DEPARTMENT OF TRANSPORTATION
NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION
There is abundant evidence that higher speeds are associated with more severe pedestrian injuries and death (Leaf and Preusser, 1999). One study has reported that 5 percent of pedestrians will die if struck by a vehicle traveling at 20 miles per hour, 45 percent will die if the striking vehicle is traveling at 30 miles per hour, 85 percent will die if the vehicle is traveling at 40 miles per hour, and almost all will die if the vehicle is traveling at 50 miles per hour (Department of Transport, 1997). In addition, a National Highway Traffic Safety Administration (NHTSA) national survey has revealed that drivers feel that speeding is more dangerous on residential roads than on other roadways, and that drivers believe that higher posted speeds will result in increased danger to neighborhood residents ( Boyle, Dienstfrey, and Sothoron, 1998) .
There has been significant work on engineering approaches to traffic calming as a means to reduce neighborhood speeds (c.f., Ewing, 1999; Stuster, Coffman, and Warren , 1998) . NHTSA has noted, however, that there have been few attempts to combine public information and enforcement techniques with engineering changes as a means of achieving greater speed reductions. NHTSA therefore funded the current study whose objective was to develop and test a behavior-based program to reduce motorist speed in residential areas by adding education and enforcement to engineering.
It was determined that the study should focus on identifying the following types of neighborhoods/streets for this study:
The study was conducted in two neighboring Arizona cities – Phoenix and Peoria . Three neighborhoods were selected for study in each city. In Peoria , they were the Desert Harbor/91 st Avenue neighborhood, the Bell Park neighborhood, and 95 th Avenue . In Phoenix , they were Clarendon Avenue , Sweetwater Avenue , and the Moon Valley/Coral Gables neighborhood.
Heed the Speed was selected as the slogan for the speed reduction program and a logo was designed. Several educational materials were developed for the program including street and lawn signs that displayed the slogan and logo; print materials for homeowners, parents, and drivers (including high school drivers); press releases for the two cities; inputs for homeowner's association newsletters; and radio spots. These materials provided information on the relationship between various speeds and pedestrian injury severity in the event of a crash. Earned media included write-ups in local papers and participation in local television programs. Enforcement activities included special enforcement, training citizens to conduct neighborhood watches and deployment of speed trailers and photo speed enforcement trailers. Roadway treatments included installation of speed humps and tables and application of visual treatments to the roadway surface that gave an illusion of a roadway impediment or engineering treatment.
Program logo containing program slogan
Program lawn sign
Four types of data were collected for the study – data collected by the police during routine stops made during special enforcement, information obtained from a survey of neighborhood residents, speed data on the various road segments, and process data obtained from project and neighborhood representatives. Each is described below.
Police Special Enforcement: Additional police patrol hours were incorporated in each neighborhood. During special enforcement, any vehicle traveling faster than the posted speed was subject to being stopped. When a stop was made, the officer approached the vehicle, announced the violation, and performed a routine check on the driver's license. For each driver stopped, the officer completed a motorist stop form that provided a profile on the drivers speeding in the various neighborhoods. The form showed that a majority of speeding drivers lived in or within a mile of the neighborhood in which they were stopped ( Phoenix , 63%; Peoria , 56%). Most traveled on the road on which they were stopped at least once per week ( Phoenix , 74%; Peoria , 76%). The average age of the drivers was 40. Most drivers were traveling alone ( Phoenix , 73%; Peoria , 70%). A large percentage of drivers wore their safety belts ( Phoenix males, 81%, and females, 89%; Peoria males, 75% and females, 93%). Tickets were given initially only to drivers who flagrantly disobeyed the law. As the study progressed, more traditional forms of ticketing were used. Five percent of the stopped Peoria drivers were given tickets for speeding; 52 percent of those stopped in Phoenix were given speeding tickets. Literature on the dangers of speeding was given to 97 percent of the drivers stopped in Peoria and 41 percent of those stopped in Phoenix .
Neighborhood Survey: A one-page survey form was designed to assess awareness of the program by neighborhood residents. Specifically, residents were asked if they noticed program components (publicity, police involvement, and roadway treatments) and if they noticed changes in vehicle speeds in the neighborhood. The survey was mailed to half of the neighborhood residents prior to the start of the program and to the other half as a follow-on after completion of the program (in Peoria ) and after the installation of roadway treatments (in Phoenix ). Pre-program and follow-on survey response rates in Peoria were 58 percent and 62 percent, respectively. Comparable returns in Phoenix were 60 percent and 63 percent.
All neighborhoods in both cities exhibited a significant increase in awareness of program publicity during the follow-on survey except for one neighborhood in Phoenix (Clarendon) that had no neighborhood association or vocal spokespersons and is predominately Spanish-speaking. There was a significant increase in awareness of publicity related to speeding with three neighborhoods showing a doubling or more in the percentage of people indicating they were exposed to publicity on speed control. The specific publicity item that received the largest proportion of post-program mentions in both cities consisted of the program signs (with reports ranging from 52% to 83% of the respondents), except for the Clarendon neighborhood where the percentage was 13. Flyers, bulletins, and newsletters were also mentioned, as were mail/letters, meetings, and newspapers, although the mentions for some neighborhoods were not large.
With regard to police involvement, the specific activities mentioned most frequently and considered to be definitely or probably program-related were visible enforcement (ranging from 51% to 63%) and radar checks (ranging from 14% to 46%). Deployment of the speed trailer was noted in all Peoria neighborhoods. The photo speed enforcement trailer was noted in the Moon Valley/Coral Gables neighborhood of Phoenix .
With regard to roadway treatments, signs again were mentioned frequently by respondents from all neighborhoods except for Clarendon. In addition, respondents in each neighborhood noticed the roadway visual treatments if they had been installed in the respondent's neighborhood as part of the program
Respondents were asked to rate the speeds in their neighborhoods compared to six months ago. With the exception of Clarendon, the majority of respondents in the remaining neighborhoods reported a perceived decrease in speeding.
A subjective assessment of each survey response by a project staff member led to the judgment that more than half of the survey respondents were aware of the key elements of the Heed the Speed program. This suggests that there is a basis for attributing any significant speed reductions in the test neighborhoods across the study period to the effects of Heed the Speed.
Speed Data: The six neighborhoods selected for study involved 10 studied road segments. Speed data were collected on all 10 segments by means of automated traffic counters. The devices consisted of multiple pneumatic tubes placed on the roadway to record the presence of a vehicle and its speed. Five waves of data were collected in Peoria except for 95 th Avenue where four waves were collected. Three waves were collected in Phoenix except on one segment that had five waves. The following measures were used to examine speed:
(Speed data are reported herein to three decimal places for convenience since that is the format of the output from the statistical software employed. The presence of three decimals is not intended to imply either that the measurements were that precise or that a precision of three decimal places was required for statistical inference testing.)
Since the primary objective of the statistical analysis was to assess the relationship between the various speed countermeasures and reductions in speed, ANOVA was the indicated primary statistical technique. Chi-square tests were used to evaluate effects on certain binary and nominal transforms of the speed scale, e.g., proportion driving seven or more miles over the speed limit. The speed data were analyzed as a set of 10 case studies. Results are summarized in the table on the next page and discussed briefly below together with a description of each test road:
Peoria – 84 th Avenue : The 84 th Avenue test segment is part of the Bell Park neighborhood. It is a typical residential street with a 25 mph speed limit. It received a moderate-to-high education effort and some enforcement. Both speed tables and an innovative visual treatment known as 3-D markings were installed as part of the program. The results were quite dramatic. The initial drop in mean speed following the installation of speed tables (3.088 mph) was almost completely sustained two months later (2.627 mph). Also, driver compliance to the posted speed limit more than doubled, and speeds 7+ mph over the limit decreased by about half.
Summary of Countermeasures Implemented and Speed Reductions by Test Segment
3-D markings on pavement between speed tables on 84 th Avenue in Peoria
Peoria – 85 th Lane : This is a quiet side street off 84 th Avenue in Bell Park with a 25 mph speed limit. It had three speed humps installed well before the study began, but the residents wanted another hump because they thought the existing installations were too far apart. There was some education on this street from the yard and fixed signs, but little enforcement because of the low traffic volumes and benign speed profile. The 3-D markings were installed during the program. There was no meaningful change on 85 th Lane, which is not surprising given the low baseline speed (19.675 mph) and the extremely small number of drivers going seven or more miles over the limit (5 and 6 in the first and last measurements, respectively).
Peoria – 91st Avenue : The 91 st Avenue test segment is a 78 foot wide street containing four travel lanes and a 28 foot wide landscaped center median. The street is conducive to high speeds and has a 30 mph speed limit. No roadway treatments, either physical traffic calming or visual treatment, were applied. The education program was intensive and sustained because of the high degree of interest of the neighborhood association. Enforcement was also applied because of the known propensity for speeding and the pressure on the police from the residents. The initial education and enforcement efforts produced a significant reduction in mean speed of 1.283 mph. After a return toward baseline, the renewed efforts of the neighborhood association and police revived the initial impact. Although the effect was relatively small compared with the streets that had physical changes, it appears to have come almost entirely from drivers in the highest speed class. The proportion of vehicles going over the limit but less than seven miles per hour over stayed virtually constant while the percentage at or under the limit increased and those going seven or more miles per hour over the limit decreased. The mean speed above the speed limit was also down significantly in all post waves.
Peoria – 95 th Avenue : This is a straight roadway with a 30 mph speed limit. In addition to a moderate-to-high level of education and moderate enforcement, five sets of 3-D markings were placed on the roadway during the study. The initial education and enforcement produced a noteworthy effect with a drop in mean speed of 1.369 mph and a particularly large drop in the number of high-speed drivers and the mean of their speeds. This effect increased to 1.933 mph with the addition of the 3-D markings and was still evident, although somewhat diminished (1.029 mph), at the end of the data collection period.
Phoenix – Clarendon: Clarendon is a street with a 25 mph speed limit that had existing speed humps. Residents were complaining that the existing vertical treatments were spaced too far apart thereby allowing motorists to speed between the humps. There is no neighborhood association for Clarendon, and a large proportion of the population is Spanish-speaking. As a result, the education campaign likely had moderate intensity even though many of the materials were translated into Spanish. Police enforcement was applied, but at a lower level than at other sites since the baseline speeds were well below the posted limit. The mean speeds at the two measurement locations between the humps showed a significant drop (2.506 mph) that was sustained through two post data collection waves. The reduction in drivers going seven mph or more over the prevailing 25 mph limit (from 16.7% to 5.1%) was particularly noteworthy.
Phoenix – Sweetwater : Sweetwater Avenue received intensive education and enforcement spurred by the efforts of a vocal and active neighborhood association. In addition, Tyregrip™ material (a pavement surface visual treatment) was installed during the study period. The results showed almost a 50-percent reduction in the number of drivers going seven mph or more over the posted limit. The reduction was largest in the last data collection wave. The mean speed of the speeders also was reduced markedly in the last two data collection waves when compared with the baseline (3.610 mph and 3.213 mph, respectively). The reduction in the mean speed above the limit of speeders from 8.017 mph in the next-to-last wave to 7.634 mph in the last wave was also statistically significant (p<.001). It should be noted that across the three waves of measurement the shift in speeds was largest to the lowest category (under the speed limit). This category more than doubled. At the same time, the intermediate category of people going less than seven mph above the limit went from 28.8 percent to 42.3 percent, an increase of 46.9 percent. The mean of speeders also showed continued improvement. This suggests that the Tyregrip™ installation may have yielded a further moderation in the most aberrant speeds even though the mean speed rose slightly. Since there was only a single wave of measurement after the appearance of the Tyregrip™, no inference can be drawn concerning its long-term effectiveness.
Phoenix – Moon Valley Drive East/West Segment: Moon Valley Drive is a 25 mph speed limit road with a segment that runs east and west on which only education and enforcement were applied. Significant mean speed reductions (1.578 mph) were achieved without the use of any physical measures – a reduction in average speed and a shift to the lower class intervals from the seven or more mph over the limit category. A particularly interesting aspect of the results at this site was that the effect was maintained and even increased significantly between the last two waves of data collection. This suggests the possibility of some persistence of the education plus enforcement approach.
Phoenix – Moon Valley Drive North/South Segment: This segment of Moon Valley Drive running north and south received essentially the same education and enforcement as the east/west segment but also had two speed humps installed during the study program. The speed limit on this segment is also 25 mph. The results indicated that the effect of education and enforcement alone on speed on this roadway, although significant (0.693 mph), was somewhat less than that on the east/west segment of Moon Valley Drive . The influence of the addition of a speed hump before the last data collection wave was large (3.233 mph) and consistent with what was seen on 84 th Avenue in Peoria .
Phoenix – Coral Gables Drive East/West Segment: The east/west segment of Coral Gables Drive was treated with 3-D and Tyregrip™ materials. There were also education and enforcement countermeasures. Overall speed was below the baseline in excess of two mph in all post waves. The last two waves show markedly lower speeds by speeders. The addition of Tyregrip™ was associated with a significant increase of 0.489 mph in overall speed but a significant decrease of 0.434 mph in the mean speed of speeders between the last two data collection waves.
Phoenix – Coral Gables Drive North/South Segment: The north/south portion of Coral Gables Drive is wide and lined with expensive single-family homes. This segment received enforcement and education as well as a Tyregrip™ installation between the last two data collection waves. Results showed a reduction in mean speed of as much as 1.470 mph and a decrease in the percentage of speeders traveling seven mph or more over the speed limit of 44.8 percent.
Tyregrip™ installation on Coral Gables Drive in Phoenix
Summary of Speed Results : The large samples of speed data indicated a consistent and operationally meaningful pattern of reductions in all neighborhoods and on all test segments except 85 th Lane in Peoria . It is particularly noteworthy that the changes in the desired direction represent thousands of vehicles per day at each site. Clearly, the physical changes such as speed tables and speed humps produced the largest drops in speed. This was expected from previous research findings (e.g., Ewing , 1999). The general pattern of results across all of the test roads consisted of a decrease in mean speed accompanied by a moderate-to-large reduction of drivers exceeding the speed limit by seven mph or more and a concomitant increase in drivers proceeding at or below the speed limit. The mean speed above the speed limit and the mean speed of speeders also declined significantly at all sites except 85 th Lane in Peoria , the site with low traffic volumes and preexisting speed humps.
Summary of Traffic Volume Results: Traffic volumes varied across the various study measurements without a consistent pattern. Increases, decreases, and static volumes were noted. There is nothing in the traffic volume results, however, that suggests that the observed speed reduction effects were the result of diverting higher speed drivers to other, non-studied streets where they would continue to be a hazard. On 84 th Avenue , volume dropped 29.7 percent from the first (baseline) wave to the last measurement. Over the same time period, the number of vehicles traveling 7+ mph over the speed limit decreased by 80.3 percent. On Sweetwater, the number of vehicles counted declined by 4,862 from the baseline to the last measurement while the number of vehicles measured at 7+ mph over the limit dropped by 8,606.
Process Data: Discussions on the program were held with project representatives and with representatives from the neighborhoods under study. The Heed the Speed program elicited mostly positive and constructive comments for improvement from both those who helped mount it and the citizens it affected. Together with the survey data, these follow-up commentaries on the process suggest that the concept of a multi-disciplinary, multi-faceted neighborhood speed reduction program was well accepted and could even be improved in future implementations based on the lessons learned.
Discussion: The results show conclusively that Heed the Speed produced significant changes in speeding behavior in the six test neighborhoods. Nine of the 10 individual road segment analyses showed statistically significant and operationally meaningful reductions in speed after the application of the Heed the Speed countermeasures. The one segment that did not show a reduction, 85 th Lane in Peoria , had baseline mean speeds almost 5 mph below the speed limit.
On all nine segments that showed speed reductions, speeds were still statistically lower than prior to the Heed the Speed countermeasures at the final measurement wave, almost five months after the program began. In some cases, speeds seemed to be sliding back towards baseline. In several others, however, the speed reductions appeared to be holding or even improving as the program progressed.
The design of this study was not intended to isolate the effects of the individual countermeasures employed. Rather, the objective was to mount a multi-pronged effort focused on achieving speed reductions in the test neighborhoods. The realities of the implementation of the Heed the Speed program, however, afforded some insights into the effects of some of the individual interventions:
Countermeasure Persistence: The present study was only able to examine very short-term persistence of the countermeasures. Even over this limited period there were conflicting findings. In four of the nine successful tests mean speed reductions were higher in the last measurement wave than in any of the preceding waves. In the remaining five tests, the last measurement represented some increase in speed from the lowest mean value obtained. In one of these five, although the last wave was not the slowest, it was significantly slower than the immediately preceding wave. This pattern of results suggests that the present study simply did not have a good view of countermeasure persistence – short, intermediate or long term. Additional research would be needed to yield a definitive resolution of how long Heed the Speed continued to be successful.
Program Costs: Exclusive of the evaluation and associated research, Heed the Speed was implemented in all six neighborhoods and 10 road segments for approximately $55,000 in out-of-pocket expenditures. These costs covered police overtime, acquiring lawn signs and stakes and printing education materials. Program participants in Peoria and Phoenix believe that subsequent Heed the Speed neighborhood programs can be mounted for $5,000 per neighborhood or less.
Safety Implications of the Findings: There is no simple formula to relate speed reduction results to either reduced crashes or decreased injury severity. The clear finding from the literature review is that distributional shifts in the speeds of vehicles on residential streets such as those obtained in this study should provide for both a lower crash incidence and greater survivability if a crash does occur. When these shifts are applied to thousands of vehicles per day, as was the case on these test roads, a safety benefit is highly likely. Further, the success of Heed the Speed on nine of 10 test road segments suggests that the widespread implementation of the program would have a significant societal benefit. The low cost of the program means that achieving the potential safety benefits of Heed the Speed should be highly cost effective.
Lessons Learned: The answer to the basic research question addressed by this study is that education and enforcement can add to the effectiveness of physical traffic calming. It was also shown that at least in the short term a program such as Heed the Speed can produce speed reductions of a significant and meaningful magnitude on through streets and other roads within traffic calmed areas that were not candidates for physical treatments.
HS Form 321