Abstract
Objective:
Measure the impact of automated photo speed enforcement in school zones on motorist speed and speeding violation rates during school travel.
Methods:
Automated enforcement cameras, active during school commuting hours, were installed around 4 elementary schools in Seattle, WA in 2012. We examined the effect of automated enforcement on motorist speeds and speed violation rates during the citation period (December 10, 2012 to January, 15 2015) compared to the “warning” period (November 1 to December 9, 2012). We evaluated outcomes with an interrupted time series approach using multilevel mixed linear regression.
Results:
Motorist speed violation rates decreased by nearly half in the citation period compared to the warning period (standardized incident rate ratio 0.53, 95% confidence interval [95% CI] 0.42, 0.66). The hourly maximum violation speed and mean hourly speeds decreased 2.1 MPH (95% CI=−2.88, −1.39) and 1.1 MPH (95% CI=−1.64, −0.60), respectively. The impact of automated enforcement was sustained during the second year of implementation.
Conclusion:
Automated photo enforcement of speed limit in school zones was effective at reducing motorist speed violations and also achieved a significant reduction in mean motorist speed.
Keywords: School Road Safety, Automated Speed Enforcement, Interrupted Time Series, Pedestrian Safety
INTRODUCTION:
Excessive vehicle speeds are one of the main causes of road injury and fatality in the United States and pose a significant threat to children traveling to and from school.1–4 When entering a designated school zone during peak school travel times, drivers are required to reduce vehicle speeds.5 These school zones are marked by signs instructing motorists to drive at a lower speed (typically 20 MPH) when children are present.6 While school zone speed limits are intended to reduce vehicle speeds, studies have found that 50–90% of motorists travel well above the speed limit in school zones.4,7–9 While school zone speed limits are also meant to reduce child pedestrian-motor vehicle collisions, limited evidence suggests that most child pedestrian collisions occur near schools during school travel times.10,11 Most school zone speed limits are set at 20 MPH to reduce risk for vulnerable child pedestrians.12–14 The risk of fatal pedestrian injury when struck by a vehicle traveling 20 MPH is less than 10%, and increases sharply to over 50% risk of death in a 40 MPH collision.13,14
Excessive vehicle speeds have other negative effects on child health. If parents perceive that motorists travel at high speeds in their neighborhoods or destinations, children are less likely to walk or bike.15,16 Active transport represents the easiest way for children to incorporate exercise into daily routines.17 The relationship between regular physical activity and the health and development of children has been well documented.18,19 Studies suggest that periodic exercise may improve children’s bone composition, social and mental health, and motor skills, while also acting as a protective mechanism against future chronic diseases such as cancers and cardiovascular diseases. Despite these benefits, children’s active transportation use, an important source of physical activity, has declined precipitously, falling from 48% in 1969 to 13% in 2009.20 Reducing vehicle speeds near schools may address these parental safety concerns and potentially lead to more walking and biking to school by children.21–24
There are a number of countermeasures available to reduce motorist speeds near schools. Traffic calming measures (e.g., speed bumps or narrow lanes) can significantly reduce pedestrian injuries and fatalities.25–27 Adding flashing beacons to school zone speed limit signs can significantly reduce speeds compared to non-school hour speeds, but even in areas with these beacons, most vehicles travel above the school zone speed limit during school hours.28 Automated photo enforcement of speed is another mechanism for regulating driver behavior that can reduce motor vehicle speeds, leading to a reduction of fatal and serious road injuries.29–37 Automated photo enforcement may also have some drawbacks, such as an increase in rear-end collisions. Several US municipalities have adopted photo enforcement in school zones with the goal of reducing speeds and speed violations, but the effectiveness of photo enforcement in school zones has not been evaluated.31 The primary objective of this study was to determine the effect of automated speed enforcement cameras on the rate of speed violations and motor vehicle speed during school travel hours in school zones. A secondary objective was to examine if the effect of cameras was sustained after the first school year of implementation.
METHODS
Study Design and Data Collection
Data were prospectively collected at four elementary schools in Seattle, WA from November 1, 2012 to January 15, 2015 (Supplementary Table 1). Schools were selected for photo enforcement by the Seattle Department of Transportation (SDOT) and the Seattle Police Department (SPD) based on traffic speed studies and predictions of safety benefits. The four schools for this pilot program were selected because they had higher rates of speeding during school travel hours than other schools examined (N=15). The 85th percentile speeds during school hours before camera installations ranged from 34 MPH to 36 MPH during school travel times and between 5% and 22% of all vehicles exceeded 40 MPH. Detailed hourly speed results from these studies were not available for analysis. Prior to installation of automated enforcement cameras, the selected schools already had existing flashing beacons (active during school travel hours) warning drivers that they were entering a school zone. Each school had two fixed cameras installed on adjacent arterial streets, one facing each direction of traffic. Cameras were installed and maintained by a private company (American Traffic Solutions, Inc.) contracted by the City of Seattle. Camera installation involved embedding pairs of inductive loops in each direction of the roadway which are used to measure vehicle speed and volumes. An additional sign was added to the existing flashing beacons to warn drivers that the school speed zone was “Photo Enforced” (Supplementary Material). If a vehicle exceeded a speed threshold slightly above the school zone speed limit (20 MPH), the camera took a photo of the vehicle license plate and recorded the violation speed, time and date (exact violation speed setting cannot be shared publicly per request from the Seattle Police Department). All motor vehicles passing through the school zones are counted and recorded and their speeds are recorded as well, though cameras only record vehicles exceeding the speed limit. Vehicles not violating the speed limit were stored as aggregate hourly volumes and mean vehicle speed by location and direction. For vehicles exceeding the speed limit, a count of the number of vehicles at each speed in MPH for each hour for each location and direction is stored.
To prepare drivers using the roads with the new automated photo enforcement, there was a warning period, (November 1, 2012 to December 9, 2012) where owners of vehicles exceeding the school zone speed limit during school travel times were mailed a letter indicating that they would have received a speeding ticket ($189). After the warning period, owners were issued a speeding ticket via mail with a photo of the vehicle during violation and were instructed to pay the citation. Data were collected by ATS and SPD and shared with study investigators as hourly aggregated mean vehicle speed and volume of non-speed violating vehicles and speed violating vehicles aggregated by violation speed within each hour.
Because individual motorist speeds could not be tracked, we evaluated changes in the outcomes among the population of motorists passing through the school zones during the study period. Ideally, the outcomes would be compared across the time period prior to the camera installation in addition to the warning and citation periods, but unfortunately such data were not available. While it is possible that drivers receiving a warning for violating the speed limit in a school zone may change their behaviors, a much stronger effect is expected when an actual fine is imposed, thus our hypotheses tested differences in outcomes between the warning and citation period.38
Because similar data were not collected at other schools, a before-after study design was used to evaluate the placement of automated photo enforcement.
Outcomes
To assess the effectiveness of the cameras during school travel hours, several outcomes were evaluated: hourly violation rate, mean hourly vehicle speed, mean hourly violating vehicle speed, hourly maximum violation speed and rate of hourly high-speed violations (over 35 MPH). The hourly violation rate was calculated as the number of vehicles violating the speed threshold during a given hour divided by the total number of vehicles recorded passing through that roadway during that hour.
Exposures and Covariates
The primary exposure of interest was the initiation of automated photo enforcement with mailed citations. We compared the change in outcome variables during the baseline period when drivers received a written warning (November 1, 2012 to December 9, 2012) to the period when drivers received a written citation for speed violations (December 10, 2012 to January 15, 2015). Detailed hourly data on speed violations were not available prior to the baseline warning period because the roadway was only instrumented with equipment when the cameras were installed. We also examined the sustained effectiveness of photo-enforcement after the first year of implementation. The time of day (AM or PM), day of the week (Monday to Friday) and quarter of the year (Fall [September to December], Winter [January to March], Spring [April to June]) were examined as potential confounders.
Statistical Analysis
An interrupted time-series analysis was used to assess camera effectiveness. All outcomes followed a linear pattern, indicating that this approach was appropriate. At one school, one of the two cameras malfunctioned, resulting in data loss during the warning period and early in the citation period. Thus, the analytical dataset included data from seven cameras at four schools since no baseline data were collected for the malfunctioning camera. We used mixed effects models with robust standard errors to estimate the change in outcomes after ticketing began in both univariate and multivariable models. For hourly vehicle volumes, mean speed, and mean violation speed, linear mixed effect models estimated mean differences and 95% confidence intervals (95% CI). For hourly violation and high speed violation rates, generalized linear mixed models with Poisson distribution estimated rate ratios and 95% CIs and an offset with hourly vehicle volume. The models for the primary objective were structured with two levels, with a random intercept for the week and for each school to account for temporal and location correlation. A random parameter for warning versus violation as included at the school level with unstructured covariance. To determine the effect of photo enforcement after the first school year of use for the secondary objective, we created an interaction term between the photo enforcement period and a time-dependent indicator variable using the dates December 10, 2012 to October 31, 2013 as the first year of the intervention and November 1, 2013 to January 15, 2015 as later years. Citation versus warning status was evaluated as a random parameter in multivariable models. To assess the effect of temporal trends, we compared vehicle volumes and mean speeds during school travel hours to those during non-school travel hours. Because vehicle volumes and speeds were only available in the aggregate outside of school travel hours, we were only able to compare mean vehicle speeds and volumes for analyses which compared school travel periods to non-travel periods. We also calculated the adjusted violation rate ratio using the combined population of all study hours as the standard population and multilevel mixed effects Poisson regression 39,40. Data were analyzed with Stata 13MP (Stata Corp, College Station, TX).
Ethics
This study was approved by the University of Washington Institutional Review Board.
RESULTS
A total of 26,500 hours of vehicle speed data were captured during school travel hours; 7% of the hours were observed during the baseline “warning only” period (Table 1). Nearly 30% of the 1,831 school travel hours observed during the warning period had at least one speed violation over the speed threshold (n=9,934 warnings issued), while 38% of the 24,669 school travel hours during the citation period had at least one speed violation (n=73,967 citations issued). Nearly 10% of violation speeds exceeded 35 MPH, 15 MPH over the posted limit during school travel hours. Schools had similar violation rates and speeds.
Table 1.
Descriptive statistics of hourly outcomes during school travel hours. Means and standard deviations (SD).
| Total Hours |
Hours with Violations |
Vehicle Volume |
Violation Rate (per 1,000 vehicles) |
Violation Rate >35 MPH (per 10,000 vehicles) |
Mean Violation Speed (MPH) |
Max Violation Speed (MPH) |
Mean Speed (MPH) |
|
|---|---|---|---|---|---|---|---|---|
| N | % | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |
| Camera | ||||||||
| Warning | 1,831 | 29 | 498 (232) | 13.0 (27.2) | 1.92 (15.5) | 29.4 (1.4) | 35.5 (4.3) | 25.7 (3.9) |
| Citation | 24,669 | 38 | 480 (216) | 6.4 (11.5) | 0.96 (6.97) | 29.5 (1.7) | 33.8 (4.0) | 24.2 (4.5) |
| P-value* | 0.296 | <0.001 | 0.262 | 0.227 | 0.3231 | <0.001 | ||
| Day | ||||||||
| Monday | 5,015 | 38 | 461 (208) | 7.0 (13.0) | 1.14 (8.11) | 29.6 (1.7) | 34.0 (4.0) | 24.5 (4.4) |
| Tuesday | 5,498 | 37 | 474 (219) | 6.2 (12.6) | 0.97 (8.92) | 29.4 (1.8) | 33.7 (4.2) | 24.2 (4.5) |
| Wednesday | 5,563 | 36 | 480 (217) | 6.4 (12.7) | 0.83 (6.10) | 29.5 (1.7) | 33.8 (4.1) | 24.2 (4.4) |
| Thursday | 5,385 | 39 | 486 (220) | 6.9 (13.4) | 1.10 (7.78) | 29.5 (1.7) | 33.9 (4.0) | 24.1 (4.5) |
| Friday | 5,039 | 38 | 497 (220) | 7.1 (13.6) | 1.01 (7.50) | 29.5 (1.7) | 34.0 (3.9) | 24.3 (4.5) |
| <0.001 | <0.001 | 0.190 | 0.001 | 0.0231 | <0.001 | |||
| Time | ||||||||
| AM | 12,700 | 37 | 415 (195) | 7.6 (14.7) | 1.21 (9.23) | 29.6 (1.7) | 34.1 (3.9) | 24.3 (4.8) |
| PM | 13,800 | 38 | 538 (220) | 5.9 (11.3) | 0.82 (6.01) | 29.4 (1.7) | 33.7 (4.1) | 24.3 (4.1) |
| 0.058 | 0.466 | 0.709 | 0.217 | 0.390 | 0.819 | |||
| Quarter | ||||||||
| Fall | 11,945 | 34 | 490 (227) | 6.3 (14.6) | 0.97 (8.27) | 29.5 (1.8) | 33.7 (4.1) | 24.6 (4.3) |
| Winter | 7,990 | 40 | 463 (210) | 7.2 (12.1) | 1.10 (7.65) | 29.5 (1.6) | 34.1 (3.9) | 24.2 (4.5) |
| Spring | 6,565 | 40 | 480 (207) | 6.8 (11.2) | 0.96 (6.73) | 29.5 (1.7) | 33.9 (3.9) | 23.7 (4.6) |
| <0.001 | 0.0428 | 0.185 | 0.631 | 0.150 | <0.001 | |||
P-values were calculated from univariate linear mixed effects models and robust standard errors or generalized linear mixed effects models as described in Methods
The mean hourly speed violation rate during the citation period was 6.4 (standard deviation [SD] 11.5) was about half the rate of the warning period of 13.0 (SD 27.2) citations per hour. Violation rates continued to decrease gradually over time (Figure 1) while speeds and traffic volume remained fairly constant. It is noteworthy that violation rates tended to spike after weekends, school breaks, and summer vacation periods (Figure 1, note patterns after blank spots).
Figure 1-.
Time series graph of outcomes (a) hourly volume, (b) mean hourly speed, (c) mean hourly violation speed, (d) maximum hourly speed ), (e) hourly vio;ation rate, and (f) hourly violation rate over 35 MPH. Dashed line indicates start of citation period ( December 11, 2012).
In multivariable models, photo enforcement citations were associated with significant reductions in the hourly speed violation rate, hourly maximum violation speeds, and hourly mean speeds, relative to the warning period (Table 2). The hourly citation rate per 1000 vehicles decreased by nearly 50% during the citation period compared to the warning period (adjusted, standardized incident rate ratio 0.53, 95% CI 0.42 to 0.66). Monday drivers tended to have significantly higher violation rates and speeds than drivers on other days of the week. Winter months had lower hourly traffic volumes (mean difference [MD] −31.7 vehicle per hour, 95% CI −47.4, 16.0) and lower mean vehicle speeds compared to fall months (MD −0.15 MPH, 95% CI −0.25, −0.043). Spring months had slightly higher speed violation rates per 1000 vehicles, maximum violation speeds and overall mean speeds compared to fall months. There were no significant differences observed between morning and afternoon times in multivariable models, thus hour of the day was not included in the final models.
Table 2.
Multilevel multivariable models of each hourly outcome and covariates. Mean differences (MD) and 95% confidence intervals (95% CI).
| Vehicle Volume | Violation Rate | High Violation Rate | Mean Violation Speed |
Max Violation Speed |
Mean Speed | |
|---|---|---|---|---|---|---|
| MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | |
| Fixed | ||||||
| Time Period | ||||||
| Warning | Ref | Ref | Ref | Ref | Ref | Ref |
| Citation | 37.1 (−10.4, 84.7) | −6.13 (−10.1, −2.17) | −0.83 (−2.01, 0.35) | 0.072 (−0.095, 0.24) | −2.14 (−2.87, −1.42) | −1.12 (−1.64, −0.60) |
| Day | ||||||
| Monday | Ref | Ref | Ref | Ref | Ref | Ref |
| Tuesday | 13.9 (8.44, 19.4) | −0.79 (−0.96, −0.61) | −0.19 (−0.42, 0.044) | −0.13 (−0.19, −0.065) | −0.26 (−0.43, −0.096) | −0.26 (−0.31, −0.21) |
| Wednesday | 19.3 (11.5, 27.2) | −0.62 (−0.88, −0.36) | −0.32 (−0.74, 0.098) | −0.10 (−0.18, −0.030) | −0.23 (−0.43, −0.035) | −0.28 (−0.32, −0.24) |
| Thursday | 25.2 (18.1, 32.3) | −0.13 (−0.24, −0.013) | −0.046 (−0.37, 0.28) | −0.058 (−0.17, 0.052) | −0.11 (−0.33, 0.10) | −0.40 (−0.47, −0.32) |
| Friday | 37.2 (25.4, 49.0) | 0.06 (−0.36, 0.48) | −0.15 (−0.46, 0.16) | −0.068 (−0.17, 0.032) | −0.030 (−0.27, 0.21) | −0.23 (−0.35, −0.10) |
| Quarter | ||||||
| Fall | Ref | Ref | Ref | Ref | Ref | Ref |
| Winter | −31.7 (−47.4, −16.0) | 1.73 (1.02, 2.43) | 0.26 (0.065, 0.45) | 0.029 (−0.14, 0.19) | 0.71 (0.27, 1.14) | −0.15 (−0.25, −0.043) |
| Spring | −16.5 (−33.9, 0.90) | 1.35 (0.70, 1.99) | 0.12 (−0.089, 0.32) | −0.011 (−0.11, 0.090) | 0.49 (0.10, 0.88) | −0.62 (−0.96, −0.28) |
| Intercept | 438.6 (326.0, 551.3) | 11.8 (7.34, 16.4) | 1.83 (0.44, 0.32) | 29.5 (29.0, 29.9) | 35.6 (34.4, 37.0) | 25.8 (24.7, 26.9) |
| Random | ||||||
| Camera | 51.3 (26.8, 98.4) | 4.71 (2.95, 7.53) | 1.42 (0.69, 2.94) | 0.086 (0.030, 0.24) | 0.77 (0.44, 1.34) | 0.62 (0.37, 1.03) |
| Intercept | 145.9 (101.4, 210.0) | 5.42 (3.50, 8.39) | 1.45 (0.68, 3.08) | 0.56 (0.34, 0.91) | 1.63 (0.85, 3.09) | 1.33 (0.83, 2.12) |
| Correlation | 0.33 (−0.06, 0.63) | −0.97 (−0.99, −0.92) | −0.95 (−0.99, −0.57) | −1.00 (−1.00, −0.99) | −0.43 (−0.76, 0.068) | 0.14 (−0.68, 0.80) |
| Residual | 142.9 (102.1, 200.0) | 12.8 (10.8, 15.3) | 7.70 (0.59, 1.00) | 1.66 (1.54, 1.80) | 3.71 (3.52, 3.91) | 4.18 (3.67, 4.75) |
Bolded results highlight findings whose 95% CI does not cross 0.
Automated enforcement speed cameras appeared to provide safety benefits in the first year after installation and these safety benefits were sustained in following years. Traffic volumes remained relatively constant in both the first and second citation periods compared to the warning period. The hourly violation rate decreased significantly in both years. In the first year of citations, a reduction of nearly 4 violations per hour per 1000 vehicles (95% CI −7.60 to −0.30) was observed; in the second year, violations decreased by an additional 3 violations per hour per 1000 vehicles (95% CI −4.09 to −2.36) compared to the warning period.
The rate of high speed violations (greater than 10mph over the posted speed limit) also fell after initiation of citation issuance (Table 3), with a drop of 0.32 violations per 1000 vehicles in the second year compared to the warning period (95% CI −0.54 to −0.10). The maximum violation speed decreased significantly in both citation periods, by 1.53 and 0.95 MPH respectively. While the mean hourly violation speed increased slightly (0.11 MPH, 95% CI 0.023, 0.20), the overall mean hourly speed decreased by 1 MPH during the first citation year relative to the warning period and remained steady during the second citation year (95% CI −1.62, −0.52).
Table 3.
Effect of automated enforcement one year after warning period. Multivariable multilevel mixed linear regression to calculate mean differences (MD) and 95% confidence intervals (95% CI).
| Hourly Vehicle Volumea |
Hourly Violation Ratea |
Hourly High Speed Violation Ratea |
Mean Violation Speedb |
Maximum Violation Speeda |
Hourly Mean Speeda |
|
|---|---|---|---|---|---|---|
| MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | MD (95% CI) | |
| Fixed | ||||||
| Time Period | ||||||
| Warning | Ref | Ref | Ref | Ref | Ref | Ref |
| Citation Year 1 | 29.0 (−16.6, 74.7) | −3.95 (−7.60, −0.30) | −0.61 (−1.78, 0.55) | 0.11 (0.023, 0.20) | −1.53 (−2.24, −0.82) | −1.07 (−1.62, −0.52) |
| Citation Year 2 | 11.9 (−0.54, 24.3) | −3.22 (−4.09, −2.36) | −0.32 (−0.54, −0.10) | −0.050 (−0.17, 0.075) | −0.95 (−1.36, −0.54) | −0.072 (−0.46, 0.31) |
| Intercept | 439 (326, 551) | 11.9 (7.36, 16.44) | 1.83 (0.44, 3.21) | 29.5 (29.1, 29.9) | 35.7 (34.4, 37.0) | 25.8 (24.7, 26.9) |
| Random | ||||||
| Random Parameters | ||||||
| Citation Year 1 (SD) | 49.6 (26.2, 93.9) | 4.7 (2.96, 7.55) | 1.42 (0.69, 2.94) | c | 0.70 (0.39, 1.28) | 0.64 (0.42, 0.98) |
| Citation Year 2 (SD) | 16.7 (8.6, 32.3) | c | c | 0.131 (0.059, 0.29) | 0.44 (0.24, 0.81) | 0.45 (0.25, 0.82) |
| Intercept (SD) | 146 (101, 210) | 5.4 (3.50, 8.39) | 1.45 (0.68, 3.08) | 0.54 (0.33, 0.88) | 1.62 (0.85, 3.09) | 1.33 (0.83, 2.13) |
| Correlation | ||||||
| Year 1 & Year 2 | 0.080 (−0.50, 0.61) | c | c | c | 0.14 (−0.57, 0.73) | −0.31 (−0.75, 0.33) |
| Year 1 & Intercept | 0.33 (−0.085, 0.65) | −0.97 (−0.99, −0.92) | −0.95 (−0.99, −0.58) | c | −0.15 (−0.62, 0.40) | 0.23(−0.45, 0.74) |
| Year 2 & Intercept | 0.029 (−0.52, 0.56) | c | c | −0.72 (−0.97, 0.38) | −0.86 (−0.97, −0.49) | −0.24 (−0.80, 0.54) |
| Residual (SD) | 143 (102, 200) | 12.7 (10.69, 15.2) | 7.70 (5.90, 10.0) | 1.66 (1.54, 1.79) | 3.68 (3.50, 3.87) | 4.17 (3.66, 4.75) |
Bolded results highlight findings whose 95% CI does not cross 0.
Adjusted for day and quarter
Adjusted for day
Not included in model
We also compared mean vehicle speeds during school travel periods to non-travel periods, controlling for the day of the week. Drivers traveled nearly 5 MPH less during school travel periods, relative to non-school travel times (MD −4.75 MPH, 95% CI −5.65 to −3.86) and during school travel times, a significant decrease in mean speed was still observed (MD −0.80, 95% CI −1.06 to −0.54) during the citation period compared to the warning period. In terms of vehicle volumes, during non-school travel hours the citation period vehicle volume increased by 6% (95% CI 1.01–1.12).
DISCUSSION
In our study, automated photo enforced speed camera citations in school zones decreased both the rate of speed violations and motorist speeds during school travel times compared to the warning phase. In the absence of speed enforcement citations, it was common for vehicles to travel in excess of 30 MPH, raising the risk of fatal pedestrian collisions. In the warning phase, maximum violation speeds reached 50 MPH, a speed at which most collisions would result in a child being killed if struck.
There are several implications of these findings. First, though violation rates and speeds decreased during the citation period, vehicle volumes varied little and may have even slightly increased. Some critics of automated photo enforcement assert that motorists may seek alternate routes without cameras, but traffic patterns remained fairly constant throughout our study.
We also found that the safety benefits of automated speed enforcement citations were sustained beyond the immediate period after implementation and into subsequent years. There were notable spikes in violation rates immediately following long school breaks; however these spikes decreased in magnitude over time. Post-summer break increases in speed also suggest that automated enforcement should not be removed after lower speeds have been achieved, as motorists need ongoing reminders to follow the speed limit in school zones. It should be noted that while the hourly violation rate decreased during the subsequent year time period, there was a slight increase in the mean violation speed, though the overall mean hourly speed decreased during the initial year of the camera citations. This may be due to motorists traveling slightly above the speed limit subsequently decreasing their speed after receiving a fine, while those who travel at higher speeds may not reduce their speeds as quickly or initially as those already traveling at lower speeds.
The observed decreases in maximum violation speed and hourly mean speed were small but statistically significant. These decreases were comparable in magnitude to the reductions observed following the implementation of automated speed cameras in areas other than school zones, such as highways and other arterial roads.33 Achieving lower speeds is important for reducing collisions and the resulting injuries and fatalities.37 Even decreasing the speed by 5 MPH could result in a significant reduction in severe injuries or fatalities to children.13
This study had several limitations. First, we were not able to compare the warning and citation periods to the period prior to camera installation, as there was no baseline with detailed hourly data on individual vehicle speed. If detailed, hourly, pre-warning period data had been available, we might have observed an even stronger intervention effect on violation rate and motorist speeds. We cannot, however, exclude the possibility that the higher violation and speeds during the warning period drivers were due to confusion by the novelty of the cameras, but considering the constant decrease in speeds and violations over time this possibility does not seem likely. Second, this study provided a conservative estimate of the true change. We did not have a co-synchronous control group, which limited our ability to control for temporal trends. To explore the possible temporal changes, we compared aggregate school travel hourly speeds and volumes with aggregate measurements during non-school travel hours at each sites. While speeds decreased during both school travel hours and non-school travel hours, the mean hourly speed decreased significantly more during school travel hours than non-school travel hours with relatively little impact on vehicle volume. A third limitation is that cameras were installed at four school locations with fairly similar road characteristics. It is possible that the effectiveness of automated enforcement may vary in locations with different road infrastructure from those evaluated. The effect of the cameras on violation rates and speeding was similar across all schools, indicating that automated speed enforcement can work in multiple school, road and neighborhood environments. The City of Seattle has expanded the number of schools with automated enforcement, thus further exploration of new environments will soon be feasible. A final limitation of our study is that we were unable to evaluate the effect of automated enforcement on motor vehicle collisions due to the infrequency of collisions during the study period and the length of the study period. A larger number of schools and a longer enforcement time period may enable this form of evaluation in the future.
Despite these limitations, our findings indicate that the use of automated enforcement speed cameras in school zones may improve child pedestrian safety around schools and may also provide additional benefits for families considering active transportation options. A commonly cited reason for preventing children from walking or biking to or from school is concern about fast moving vehicles.21–23 Reducing motorist speeds near schools may help address those parental concerns, thus could help parents feel better about allowing their children to walk or bike to school. Decreasing the number of speeders and slowing down motorists will increase safety for child pedestrians and cyclists by decreasing the risk of collision and the severity of trauma caused if a crash were to happen.14 Child passengers of motor vehicles traveling through these school zones may benefit from lower motorist speeds as well.
An additional and important result of this work was the generation of new resources to invest in safe and active school transportation. In Seattle, automated speed enforcement in school zones has provided a new revenue stream to make road safety improvements around schools. Seattle required that photo enforcement revenues be used for school road safety improvements and for school road safety education programs. Reducing motorist speed near schools during school hours is an important public health goal that can potentially be achieved with automated speed enforcement cameras.
Supplementary Material
What is already known on this subject?
Motor vehicle speeds have a negative effect on children’s health by increasing the risk of serious injuries and reducing active transportation
Automated speed enforcement has been shown to be effective at reducing motor vehicle speeds in settings such as highways and arterial roads
What this study adds.
Automated photo enforcement of motor vehicle speed significantly reduced the rate of speeding violations by nearly 50%
The mean vehicle speed significantly decreased by 2 MPH during automated photo enforcement
The effects of automated photo enforcement were sustained after 1 year of implementation
Acknowledgements
The City of Seattle and the Seattle City Council approved and initiated this program. American Traffic Solutions, Inc. was contracted by the City of Seattle to collect data and provide the material and technical support and provided raw data to the study authors. Chris Steele and Greg Doss at the Seattle Police Department and Kate Coulson, Ray Pedrosa, Servando Rojas and Bob Lukas at American Traffic Solutions, Inc. provided assistance acquiring the data.
Note: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Funding
Research reported in this article was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under award number T32HD057822.
Footnotes
Competing Interests
None to declare.
Patient Consent
Not required – no human subjects involved.
Ethic Approval
Institutional Review Board of the University of Washington.
REFERENCES
- 1.National Center for Statistics and Analysis. Traffic Safety Facts 2013 Data: Speeding. Washington, DC: National Highway Traffic Safety Administration, 2015. [Google Scholar]
- 2.National Center for Statistics and Analysis. School-Transportation-Related Crashes. Traffic Safety Facts. Washington, DC: National Highway Transportation Safety Administration, 2014. [Google Scholar]
- 3.Dumbaugh E, Frank L. Traffic Safety and Safe Routes to Schools. Transportation Research Record: Journal of the Transportation Research Board 2018;2009(1):89–97. doi: 10.3141/2009-12 [DOI] [Google Scholar]
- 4.Kattan L, Tay R, Acharjee S . Managing speed at school and playground zones. Accid Anal Prev 2011;43(5):1887–91. doi: 10.1016/j.aap.2011.04.009 [published Online First: 2011/06/11] [DOI] [PubMed] [Google Scholar]
- 5.University of North Carolina Highway Safety Research Center Safe Routes to School Online Guide Development Committee. Safe Routes to School Online Guide: Around the School: Pedestrian and Bike Information Center; 2015. [Available from: http://guide.saferoutesinfo.org/engineering/the_school_zone.cfm accessed 11 September 2015.
- 6.National Committee on Uniform Traffic Control Devices. The Manual on Uniform Traffic Control Devices (MUTCD) for Streets and Highways, 2009 Edition Including Revision 1 dated May 2012 and Revision 2 Date May 2012. 2009 ed. Washington, DC: Federal Highway Administration 2012. [Google Scholar]
- 7.Ellison AB, Greaves S, Daniels R. Capturing speeding behaviour in school zones using GPS technology. Road & Transport Research 2013;22(4):30–42. [Google Scholar]
- 8.Fitzpatrick K, Brewer MA, Park ES. Suggestions Concerning School Traffic Control Devices. Transportation Research Record 2010(2149):1–10. doi: 10.3141/2149-01 [DOI] [Google Scholar]
- 9.Tay R Speed Compliance in School and Playground Zones. Ite Journal-Institute of Transportation Engineers 2009;79(3):36–38. [Google Scholar]
- 10.Warsh J, Rothman L, Slater M, et al. Are school zones effective? An examination of motor vehicle versus child pedestrian crashes near schools. Inj Prev 2009;15(4):226–9. doi: 10.1136/ip.2008.020446 [published Online First: 2009/08/05] [DOI] [PubMed] [Google Scholar]
- 11.Clifton KJ, Kreamer-Fults K. An examination of the environmental attributes associated with pedestrian-vehicular crashes near public schools. Accident Analysis and Prevention 2007;39(4):708–15. doi: 10.1016/j.aap.2006.11.003 [DOI] [PubMed] [Google Scholar]
- 12.Kröyer HRG, Jonsson T, Várhelyi A. Relative fatality risk curve to describe the effect of change in the impact speed on fatality risk of pedestrians struck by a motor vehicle. Accident Analysis & Prevention 2014;62(0):143–52. doi: 10.1016/j.aap.2013.09.007 [DOI] [PubMed] [Google Scholar]
- 13.Tefft BC. Impact speed and a pedestrian’s risk of severe injury or death. Accid Anal Prev 2013;50:871–8. doi: 10.1016/j.aap.2012.07.022 [published Online First: 2012/09/01] [DOI] [PubMed] [Google Scholar]
- 14.Rosen E, Stigson H, Sander U. Literature review of pedestrian fatality risk as a function of car impact speed. Accid Anal Prev 2011;43(1):25–33. doi: 10.1016/j.aap.2010.04.003 [published Online First: 2010/11/26] [DOI] [PubMed] [Google Scholar]
- 15.Oluyomi AO, Lee C, Nehme E, et al. Parental safety concerns and active school commute: correlates across multiple domains in the home-to-school journey. Int J Behav Nutr Phys Act 2014;11(1):32. doi: 10.1186/1479-5868-11-32 [published Online First: 2014/03/08] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Stewart O Findings from Research on Active Transportation to School and Implications for Safe Routes to School Programs. Journal of Planning Literature 2011;26(2):127–50. doi: 10.1177/0885412210385911 [DOI] [Google Scholar]
- 17.Smith L, Sahlqvist S, Ogilvie D, et al. Is a change in mode of travel to school associated with a change in overall physical activity levels in children? Longitudinal results from the SPEEDY study. International Journal of Behavioral Nutrition and Physical Activity 2012;9 doi: 10.1186/1479-5868-9-134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Flynn MAT, McNeil DA, Maloff B, et al. Reducing obesity and related chronic disease risk in children and youth: a synthesis of evidence with ‘best practice’ recommendations. Obesity Reviews 2006;7:7–66. doi: 10.1111/j.1467-789X.2006.00242.x [DOI] [PubMed] [Google Scholar]
- 19.Janssen I, LeBlanc AG. Systematic review of the health benefits of physical activity and fitness in school-aged children and youth. International Journal of Behavioral Nutrition and Physical Activity 2010;7 doi: 10.1186/1479-5868-7-40 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.McDonald NC, Brown AL, Marchetti LM, et al. U.S. School Travel, 2009: An Assessment of Trends. American Journal of Preventive Medicine 2011;41(2):146–51. doi: 10.1016/j.amepre.2011.04.006 [DOI] [PubMed] [Google Scholar]
- 21.Carver A, Timperio A, Crawford D. Playing it safe: The influence of neighbourhood safety on children’s physical activity—A review. Health & Place 2008;14(2):217–27. doi: 10.1016/j.healthplace.2007.06.004 [DOI] [PubMed] [Google Scholar]
- 22.Panter JR, Jones AP, van Sluijs EMF. Environmental determinants of active travel in youth: A review and framework for future research. International Journal of Behavioral Nutrition and Physical Activity 2008;5 doi: 34 10.1186/1479-5868-5-34 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mitra R Independent Mobility and Mode Choice for School Transportation: A Review and Framework for Future Research. Transport Reviews 2013;33(1):21–43. doi: 10.1080/01441647.2012.743490 [DOI] [Google Scholar]
- 24.Evers C, Boles S, Johnson-Shelton D, et al. Parent safety perceptions of child walking routes. Journal of Transport & Health 2014;1(2):108–15. doi: 10.1016/j.jth.2014.03.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Elvik R Area-wide urban traffic calming schemes: a meta-analysis of safety effects. Accid Anal Prev 2001;33(3):327–36. doi: 10.1016/s0001-4575(00)00046-4 [published Online First: 2001/03/10] [DOI] [PubMed] [Google Scholar]
- 26.Retting RA, Ferguson SA, McCartt AT. A Review of Evidence-Based Traffic Engineering Measures Designed to Reduce Pedestrian-Motor Vehicle Crashes. Am J Public Health 2003;93(9):1456–63. doi: 10.2105/ajph.93.9.1456 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bunn F, Collier T, Frost C, et al. Area-wide traffic calming for preventing traffic related injuries. Cochrane Database Syst Rev 2003(1):CD003110. doi: 10.1002/14651858.CD003110 [published Online First: 2003/01/22] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Simpson CL. Evaluation of Effectiveness of School Zone Flashers in North Carolina. Transportation Research Record 2008(2074):21–28. doi: 10.3141/2074-03 [DOI] [Google Scholar]
- 29.Hirst WM, Mountain LJ, Maher MJ. Are speed enforcement cameras more effective than other speed management measures? An evaluation of the relationship between speed and accident reductions. Accid Anal Prev 2005;37(4):731–41. doi: 10.1016/j.aap.2005.03.014 [published Online First: 2005/05/07] [DOI] [PubMed] [Google Scholar]
- 30.Jones AP, Sauerzapf V, Haynes R. The effects of mobile speed camera introduction on road traffic crashes and casualties in a rural county of England. J Safety Res 2008;39(1):101–10. doi: 10.1016/j.jsr.2007.10.011 [published Online First: 2008/03/08] [DOI] [PubMed] [Google Scholar]
- 31.Morain SR, Gielen AC, Bhalla K. Automated speed enforcement systems to reduce traffic-related injuries: closing the policy implementation gap. Inj Prev 2016;22(1):79–83. doi: 10.1136/injuryprev-2014-041507 [published Online First: 2015/05/27] [DOI] [PubMed] [Google Scholar]
- 32.Pilkington P, Kinra S. Effectiveness of speed cameras in preventing road traffic collisions and related casualties: systematic review. BMJ 2005;330(7487):331–4. doi: 10.1136/bmj.38324.646574.AE [published Online First: 2005/01/18] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wilson C, Willis C, Hendrikz Joan K, et al. Speed cameras for the prevention of road traffic injuries and deaths. Cochrane Database of Systematic Reviews 2010; (11). http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004607.pub4/abstract. [DOI] [PubMed] [Google Scholar]
- 34.Carnis L, Blais E. An assessment of the safety effects of the French speed camera program. Accid Anal Prev 2013;51(0):301–9. doi: 10.1016/j.aap.2012.11.022 [published Online First: 2013/01/10] [DOI] [PubMed] [Google Scholar]
- 35.De Pauw E, Daniels S, Brijs T, et al. An evaluation of the traffic safety effect of fixed speed cameras. Safety Science 2014;62:168–74. doi: 10.1016/j.ssci.2013.07.028 [DOI] [Google Scholar]
- 36.Retting RA, Farmer CM, McCartt AT. Evaluation of automated speed enforcement in Montgomery County, Maryland. Traffic Inj Prev 2008;9(5):440–5. doi: 10.1080/15389580802221333 [published Online First: 2008/10/07] [DOI] [PubMed] [Google Scholar]
- 37.Stevenson M Safety cameras: more, not fewer, saves lives. Inj Prev 2012;18(1):1–2. doi: 10.1136/injuryprev-2011-040288 [published Online First: 2011/12/22] [DOI] [PubMed] [Google Scholar]
- 38.Nichols JL, Tippetts AS, Fell JC, et al. The effects of primary enforcement laws and fine levels on seat belt usage in the United States. Traffic Inj Prev 2014;15(6):640–4. doi: 10.1080/15389588.2013.857017 [published Online First: 2014/05/29] [DOI] [PubMed] [Google Scholar]
- 39.Cummings P Methods for estimating adjusted risk ratios. The Stata Journal 2009;9(2):175–96. [Google Scholar]
- 40.Greenland S Estimating standardized parameters from generalized linear models. Stat Med 1991;10(7):1069–74. [published Online First: 1991/07/01] [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.