6. DISCUSSION

The primary objective of the present study was to develop and test a behavior based program to reduce motorist speed in residential areas by adding education and enforcement to engineering. The Heed the Speed program that was mounted consisted of education, special police enforcement, application of innovative pavement markings and conventional vertical traffic calming treatments. It was evaluated by two waves of a knowledge and opinion survey, police data collection during stops for speeding and up to five waves of 48-hour speed measurements.

In addition to demonstrating speed reductions, in order to conclude that the Heed the Speed program was effective, it is necessary to show that one or more of its components was applied with sufficient intensity to yield a speed reduction. Further, it must be shown that the affected populations in the six test neighborhoods were aware of the countermeasures and likely responded to them. Thus, positive survey and process results are necessary for a conclusion that any observed decreases in the speed data were the result of Heed the Speed efforts.

The process data clearly indicate that the program countermeasures were applied with sufficient vigor to have been noticed by the residents of the test neighborhoods. The large number of police stop reports and high quantities of literature and signs distributed as well as the physical changes applied – both new vertical treatments and 3-D and Tyregrip™ markings – are compelling evidence of the activity levels of the Phoenix and Peoria Heed the Speed programs. This was confirmed by the survey results that showed large increases in awareness and in the perception that neighborhood speeds had moderated.

In the context of the survey, an additional point is noteworthy. The survey response rate of almost 61 percent is extremely high for a mailed questionnaire, particularly one addressed to “Resident.” This is likely the result of a strong prevailing interest in traffic safety in general and speed management in particular among those dwelling in the test neighborhoods. This is not surprising as the participating cities selected the test sites based on previously expressed interest in achieving a speed reduction.

The number of police stops, both with and without the issuance of citations for speeding, suggests that the increased awareness of police activities can be attributed to Heed the Speed . It is unfortunate that there is no accurate record of the patrol hours assigned to each test road segment as that would have supported a more detailed examination of the relationship between increased enforcement presence and speeding.

The police stop data provide two other interesting findings. First, the vast majority of those stopped for speeding were residents of the test neighborhoods or regular users of its streets. This dispels any notion that aberrant speeding behavior stems primarily from outsiders who do not have a stake in the safety of the neighborhoods through which they travel. It also suggests that localized programs such as Heed the Speed can be effective because most of the targeted drivers will be exposed repeatedly to any deployed countermeasures.

The second interesting finding from the police stop data is the high rate of safety belt use. It is counterintuitive and contrary to much of the safety literature to find 85 percent safety belt use among drivers who are speeding. The neighborhood locale of the stops and the relatively high socioeconomics of the test sites may have contributed to the high observed safety belt rate. Regardless of its origin, however, the existence of such a high safety belt use rate suggests that the target population for speed countermeasures, at least in residential neighborhoods such as the six used in this study, is likely safety conscious and desirous of being law-abiding. Thus, safety messages and the other types of activities employed by Heed the Speed should be more influential with this group than with a population less concerned with safety.

The results presented earlier 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 (compliance with the limit and speeds 7+ over the limit) were still statistically lower than prior to the Heed the Speed countermeasures at Wave 5, almost five months after the program began.

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. This was because the various countermeasures were phased in as they were ready while periodic speed measurements were being made. The discussion that follows refers back to the speed data by test segment contained in the tables in Section 5.3.3. The focus in this discussion will be on mean speed changes since the other speed measures appeared to be consistent with changes in the average observed speed.

There were no pre/post speed measurements collected when the only intervention mounted was enforcement. Several education measures, such as the traffic and lawn signs, were implemented at the outset and therefore were coincident with the special enforcement efforts. The survey and follow-up discussions with neighborhood leaders, however, suggest that the enforcement effort was visible and appreciated by the residents who were interested in better control of speeds. The large number of stops recorded also supports a conclusion that the enforcement efforts contributed meaningfully to the speed reduction results achieved.

Although not in total isolation, enforcement was a primary countermeasure in operation at the Clarendon Avenue test site in Phoenix . This neighborhood had no organized residents' association, and the predominance of Spanish-speaking residents limited the penetration of some of the educational materials. Thus, the reduction of approximately 2.5 mph in mean speed between the speed humps on Clarendon can be largely attributed to the special police enforcement and its deterrent effect.

The Peoria police used essentially a neighborhood rotation approach to enforcement while Phoenix employed a more random deployment. The evaluation data do not provide any clear insights on which of these approaches is preferable. Perhaps a combined strategy in which there is a strong effort followed by random “boosters” would be best.

As with enforcement, there was no “pure” application of education in the test. The closest situation was likely the change between Wave 1 and Wave 2 at both 91 st Avenue and 95 th Avenue in Peoria . During this period, the Peoria police were focusing their enforcement primarily on 84 th Avenue in Bell Park . Taking this view, the education efforts on 91 st and 95 th Avenues can be credited with the observed reductions of 1.283 mph and 1.369 mph, respectively, in mean speeds and reductions of 37.8 percent and 45.9 percent , respectively, in the percent of drivers going 7+ mph over the speed limit. It is therefore reasonable to conclude that the Heed the Speed education materials contributed positively to the overall program effectiveness.

The data provide many replications of wave-to-wave speed measurements in which the operative countermeasures applied over the transition were education and enforcement in combination. These were the only countermeasures deployed at the outset of Heed the Speed because there was a delay in achieving the installation of the planned vertical treatments and innovative street markings.

The combination of education and enforcement only failed to yield a significant speed reduction in one instance – between Waves 1 and 2 at 84 th Avenue in Peoria . This may have been a result of the almost single-minded focus of the neighborhood leaders on getting speed tables installed as soon as possible. As a result, they devoted most of their effort to lobbying the city rather than on distributing literature to the neighborhood. In all of the other cases in which education and enforcement operated alone, a significant mean speed reduction was achieved. The largest observed decline in mean speed under these conditions was 3.610 mph on Sweetwater Avenue in Phoenix between baseline (Wave 1) and the next measurement wave (Wave 4). The smallest (but still statistically significant) decline in mean speed for education and enforcement alone was 0.693 mph observed between the baseline (Wave 1) and second measurement wave (Wave 4) on the North/South segment of Moon Valley Drive in Phoenix.

These results lead to the conclusion that combining education and enforcement can provide synergism under the proper conditions. The decline in mean speed in excess of 3.5 mph on Sweetwater is similar in magnitude to the reductions achieved by vertical treatments, although likely not as enduring. It is greater by far than what was apparently achieved by either education or enforcement alone.

The study included two new installations of vertical roadway treatments – speed tables on 84 th Avenue in Peoria between Waves 2 and 3 of the speed measurements and speed humps on the North/South segment of Moon Valley Drive in Phoenix between Waves 4 and 5. The former was associated with a drop in wave-to-wave mean speed of 3.163 mph while the latter wave-to-wave reduction was 2.540 mph. It must be remembered that these reductions are in addition to any changes in speed from the previous waves. On 84 th Avenue , there was no significant change in speed from Wave 1 to Wave 2. On Moon Valley , however, there was a significant drop of 0.693 mph in mean speed from baseline (Wave 1) to Wave 4 (the first wave after the program start). Thus, the 2.540 mph decline after the installation of the speed humps was actually a total mean speed reduction of 3.233 mph from baseline.

It is not surprising to find that vertical engineering treatments are effective in achieving traffic calming. This has been reported before (e.g., Ewing , 1999). Their application within this study, however, provides both a replication of the effects reported in the literature and a benchmark against which to compare the other countermeasures used.

The study employed two types of innovative pavement markings – 3-D markings and Tyregrip™ surfacing. The phasing of the installations provided some opportunities to examine the additive effects of these materials in some degree of isolation.

The 3-D marking was the only physical change installed on 85 th Lane where there was no significant change in speeds as discussed earlier. The baseline speeds on this quiet road were so low, that no reasonable speed reduction approach could be expected to work. The same material, however was installed on 95 th Avenue in Peoria and on the East/West segment of Coral Gables Drive in Phoenix after Wave 2 of measurements. For 95 th Avenue , the first available post-installation measurement was Wave 4 because of a failure of the data collection equipment during Wave 3. The Wave 4 measurements show a statistically significant decrease from Wave 2 of 0.564 mph on top of the previously observed decline of 1.369 mph for a total reduction in mean speed of 1.933 mph. On Coral Gables , the decline after Wave 2 is 0.215 mph on top of the 2.174 mph decline observed between Wave 2 and baseline (total decline of 2.389 mph). The apparent effect of the 3-D installation, however, continues to climb with an additional reduction of 1.777 mph in mean speed from Wave 3 to Wave 4 for a total Wave 4 reduction of 4.166 mph.

The Tyregrip™ material was installed between Waves 4 and 5 of speed measurement on three test segments in Phoenix – Sweetwater Avenue and both segments (East/West and North/South) of Coral Gables Drive in Phoenix . On Sweetwater, Wave 5 showed slightly higher mean speeds than Wave 4 but was still 3.213 mph below baseline (Wave 1). Although no additional speed reduction was achieved, the Tyregrip™ installation may have slowed a return to higher speeds by keeping the increase to 0.397 mph. A similar phenomenon was observed on the East/West segment of Coral Gables Drive where a small (0.489 mph) increase in mean speed was observed after Tyregrip™ was installed. It must be remembered that this segment was already associated with the previously discussed decline of 4.166 mph and the prior installation of 3-D markings.

Perhaps the least confounded test of the Tyregrip™ material took place on the North/South segment of Coral Gables Drive . On this test road where Tyregrip™ was the only installed physical change, it was associated with a mean speed reduction of 1.045 mph from Wave 4 to Wave 5 (1.470 mph decline from baseline).

Based on these findings, it is reasonable to conclude that innovative pavement markings such as the ones tested as part of Heed the Speed are a viable addition to the speed reduction techniques available to program managers and merit serious consideration, particularly when used in combination with other types of speed reduction countermeasures. It might have been even more effective to integrate the innovative markings more closely with the remainder of the program, particularly the education materials. This could help give the innovative markings a dual purpose as both the illusion of an impediment in the roadway and as a symbolic reminder to slow down.

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. New countermeasures were still being implemented on test segments just prior to the last speed measurement. Comments from the members of the Heed the Speed steering committee indicated that they believed reapplications of enforcement and education would be needed to sustain their effects. The consensus was that the vertical treatments would continue their effectiveness as they represent a true physical barrier to excess speed. Feelings were split with respect to the innovative markings. Some believed their effect would fade rapidly as motorists learned that they were merely illusions and not true barriers. Others felt that they would continue to have some effectiveness as a reminder that the neighborhood was trying to control speeds. Additional research would be needed to yield a definitive resolution of how long Heed the Speed continued to be successful.

One potential counterproductive effect of traffic calming is the diversion of traffic to neighboring streets that have not been calmed. This is a natural behavior as drivers seek the path of least resistance to get to their destination quickly. In this study, there was no consistent pattern of reduced traffic flow after the implementation of any of the types of countermeasures deployed. Traffic on a few road segments went down while traffic on others increased. Most importantly, however, in the context of the current research is the clear finding that the speed reduction results observed could not possibly have been the result of diversion of higher speed traffic. Simply, in all cases, the reduction in cars going 7+ mph over the limit was either in excess of any concomitant decrease in traffic volume or so large that it would have necessitated all or most of the fastest vehicles to have been diverted. Thus, while there may have been some minor diversion, particularly on 84 th Avenue in Peoria after the installation of speed tables, there is no reasonable possibility that this diversion confounded the speed results reported herein.

It is also worth noting that no speed measurements were taken on the streets bordering the 10 test segments. It is theoretically possible that the fastest drivers simply changed their travel routes to avoid the enforcement activities or physical changes on the test segments. However, since there was not a major reduction in traffic volumes, this hypothesis could only be supported by a net increase in traffic in the test neighborhoods together with a significant diversion of high speed drivers from the test segments. The process data and anecdotal reports from the police strongly suggest that this did not happen. In particular, it must be remembered that survey respondents were drawn generally from the six neighborhoods and not just from the streets that were test segments. These survey respondents generally reported that they perceived lower speeds in their neighborhoods. It is likely that at least some responses would have noted a shift in the locus of speeding if, in fact, one had occurred.

It is of interest to attempt to assess the safety benefits of the observed speed reductions. Unfortunately, as discussed in Section 2.1.2, there do not appear to be any simple formulas for estimating either crash reduction or the lessening of injury severity that will result from lowered vehicle speeds in residential neighborhoods. Even the fatality probabilities cited from the study in the United Kingdom (Department of Transport, 1997) cannot be directly applied to aggregate data such as those collected by this study because they only deal with the conditional case in which a crash has occurred. Although the current study documented a significant and operationally meaningful reduction in the number of vehicles traveling at the highest speeds on the test segments, there is no way to estimate how many of these vehicles would actually have been crash-involved. Any such estimate would have to include consideration of many other factors such as the environment on any particular street (e.g., the existence of visual screens, the prevailing speeds before the application of countermeasures). It would also have to include some measure of the likely pedestrian behavior on the street and the age distribution of the pedestrians.

The absence of any formulas to translate shifts in the speed distribution of vehicles to crash and injury effects means that any safety implications must necessarily be subjective. One way to accomplish this is by determining which of the previously proved conditions for safety improvements were, in fact, achieved by Heed the Speed . For example, Tester et al. (2004) showed that the speed reductions associated with traffic calming engineering treatments were associated with reduced child pedestrian crash risk. The level of speed change studied by Tester et al. (2004) was likely achieved on 84 th Avenue in Peoria and the Moon Valley North/South segment as a result of the installation of speed tables or speed humps. Likewise, speeds on Clarendon were reduced below what had been achieved from speed humps alone. Even 85 th Lane in Peoria , although its speeds did not change during the study, was already at a calmed level and therefore met the conditions studied by Tester et al. (2004). Thus, at least four of the test segments achieved or exceeded the speed reduction proved to alter child pedestrian crash probability favorably.

The balance of the test segments also exhibited significant speed reductions. In at least two cases (the East/West segment of Coral Gables and Sweetwater) the reduction in mean speed was even greater than on the segments that had been calmed with engineering changes, but the starting mean speeds were also notably higher. Based on the findings of Kallberg (1997), it would appear likely that these segments would also have experienced reductions in both crash probability and resulting injury severity if a crash did occur.

In summary, while there is no readily apparent way to make a quantitative estimate of the safety benefits of the Heed the Speed efforts, the guidance from the literature suggests that the results were of a sufficient magnitude to yield meaningful reductions in both pedestrian crash risk and likely injury severity if a crash occurs. Since the desirable speed changes obtained typically involved thousands of vehicles per day on each segment, it is reasonable to conclude that safety was enhanced. It is also noteworthy that the evidence from Kallberg (1997) and Lindenmann (2004) among others is that the speed reduction effects would also have yielded a safety benefit for all traffic crashes on the treated road segments not just pedestrians.

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.

In the process of conducting this study, several principles emerged that appeared to be strongly associated with the success of Heed the Speed . These were:

• A multi-disciplinary steering committee of individuals who truly believe in the benefits of reducing speeds in neighborhoods is essential. This committee must have representation from enforcement, engineering, and public relations and be fully supported by the local government.
  
  
• For maximum impact, it is beneficial to have a well-organized and firmly committed neighborhood organization in any community where the program is to be applied.
  
  
• Yard signs can be effective in getting the word out to residents, but participating homeowners need to be cautioned about potential vandalism and spare yard signs should be provided.
  
  
• Using special enforcement stops for speeding as a way to train new police officers in the techniques for making traffic stops provides an additional benefit from the program.
  
  
• Multi-lingual education materials are essential in communities where there are significant subgroups of non-English-speaking residents.
  
  
• Addressing special groups such as car dealers or high school students can be productive and is worth the effort if there are data to indicate they may be part of the problem.
  
  
• If the program is going to have a limited duration, implement all countermeasures together whenever possible.
  
  
• In order to potentially give the program extended life after the major enforcement, education, and engineering countermeasures have waned, include some aspect that is durable in the environment. This was accomplished in Heed the Speed by including the fixed street signs that were left in place.
  
  
• Consider the use of innovative approaches such as the pavement markings employed during Heed the Speed . If these types of approaches are used, integrate them as completely as possible in the program's education material at the inception of the program.
  
  
• Consider including a long-term plan for continuing to monitor and control speeds after the initial campaign. This should overcome any regression to the mean types of phenomena.
  
  
• Include multiple, longitudinal speed measurements as part of the effort. These will provide valuable feedback on program effectiveness. The survey for this study also suggested that the deployment of speed measurement tubes is noticed and may even serve as a countermeasure by itself.

The lessons learned from this research application of Heed the Speed have been turned into a guide for future program implementations. This Guide is included as Appendix F to this report. It is important to view the recommendations in the Guide in the proper context. In particular, the application of Heed the Speed in six separate neighborhoods in two cities as part of this research clearly showed that localization is of the utmost importance. For example, although there was a clear benefit to the availability of an active neighborhood group, significant improvements were obtained even when there was no active community body. Therefore, it is prudent to work with the available resources and adapt Heed the Speed rather than attempting to change the local organization to meet any strict definition of a Heed the Speed implementation.

All of the tests in this study were guided and overseen by local representatives from the engineering, enforcement, and education communities. Thus, for example, even when new engineering was not implemented, the countermeasure planning process had the benefit of engineering input. This multi-disciplinary focus appears to have been one of the factors that contributed to the success of the Peoria and Phoenix Heed the Speed implementations. The question therefore arises whether the involvement of all three disciplines is a necessary condition for future Heed the Speed implementations. This is difficult to answer in the absolute. On the one hand, a broad base of experience and resources should aid success. On the other hand, there almost certainly will be locales in which a viable speed reduction program can be implemented without one or more of the central disciplines. For example, a strong enforcement program by itself should result in a reduction in speeding even if it is not as effective as the same enforcement effort coupled with education and/or engineering countermeasures.

Perhaps the most universal advice for using the guide is, therefore, to begin the entire process with a catalog of local resources. While the ideal is for every relevant discipline to join willingly and eagerly in Heed the Speed efforts, the reality is that local conditions will vary widely. As long as some involved group – citizens, enforcement, engineering or safety education – is eager to proceed, a successful Heed the Speed program is likely possible. Soliciting the help of all of the other participants, however, is clearly desirable and should be attempted. Once all of the players are identified and fully committed, the information in the Guide in Appendix F can be viewed in the proper context and adapted appropriately.