Effectiveness of Behavioral Highway Safety Countermeasures (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

22 The overall goal of this project is to help states allocate their money effectively. This chapter summarizes the infor- mation developed thus far into guidelines for doing so, as follows: 1. It is critically important to focus on countermeasures that have been proven to be effective (i.e., Proven countermea- sures); that can be accomplished within existing financial and political constraints; and that will provide the greatest benefit/cost return on invested resources. 2. It is appropriate to examine those countermeasures that are judged Likely to be effective, even though there may not yet be firm evidence that they reduce crash injury. Some of these may be quite appropriate for inclusion in your high- way safety plan, although it may be necessary to conduct additional evaluations of their effectiveness, since cur- rently available data are not definitive. 3. To the extent possible, a state should avoid countermea- sures that are unlikely to be effective or for which the effects are unknown (i.e., Unknown/Uncertain/Unlikely effective countermeasures). This will be quite a challenge since nearly half of the 104 listed countermeasures are in this category. To the extent countermeasures in this group are considered, priority should be given to those with plus (+) ratings, indicating there is some basis sug- gesting they may work, although not enough to place them in the Likely group. 4. It is important to stay away from measures that are known to have negative consequences (i.e., those Proven Not to Work). This may also be difficult due to commonly accepted views of such actions (e.g., providing driver ed- ucation as a more expedient way to license young novice drivers). 5. A state may want to explore newly developing, innovative countermeasures that are untried. Guidelines are suggested for how to assess such new countermeasures. 1. Identify Proven Injury Reducing Countermeasures That Can Be Implemented There are 23 countermeasures that are rated as Proven and for which injury reduction estimates are available (see Appendix B). Any of these measures that have not yet been implemented in a given state should be consid- ered first for inclusion in a state’s highway safety plan. The countermeasures in this Proven classification are as follows: 1. School pedestrian training, 2. Booster seat programs, 3. Bike helmet law for children, 4. Motorcycle helmet use law, 5. Primary seat belt law, 6. Short, high-visibility belt law enforcement, 7. GDL, 8. Extended learners permit, 9. Night restrictions for young novice drivers, 10. Passenger restrictions for young novice drivers, 11. Administrative license revocation, 12. Mass media in support of alcohol (or other) enforcement, 13. Mandatory attendance at alcohol treatment programs, 14. Sobriety checkpoints, 15. Alcohol safety interlocks, 16. Multi-component community programs to address under- age drinking, 17. Speed cameras (automated speed enforcement), 18. Red-light cameras (automated red-light enforcement), 19. License suspension, 20. Individual meetings for traffic violators, 21. Group meetings for traffic violators, 22. Warning letters for traffic violators, and 23. Reduced speed limits (for pedestrian safety). C H A P T E R 7 Using this Guide

It is recognized that it may not be feasible to implement some of these measures at the present time. There may be political, legal, or financial constraints that make current implementation difficult, if not impossible. Alternatively, im- plementation may not be possible at the present time because the likely quality or intensity of activity that your state can pro- vide is not sufficient to fully implement the countermeasure. For example, in the Voluntary Action Group, booster seat promotions and parent management programs are rated as Proven, but this rating is based on programs that are of very high quality. Successful programs in these areas have been quite sophisticated, involving techniques associated with more effective PI&E programs (e.g., interactive methods, in- puts from multiple sources, long-term programs, financial incentives, etc.). Lesser-quality programs that are short term, or that are based on didactic approaches or posters in schools, for example, are not likely to increase booster seat use, or to influence parents to better manage their teenagers’ driving practices. In general, every PI&E program proposed, whether it is used alone or in conjunction with other measures, needs careful formative evaluation and development to ensure that it is of sufficiently high quality and that it is based on behav- ior change principles. There is a second-level judgment to be made as well. PI&E programs that work tend to have high development and implementation costs, and one consideration in funding such programs is whether or not they are intended to be repeated. For example, a booster seat program run one time in one community may increase use, but can it be repeated and/or sustained, and is it a program that other communities can adopt? This is somewhat of a Catch-22 situation, because in order to work, such a program may have to be so extensive and thorough (with associated costs) that it is outside the reach of other communities. On the other hand, there are more modest PI&E programs for children, such as the Willy Whistle pro- gram for teaching young children how to cross streets, that have modest costs associated with them and could readily be applied in school systems across the state. Still, it should be remembered that the savings associated with effective programs are substantial and that high-cost programs that produce substantial effects can be a stimulus for other such programs. There have been demonstration programs, for example, that have had high costs but that have also be- come models for other, often large-scale, effective efforts. For example, the seat belt enforcement program in Elmira, New York, was a high-cost effort, but it demonstrated that vigor- ous enforcement of belt use laws could be done with high public acceptance. This paved the way for the North Carolina CIOT Program and eventually led to the nationwide launch of high-intensity belt use enforcement programs (i.e., the national CIOT mobilizations). Another issue that must be considered in selecting Proven countermeasures is how long the program must be maintained to be effective or cost effective. This is a major consideration for some of the programs in the Laws Plus Enhancements Group. For example, sobriety checkpoints are a proven tech- nique, but to work they have to be sustained over time. If checkpoints are run for just a short period of time, such as during a single holiday period, they are likely to have little or no long-term effect. Short-term media bursts to supplement laws also have very limited effects. In establishing an effective countermeasure program (and selecting countermeasures for that program), it is also important to take advantage of favorable trends. Graduated licensing provides a good example of a trend that is both popular and effective at the present time. Research is now avail- able that has established the importance of the core elements of graduated licensing (extended learner permit, night and passenger restrictions, are all rated Proven). Some states do not have these provisions or have weak versions of them. Estab- lishing or upgrading the key provisions of GDL is an impor- tant step forward. Because of the current trend of activities and public support, this may be an opportune time for a state to assess and upgrade its novice driver GDL program. Some countermeasures are both highly cost effective and highly controversial. Automated speed and red-light enforce- ment provides a case in point, especially in reference to speed cameras. Controlling high speeds is an important goal, and speed cameras are a proven countermeasure in accom- plishing this goal. It is important to note, however, that community programs involving camera technology can be designed in ways that are publicly acceptable. Such programs exist in Scottsdale, Arizona; Charlotte, North Carolina; and Montgomery County, Maryland, and can be used as models. General guidelines also are available for conducting speed- control programs that are effective and acceptable, using such techniques as focusing on “extreme” speeders and speeding in school zones (Harsha and Hedlund, 2007). Some highly effective countermeasures simply may not be possible given existing state law or political climate. Sobriety checkpoints, for instance, are not legal in some states, and a motorcycle helmet law or a primary seat belt law simply may be out of reach given the current legislature. The result of this process will be the selection of Proven countermeasures for inclusion within the state’s highway safety plan that can be accomplished within the political, legal, and financial conditions that currently exist within the state. In addition, to be effective, the state must recognize that these countermeasures must share two characteristics: (1) that they can be implemented; and (2) that they can be implemented to the level necessary to achieve crash and in- jury reduction. 23

Implementation, in some cases, need not be statewide. Urban jurisdictions will have a greater need for pedestrian countermeasures than their rural counterparts; rural areas may be more concerned with speeding than are urban areas; college communities will be more concerned with underage drinking than other parts of the state. Problem identification (i.e., identification of the who, what, when, where, and why of any particular target group of fatal and nonfatal injuries) should be used to cost effectively deploy programs. 2a. Use Countermeasures That Are Likely To Be Effective Thirteen of the 104 countermeasures are listed as Likely to be effective. Each of these countermeasures fits within the effectiveness guidelines for its classification (i.e., Voluntary Action; Laws; Laws Plus Enhancements; or Sanctions and Treatments Groups). This means that, based on what has been seen with other similar countermeasures, these countermea- sures should be effective, if properly or fully implemented. These Likely to be effective countermeasures include the fol- lowing (references are provided in Appendix C): 1. Responsible beverage service; 2. Parent guiding teen licensing; 3. Ice cream vendor ordinances; 4. Local primary seat belt laws or ordinances; 5. Adult bicycle helmet laws; 6. License renewal policies for elderly drivers; 7. License actions for underage alcohol violations; 8. Integrated enforcement (e.g., alcohol, seat belts, speeding); 9. Zero-tolerance enforcement; 10. Vendor compliance checks for age-21 enforcement; 11. Sustained seat belt enforcement; 12. Increased belt use law penalties; and 13. Simplifying and streamlining DUI statutes. Each of these Likely countermeasures has been imple- mented in one or more jurisdictions in the past and most have been evaluated with positive results. However, the evaluations of these countermeasures have typically focused on intermedi- ate measures of effectiveness (e.g., number of bicycle helmets distributed, number of drinks served, number of arrests, in- creases in seat belt use, etc.), often within limited areas of a state or community. These countermeasures may be pre- sumed to be effective, although the full benefit they provide as measured by injury reduction is not fully established. It is suggested that Likely countermeasures be considered side by side with Proven countermeasures for which benefit information is not available. Both tend to positively impact some intermediate measure of highway safety that should, eventually, reduce injury. 2b. Consider Proven Countermeasures with No Effectiveness Estimates There are some countermeasures that are Proven yet the evaluation evidence does not provide estimates of crash re- duction. These countermeasures are discussed and refer- enced in Appendix D. 1. Speed limits, 2. BAC test refusal penalties, 3. Saturation patrols for alcohol-impaired driving, 4. Preliminary breath test devices, 5. Passive alcohol sensors, 6. Restrictions on plea bargaining, 7. Court monitoring, 8. Brief interventions (alcohol), 9. Vehicle immobilization, 10. Vehicle impoundment, 11. Close monitoring of DUIs, 12. PI&E supporting belt law enforcement, and 13. Child bicycle helmet promotions. Speed limits, or more particularly, “rational speed limits,” can limit the number of speed violators and reduce the variance between vehicles traveling at the fastest speeds and vehicles traveling more slowly on a particular stretch of road. This will make speed-related crashes less likely. However, there is no direct formula for creating a crash reduction esti- mate. Moreover, for most jurisdictions, the number of miles of roadway for which rational speed limits are applicable is limited. Therefore, statewide implementation of this coun- termeasure is not possible. This should be viewed as a local countermeasure. As indicated, the next four countermeasures on the list shown above all deal with increasing the arrest rate of drink- ing drivers and/or collecting more complete evidence. Each of these should be thought of as enhancements to the process of finding, arresting, and prosecuting offenders. Obviously, improvements in the process are desirable and this process has strong general deterrence potential. However, as above, there is no direct formula for creating a crash reduction estimate. The next group of six countermeasures on the list deal with increasing the probability of conviction and the severity of sanction for drinking drivers. Certainty and severity of sanc- tion are the desirable goal. However, again, there is no direct formula for creating a crash reduction estimate from these countermeasures. Child bicycle helmet promotions increase use to varying degrees, but do not yield estimates of injury reductions. Sim- ilarly, PI&E in support of belt law enforcement can increase use but does not readily convert to injury reductions. 24

Decisions to implement these countermeasures will need to be made without direct estimates of their potential crash reduction benefits. Many will play an important role in the development of a highway safety plan. However, final deci- sions will need to be made on identified need for process improvements and judgment as to how these measures may be used to fill those needs. 3. Avoid Countermeasures with Unknown and Unlikely Effectiveness Many of the Unknown/Uncertain/Unlikely countermea- sures are in the Voluntary Action Group. The education and information approach that characterizes these countermea- sures is popular. This group involves more than one-third of all the countermeasures. It has, by far, the highest percentage of countermeasures of unproven effectiveness (84%). Many voluntary action programs have long been known to be ineffective, but their popularity remains. Exhorting people to take some preferred health action for their own benefit is easy to do and gives the appearance of doing something important for society; sometimes these efforts are referred to as feel-good programs. A somewhat cynical appraisal of this situation was presented by Stone (1989, p. 891), who noted that “. . . health education is cheap, generally uncontrover- sial, and safe: if it works, the politicians take the credit, and if it does not, the target population takes the blame.” In this context, the following conclusions of Williams (2007a, p. 9) bear repeating: Never assume that a PI&E program will be successful. In fact, most PI&E programs do not lead to a measurable reduction in crashes or injuries. . . . Never assume that a PI&E program will do no harm. Some well-meaning educational programs, albeit a very few, actually lead to more crashes and injuries. Moreover, the implementation of a program that does not work will limit the amount of resources available for programs that can make a difference. Avoidance of ineffective countermeasures is by no means limited to the voluntary action group, but applies to the other three groups as well. Passing laws that have limited public acceptability or that are not readily enforceable are two ex- amples. Enforceability is a key issue. Much current attention in the United States has been given to distracted, fatigued, and aggressive driving. However, enacting laws against these behaviors, although it may be a popular approach, is likely to have little effect because they are difficult for police to enforce. Laws against hand-held cell phones constitute a separate distracted driving issue. General cell phone laws are listed as having uncertain impact. The behavior is observable by police, but the evidence indicates minimal decreases in use unless the law is enforced (McCartt, Hellinga, and Geary, 2006). How- ever, research has indicated that the distraction is the cell phone call itself, not whether it is hands free or hands on. States choosing to implement countermeasures in the Unlikely/Unknown group need to emphasize those listed with plus (+) ratings. Whatever the choice, states will need to conduct a comprehensive evaluation and take corrective action at the first sign of trouble. Again, resources spent for measures that do not work will limit the resources available for things that will work. 4. Do No Harm Some programs with potentially negative effects (Proven Not to Work) have become popular. For instance, programs that teach teenagers advanced driving maneuvers, typically how to handle skids. The logic here is that standard driver education courses generally concentrate on basic driving skills. However, it is known that graduates of these courses are prone to being involved in crashes that reflect driving inex- perience (e.g., running off the road and not being able to recover) so it is important to teach new drivers skills that will help them avoid these types of crashes. That is the premise, but research in Scandinavia and the United States has found that skid courses have a negative effect. Drivers who take these courses—especially young males—have more crashes than untrained drivers (Jones, 1993; Glad, 1988). This is perhaps because of the age factor. That is, drivers who get this train- ing may become overconfident and drive more aggressively, showing off to their friends, etc. (see Williams and Ferguson, 2004). Despite these research findings, skid courses offered by manufacturers, police organizations, and others have prolif- erated in recent years in the United States. Another example is offering traffic violator school, in lieu of fines and points on the license. The “in lieu of” qualifier is important here. These schools have not been shown to pro- vide a safety benefit that outweighs the benefits provided by traditional “fine and point” sanctions (Peck and Gebers, 1991; McCartt and Solomon, 2004). School in addition to fines and points is a different matter and can be effective. A third example is driver education that speeds the process of young driver licensing or programs that provide a “time dis- count” for their completion (i.e., earlier licensure for those who have taken driver education). Speeding the licensure process and therefore increasing exposure to crashes at an early age has, repeatedly, been shown to increase the number of young driver crash involvements (Vernick, Li, Ogaitis, et al., 1999; Roberts, Kwan, and Cochrane Injuries Group, 2006; Mayhew, 2007). Any “value added” that the completion of driver education may provide does not outweigh the negative effects of licensing teens at a younger age. 25

Selection of Countermeasures The relative effectiveness of various countermeasures and countermeasure groups has been described, along with con- siderations regarding their selection. This information should be considered by HSOs in the development of their highway safety plans. The next step is to actually select countermeasures from these lists of potential countermea- sures. One way to begin is to first identify total highway loss, as well as the proportion of that loss represented by each target group for which a Proven countermeasure can be implemented. Table 9 shows the total fatal and nonfatal highway loss for a typical 600-fatality state. This table represents national averages. Table 9 can, and probably should, be generated specifi- cally for your state based on your fatality data. Small states should average the last 5 years; large states might use the last 3 years. The advantage to averaging across several years is that it provides a more statistically stable estimate for each target group. The disadvantage associated with using too many prior years, however, is that the further back in time a state goes, the less representative the numbers will be of its current situation. This is particularly important if the state has recently implemented an important, potentially impacting countermeasure, such as a primary belt law, the effects of which are only apparent for the recent past (e.g., the last 18 months). Even if a state does not use its own data averaged across the last few years, it will still need to scale Table 9 to reflect its annual number of fatalities. Thus, if a state has 1,200 fatalities per year, then all of the figures in Table 9 need to be doubled. States with 300 fatalities need to cut the figures in Table 9 by half. Scaling is easily accomplished by taking your number of fatalities (say 450); dividing by 600 (450/600 = 0.75); then multiplying this number (0.75) times every dollar estimate in the table. Similarly, a state with 750 fatalities would get a mul- tiplier of 1.25 (750/600 = 1.25). The next step, after the dollar estimates in Table 9 have been adjusted for your state, is to array the potentially viable countermeasures against the respective target groups. Twenty-three Proven countermeasures are described in Appendix B, along with percent crash reduction estimates. Each has an intended target group and each indicates the expected effectiveness when fully and properly imple- mented. Simply array the countermeasures against the ap- propriate target group; then go to the last column in the table (total highway loss) and multiply the total by the expected level of effectiveness. The result is the expected benefit for your state. For example, alcohol-impaired drivers cost the typical 600-fatality state $537,874,321 per year. Sobriety checkpoints that are implemented across the state for the full year, with full media support, would be expected to pro- vide a 20% injury reduction. Multiply 20% by $537,874,321 (or your state’s adjusted total cost of deaths and injuries) to calculate the estimated savings, which in this case is $107 million. Sobriety checkpoints were selected as the first example since they are one of the most expensive countermeasures a state might consider. Proven countermeasures that can be implemented at relatively less cost include child bike hel- met law, GDL, extended learner permit, night restrictions, 26 Type of highway loss Crashes involving Fatal injuries Nonfatal injuries Fatal Nonfatal Total All persons 600 75572 $669,492,000 $2,285,140,088 $2,954,632,088 Pedestrians 67 2116 $75,110,214 $63,971,882 $139,082,096 Distracted 60 13997 $66,663,455 $423,239,815 $489,903,270 Drowsy 19 1412 $21,101,197 $42,704,002 $63,805,199 Speed 162 14178 $180,836,003 $428,711,216 $609,547,219 Aggressive 164 14770 $183,442,252 $446,600,914 $630,043,166 Alcohol related 248 11583 $276,723,360 $350,234,416 $626,957,776 Alcohol impaired 213 9937 $237,404,168 $300,470,153 $537,874,321 Teen drivers 32 7681 $35,969,377 $232,263,878 $268,233,255 Age-16 drivers 12 3318 $13,816,260 $100,328,595 $114,144,855 Elderly drivers 44 4232 $49,220,426 $127,978,295 $177,198,721 Motorcycles 65 2245 $72,528,300 $67,871,854 $140,400,154 Nighttime 214 14109 $238,550,290 $426,621,478 $665,171,767 Child in car 14 9168 $15,590,393 $277,215,254 $292,805,648 Front seat occupants 376 58720 $419,433,394 $1,775,583,025 $2,195,016,419 Bicycles 11 1153 $12,089,228 $34,859,723 $46,948,950 Table 9. Fatal and nonfatal highway loss for typical 600-fatality state.

passenger restrictions, license suspension, and warning letters. Countermeasures that can be implemented at comparatively little direct cost to the state (e.g., offender-pay counter- measures) include mandatory attendance at alcohol treat- ment, alcohol interlocks, and red-light running cameras. All of these efforts should be given serious consideration in the development of a state highway safety plan. Cost benefit ratios for these countermeasures, if calculated, would be enormous. Next, there are several countermeasures for which the benefit/cost ratio measured in terms of dollars is large but that may be difficult to implement given the current political climate. These countermeasures include universal mandatory motorcycle helmet law, primary seat belt law, speed cameras, and reduced speed limit. Each of these countermeasures, if not already in place, should be implemented as soon as it is feasible to do so. Several countermeasures fall between these two cost ex- tremes. These include child pedestrian training, booster seat promotions, high-visibility enforcement, ALR, mass media in support of alcohol enforcement, community programs for under age 21 enforcement, individual meetings, and group meetings. For each of these, it will be necessary for a state to estimate its own costs for “fully” implementing these programs. Once the costs have been estimated, the benefit/cost ratio can be estimated using the expected effec- tiveness (shown in Appendix B) and the cost data shown in Table 9. A state may conclude that it would be better served by im- plementing a program on a regional basis instead of statewide. All of the benefit and implementation cost numbers can be scaled accordingly just by entering the total number of deaths for that region, relative to those of the “typical” state (to get an estimate of total costs) and then entering the number of deaths for the targeted population in the region involved. Programs that may fall into this category of regional or sub-group im- plementation include child pedestrian safety (urban), red-light running (urban and suburban), community under age-21 enforcement (college towns), and reduced speed limit for pedestrians (urban). This exercise will result in a series of countermeasures that the state intends to include in its highway safety plan. At least some of these countermeasures will be accompanied by benefit/cost estimates. Other countermeasures may be implemented with little cost or on a user-pay basis such that the benefit/cost ratio is not relevant. Note that all of the selected countermeasures—Proven, Proven without injury reduction estimates, and Likely—can be assessed with respect to the target groups listed in Table 9. Measures attempting to deal with a large target group will be favored, other things being equal, over measures dealing with a much smaller target group. Shift in Strategy It should be noted that the above strategy for countermea- sure selection differs from current guidelines. Currently, the recommended approach is to (1) conduct problem identifi- cation, (2) identify priority areas, (3) formulate goals and objectives for each priority area, then (4) select countermea- sures that will lead to the accomplishment of those goals. The present analysis suggests that this is not the optimal approach. Specifically, it is highly likely that there will be no effective and cost-effective countermeasures available to achieve significant impact in some program/problem areas. Even if a potentially effective countermeasure is available, it may not be possible to implement that countermeasure at this time. For instance, it may not be possible at this time to enact a univer- sal mandatory helmet law for all motorcycle riders. Or, in your state it may not be possible to conduct sobriety checkpoints. Similarly, speed cameras, a primary seat belt law, or restrictions on plea bargaining may not be viable options at this time. The present analysis suggests that the availability of effective/ cost-effective countermeasures that can realistically be imple- mented at this time is a major limiting factor. Highway safety plans should attempt to avoid objectives that are unattainable, given current technology and political climate. Rather, they should focus funds and other resources primarily on those objectives that can be addressed. Therefore, the strategy rec- ommended in this document is for the state to (1) conduct problem identification (as above), (2) identify priority areas (as above), (3) identify cost-effective countermeasures that can be implemented, and then and only then (4) formulate realistic objectives based on the expected effectiveness of the identified countermeasures. Consider, for instance, the objectives of the health care delivery system, which includes doctors, hospitals, medical insurance companies, etc. This system is presented with a range of medical conditions. Some of these conditions can be treated, given current technology, and some cannot. The ob- jective of the system is to treat those that can be treated. The best that can be accomplished for the remainder is to attempt to deal with the symptoms since the underlying problem can not be solved given current medical technology. The limiting factor is the availability of a course of treatment. Although the desire may be to cure everyone, the system can only realisti- cally cure those medical conditions that are treatable now, at this time, in this place, with available technology. Similarly, although highway safety professionals may wish to save everyone, they can only realistically deal with those deaths and injuries for which there is an available counter- measure that can be implemented in their state, at this time. This list of effective countermeasures is the limiting factor and, as such, it is the driving consideration in the develop- ment of a highway safety plan. 27

New and Emerging Countermeasures The final highway safety plan will leave some target groups that are not well addressed by the selected countermeasures. This could lead to the possibility that the state may have an interest in pursuing countermeasures that are not yet fully developed, let alone implemented and evaluated. Such new countermeasure development may take the form of an unso- licited proposal from some vested constituency or it may result from innovative efforts from staff or perhaps serendipitous effects from other ongoing efforts. The potential impact of new and emerging countermea- sures should immediately be assessed by first classifying them as: Voluntary Action; Laws, Regulations, Policies; Laws Plus Enhancements; or Sanctions and Treatments. Then, applying the same principles as we have for known existing counter- measures, the state can estimate how the implementation of such countermeasures is likely to affect the fatality, injury, and cost aspects of its traffic safety problem. Some of the most important characteristics, requirements, and opportunities associated with the above listed categories are as follows: • Voluntary Action – Must be of high quality and intensity; – Works best when:  Targeting children;  Allowing communicator some control over audience;  Communicating new knowledge; and  Serving as part of some larger community-based effort. • Laws, Regulations, Policies – Must be well known to the public; – Must be enforceable, based on easily observable and objective criteria; and – Must apply to entire targeted population, not to just a subset of the population. • Laws Plus Enhancements – Enhancement must be well publicized; – May involve special equipment to aid officers, prosecu- tors, probation officers, etc.; and – Generally involves intense selective and concentrated enforcement. • Sanctions and Treatments – Sanction must be well known to violators; – There must be an immediacy and certainty to imposi- tion of the sanction; and – There should be a high degree of intrusiveness to the violator (either through penalty or extent of mandated treatment). New countermeasures that fit with the above guidelines and/or are similar in design to countermeasures that are known to be effective are much more likely to work than measures that do not conform to these guidelines. Note that the majority of countermeasures that have ever been tried have not stood the test of time. Most have not met their ob- jectives. Some have had negative consequences. States imple- menting new and untried efforts must evaluate progress and take corrective action including halting those countermea- sures that are not producing the desired behavioral change. It is important to redeploy resources to effective programs as soon as possible. Conclusion This report has focused on moving resources from ineffec- tive and/or marginally effective countermeasures to those with the largest benefit/cost payoff. Sometimes, there are other goals and objectives that need to be considered for at least a portion of available highway safety dollars. It may be found, for instance, that a particular counter- measure is not likely to work alone, nor will it enhance the impact of another measure, but it could aid in the recruit- ment of additional advocates and, as such, facilitate adoption of something that does work. One example is the networking and outreach that occurred just prior to the enactment of seat belt use laws. There is little evidence that such activity mea- surably increased seat belt use, but there are indications that it led to the public awareness and acceptability of seat belt laws before and after enactment. Perhaps the “take one” pam- phlet, police overtime, or out-of-state training (and associ- ated travel) should be funded, if it is not very expensive and if it will help cement ties between highway safety and some other key agency. There will always be a fine line between implementing Proven countermeasures, satisfying key constituencies, and developing new countermeasures to meet emerging issues. This guide focuses on Proven strategies in the hope that states will shift resources as much as is reasonably possible and pru- dent to countermeasures that work and away from counter- measures that don’t work or are unlikely to work given our current understanding of behavioral change strategies. 28