speeding in metropolitan Adelaide: a situation analysis. In: Paper Presented at the Road Safety Research, Policing and Education Conference. Adelaide.
car drivers. In: Proceedings of the Australasian Road Safety Research, Policing and Education. Sydney. Fleiter, J., 2010. Examining Psychosocial Influences on Speeding in Australian and Chinese Contexts: a Social Learning Approach. Queensland University of Technology PhD thesis. Freeman, J., Watson, B., 2009. Drink driving deterrents and self-reported offending behaviours among a sample of Queensland motorists. J. Safety Res. 40 (2), 113–120. http://dx.doi.org/10.1016/j.jsr.2008.12.009. Freeman, J., Liossis, P., Schonfeld, C., Sheehan, M., Siskind, V., Watson, B., 2006. The self-reported impact of legal and non-legal sanctions on a group of recidivist drink drivers. Transp. Res. F: Traffic Psychol. Behav. 9 (10), 53–64. Helman, S., Reed, N., 2015. Validation of the driver behaviour questionnaire using behaviour data from an instrumented vehicle and high-fidlity driving simulator. Accid. Anal. Prev. 75, 245–251. http://dx.doi.org/10.1016/j.aap.2014.12.008. Homel, R.J., 1988. Policing and Punishing the Drinking Driver: A Study of Specfic and General Deterrence. Springer-Verlag, New York. Job, S., Sakashita, C., Mooren, L., Grezbieta, R., 2013. Community perceptions and beliefs regarding low-level speeding and suggested solutions. In: Paper Presented at the 2013 Annual Meeting of the Transportation Research Board. Washington, U.S.. Kaye, S.-A., Lewis, I., Algie, J., White, M.J., 2016. Young drivers' responses to antispeeding advertisements: comparison of self-report and objective measures of persuasive processing and outcomes. Traffic Inj. Prev. 17 (4), 352–358. http://dx.doi. org/10.1080/15389588.2015.1084419. Lawpoolsri, S., Li, J., Braver, E.R., 2007. Do speeding tickets reduce the likelihood of receiving subsequent speeding tickets? a longitudinal study of speeding violators in Maryland. Traffic Inj. Prev. 8 (1), 26–34. http://dx.doi.org/10.1080/ 15389580601009764. Lebel, C., Beaulieu, C., 2011. Longitudinal development of human brain wiring continues from childhood into adulthood. J. Neurosci. 31 (30), 10937–10947. http://dx.doi. org/10.1523/JNEUROSCI.5302-10.2011. McCarroll, C., 2017. Looking the past in the eye: distortion in memory and the costs and benefits of recalling from an observer perspective. Conscious. Cogn. 49, 322–332. Nagin, S., Pogarsky, G., 2001. Integrating celerity, impulsivity: and extralegal sanction threats into a model of general deterrence: theory and evidence. Criminology 39, 865–889. Newstead, S., Cameron, M., 2003. Evaluation of the Crash Effects of the Queensland Speed Camera Program. Retrieved from Monash University Accident Research Centre: http://www.monash.edu/__data/assets/pdf_file/0008/216368/muarc204. pdf. Paternoster, R., Piquero, A., 1995. Reconceptualizing deterrence: an empirical test of personal and vicarious experiences. J. Res. Crime Delinquency 32 (3), 251–286. Pennay, D., 2004. Community Attitudes to Road Safety: Community Attitudes Survey Wave 16. Retrieved from: https://infrastructure.gov.au/roads/safety/publications/ 2004/pdf/Comm_Att_16.pdf. Petridou, E., Moustaki, M., 2000. Human factors in the causation of road traffic crashes. Eur. J. Epidemiol. 16 (9), 819–826. http://dx.doi.org/10.1023/a:1007649804201. Petroulias, T., 2014. Community Attitudes to Road Safety – 2013 Survey Report.
208