The Impact of Legislation, Enforcement, and Sanctions on Safety Belt Use (2007)

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.

14 Canadian Influence During the 1980s, several Canadian studies showed signifi- cant increases in observed usage associated with enforcement. Some of these programs were conducted locally, such as in Ottawa, Ontario (Jonah, Dawson, and Smith, 1982; Jonah and Grant, 1985; and Grant, 1991) and some were conducted at the provincial level, such as in Quebec (Dussault, 1990; Grant, 1991). Nearly all of the programs evaluated were STEPs, periodic waves of intensified enforcement accompanied by publicity to make the public aware of the enforcement. This approach, sometimes called HVE, would come to have an important influence on U.S. efforts to increase SBU. Daytime STEPs in Ontario Some of the earliest Canadian studies were conducted in the regional municipality of Ottawa-Carleton (RMOC), Ontario. A series of three studies examined the impact of enforcement waves of varying duration. In the first STEP (4 weeks), usage increased from 58% to 80% (+22 points), then declined to 70% after 6 months (Jonah, Dawson, and Smith, 1982). In a second study (see Figure 5), a series of three STEPS of varying length (1 month, 4 days, and 1 week) increased usage from 66% to 84% (+18 points). Usage declined to some extent between each wave, but then increased to slightly higher levels with each subsequent implementation.36 Figure 5 shows the cumulative impact of successive waves of activity. Longer periods of enforcement (4 weeks and 1 week) were reported to be more effective than the shorter period (two waves of 2 days each). Based on their results, Jonah and Grant suggested that an optimal program might begin with a 4-week enforcement effort, followed by quarterly follow-ups, each involving 1 week of enforcement. They also suggested that higher fines and/or demerit points may be needed to reach those motorists least likely to buckle up. A third study, conducted in 1987, found that another 4-week STEP was associated with an 8-point increase in usage, from 79% to 87% (Grant, 1991). Usage increased during day- light hours and in the evening, but, as Figure 6 shows, this day- time enforcement effort had no effect on drivers leaving drinking establishments late at night (10:00 p.m. to 1:30 a.m.). Their use remained at about 61%, much lower than that of daytime and evening use. Nighttime Enforcement in Nova Scotia Another Canadian study was conducted in Nova Scotia and reported by Malenfant and Van Houten (1988). It involved nighttime enforcement designed to influence patrons leav- ing drinking establishments in two cities, Moncton and Hali- fax. Enforcement occurred on Thursday and Friday nights, between 9:00 p.m. and 2:00 a.m. Although checkpoints were conducted near the drinking establishments, only a small pro- portion of violators received citations.37 The most prevalent C H A P T E R 3 Enforcement 36A control city, Kingston, was identified for comparison purposes. Usage changed very little in Kingston over the study period. 37Checkpoints (in this case, sobriety checkpoints) constitute an enforcement approach whereby vehicles pass through a designated segment of the roadway and vehicles are stopped according to some predetermined rate or pattern (e.g. every vehicle, every other vehicle, every fifth vehicle, etc.). Drivers of stopped vehicles are observed for evidence of alcohol use and/or impairment. If such evidence exists, a standardized field sobriety test (SFST) may be administered and a breath test may be requested. Fail- ure of such tests would generally result in an arrest for Driving Under the Influence (DUI) of alcohol or other drugs or a violation of an adminis- trative law such as Driving with an Illegal (per se) Blood/Breath Alcohol Concentration, or both. Safety checkpoints (or safety checks) are similar to sobriety checkpoints, except that stated purpose of the operation is to search for inoperative or malfunctioning safety equipment (e.g., horn, lights, etc.). Proper licensing and registration is also examined and, if there is probable cause, impaired driving laws may be enforced. The term roadblock is a more generic term that may refer to either form of check- point or to stops made for some other purpose.

enforcement action was a verbal warning. Thus, in spite of the use of checkpoints, this was not a “zero tolerance” enforcement effort. Further, public information was of modest intensity, involving a press release, 20 signs placed in each of the two cities, and posters placed at tavern exits and in shopping malls. Figure 7 shows that baseline nighttime usage among bar patrons in Halifax (the larger city) was more than 30 points lower than among daytime motorists (54% late night versus 86% daytime). There was less difference in Moncton (58% late night versus 63% daytime).38 Late-night usage among bar patrons increased substantially during this 4-week nighttime effort (+9 points in Halifax and +16 points in Moncton) and remained elevated after 1 month of follow-up (+8 points in both cities). Just as daytime enforcement in Ottawa did not impact late usage among bar patrons, nighttime enforcement in Halifax did not impact daytime usage. In Moncton, with a lower daytime rate, both day and night usage increased. A STEP and Incentive Program in Quebec Another Canadian STEP evaluation examined a combined enforcement and incentive program. This month-long effort was implemented in Quebec in 1986 (Dussault, 1990). It involved extensive training of police in every sector of the province; increased ticketing to three to four times that of pre-STEP levels;39 earned media; approximately $1 million (Canadian) in paid advertising (about 15¢ per capita); and a provincewide incentive program. Small, immediate rewards were given for observed SBU and larger, delayed rewards were provided in the form of entry into a lottery. Figure 8 shows usage increasing from 53% (in 1985) to 86% (in 1987), accord- ing to surveys conducted by Transport Canada (+33 points). Subsequent, semiannual STEPs resulted in a 94% usage rate by 1990 (Grant, 1991). In summary, the early Canadian experience involved exten- sive use of STEP, “blitz,” or “wave” enforcement. Evaluations of these efforts suggested that a) successive waves of enforce- ment would be needed to sustain usage gains; b) longer peri- ods of blitz enforcement appeared to be more effective than 15 84 7976 66 44 48 43 44 30 40 50 60 70 80 90 Baseline STEP 1 (4 wks) STEP 2 (4 days) STEP 3 (1 wk) Pe rc en t U se RMOC (study area) Kingston (control) Figure 5. Changes in SBU rates in the regional municipality of Ottawa-Carleton (RMOC), following a series of STEP programs, and in a control city. 89 82 76 82 6161 50 60 70 80 90 100 Pre-Program Post Program Pe rc en t U se Day Evening Late Night (bar patrons) Figure 6. Percent driver daytime and nighttime SBU in RMOC following a 1-month STEP (Source: Grant, 1991). 38Actually, it may be more appropriate to say that baseline daytime usage was much higher in Halifax (86%) than in Moncton (63%) but nighttime usage was relatively low in both cities (54% to 58%). The high daytime rate in Halifax is likely why there was little change in that mea- sure associated with the program. 39The number of citations increased to nearly 1,500 per day or about 67 citations per 10,000 population over a period of 1 month.

shorter periods;40 and c) nighttime enforcement impacted usage among late-night road users, including bar patrons. Based on these results, Canadian officials recommended that each province conduct two STEPs per year and complement such activities with a minimum of $300,000 in paid media per mil- lion residents (about 30¢ per capita per year).41 Background to U.S. Enforcement Efforts Very soon after the first SBU laws were enacted in the United States, evidence of the need to enforce such laws began to accu- mulate. Many states experienced initial postlaw increases in usage, followed by subsequent declines and stabilization in the absence of enforcement. Further, Campbell (1988) reported a positive relationship between enforcement levels and observed usage in both primary and secondary law states.42 In part due to the mix of primary and secondary enforcement laws, a variety of enforcement procedures have been imple- mented in this country. Such programs varied on dimensions such as periodic versus sustained efforts, use of primary versus secondary enforcement procedures, use of special patrols ver- sus integration of SBU enforcement into regular patrols, issuance of citations versus warnings, extensive publicity ver- sus little or no publicity, and “hard” enforcement messages versus “soft” nonenforcement messages, etc. Another important variable has been the size and complex- ity of jurisdictions where the enforcement program was imple- mented. Early programs were nearly always confined to local areas, later expanding to statewide, regional, and nationwide 16 86 66 86 86 73 63 54 63 62 74 58 45 50 55 60 65 70 75 80 85 90 Pre-Program Mid-Program Post-Program Pe rc en t B el t U se Halifax (day) Moncton (day) Halifax Bar Patrons (night) Moncton Bar Patrons (night) Figure 7. Daytime and nighttime SBU in Halifax and Moncton, Nova Scotia (Source: Malenfant and Van Houten, 1988). 5354 60 68 4139 94 828286 68 0 20 40 60 80 100 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Pe rc en t U se Pre-STEP Post-STEP 40Less that 1 week of enforcement was associated with the smallest impact but there may have been some confounding with order effects (i.e., early waves were associated with greater impact than later waves). 41This was in 1992, just 2 years after NHTSA, recognizing a stabilization of usage rates at just below 50%, implemented STEP-like enforcement as part of its national 70% by 1992 program. The Canadian success with STEPs was an important stimulus for this U.S. initiative. 42Annual citation rates, as reported in the Campbell report, ranged from 1 to 88 per 10,000 residents. Figure 8. SBU in Quebec, before and after the implementation of STEPs (Sources: Dussault, 1990; Grant, 1991; Transport Canada, 1992).

efforts. Because of increasing difficulty involved in implement- ing enforcement and publicity programs in larger and more complex jurisdictions, state and national programs were nearly always implemented with less intensity than local programs. Local Enforcement Benchmark: Elmira, New York (1985–1986) One of the earliest evaluations of a local enforcement program was conducted in Elmira, New York, about 1 year after this first state law was implemented. Influenced by Cana- dian efforts, the Elmira program quickly became a local benchmark for program intensity and impact. The enforce- ment and media effort constituted two campaigns. The first (Elmira I) was a 3-week effort implemented in November 1985 (Williams, Lund, Preusser, and Blomberg, 1987). It included 1 week of publicity, 1 week of publicity and warn- ings, and 1 week of publicity and ticketing. The second wave (Elmira II) was implemented as a reminder campaign in April 1986 (Williams, Preusser, Blomberg, and Lund, 1987). It con- sisted of 3 weeks of publicity, combined with checkpoints and extensive warnings, particularly in the final week. Overall, there were many more warnings than citations,43 and the rate of warnings increased from 140 per 10,000 residents in the first week of enforcement (Elmira I) to 600 per 10,000 resi- dents in the final week (Elmira II). During the first week in which tickets were issued (week 3 of the first campaign), the citation rate was about 143 per 10,000 residents. About 26¢ to 31¢ per resident (per wave) was spent on paid advertising, which was supplemented by earned (news) media and extensive distribution of printed materials directly to households. At least 250 ads per 10,000 residents were aired during the first campaign and about twice as many were aired during the second campaign (514 per 10,000 residents). Fol- lowing the second wave, telephone surveys found that about 40% of Elmira’s residents were aware of the program. The results of these two campaigns are shown in Figure 9. There were large increases in observed usage associated with the intense publicity and enforcement (consisting mostly of warnings). Driver belt use increased by 28 points by the end of the first campaign (+57%), then declined by 8 points after 2 weeks and by 11 points after 2 months, losing nearly 40% of its gain. In spite of this decline, usage remained 17 points (35%) above its baseline. Following the reminder campaign, usage increased by 14 points, reaching 80%, and then declined by 20 points over a period of 8 months. The net gain was 11 points (22%) over the original 49% baseline. Thus, while the gains were substantial, subsequent declines clearly pointed to a need for repeated implementations.44 Subsequent Local Programs and Evaluations (1985–1999) Since the Elmira (I and II) benchmark campaigns, at least 12 studies of local enforcement efforts have examined programs implemented in the following locations: • Albany and Greece, New York (1986), where blitz enforce- ment was compared with sustained and integrated enforce- ment. Both cities had modest levels of earned and public service media that appeared to decline over the 5– to 17 65 77 60 69 8078 71 6666 69 77 63 49 35 40 45 50 55 60 65 70 75 80 85 * Wave 2 Program and Follow-up Pe rc en t u sa ge Elmira Glens Falls (control) Wave 1 Program and Follow-Up Figure 9. Changes in observed usage in Elmira, New York, following a benchmark local STEP involving two waves of activity (Sources: Williams, Lund, Preusser, and Blomberg, 1987; Williams, Preusser, Blomberg, and Lund, 1987). 43In fact, over the two waves of the Elmira program, only warnings were issued in 4 of the 5 weeks of enforcement activity. 44Both campaigns included a control community that experienced very little change in usage during either campaign.

6–month project period. Baselines: 49%–52%; peak in- creases: 13 points in Albany (blitz); 17 points in Greece (sus- tained/integrated) (Rood, Kraichy, and Carmen, 1987). • Modesto, California (1986), where blitz enforcement was adapted to a secondary law state, along with paid, earned, and public service media over a 2-month program, ending with a publicity-only wave. Baseline: 32%; peak increases: 24 points (daytime), 17 points (nighttime) (Lund, Stutser, and Fleming, 1989). • Austin, Beaumont, Brownsville, Dallas, Laredo, and Tyler, Texas (1988–1990), where sustained/integrated enforce- ment was implemented in five cities over a period of about two years. Three, 2-week blitzes were implemented over a period of 8 months in Dallas.45 Public information con- sisted primarily of earned and public service media, which declined over time. Baselines: from 32% in Laredo to 61% in Dallas and 71% in Austin; peak increases: from 7 points in Dallas to 41 points in Laredo (Mounce, Brackett, and Womack, 1990). • Binghamton, New York (1988–1990), where blitz safety belt enforcement was combined with impaired-driving enforce- ment. Safety belt checkpoints were conducted during the day and impaired driving checkpoints were conducted at night over a period of about 2 years. Over time, safety belt enforce- ment changed from mostly warnings to mostly citations and was accompanied by extensive earned and public service media. Baseline: 46%; peak increases: 16 points (daytime); 24 points (late-night)46 (Wells, Preusser, and Williams, 1992). • Rantoul and Galesburg, Illinois (1988), where periodic waves of secondary/integrated enforcement were combined with modest publicity in both cities and with incentives (in Galesburg). Baselines: 42% in Rantoul, 25% in Galesburg; peak increases: 8 points in Rantoul (enforcement only); 6 points in Galesburg (enforcement and incentives) (Mortimer, Gold- steen, Armstrong, and Macrina, 1990). • Galesburg, Rock Falls/Sterling, and Danville, Illinois (1989), where 4 months of continuous or sustained safety belt enforcement was integrated into ongoing traffic enforce- ment activities over a period of 4 months, with apparently modest earned and public service media. Baselines: 31% in Galesburg, 28% in Rock Falls, 38% in Danville; peak increases: 19 points in Galesburg; 11 points in Danville; 9 points in Rock Falls/Sterling (Mortimer, 1992). • Allegan, Muskegon, and Ottawa Counties in western Michigan (1989–1990), where continuous safety belt enforcement was integrated into an ongoing traffic law enforcement campaign over a period of about 1 year, with apparently modest public information activity. Baseline: about 56%; peak increase: 13 points (Streff, Molnar, and Christoff, 1992). • High Point, Elizabeth City, and Haywood County, North Carolina (1993), where intensive blitzes involving check- points, extensive ticketing, and intensive media (public ser- vice, earned, and paid) were implemented over a period of 2 months. Baselines: 78%–79% in Elizabeth City and High Point, 43% in Haywood County; peak increases: 38 points in Haywood County; 13 points in High Point; 10 points in Elizabeth City (Williams, Hall, Tolbert, and Wells, 1994). • Elmira, New York (1999), where an intensive blitz pro- gram involving checkpoints, extensive ticketing, and intensive public service, earned, and paid media, were implemented over a period of 3 weeks. Baseline: 72%; peak increase: 21 points (Williams, Wells, McCartt, and Preusser, 2000). • Reading, Pennsylvania (2004), where enforcement included a combination of safety checkpoints and roving patrols conducted at night, using night-vision equipment. In addi- tion, police cars with flashing lights were positioned near signs encouraging seat belt use (“mini-cades”). Although described as “zero tolerance” enforcement, more warnings (225) than citations (42) were given. However, more than 5,500 motorists were contacted via the various enforce- ment approaches. Publicity was via earned media only with no paid media, but the night-vision equipment generated much media interest. Baseline: 56% day, 50% night; peak increases: +3 points daytime; +6 points nighttime (Chaudhary, Alonge, and Preusser, 2005). Key characteristics of these programs are summarized in Table 5. Following is a general description of results relative to such characteristics. 18 45The following combinations of enforcement procedures and law type were evaluated: Elmira (I and II), Albany, Dallas, and Binghamton were examples of periodic, primary enforcement procedures used in primary law states, with reasonably frequent use of warnings, whereas the programs in Elizabeth City, High Point, Haywood County, and Elmira (III) were examples of primary law enforcement procedures, in primary law states, with a near total focus on issuing citations. The programs in Greece, New York, and in five of the six Texas cities involved sustained, secondary enforcement procedures, integrated with other traffic law enforcement, in primary law states. The program in Modesto, California, and the initial programs in small Illinois cities were periodic or blitz efforts implemented in secondary law states. The second group of Illinois demonstrations and the Western Michigan program involved sustained secondary enforcement procedures (integrated with other traffic law enforcement) in secondary law states. Of the four jurisdictions where nonsanction approaches were imple- mented, two were in a secondary law state (Florida) and two were in a primary law state (North Carolina). 46In Binghamton, in addition to eight daytime checkpoints for safety belt nonuse, warnings were also issued during 54 nighttime checkpoints focused on impaired driving.

19 Location Year (Prog. Length ) Enf. Type Program/ Enf. Duration Tickets/ Warnings Enf. Weeks Enf. Rate /10K Media Type Media $ Media Rate / Capita Elm ira, NY I (35,000) 1985 3 wks Blitz/ Prim ar y c 1 wk warn 1 wk tckt 500 warn 189 tckt 1 wk 1 wk 143 warn 54 tckt Earned + Ads $9K 26¢ 9¢/wk Elm ira, NY II (35,000) 1986 3 wks Blitz/ Prim ar y c 3 wks of warnings 2,800 warn 3 wks 800 w 267 w/wk Earned + Ad s $11K 31¢ 10¢/wk Albany, NY 102,000 1986 (6 mo ) Blitz/ Prim ar y c 4 blitzes + integrated 1,440 tckt 6 mo 4 blitz 141 40/wk Earned + Ads $0 $0 Greece, NY 98,000 1986 (5.5 mo ) Sustained / Integrated 5.5 mo . routine patrol 163 tckt 5.5 mo . 17 .7/wk Earned + Ads $0 $0 Modesto, CA 160,000 (Co) 1986 (8 wks) Blitz/ Secondary 1 wk warnings 1 wk tckts 530 warn 209 tckt 1 wk 1 wk 33 warn 13 tckt Earned + Ads $52K 33¢ 8¢/wk Austin, TX 508,000 1988-90 (2+ yrs) Sustained / Integrated 2+ yrs (25 hrs/wk) 11,500 t/yr 52/yr (25 hrs) 226/yr 4/wk Earned + Ads $0 $0 Beaum ont, TX 117,000 1988-89 (19 mo ) Sustained / Integrated 19 mo (28 hrs/wk) 6,000 t/yr 52/yr (28 hrs) 513/yr 10/wk Earned + Ads $0 $0 Brownsville, TX 107,000 1988-89 (21 mo ) Sustained / Integrated 19 mo (24 hrs/wk) 6,000 t/yr 52/yr (24 hrs) 561/yr 11/wk Earned + Ads $0 $0 Dallas, TX 1,060,000 1988-89 (12 mo ) Blitz/ Prim ar y c (3 blitzes) 2 wks each 6,600 tck t 6 wks 62/6 wks 10/wk Earned + Ads $0 $0 Laredo, TX 117,000 1988-90 (2+ yrs) Sustained / Integrated (3 yrs) (39 hrs/wk) 6,000 t/yr 52/yr (39 hrs) 513/yr 10/wk Earned + Ads $0 $0 Tyler, TX 82,000 1988-90 (2+ yrs) Sustained/ Integrated (3 yrs) (36 hrs/wk) n/ a 52/yr (36 hrs) n/ a Earned + Ads $0 $0 Bingha mt on, NY 56,000 1988-90 (2 yrs) Blitz/ Prim ar y c (2 yrs) 6 sets (waves ) 5000 warn 864 tckt 6 sets (waves ) 893 wrn 154 tckt 26/se t Earned + Ads “small” amount n/ a Rantoul, IL 18,000 1988? (8 wks) Blitz/ Secondary 2 x 2 wks = 4 wks 35% of all violations 4 wks 35% of all viol. Earned + Ads ? ? Galesburg, IL 34,000 1988? (8 wks) Blitz/ Secondary 2 x 2 wks = 4 wks 33% of all violations 4 wks 33% of all viol. Earned + Ads ? ? ? Rock Falls, IL 25,000 1989 (4 mo ) Sustained / Integrated 4 mo (40 hrs/wk) 30% of all violations 4 mo (40 hrs/wk) 30% of all viol. Earned + Ads ? Galesburg, IL 30,000 1989 (4 mo ) Sustained / Integrated 4 mo (40 hrs/wk) 30% of all violations 4 mo (40 hrs/wk) 30% of all viol. Earned + Ads $0 $0 $0 Danville, IL 35,000 1989 (4 mo ) Sustained / Integrated 4 mo (40hrs/wk) 22% of all violations 4 mo (40 hrs/wk) 22% of all viol. Earned + Ad s $ 0 Western MI 500,000 (3 Co) 1989 (11 mo ) Sustained / Integrated 11 mo (140 hrs/ mo ) 2,635 t+w (60 tckts/ 100 speed tckts ) 11 mo (140 hrs/ m onth) 53 (t+w) 1 (t+w) per wk Earned No ads? $0 $0 Escam bia/ Santa Rosa, FL 240,000 1989 (3 mo ) Sustained / Integrated / no sanction 3 mo (no sanction) 39 tckts per mo 3 mo 2/ mo 0.4/wk Earned + Ads $0 $0 Hernando/ Pasco Co, FL 380,000 1989 (3 mo ) Sustained / Integrated / no sanctio n 3 mo (no sanction) 148 tckt per mo 3 mo n/ a 4/ mo 1/wk Earned + Ads $0 $0 Albem arle, NC 16,000 1989 (7 mo ) Sustained / Integrated / no sanctio n 7 mo (no sanction) n/ a Mostly Warnings 7 mo n/ a mo stly warn Earned + Ads $0 $0 Gastonia, NC 60,000 1989 (7 mo ) Sustained / Integrated / no sanctio n 7 mo (no sanction) Mostly Warnings 7 mo mo stly warn Earned + Ads $0 $0 Eliz. City, NC 14,000 1993 (8 wks) Blitz/ Prim ar y c (4 wks) 450 tck t (4 wks) 315/4 wk 79/wk Earned + Ads $5K 38¢ 8¢/wk High Point, NC 71,000 1993 (8 wks) Blitz/ Prim ar y c (4 wks) 650 tck t (4 wks) 92/4 wk 23/wk Earned + Ads $27K 38¢ 8¢/wk Haywood Co, NC 48,000 1993 (8 wks) Blitz/ Prim ar y c (4 wks) 1459 tck t (4 wks) 304/4 wk 76/wk Earned + Ads $18K 38¢ 8¢/wk Elm ira, NY 91,000 (Co) 1999 (3 wks) Blitz/ Prim ar y c (3 wks) (12 da) 474 tck t 12 da 52/12 da 30/wk Earned + Ads n/ a n/ a Reading, PA 81,000 2004 (30 da) Blitz/ Secondr y c 30 da 42 30 da 5.2/4 wk 1.3/wk Earned no ads $0 $0 “K” = 000; “t” or “tckt” = ticket(s); “w” or “warn” = warning(s); “da” = day(s); “viol” = violation(s); “c” superscript = check points Table 5. Key characteristics of local blitz and integrated enforcement programs.

Findings Regarding Various Program Characteristics47 Peak Increases. The median peak increase of the 27 out- comes provided by the above studies was about 13 points (range: −7 points to +41 points). The 31-point gain, from baseline through wave 2 of the benchmark Elmira program, was the third largest increase among these outcomes. The largest gain was 41 points, which was associated with the 2-year, sustained and integrated enforcement effort in Laredo, Texas, implemented from a very low baseline usage rate of 32%. The second highest increase was a 38-point gain associ- ated with the pilot program implemented in Haywood County, North Carolina. Blitz Versus Sustained Enforcement. At the local level, sustained or continuous enforcement efforts received as much evaluative attention as blitz efforts. Sustained enforcement, implemented as part of special patrols or as an integrated component of regular patrols, was generally associated with impacts that were comparable to those associated with blitz programs.48 Comparing 12 blitz outcomes and 14 sustained- program outcomes, the median peak increase was 13 to 14 points for both approaches. However, in each of three comparisons from similar baselines, sustained and integrated enforcement programs were associated with modestly higher increases. These included a Dallas blitz (+6 points), compared with a sus- tained effort in Austin (+9 points); a Galesburg, Illinois, blitz (+6 points), compared with a subsequent sustained and inte- grated enforcement in that city (+19 points); and an Albany, New York, blitz (+13 points), compared with a sustained and integrated effort in Greece, New York (+17 points). Daytime Versus Nighttime Enforcement. Most safety belt enforcement programs have been implemented during the day. Such enforcement often impacts daytime and “early evening” usage but it has less often been shown to impact “late-night” usage, particularly among high-risk drivers such as drinking drivers. Daytime STEPs in Ontario, for example, failed to affect late-night usage among bar patrons (Grant, 1991). A combined alcohol and safety belt enforcement program was conducted in Binghamton, New York, that did impact late-night usage (Wells et al., 1992). It consisted of daytime checkpoints to enforce safety belt laws and nighttime checkpoints to enforce impaired driving laws. However, as part of the nighttime checkpoints, warnings were issued to unbuckled motorists. Thus, part of the nighttime effort was focused on SBU. At least two nighttime enforcement programs that were specific to SBU resulted in significant impacts on high-risk, late-night motorists. The Nova Scotia study already described (Malenfant and Van Houten, 1988) affected late-night bar patrons in two cities and a program in Reading, Pennsylvania (Chaudhary et al., 2005) increased usage among late-night motorists in that city. Both programs affected nighttime usage more than daytime usage (see Figures 7 and 10). Postprogram Decay in Usage and Enforcement. Gener- ally, there has been evidence of less decay in usage following cessation of sustained/integrated programs than following blitz efforts. Results from eight sustained programs found little or no decay in usage after 2 months, whereas results from six blitz efforts showed a decline of about 30% during similar periods of time. There also have been indications of less abrupt de- creases in enforcement following cessation of sustained pro- grams, compared with blitz efforts.49 These results have come 20 47These summaries describe the relative magnitude of reported impact of various efforts and categories of efforts. Unless the term “significant” is used to describe a comparison or results, the comparison or result is not the result of a statistical test. 48Some caution is advised in interpreting these results as sustained efforts generally involved programs implemented at low baselines and they were generally implemented over longer periods of time than blitz programs. 49Evidence of abrupt decreases in enforcement following blitz programs were found in Dallas, Texas, and Albany, New York. 59 56 50 56 45 50 55 60 65 Pre-Program Post Program Pe rc en t U se Day Night Figure 10. Percent daytime and nighttime SBU in Reading, Pennsylvania, following nighttime enforcement (Source: Chaudhary, Alonge, and Preusser, 2005).

from several locations, including Albany and Greece, New York, six cities of varying size in Texas, and small-to-modest size cities in Illinois. Following their evaluations in Albany and Greece, New York, Rood et al. recommended that programs begin with blitz procedures and then integrate safety belt enforce- ment into regular patrol activities. Alternatively, advocates of blitz programs have recommended repeated blitz efforts to sustain and increase gains in usage.50 Nonsanction Enforcement. Some enforcement programs have focused on nonsanction approaches, employing positive reminders by police officers to buckle up and little or no issuance of warnings or citations. Examples of such programs were found in four Florida counties (Kaye et al., 1995) and two North Carolina cities (Hunter et al., 1993). The key enforce- ment activity in each of these programs involved police officers giving a “thumbs up” gesture to motorists to remind them to buckle up.51 Although some programs were accompanied by substantial amounts of earned and public service media, they generally were not associated with significant increases in usage. The median change associated with these four nonsanction programs was +3 points (range: −7 points to +8 points, from similar baselines of about 50% usage). Although their impact was small in every case, there were indications in both Florida and North Carolina that such approaches may have more potential in smaller, less complex media markets than in larger and more complex markets.52 The issuance of warnings versus citations is another aspect of the nonsanction issue. Several programs, including the benchmark Elmira (I and II) programs focused more on warnings than on the issuance of citations. While adequate comparisons of impact have not been documented, most pro- grams that began with an emphasis on warnings eventually shifted to a focus on citations. Examples include Modesto, California, Binghamton, New York, and Elmira III (compared with Elmira I and II). Paid Media. Paid media was generally associated with sub- stantial increases in observed usage (median gain = 16 points). Examples of such increases were found in the Elmira blitzes, in the combined driving under the influence (DUI)/safety belt effort in Binghamton, New York, and in the three North Carolina pilot programs. Where information was provided, spending on paid media ranged from 26¢ per capita (in Elmira) to 37¢ per capita (in the North Carolina pilot studies), or about 8¢ to 10¢ per capita per week of paid media. Earned and Public Service Media. Publicity efforts involv- ing earned media, public service ads, or both were part of nearly every program evaluated, albeit to varying degrees. In several studies, including those conducted in New York (Rood et al., 1987) and Texas (Mounce et al., 1990), evaluators reported that programs relying on earned (i.e., news) media experienced initial interest that declined rapidly as the program continued.53 In the Florida and North Carolina nonsanction programs, earned and public service media, along with the distribution of educational materials, were more visible in smaller, less complex media markets than in larger, more complex markets. The median increase for programs that depended solely on earned and public service media (i.e., those with no paid media) was about nine points (range: −7 to +41). In spite of limitations when used alone, earned media has proven to be an important component of HVE efforts. Evi- dence of intensive earned media coverage was found in all three Elmira blitz programs and in the three North Carolina pilot blitzes. All six programs were among the most effective implemented. Enforcement accompanied by little or no media was gener- ally associated with smaller increases in usage than programs with intense publicity. Two examples included the nonsanc- tion programs conducted in larger jurisdictions in Florida and North Carolina. However, perhaps the best example was reported for the comparison city (Tonawanda) in the New York demonstrations. There was a substantial increase in tick- eting in Tonawanda both during and after the study period. In fact, citation rates were greater in Tonawanda than in one of the study communities. Without publicity, however, there was no measured increase in usage (Rood et al., 1987). Feedback Signs. One unique form of public information involves the use of feedback signs to inform motorists of levels and changes in SBU and enforcement efforts. Feedback signs were evaluated in two North Carolina cities in 1985 (Asheboro: population 18,000, and Greensboro: population 180,000) and reported by Malenfant, Wells, Van Houten, and Williams (1996). Results suggest that these signs were associated with increases in usage from 75% to 89% in Asheboro, the smaller city (+11 points). However, impact was more modest in 21 50It should be remembered that the above results come from evaluations of local programs, most of which were implemented in small jurisdic- tions and from relatively low baselines, compared with current rates of usage. Also, it should be noted that it is more difficult to implement a statewide, rather than a local, sustained enforcement effort that is of suf- ficient intensity to impact usage. As a result, there are few examples of sus- tained/integrated enforcement programs implemented at the statewide level. One exception may be provided by California, which has histori- cally reported relatively high levels of sustained enforcement activity. 51The “thumbs up” gesture involved a police officer tugging on his safety belt with his thumb to remind unbuckled motorists to buckle up. 52This finding of greater potential for impact in smaller, less complex environments likely holds for the use of incentives (Hunter et al., 1993) and feedback signs (Malenfant et al., 1996) in conjunction with enforcement programs as well. 53This decline in interest on the part of the news media is a key reason why a paid media component is currently considered to be critical to the success of enforcement efforts.

Greensboro, the larger city, where usage increased from 80% to 86% (+6 points). Feedback signs were also used in the 1999 Elmira “Buckle Up Now!” program, where usage increased from 69% to 90% (+21 points), but these signs constituted only one of several components of an intense enforcement and publicity program. As with earned media, incentives, and nonsanction approaches, the use of feedback signs appears to be most effectively implemented in smaller, less complex media markets and jurisdictions. Checkpoints and Roadblocks. As was the case with paid media, checkpoints and roadblocks54 were generally associated with greater impact than other enforcement approaches. The median increase among programs using checkpoints or simi- lar programs (excluding nonsanction programs) was 15 points, with a range of +8 points to +38 points.55 By comparison, programs that did not use such techniques had a median increase of about nine points (range: −7 to +41 points). Williams et al. (1994) pointed out that checkpoints were a key factor in the 38-point increase in Haywood County, North Carolina, one of the largest impacts found in this review. Intensity Measures. It is difficult to compare enforcement levels from one project to another, particularly with regard to blitz and sustained efforts. Rates of ticketing in STEPs are usu- ally reported only for the period during which the enforcement is intensified (usually 2 to 4 weeks). Rates of ticketing for sus- tained programs, on the other hand, are usually reported for the entire program period (ranging from 4 to 5 months to 2+ years). Overall, the weekly rate of ticketing per 10,000 resi- dents varied from near zero (in nonsanction sites) to 79 in Elizabeth City, North Carolina, with a median of about 10 tick- ets per 10,000 residents per week. Looking only at blitz efforts, the median rate during the blitz periods was the benchmark rate of 54 tickets per 10,000 residents (1 week) issued during the Elmira I program (range: 20 to 140).56 Summary of Local Program Results Overall, local program findings are summarized as follows: • The largest peak effect size was 41 points, with a median increase of 13 points. • Sustained programs had impacts that were at least compa- rable to those of blitz programs and they were associated with less postprogram decay. • In spite of postprogram decay, it is possible to sustain impact over long periods of time by repeated program implementations. • Checkpoints, roadblocks, and other HVE efforts were generally associated with some of the larger increases in observed use. • Enforcement accompanied by high levels of publicity was most effective. • Paid media helps to counter waning media interest over time and, as such, was generally associated with larger impacts on usage. • The use of feedback signs facilitated enforcement and publicity efforts, particularly in smaller, less complex jurisdictions. • Nonsanction enforcement approaches generally were not effective. • Incentive programs were shown to have some potential for sustaining (and possibly adding to) the impact of enforce- ment efforts (at low usage rates). • Median levels for various indices of media and enforce- ment were as follows:57 a. The median citation rate in blitz programs was about 30 citations per 10,000 residents per week (Elmira III), 22 54In this category, the researchers also include the practice of officers stationed at signalized intersections for the purpose of issuing warnings or citations for safety belt violations. 55Checkpoints or roadblocks were used in Elmira, New York; in the three North Carolina pilots; and in blitz efforts conducted in Albany and Binghamton, New York. The approach used in Dallas, Texas (police situated at intersections and observing for violations) was also included in this category because of its visibility. 56This weekly rate of ticketing is just one way of looking at intensity. In Elmira this rate was achieved in a program where there was only 1 week of ticketing; other programs may have issued tickets at a slightly lower rate over a period of several weeks. The total number of persons cited (per 10,000) may be a more relevant index of intensity when enforce- ment efforts are contiguous. 57Median values vary, depending on how waves of activity are counted and how warnings are viewed. In Elmira, New York, one could consider the two waves of activity as a single activity or as two separate activities. Because these two waves were separated by nearly 5 months, they were viewed as separate activities. It may be most relevant to view contigu- ous enforcement efforts as a single activity and to view noncontiguous efforts that are separated by time as separate activities. In spite of these definitional issues, median values were relatively robust as indicators of intensity. One exception is found in the Elmira I and II projects where road checks, combined with warnings, constituted the main enforcement activity. Because of their visibility (i.e., road checks plus 3,000 warnings in Elmira II), the intensity of this effort is underestimated by the zero citation rate assigned to it. While warnings are not as effective as cita- tions, they do not have zero deterrence value, particularly when used in conjunction with road checks. If a warning had been assigned the same value as a citation, Elmira II would have had the highest intensity (286 actions per 10,000 residents per week) and Elmira I would have the third highest rate (98 actions per 10,000 residents per week). The median level of intensity would then be 76 citations per 10,000 residents per week.

ranging from zero (Elmira II—with warnings only) to 79 (Elizabeth City, North Carolina). b. Enforcement in secondary law states was sometimes mea- sured in terms of the safety belt proportion of total cita- tions. The median of this index was 32%, with a range of 22% (Danville, Illinois) to 35% (Galesburg, Illinois). c. The median expenditure for paid media was about 36¢ per resident (overall program), with a range of 26¢ (Elmira I) to 38¢ (North Carolina pilots).58 Expressed as per capita, per week, the median rate was 8¢. d. The median awareness of enforcement level was 49% (Albany, New York), ranging from 10% (Greece, New York) to 90% (Elmira III). Statewide, Regional, and National Enforcement Programs Local demonstrations provide an opportunity to implement relatively intense enforcement and media efforts, using differ- ent combinations, intensities, and patterns of enforcement and publicity. However, the most relevant results for states attempt- ing to increase usage likely come from studies of programs that have been implemented statewide. It is here that the greatest potential exists for affecting usage among large populations, even though local programs are likely to be more fully imple- mented. Following are descriptions of an initial benchmark pro- gram, followed by subsequent statewide, regional, and national efforts. Most of the recent examples of statewide STEP pro- grams come from participation in Operation ABC and national CIOT mobilizations implemented since 1997. Benchmark: North Carolina (1993–1994) Background. Beginning in 1993, the State of North Car- olina initiated a series of statewide efforts to increase SBU. This program, called CIOT, was implemented about 8 years after North Carolina enacted its primary SBU law (1985) and experienced increases in usage that eventually stabilized at 60% to 65%. Early enforcement efforts had been imple- mented in 1987 and those efforts resulted in significant increases in usage (Reinfurt, Campbell, Stewart, and Stutts, 1990). In addition, nonsanction approaches had been imple- mented in Albemarle and Gastonia in 1989, with little mea- sured impact (Hunter, Stewart, Stutts, and Marchetti, 1993). In 1993, just prior to the statewide CIOT program, three inten- sive pilot blitzes were conducted in Elizabeth City, Haywood County, and High Point. These pilot programs were accompa- nied by large and significant increases in observed usage, with gains ranging from 10 to 38 percentage points (Williams, Hall, Tolbert, and Wells, 1994). Program Characteristics. The statewide CIOT program was planned as a 5-year effort that began with two waves of enforcement and media activity in November 1993 and July 1994 (Williams, Reinfurt, and Wells, 1996). This was one of the first statewide applications of the STEP model in the United States and it was implemented with great intensity. The waves of activity were 8 weeks and 7 weeks in duration, respectively. The first included four phases: 1) an initial 2 weeks of public information; 2) 3 weeks of blitz enforcement using both checkpoints and roving patrols; 3) 2 weeks with no enforcement; and 4) intensified enforcement in the final week. Throughout this period there was extensive paid, earned, and public service media, featuring local police officers and the CIOT slogan. As in Elmira, the second wave was intended as a “booster.” It was similar in design to the first wave except that it entailed only 3 weeks of enforcement, rather than 4. Intensity. Over the two waves, more than 6,300 check- points were implemented and nearly 59,000 citations were issued for safety belt violations, about 81 citations per 10,000 residents over 7 weeks. A total of $600,000 was spent on paid advertising, about 8¢ per resident over the 15 weeks covered by the two waves. This expenditure resulted in just over 7,100 ads placed during the 15 weeks, about 10 per 10,000 residents. The spring 1994 wave, which was intended as a reminder, was some- what less intense in terms of citations, paid media, and ads. Impact. As a result of this statewide effort, combined driver and passenger usage increased from 63% to 79% dur- ing the first wave (+16 points), then declined by 7 points to 72%, losing about 40% of its gain. Following the second wave, usage increased by 7 points, reaching its previous peak of 79% (see Figure 11). One often-missed aspect of this program is that it was peri- odically repeated over a 5-year period. As a result, statewide usage rates in North Carolina, which increased by 15 percent- age points within a period of several weeks, stabilized at about 80%, 15 points above its preprogram baseline. Statewide usage remained at about that level through 2000, after which it began to increase again, likely associated with North Carolina’s par- ticipation in national Operation ABC and CIOT mobilizations. As of 2006, the North Carolina usage rate was 89%. Statewide Demonstration Programs (1993–1999) Following the benchmark CIOT program implemented in North Carolina and prior to 2000, when paid media efforts 23 58The amount spent on paid media in the three North Carolina pilots ($50,000) was distributed across sites by population. It is not known if that is the exact proportion spent in each site.

were introduced to Operation ABC mobilization efforts, there were a number of demonstrations funded to emulate the North Carolina program. These efforts, called Occupant Protection Special Traffic Enforcement Programs (OP-sTEP) were imple- mented with relatively modest levels of funding. They included: • Five statewide OP-sTEPs59 funded by NHTSA in 1993 and 1994, and • An expanded series of 20 OP-sTEPs implemented from 1995 through 1998 and funded by NHTSA, with supplemental funding from the private sector. OP-sTEP Demonstrations in Five States (1993–1995). An initial series of OP-sTEP programs were funded by NHTSA, beginning in 1993. Approximately $1.5 million in federal funds, $125,000 in private sector funds,60 and $1.03 million in state funds were identified to conduct two to three sTEP waves per year. These were conducted initially in New Mexico, Oregon, South Carolina, Washington, and Vermont.61 Each wave was to consist of 10 to 14 days of intensified enforcement accompanied by public service and earned media efforts. No funds were provided for paid media. Police agency participa- tion was obtained by providing overtime and/or officer and agency incentives. Former police officers, known as Operation BuckleDown (OBD) Spokespersons, were used as liaisons to contact agencies within their state and persuade them to par- ticipate. These funds, incentives, and liaison activities re- sulted in programs of modest intensity in comparison with the intense North Carolina program. Results. On average, two to three waves of enforcement were conducted in each state, with a median rate of about four citations per 10,000 residents per week (range: 2 to 21 cita- tions). No estimates of the intensity of public information activities were provided and no paid media was involved. About 45% of the police agencies in each state participated in these sTEPs (range: 30% to 71%). Usage increased by an aver- age of about four percentage points from 1993 to 1994 (range: 3 to 5 points).62 (NHTSA, 1995). Expanded OP-sTEP Demonstrations (1996–1999). In conjunction with a new initiative called Campaign Safe and Sober, NHTSA expanded the OP-sTEP grant program to about 20 states from 1995 through 1998. Approximately $5.5 million in federal (Section 403)63 funding was provided, along with 24 82 63 79 72 79 81 80 50 55 60 65 70 75 80 85 * Oct-92 Nov-93 May-94 Aug-94 Aug-95 Aug-96 Aug-97 * Us ag e (% ) Driver Combined Passenger Figure 11. SBU in North Carolina: Before and after implementation of a 5-year CIOT enforcement program [Sources: Williams, Reinfurt, and Wells, 1996 (1992–1994) and annual usage rates reported to NHTSA (1995–1997)]. 59The acronym STEP, which originally referenced Selective Traffic En- forcement Programs used in several areas of traffic safety (e.g., speed, impaired driving, etc.), was modified to OP-sTEP to refer to Special Traffic Enforcement Programs focused on Occupant Protection. 60The Automotive Coalition for Traffic Safety, an auto industry-funded organization, provided $125,000 to these states to spend on the devel- opment and distribution of materials to support their enforcement efforts. 61Indiana was later included as a sixth state funded under this program. 62In Vermont, the increase was 14 points but a portion of that increase (9 points) was attributed to implementation of its safety belt law (from a baseline rate of 54%). 63Section 403 refers to that section of the Highway Safety Authorization (U.S.C. 23) that provides for funding for research and demonstration efforts. Another section of the Highway Safety Act, Section 402, pro- vides funding for state and local programs.

approximately $6 million in private sector funds.64 Section 402 funding provided by states may have approximated the Section 403 funds. As with the initial OP-sTEPs, states were required to conduct waves of intensified enforcement, accompanied by earned and public service media. An evaluation of this program was conducted by Solomon, Nissen, and Preusser (1999), using quarterly reports provided by grantee states.65 Based upon this state-reported information, the median citation rate per 3-month period was about five cita- tions per 10,000 residents (range: <1 to 43).66 Another measure of enforcement intensity was the safety belt proportion of total citations issued, a measure frequently used in secondary law states. Safety belt violations accounted for about 22% of total citations, compared with 32% for speeding and 3% for alcohol- impaired driving, somewhat lower than the 33% median rate reported for local programs where this measure was used. Only qualitative data were provided regarding the level and type of public information activity, which typically consisted of a com- bination of public service and earned media activities. Results. Figure 12 shows results from preprogram and postprogram observations. It suggests a step-like or ratcheting effect associated with successive waves of activity, with usage increasing with enforcement and publicity and then declin- ing prior to the start of the next wave. There was an average gain of about 7 points after two waves of activity and a peak gain of about 10 points after three or more waves. Gains diminished after three to four waves. Solomon suggested that the states may have reached a “ceiling” at this point. It also is possible that the intensity of enforcement and media declined over time or that, following repeated implementations, greater intensity was needed to retain public awareness. Several local studies of blitz and sustained enforcement efforts have men- tioned burnout among participating enforcement agencies and in media attention. The Canadian STEPs also showed declin- ing gains with repeated applications.67 Operation ABC Mobilizations (1997–2003) Background. At about the same time as the OP-sTEP pro- grams and Campaign Safe and Sober were being implemented, a crisis was emerging with regard to injuries and deaths occur- ring among young children and short-stature adults riding in the front seat of passenger vehicles. These injuries and deaths were associated with the deployment of passenger-side air bags in low speed crashes and frequently involved unrestrained and “out-of-position” occupants who were positioned directly 25 66.6 63.7 66.8 63.9 66.4 62 65.1 61.2 62.6 56.4 54 56 58 60 62 64 66 68 pre (1) po st(1 ) pre (2) po st(2 ) pre (3) po st(3 ) pre (4) po st(4 ) pre (5) po st(5 ) Us ag e (% ) Figure 12. Changes in SBU in States Participating in OP-sTEP Demonstration Programs (Solomon, Nissen, and Preusser, 1999). 64The private sector funding came from General Motors Corporation as part of a settlement agreement with NHTSA. 65Twenty states participated, but there was considerable variation in the level of reporting. 66It is not known exactly how many days of enforcement were included in these rates but it is likely that they refer to just over one wave of activ- ity or about 15 days of enforcement. Thus, the rate of five citations per 10,000 residents most likely translates to about 2.5 citations per 10,000 residents per week. This can be compared with the rate of 12 citations per 10,000 per week in the North Carolina benchmark. 67This is an interesting dilemma. Even at relatively low levels of usage, repeated efforts often result in declining gains over time. Yet, over the longer term, nearly all states that have implemented blitz efforts have achieved usage rates above 70% to 80%. Frequently, “breakthrough” gains (associated with blitzes) have occurred in secondary law states after a primary law upgrade has been enacted and implemented.

within the deployment area of the air bag. As part of the response to this problem, a national “Call to Action” in early 1996 resulted in the formation of a public-private partner- ship involving NHTSA, the National Safety Council (NSC), the National Transportation Safety Board (NTSB), the CDC, automobile manufacturers, insurance companies, air bag sup- pliers, advocacy groups, public health and medical groups, and others. Out of that initial partnership evolved an organi- zation called the Air Bag and Seat Belt Safety Campaign (known as the Air Bag Campaign or ABSBSC)68 and a three- point program to increase SBU that included: a) intensive public information, b) upgraded legislation, and c) intensified enforcement.69 Operation ABC Mobilizations. In 1997, the Air Bag Campaign implemented what became a decade-long series of safety belt enforcement mobilizations. The first, imple- mented in November of 1997, reportedly included 1,000 par- ticipating enforcement agencies. Such mobilizations were then implemented in the spring and fall of each year. By 2002, more than 12,000 enforcement agencies reported par- ticipating in these efforts. The media component of each campaign included extensive publicity generated by a combi- nation of public service ads, earned media, and paid adver- tising, usually initiated by the ABSBSC. Individual state par- ticipation in the mobilizations prior to 2000 is not well documented. However, there was evidence of considerable activity among state and local police agencies. Further, based on NOPUS,70 there was an increase in national usage, from 62% in May 1997 to 70% in December 1998 (Glassbrenner and Utter, 2001). A substantial portion of this increase ap- peared to be associated with greatly intensified mobilizations in 1998. Innovative Grants and Expanded Use of Paid Media. Beginning in 2000, there was a substantial increase in em- phasis placed on paid advertising and on the use of “hard” enforcement messages (e.g. CIOT). This increased use of paid media began with the May 2001 Operation ABC mobilization in the southeast region and was enabled by funding provided by Section 157 of the Transportation Equity Act for the 21st Cen- tury (TEA-21).71 Effective allocation of these funds became a priority for NHTSA and the funds were increasingly used to support state participation in Operation ABC mobilizations. These funds enabled states to reach levels of media intensity that were comparable to those of the 1993 North Carolina CIOT benchmark. State, Regional, and National CIOT Programs (2000–2006) Several key enforcement programs and evaluations resulted from the national Operation ABC mobilizations implemented by ABSBSC in cooperation with NHTSA and its Buckle Up America program. They included: a) the South Carolina CIOT mobilization implemented in 2000; b) the regionwide expan- sion of the CIOT program to eight southeastern states in 2001; c) the funding of fully-implemented STEPs in 10 states across the United States in 2002; and d) the national CIOT mobilizations conducted in May 2003 and 2004. They also included several targeted mobilizations directed at pickup truck occupants (in the Southeast and South central regions), rural motorists (in the Great Lakes region) and, more recently, nighttime occupants (currently being implemented in North Carolina and West Virginia). Following are brief descriptions of these efforts. South Carolina CIOT Program (Fall 2000). Character- istics. As part of the Operation ABC mobilization conducted in November 2000, the Air Bag Campaign provided the State of South Carolina with a $500,000 grant for paid media using the CIOT slogan. NHTSA provided additional Section 157 funds and the South Carolina Highway Patrol committed to conducting more than 3,000 checkpoints throughout the state. As in the North Carolina benchmark, publicity involved a combination of paid, earned, and public service media. Intensity. Based on the expenditure of $500,000 for paid media, the State spent approximately 12¢ per capita over a 4-week advertising period (about 3¢ per capita per week). 26 68The ABSBSC was initially called the National Automotive Occupant Protection Campaign (NAOPC). 69A key objective of the three-point program was to increase adult SBU as several studies had shown that one of the surest ways to get children restrained was to get the adults in the vehicle restrained. 70NOPUS are national-probability-sample, observational surveys con- ducted annually to monitor usage across the United States. Beginning in 1998, several smaller-sample (mini-NOPUS) surveys were conducted before and after the May and November mobilizations to estimate changes associated with these enforcement efforts. 71Section 157(f) innovative grant funds were part of a larger Section 157 program that authorized approximately $100 million per year as incen- tive grants. Such grants were provided to states based on having higher than average usage rates (compared with the national average) or in- creases in their rates over a 2-year period, after 1996. Funds not allo- cated under the incentive grant provision of this section were allocated under Section 157(f) for innovative (enforcement) efforts to increase statewide usage rates (Federal Register, Vol. 64, No. 4, pp. 1062–1067, Jan. 7, 1999). In Fiscal Year (FY) 2000, about $25.5 million was allocated to states under Section 157(f), and in FY 2001, approximately $38.2 mil- lion was allocated to states under this provision. The Air Bag Campaign supplemented these efforts by coordinating the nationwide mobiliza- tions, purchasing national advertising, and generating news events (i.e., earned media).

Two weeks of enforcement yielded 3,300 checkpoints72 and 21,321 citations for safety belt and child restraint violations, more than three times as many citations as in past mobiliza- tions in the state. This level of enforcement translated to 53 citations per 10,000 residents over 2 weeks, compared with 81 citations per 10,000 residents over 7 weeks in the 1993 North Carolina benchmark program. Awareness of the CIOT program increased significantly over time. Surveys conducted at licensing centers across South Car- olina found that: a) awareness of recent safety belt messages increased from 67% to 95% (+28 points); b) awareness of the CIOT slogan increased from 18% to 80% (+62 points); and c) messages were received primarily from television, both before and after the program (followed by newspapers and radio). Awareness of enforcement increased as well. Statewide telephone surveys conducted before and after the mobiliza- tion found that awareness of enforcement efforts increased from 22% to 73% (+51 points) and awareness of checkpoints increased from 28% to 86% (+58 points). These results sug- gested that publicity regarding the program had reached the majority of the residents of the state. With regard to the source of information, most residents became aware of safety belt messages via television (67%), compared with radio (38%) and newspapers (42%). Impact on SBU. According to South Carolina’s statewide survey, usage increased from 66.5% to 73.9% (+7.4 points). Figure 13 shows the results of a smaller, subsample survey, con- ducted in six counties across the state. It found that usage in these counties increased from 65.4% to 78.9% (+13.5 points). In the latter survey, all categories of gender and race, as well as both urban and rural counties, experienced significant increases. In fact, usage increased among rural motorists, males, and blacks to a greater extent than among urban motorists, females, and whites. The South Carolina CIOT program provided a second example of an intense statewide enforcement program. Fur- ther, this effort was implemented at a time when the avail- ability of resources for such programs was increasing in the form of Section 157 Innovative Grants.73, 74 Thus, it became a model for similar program in other states. Southeast Regional CIOT Mobilization (May 2001). Fol- lowing the South Carolina program, an eight-state CIOT mobilization was implemented in the southeast region of the United States (Solomon, 2002). States included in this effort were: Alabama (newly upgraded primary law), Florida (sec- ondary law), Georgia (upgraded primary law), Kentucky (sec- ondary law), Mississippi (secondary law), North Carolina (original primary law), South Carolina (secondary law/primary law),75 and Tennessee (secondary law). This program, summarized in Table 6, and most subse- quent statewide mobilizations had essentially the same pro- gram design, which called for: a) four weeks of program activ- ity; b) earned media throughout the program; c) two weeks of paid media (weeks 2 and 3); and d) two weeks of enforcement (weeks 3 and 4). Observational and awareness surveys were conducted at the baseline; during earned media (week 1); dur- ing earned and paid media (week 2); at peak media and enforcement (between weeks 3 and 4); and postprogram. 27 78.9 72.3 65.4 60 65 70 75 80 Publicity Enforcement Enforcement Program Phase Us ag e (% ) Figure 13. Changes in SBU in six South Carolina counties: Before and after the 2000 CIOT mobilization (from Solomon and Preusser, under review). 72Thirty-three hundred checkpoints, in this state with just over 4 mil- lion residents, constituted a rate of eight checkpoints per 10,000 resi- dents (over a period of 2 weeks), compared with a rate of nine check- points per 10,000 residents in the North Carolina benchmark (over a period of 7 weeks). 73Section 157(f) innovative grant funds were not the only sources of funding for these programs but they were funds that were specifically used to support these enforcement efforts. 74Unfortunately, the South Carolina program was followed by a politi- cal controversy that greatly reduced subsequent enforcement efforts in the state and, as a result, much of the increase in usage diminished quite rapidly. This is in stark contrast with the North Carolina program, where waves of enforcement were repeated over a 5-year program period and usage stabilized at around 80%. 75South Carolina’s safety belt law allowed for primary enforcement as part of safety checkpoints.

Expenditures for paid media averaged about 6¢ per capita for the 2-week effort. This was about half the per capita expen- diture in the 2000 South Carolina mobilization (12¢ per capita over 2 weeks). By comparison, the expenditure for Wave 1 of the 1993 North Carolina program was about 6¢ per capita for an 8-week effort and the expenditures in the early Elmira pro- grams were 26¢ to 31¢ per resident over 3 weeks. Still, aware- ness of enforcement efforts reached higher levels in these more recent statewide programs than that measured in Elmira. Some of this elevated awareness may have resulted from re- peated participation in STEPs and/or national mobilizations. The average citation rate in this regional effort was 21 cita- tions per 10,000 residents over 2 weeks of enforcement. This was lower than in North Carolina, where 51 citations (per 10,000) were issued over 4 weeks in Wave 1 and 30 citations (per 10,000) over 3 weeks in Wave 2. It is also lower than in the South Carolina program, where 52 citations were issued (per 10,000) over 2 weeks of enforcement. Looking back to the more effective local blitz programs, the median citation rate (in Elmira, Modesto, and the three North Carolina pilots) was 92 (per 10,000) over enforcement periods ranging from 1 to 4 weeks. These comparisons suggest that enforcement intensity may have declined with expansion of the STEP model from the local level to statewide and regional applications. Per capita media spending was also less than in the local programs that used paid media. This expansion from statewide implementation in South Carolina to a regional effort generally resulted in a smaller aver- age per capita expenditure for media (compared with South Carolina). Perhaps more significant is the fact that the median awareness of enforcement (checkpoints) in the southeast regional program reached only 62%, compared with 73% to 82% in South Carolina and 85% in North Carolina.76 Impact on Observed Usage. In spite of lower awareness of enforcement, Figure 14 shows a substantial increases in observed usage in all eight states, ranging from 4 points in North Carolina (which had the highest baseline rate) to 20 points in Tennessee (which had nearly the lowest baseline rates). Regionwide, there was an average increase of about nine percentage points. Among the states with comparable baseline rates (60% to 68%), there was a modest positive cor- relation between media intensity (as measured by per capita expenditure) and usage rate increase (r = .50). Model CIOT Programs (2002) In 2002, full implementation of the CIOT model was attempted in 10 states located in various regions of the United States. These states were Alabama, Florida, Illinois, Indiana, Mississippi, Nevada, Texas, Vermont, Washington, and West Virginia.77 In addition, more limited paid media and enforcement efforts were implemented in four states: Colorado, Michigan, Ohio, and Rhode Island. By comparison, little or no paid adver- tising was included in four state programs: Iowa, New York, Oregon, and western Massachusetts.78 It should be noted that the intensity of enforcement in these four states was compara- ble to that of the 10 fully implemented states (Solomon, Ulmer, and Preusser, 2002). Paid Media. The full implementation states spent an aver- age of 14¢ per capita on paid media (over 2 weeks), compa- rable to the 12¢ per capita spent in the 2000 South Carolina 28 State Law Type Min. Fine Pop. (m) Media Costs (K) Media per Capita 2 wks/1wk Number Tickets Issued Tickets per 10K 2 wks/1wk Pre- Use (%) Post- Use (%) Pct. Pt. Chg. AL P $25 4.5 $500 11¢ / 6¢ 12,257 27 / 14 68 76 +8 FL S $30 16.3 $708 4¢ / 2¢ 29,724 18 / 9 61 70 +9 GA P $15 8.4 $500 6¢ / 3¢ 33,208 40 / 20 72 82 +10 KY S $25 4.1 $600 15¢ / 7¢ 5,806 14 / 7 60 70 +10 MS S $25 2.9 $250 9¢ / 4¢ 2,450 8 / 4 49 62 +13 NC P $25 8.2 $250 3¢ / 2¢ 20,055 25 / 12 80 84 +4 SC S/P $10 4.1 $250 6¢ / 3¢ 7,115 17 / 9 65 70 +5 TN S $10 8.2 $500 6¢ / 3¢ 9,190 11 / 6 53 73 +20 Region 56.5 $3,558 6¢ / 3¢ 119,805 21 / 11 65 74 +9 Regional rate Regional rate wt. ave. “P” = primary law; “S” = secondary law; “Min.” = minimum; “Pop” = population; “m” = million; “K” = thousand; “wk(s)” = week(s); “Pct. Pt.” = Percentage points; “Chg.” = Change; “wt” = weighted Table 6. A summary of key characteristics of the southeast regional CIOT program implemented in 2001 (Solomon, 2002). 76Slightly different awareness levels were found in telephone surveys and driver licensing center surveys. 77The term fully implemented refers to the fact that, in addition to inten- sified enforcement, each of these states also invested heavily in paid media and they used the CIOT slogan or a similar “hard enforcement” slogan. 78It should be noted that, while these latter states are used for compari- son purposes, three of the four (all except western Massachusetts) had very high baseline rates, thus limiting the potential for large gains to some extent.

CIOT (2 weeks), and considerably more than the 6¢ per capita spent during Wave 1 of the 1993 North Carolina benchmark (8 weeks). The four partial implementation states spent only about 4¢ per capita on paid media (2 weeks). Awareness. Postprogram awareness of recent safety belt mes- sages, as measured by telephone and licensing center surveys, was higher in the 10 fully implemented states (86%) than in the 4 states with lower expenditures for paid media (81%) and the 4 states with no paid media (71%).79 There were even greater differences with regard to awareness of enforcement. By the end of the program, 59% of residents in the fully implemented states were aware of special safety belt enforcement efforts, compared with only 34% in the states with less paid media and 25% in the four states with no paid media. The only previ- ous programs with higher levels of enforcement awareness, compared with these fully implemented states, were the North Carolina benchmark (85%) and the 2000 South Carolina pro- gram (73%). Both had very high citation rates. Enforcement intensity, as measured by citations per 10,000 residents, was similar in the fully-implemented states and in the states with no paid media. Both groups had an average of about 22 citations per 10,000 residents (2 weeks). States with modest levels of paid media also had modest citation rates (about 11 per 10,000). The fact that the citation rates in the fully implemented and no paid media states were comparable suggests that the lower level of enforcement awareness in the latter group was due to the absence of paid media, since enforcement intensity was comparable. Publicity in the four non-paid-media states appar- ently resulted in less public awareness of the enforcement effort, possibly due to less control of when and where messages were aired and possibly due to the use of “softer” enforcement messages. Table 7 provides a summary of key characteristics of each group of states. Compared with previous efforts, the 2-week citation rates in the fully-implemented states and in the states with no paid media were nearly identical to those found in the southeast regional program (i.e., 21 to 22 citations per 10,000 residents), but not as high as in the 2000 South Carolina CIOT (52 per 10,000 over 2 weeks) or in the 1993 North Carolina CIOT (51 per 10,000 over 4 weeks during Wave 1). Impact on Observed Usage. Figure 15 shows that the aver- age increase in usage in the fully implemented states was 8.6 points, nearly three times as great as in the partially imple- mented states (3 points) and about nine times as great as in the states with no paid media (less than 1 point).80 This is con- sistent with a hypothesized positive relationship between paid media and level of impact. In fact, looking only at the 14 states with reasonably similar baseline levels (±10 percentage points of the overall mean baseline of 71%), there was a reasonably strong correlation between media intensity and postprogram change in usage (r = .77). In comparison with past benchmark programs, the 9-point average increase in the fully implemented states was compara- ble to the 9-point increase found in the 2001 southeast regional program and the 7-point increase in South Carolina. As with levels of media intensity and awareness, these increases in usage were smaller than the increases observed following Wave 1 of the North Carolina benchmark (+16 points). National CIOT Mobilizations Beginning in 2003, the CIOT model was implemented on a nationwide basis, using essentially the same phasing of earned 29 74.2 66.6 67.3 64.9 60 62 64 66 68 70 72 74 76 Baseline Earned Paid Enforce 1 Enforce 2 Program Phase Us ag e (% ) Figure 14. Changes in SBU, by phase, in the 2001 CIOT program implemented in eight southeastern states in 2001 (Solomon, 2002). 79By comparison, awareness of safety belt messages was 86% in the 2001 southeast regional program and 92% in the 2000 South Carolina CIOT program. 80Again, this comparison with the states with no paid media has to be tempered due to the very high baseline in three of the four states that did not purchase media time.

30 State and Characteristics Media Costs Enforcement Levels Usage Rates State Law Min. Fine Pop. (m) Media Costs $ (000) Per capita 2 wks Cites Tckts/ 10K 2 wks Tckts/ 10K 1 wk Base Use (%) Peak Use (%) Chg. Pct. Pts. Full Implementation States (Paid Media) AL P $25 4.50 $252 6¢ 13,664 31 70.3 78.7 +8.4 FL S $30 17.02 $2,112 12¢ 37,063 23 66.5 75.1 +8.6 IL S $25 12.65 $1,000 8¢ 22,073 18 70.6 74.3 +3.7 IN P $25 6.20 $963 16¢ 24,697 40 69.2 72.2 +3.0 MS S $25 2.88 $330 12¢ 2,486 9 53.8 61.5 +7.7 NV S $25 2.24 $290 14¢ 3,570 17 70.6 76.4 +5.8 TXa P $25+ 6.54 $1,046 16¢ 27,260 40 80.5 86.4 +5.9 VT S $10 0.62 $200 33¢ 1,304 21 66.2 84.9 +18.7 WA S/P $86 6.13 $500 8¢ 5,505 9 80.8 89.5 +8.7 WV S $25 1.81 $250 14¢ 3,104 17 15 11 9 20 4 8 21 1 5 9 56.5 71.6 +15.1 Totals Averages 60.59 $7,718 14¢ 140,726 22 11 68.5 77.1 +8.6 Partial Implementation States (Less Paid Media) CO S $15 4.55 $100 2¢ 3,026 7 3 72.1 73.2 +1.1 MI P $25 10.08 $650 6¢ 5,463 5 3 82.3 80.0 -2.3 OH S $25 11.44 $433 4¢ 21,790 19 10 64.2 70.3 +6.1 RI S $75 1.08 $27 3¢ 1,301 12 6 62.6 68.6 +6.0 Totals Averages 27.15 $1,210 4¢ 31,580 11 6 70.3 73.0 +2.7 Comparison States (No Paid Media) IA P $25 2.94 $0 0¢ 3,033 10 5 81.4 83.0 +1.6 NY P $50+ 19.19 $0 0¢ 69,034 36 18 78.3 82.8 +4.5 OR P $94 3.56 $0 0¢ 5,745 16 8 88.5 87.8 -0.7 West MA S $25 0.34 $0 0¢ 818 24 12 60.6 57.2 -3.4 Totals Averages 26.03 $0 0¢ 78,630 22 11 77.2 77.7 +0.5 a) The 10 largest cities in Texas participated b) Legend: P = primary law; S = secondary law; Tckts = Tickets; m = million; K = thousand Table 7. A summary of key characteristics of states included in an evaluation of model CIOT programs conducted during the May 2002 mobilization. 77.1 68.5 70.3 73 77.7 77.2 60 62 64 66 68 70 72 74 76 78 80 Pre-Mobilization Post-Mobilization Us ag e (% ) Fully Implemented Partially Implemented No Paid Media Figure 15. Changes in SBU in 10 states with fully implemented media and enforcement efforts, compared with changes in four states with less paid media and four states with no paid media: May 2002 Operation ABC mobilization (Solomon, Ulmer, and Preusser, 2002).

and paid media to publicize enforcement activity as in the statewide and regional applications already described. 1) The 2003 National CIOT Mobilization (May 2003). The May 2003 mobilization was the first national mobilization to be called CIOT, signifying that most states were now using either the CIOT slogan or some other “hard enforcement” slogan (Solomon, Chaudhary, and Cosgrove, 2004). Key characteris- tics of this nationwide effort are summarized in Table 8. Paid Media. In conjunction with the 2003 effort, there was an $8 million national media purchase, in addition to approxi- mately $15.7 million spent on paid media by 45 reporting states. This paid media was supplemented by extensive earned media, including national network coverage, more than 1,000 local tel- evision news stories, an audio news release that was aired on several thousand radio stations, and articles and editorials in nearly 100 publications. Citations. About 7,125 law enforcement agencies partici- pated in the 2003 national program in the 44 states that pro- vided enforcement data. These agencies reported issuing about 508,492 citations for safety belt violations (about 20 per 10,000 residents over 2 weeks). The citation rate was 70% higher in primary law states (about 24 per 10,000 residents) than in secondary law states (about 14 per 10,000 residents). Awareness. Nationwide telephone surveys measured 82% awareness of recent safety belt messages (an increase of 9 points from preprogram levels) and 40% awareness of safety belt enforcement efforts (an increase of 24 points). The 82% aware- ness of safety belt messages was generally somewhat lower than levels found in previous statewide and regional benchmarks (ranging from 84% in the southeast region mobilization to 95% in the South Carolina mobilization), but these were the highest levels and the greatest increase documented at the national level. Awareness of enforcement, as indicated, was measured at about 40% in national telephone surveys. This was consider- ably lower than in prior programs, which ranged from 53% in the southeast (2001) to 85% in North Carolina (1993). As with media and enforcement indices, these results may reflect the difficulty of maintaining intensity as programs are expanded to larger and more complex geographical areas. On the positive side, the 40% awareness of enforcement was greater than the 35% awareness measured in the 2002 partial implementation states and much greater than the 28% aware- ness found in the 2002 no-paid-media states. Further, this in- crease in enforcement awareness was 24 points (50%) greater than the 16-point increase measured in the fall 2002 national mobilization and 33% greater than the 18-point gain measured in the May 2002 national mobilization. Thus, there is evidence that a substantial effort was initiated nationwide in May 2003. Impact on Observed Usage. The national SBU rate, as measured by the NOPUS, increased from 75% to 79%, a significant 4 percentage point increase. This was the largest nationwide increase measured by NOPUS since 1998, when two intensive Operation ABC mobilizations were imple- mented in the same year. Figure 16 shows that the average statewide usage rate, as meas- ured by mini-surveys conducted in 28 states just prior to the May 2003 mobilization, had declined to 72.8%, from its 2002 postmobilization level of 75.2% (−2.4 points). These declines were found in both primary law states (−3.2 points, from 82.6% to 79.4%) and in secondary law states (−1.8 points, from 71.0% to 69.2%). However, usage in these 28 states increased by an average of 5.7 points by the end of the 2003 mobilization, about 5 points in primary law states and about 7 points in secondary law states. Thus, the gain associated with the May 2003 mobi- lization may have been greater than the 4 points measured by NOPUS (which did not measure immediate preprogram rates). 2) The 2004 National CIOT Mobilization. An even greater level of media and enforcement intensity was implemented during the May 2004 national CIOT mobilization (Solomon and Chaffe, under review). The characteristics of this second national effort are summarized in Table 9. Paid Media. The 2004 effort included approximately $32 million for paid advertising (about 11¢ per resident over 2 weeks), including a $12 million national media pur- chase and approximately $20 million expended by the states. As in the previous year, paid media was supplemented by exten- sive earned media, including national network coverage, local radio and television news stories, articles and editori- als, and press events. 31 # States Law Min. Fine Media Cost Cost per Capita # SBU Citations Cites/ 10K 2 wks Cites/ 10K 1 wk Base Usage (%) Peak Usage (%) Chg Pct. Pts. 44 17 P 27 S varies $25 m 8¢ / 2 wks 508 K 20 10 75 79 +4 Media rates include both state and national expenditures. Citation rates are based on rates in reporting states; these rates ranged from 14 per 10,000 residents in secondary law states to 24 per 10,000 residents in primary law states. Legend: P = primary law; S = secondary law; Cites = citations; m = million; K = thousand Table 8. A summary of key characteristics of the 2003 national CIOT program (Solomon, Chaudhary, and Cosgrove, 2004).

32 84 79.4 82.6 75.2 72.8 78.5 75.8 69.2 71 60 62 64 66 68 70 72 74 76 78 80 82 84 86 2002 (post-CIOT) 2003 (pre-CIOT) 2003 (post-CIOT) Survey Period Us ag e (% ) Primary Law States Combined (28 States) Secondary Law States Figure 16. Observed change in SBU in 28 states: Postmobilization 2002 to postmobilization 2003 (from Solomon, Chaudhary, and Cosgrove, 2004). # States Law Type Min. Fine Paid Media ($) Per Capita Cost 2 wks # SBU Tickets Issued Tckts /10K 2 wk Tckts /10K 1 wk Preprog. Usage (%) Postprog. Usage (%) Usage Chg. (Pct. Pts.) 44 23 P 30 S Varies $32 m 11–12¢ 657,000 30 P 15 S 15 P 8 S 76.6 79 +2.4 Media costs include both state and national expenditures. Citation rates are based on 44 reporting states. Legend: P = primary law; S = secondary law; Tckts = Tickets; m = million; K = thousand; Prog. = Program Table 9. A summary of key characteristics of the 2004 national CIOT program (adapted from Solomon and Chaffe, under review). Citations. Across the 50 states, the District of Columbia, Puerto Rico, and the Virgin Islands, about 7,500 law enforce- ment agencies participated in the 2004 program, slightly more than in 2003 (about 7,125). These agencies reported issuing 657,300 citations for safety belt violations (about 22 per 10,000 residents over 2 weeks), compared with 508,492 citations in 2003 (about 20 per 10,000 residents). As in the previous year, more citations were written in primary law states (74%) than in secondary law states (26%), resulting in a rate of about 30 citations per 10,000 residents in primary law states and 15 per 10,000 in secondary law states. Awareness. Nationwide telephone surveys measured 83% awareness of recent safety belt messages (+ 13 points over baseline) and 41% awareness of safety belt enforcement efforts (+25 points over baseline). These were nearly identical to the postlaw rates and increases found in 2003. As in 2003, the awareness of recent safety belt messages (83%) was compara- ble to levels in benchmark programs, but awareness of enforce- ment efforts (41%) was considerably lower than in benchmark programs (ranging from 63% in the southeast to 73% in South Carolina and 85% in North Carolina). Impact on Usage. The national SBU rate, as measured by the NOPUS, increased from 79% in 2003 to 80% in 2004, a non- significant change. As with the 2003 national effort, Solomon and Chaffe reported pre-to-post program changes based on statewide surveys. They suggested that usage had declined by about 2 points from its 2003 peak of 79%, and then increased by about 2.4 points with the 2004 mobilization. These results suggested that, while the intensified paid media and enforcement efforts implemented as part of annual enforcement mobilizations have been associated with past in- creases in nationwide observed usage, further progress may be more difficult without some change in strategy. Subsequent mobilizations, conducted in 2005 and 2006, have reinforced that suggestion. 3) Efforts to Reach High-Risk, Low-Use Groups. Based on the fact that usage among high-risk, low-use target audiences remained lower than in the general population, several tar- geted media and enforcement efforts were planned and imple- mented as part of the 2004 and 2005 CIOT mobilizations. One of these was an effort to reach pickup truck drivers in the south central region of the United States. Another was an effort to reach rural motorists in the Great Lakes region. a. South Central Regional Effort to Reach Occupants of Pickup Trucks. As part of the 2004 CIOT mobilization, a tar- geted program to increase use among pickup truck occupants was implemented in five states in the south central region of the United States (Solomon and Chaffe, under review).

Paid Media. This program, called “Buckle Up in Your Truck” (BUIYT), included an additional 2 weeks of paid media, implemented prior to the national CIOT mobilization. Approximately $688,000 was expended on BUIYT advertise- ments during this initial phase of the mobilization. This trans- lated to about 2¢ per resident in these five states (Arkansas, Louisiana, New Mexico, Oklahoma, and Texas). The messag- ing during this initial phase of the mobilization was not enforcement-related. However, an additional $2.9 million (8¢ per resident) was spent on ads during the CIOT phase of the mobilization that followed. This messaging, as in all CIOT mobilizations, was specifically designed to publicize enforce- ment efforts. Awareness. As a result of these two media campaigns and the CIOT enforcement, awareness of messages to buckle up in pickup trucks among pickup truck drivers increased by 16 points, from 30% (preprogram) to 41% (post-BUIYT) to 46% (post-CIOT). Awareness of enforcement efforts increased by 23 points, from 55% (preprogram) to 58% (post-BUIYT) to 78% (post-CIOT). As these data show, most of the increase in awareness regarding SBU and pickups occurred during the BUIYT phase and most of the increase in awareness of enforcement occurred during the CIOT phase. This would be expected since the BUIYT advertising did not include an enforcement message and there was no enforcement during the BUIYT phase.81 Impact on Usage. Following 2 weeks of BUIYT media, there was a 2-point increase in observed usage among pickup truck occupants (63.7% to 75.3%) and a 1-point increase among passenger car occupants (74.6% to 80.8%). Associated with the CIOT phase, there was a 6-point increase among pickup truck occupants (65.9% to 71.6%) and among passenger car occupants (75.3% to 80.8%). Overall, belt use in pickup trucks increased by 8 points and use in passenger cars increased by 6 points (rounded to nearest whole percent). The significance of these increases was not reported. b. A Local Implementation Using Paid Media and the BUIYT Message. Using a local implementation to test the potential for target-specific messaging, BUIYT ads were purchased in Ama- rillo, Texas, during the November 2004 mobilization, with Wichita Falls serving as a control community. Solomon and Chaffe describe the results of this program as well. While the intensity of the media purchase was not described, they reported that awareness of the BUIYT message increased by 21 percentage points among Amarillo pickup truck drivers (from 43% to 64%) and awareness of enforcement increased by 5 points (from 69% to 74%). Interestingly, baseline and post- program awareness of enforcement was higher than the aware- ness of the buckle up message. Most importantly, Figure 17 shows that observed usage among pickup truck occupants increased by 12 points in Amarillo (from 72% to 84%), compared with 5 points in Wichita Falls (from 77% to 82%), a net 7-point difference. Usage among passenger car occupants increased by 8 points in Amarillo (from 81% to 89%) and by 5 points in Wichita Falls (from 84% to 88%, rounded), a net 4-point difference. From these reported results, it appears that, while there may have been some statewide increase in usage that affected both Amarillo and Wichita Falls, the BUIYT program in Amarillo had a modest additional impact. No additional enforcement activity, beyond whatever was implemented during the Fall CIOT program, was reported for Amarillo. 33 81A similar finding was reported by Nichols, Ledingham, and Preusser (2007) with regard to an RDP implemented in the Great Lakes Region. In this program, awareness of rural messages increased most during an initial RDP phase while awareness of enforcement increased most dur- ing the CIOT phase. 72 84 82 77 65 70 75 80 85 Premedia Postmedia Us ag e (% ) Amarillo (media) Wichita Falls (control) Figure 17. Observed change in usage among pickup truck occupants in Amarillo, Texas, and Wichita Falls, Texas (Solomon, Chaudhary, and Cosgrove, 2004).

c. Great Lakes Regional Effort to Reach Rural Motorists. Sev- eral targeted demonstrations, such as the 2004 BUIYT program in the south central region, have been implemented to target higher-risk, lower-use target populations. Additional BUIYT demonstrations, for example, were implemented in the south- east region and south central regions in 2005. Another pro- gram, which targeted rural belt use, was implemented in the Great Lakes region in 2005 and 2006. Final results are available for the first year of this Rural Demonstration Program (RDP) and preliminary results are available for the second year. Paid Media. The 2005 RDP was implemented in six states (Illinois, Indiana, Michigan, Minnesota, Ohio, and Wisconsin) and evaluated by Nichols, Ledingham, and Preusser (2007). Like the BUIYT program, the RDP included an additional 2 weeks of paid media, implemented prior to the national CIOT mobilizations conducted in May 2005 and May 2006. About $1.3 million was spent for paid media during the 2-week RDP (12¢ to 14¢ per capita). Following the RDP, an additional $4.4 million was spent on CIOT advertising (6¢ per capita) for the 2-week CIOT campaign. About two-thirds of this funding was spent on television. Although both the RDP and CIOT media campaigns were enforcement oriented, the CIOT cam- paign conveyed a “harder” message. Enforcement. In 2005, only three of the six states intensified enforcement during the RDP, but all states did so during the CIOT phase that followed. In 2006, all states intensified enforcement during both phases of the mobilization. Citations were issued at the rate of about 18 per 10,000 residents (in the rural targeted areas) during the RDP and 28 per 10,000 resi- dents (statewide) during the CIOT phase. Awareness. In general, awareness of safety belt messages in the rural targeted areas increased most during the RDP and aware- ness of enforcement activity increased most during CIOT. Dur- ing the RDP phase, awareness increased more in the rural tar- geted areas than it did statewide, suggesting that the rural messaging was reaching its targeted audience. By the end of both phases, however, rural and statewide awareness levels were similar, with about 85% awareness of safety belt messages and 55% awareness of enforcement in 2005. Impact on Usage. One of the key findings from the 2005 campaign was that only the three states that intensified enforce- ment during the RDP (Illinois, Indiana, and Ohio) experienced significant increases in usage in rural targeted areas during that phase. Further, only in these three states did the rural areas have larger overall gains, compared with statewide gains, after both phases of the program were completed (RDP + CIOT). In 2005, there was a median 7-point increase in the rural tar- geted areas and a median 5-point increase statewide. Based on the results during this initial year, it appeared that the combi- nation of 2 additional weeks of media (beyond that imple- mented as part of CIOT) and 1 additional week of enforcement were associated with greater impact in the rural targeted areas. However, rural and statewide usage increased significantly in most states and the differences between rural and statewide gains, while frequently greater in the rural areas, were not sta- tistically significant. Figures 18 and 19 show the 2005 trends for the six partici- pating states. There are three measurement points for each state (pre-RDP, post-RDP, and post-CIOT). Figure 18 shows trends for the three primary law states (Illinois, Indiana, and Michigan, from left to right), and Figure 19 shows the trends for the three secondary law states (Minnesota, Ohio, and Wisconsin, from left to right). Again, key findings of this first- year effort were that these states generally experienced signif- 34 60 65 70 75 80 85 90 95 IL1 IL2 IL3 * IN1 IN2 IN3 * MI1 MI2 MI3 Measurement Periods Us ag e (% ) Rural Targeted Areas Statewide Figure 18. Changes in rural and statewide observed usage in three primary law states participating in a 2-year RDP. (First-year data from Nichols, Ledingham, and Preusser, 2007; second-year data preliminary).

icant overall gains in rural targeted areas (median = 7 points), as well as statewide (median = 5 points), but that increases were observed only when enforcement was present. Cost Savings Estimates North Carolina Benchmark Impact on Fatalities. In order to meaningfully estimate cost savings associated with HVE efforts, these efforts must be implemented in a manner such that gains are sustained over time. This appears to have been the case with regard to the benchmark North Carolina program. Based on time- series analyses, Williams et al. (1996) reported a reduction of 45 deaths and 320 serious injuries during a 6-month period that followed implementation of the program, which was about a 7% decrease in driver deaths. These researchers also reported an estimated savings in medical and EMS costs of about $14.6 million.82 Impact on Injuries and Total Savings. Based on these esti- mates, it was calculated that about 1,100 moderate-to-serious (MAIS 2–5) injuries and about $115 million in total costs were prevented over this 6-month period (about $160 million in 2007 dollars). Based on the number of injuries prevented, there was an estimated savings of about $23 million in medical and emergency services alone ($32 million in 2007 dollars). This component constitutes about 20% of total savings. Sav- ings per death prevented (including the prevention of an esti- mated 25 MAIS 2–5 injuries per death) were about $2.6 mil- lion ($3.6 million in 2007 dollars). If impact was maintained over a full year, total savings would be about $320 million (in current dollars).83 Program Costs. All of the costs associated with the North Car- olina CIOT program are not documented by Williams et al. However, based on data provided in their report and in vari- ous other project reports, and on personal communications with the principal investigator and regional and state repre- sentatives, an estimated $2 million was spent on the two waves of activity, including media and enforcement-related efforts in the fall of 1993 and summer of 1994.84 Based on that estimate, Table 10 shows $117 million in (6-month) cost sav- ings (just over $230 million in annual savings), a return of 35 60 65 70 75 80 85 MN1 MN2 MN3 * OH1 OH2 OH3 * WI1 WI2 WI3 Measurement Periods Us ag e (% ) Rural Targeted Areas Statewide Figure 19. Changes in rural and statewide observed usage in three secondary law states participating in a 2-year RDP (First-year data from Nichols, Ledingham, and Preusser, 2007; second-year data preliminary). 82Serious injuries (reported by Williams et al.) are generally classified as MAIS 3–5 injuries. In order to estimate the impact on the broader cate- gory of moderate-to-serious injuries (i.e., MAIS 2–5), we multiplied the number of serious (MAIS 3–5) injuries by 3.5, resulting in an estimate of 1,100 MAIS 2–5 injuries prevented. This increased medical and emergency medical services (EMS) savings from $14.6 million to $23.4 million. Medical and EMS savings generally account for about 20% of total savings (associated with deaths and injuries prevented). Based on this relationship, the estimated total savings was $116.8 million. Using the unit costs for deaths and the proportions of costs accounted for by var- ious components (i.e., lost productivity, medical and EMS, and “other” costs) provided in Appendix E, we calculated values for the remaining cells in Table 12. Estimates in terms of current (2007) dollars were obtained by multiplying by a factor of 1.39, based on the inflation index available on the Bureau of Labor statistics Web site (www.bls.gov/cpi/). 83There is evidence that usage increased and then declined following each wave of the CIOT program. On the other hand, subsequent annual statewide surveys seldom measured a use rate less than 80% after 1994. Thus, it appears that, over the long term, increases associated with this HVE program were sustained. 84This estimate was based upon personal communications with persons involved with the program, including the Director of the Governor’s Highway Safety Program (GHSP) and the principal investigator.

about 115:1 on the funds expended. Medical and EMS sav- ings alone provided a return of about 23:1.85 Relevance in the Current Environment. The North Carolina CIOT program remains a benchmark for statewide HVE pro- grams. The substantial impact experienced in this program was likely the result of the intensity and duration of media and enforcement efforts, as well as the extensive use of road- blocks or checkpoints. Based on the results of telephone sur- veys, few residents were unaware of this enforcement effort. Although impacts of this magnitude (+16 points) have not been observed in most subsequent efforts, impacts of 8 to 10 points remain feasible, particularly in the 20 or more states with usage rates less than 80%. However, such states would likely require a primary law and an HVE mobilization with media and enforcement intensity (including checkpoints) compa- rable to that implemented in North Carolina.86, 87 If such a program were repeatedly implemented and if gains were sustained over the long term, fatality and injury reduc- tions of about 4% to 5% could be achieved and, for each death (and associated injuries) prevented, approximately $3.8 mil- lion in cost savings would result. However, the literature clearly suggests that HVE interventions must be repeated (even within each year) to maintain impact. Most HVE pro- grams have been implemented only annually, usually in con- junction with national mobilizations. Immediate Versus Long-Term Impact It is important to examine longer-term gains associated with repeated HVE implementations, particularly when such efforts supplement (or are supplemented by) primary law 36 Type of Loss Prevented Number of Prevented Incidents b Unit Costs Per Fatality c Lost Produc- tivity ($) Medical & EMS ($) Other Injury- Related ($) Total Economic Cost Savings Fatalities (2007 dollars) 45 $0.8 m ($1.1 m) $32 m 86% ($44 m) $1 m 2% ($1 m) $5 m 12% ($6 m) $37 m ($51 m) MAIS 2-5 Injuries (2007 dollars) 1,131 $1.8 m (per fatality) ($2.5 m) $40 m 50% ($55 m) $23 m 29% ($32 m) $18 m 22% ($24 m) $80 m ($111 m) Fatalities & Injuries (2007 dollars) 1,176 $2.6 m ($3.6 m) $71 m 61% ($99 m) $23 m 20% ($33 m) $22 m 19% ($31 m) $117 m ($162 m) Notes: a) Cost estimates were derived from information provided in Appendix E. Information in Appendix E derived from Blincoe and Faigin, 1992; Blincoe, 1996; and Blincoe, Seay, and Zloshnja et al., 2002. b) Fatalities and injuries prevented are based on the number of fatalities and “serious” injuries avoided, as reported by Williams et al., and translated to MAIS 2-5 injuries using ratios contained in Appendix E. c) Non-injury-related components (travel delay and property damage) are not included in these estimates. d) The estimate of moderate-to-serious injuries is based on the 320 reported serious injuries prevented. e) Total savings are estimated from reported medical and EMS savings, estimated to be 20% of total. f) Productivity, medical, and “other” component proportions are from tables in Appendix E. Table 10. Estimated 6-month cost savings in North Carolina following the implementation of a CIOT enforcement and publicity program (based on time series analysis and on fatality, injury, and cost data from NHTSA crash cost reports, which are summarized in Appendix E).a 85This 6-month estimate includes deaths prevented while usage was at its highest, immediately following the CIOT program in November 1993, and over much of the period prior to the next wave of activity, which occurred in July 1994. Thus, it appears to represent impact over a period when usage varied between 80% and 72%. Since the July wave was fol- lowed by another increase to about 80% usage, it would be assumed that that these 6-month estimates would be repeated over the next 6 months as well. However, only the initial impact was reported by Williams and his coinvestigators. 86Wave 1 included a media expenditure of nearly $450,000 (6¢ per capita over 8 weeks) and just over 36,000 citations (about 51 per 10,000 residents over 4 weeks). 87A similar number of fatalities and injuries avoided would also assume a comparable baseline number of occupant fatalities as in North Carolina, which is about 1,180 per year.

upgrades. Using data from three recent upgrade states in the Great Lakes region as examples (i.e., Illinois, Indiana, and Michigan), it becomes clear that the longer-term impact of repeated implementations eclipses the short-term gains so frequently reported in the literature, including reductions in deaths, injuries, and costs. These three states were chosen as examples because they represent states that have been partici- pating in annual Operation ABC/CIOT mobilizations since 2002 and each enacted a primary law upgrade either during that time period or immediately prior to it. Indiana implemented its upgrade in 1998, Michigan in 2000, and Illinois in 2003. An examination of only the short-term changes associated with the 2005 RDP effort in these three states (from 2004 to 2005) shows relatively small effects on usage among crash victims (+1.1 point in Illinois, −0.2 points in Indiana, and +2.6 points in Michigan). Out of context, this suggests that the mobilization (or HVE in general) did not have a major impact. However, as Figure 20 shows, each of these states experienced substantial long-term increases in usage among crash victims associated with HVE (and its law upgrades) in the years immediately prior to the 2005 mobilization and such increases continued in two of the three states. Large increases in Indiana from 2001 to 2003 were likely associated with the implementation of enforcement zones, a pio- neering enforcement approach similar to the use of checkpoints for impaired drivers. A 12-point increase in Michigan followed its 2000 standard enforcement upgrade and additional increases were associated with participation in subsequent mobilizations. Finally, a 10-point increase in Illinois followed its 2003 upgrade and participation in the 2004 CIOT mobilization. Looking at the longer term, a very different picture emerges of impact, the role of primary law upgrades, and the continu- ing impact of subsequent HVE efforts. Usage among victims in these three states increased substantially following their respec- tive upgrades and in conjunction with subsequent HVE efforts, culminating in the 2005 RDP, which is represented by the high- lighted area of Figure 20. In Indiana, FARS use increased from a 4-year preupgrade average of 32% to an 8-year postupgrade average of 43% (+11 points); usage in Michigan increased from 45% to a 6-year postupgrade average of 58% (+13 points); and FARS use in Illinois increased from 38% to a 3-year post- upgrade average of 46% (+8 points). Translated into changes in UPFC, these gains suggest an average of 63 fewer deaths per postupgrade year in Indiana, 167 fewer in Michigan, and 59 fewer in Illinois. The estimated cost savings associated with these reductions (and reductions in MAIS 2–5 injuries) aver- age more than $200 million per state, per year of postupgrade effort (about $238 million in current dollars). This examination is intended to illustrate several points: a) impact associated with interventions such as primary law upgrades and repeated HVE implementations should be viewed in the long term as well as in the short term; b) the combination of primary law upgrades and HVE is a powerful combination for impacting safety belt nonuse among persons involved in potentially fatal crashes; c) such combinations in the three states used as examples provided long-term cost savings that averaged more than $200 million per state, per year; and d) any state with a secondary law and a modest usage rate of 65% to 80% should be able to experience similar results. For primary law states with higher rates, additional targeting, greater intensity, and possible increases in sanctions may be needed. A Summary of the Impact of Safety Belt Enforcement Efforts Local Programs This review examined the outcomes from studies involving 30 city or county enforcement efforts implemented from 1985 through 1999, plus two targeted-population programs imple- mented in 2004 [i.e., the Amarillo, Texas, pickup truck effort (Solomon et al., 2004) and the Reading, Pennsylvania, night- time enforcement effort (Chaudhary et al., 2005)]. Additional summary information as it relates to various enforcement char- acteristics and dimensions are provided at the end of the “Local Enforcement” section. Important local benchmark programs included: a) the three blitz programs implemented in Elmira, New York, in 1985, 1986, and 1999, respectively; b) the comparison of blitz and sustained enforcement conducted in Albany and Greece, New York, in 1986; and c) the three blitz programs implemented in Albemarle, High Point, and Haywood County, North Carolina, in 1993, just prior to the first statewide CIOT program. 1. The three Elmira blitzes were associated with increases of 28, 14, and 18 percentage points, respectively, from baseline rates of 49%, 66%, and 72%, respectively (Williams et al., 37 20 25 30 35 40 45 50 55 60 65 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 FA R S Us e (% ) Indiana Michigan Illinois Upgrades Figure 20. Changes in usage among fatally injured occupants in three primary law states participating in the 2005 Great Lakes RDP (solid symbols indicate primary law in effect).

1987, 1987, 2000). Enforcement activity included both warnings and citations issued primarily at roadblocks or checkpoints. Peak intensity was measured at 267 warnings per 10,000 residents in a 1-week wave that focused only on warnings and 52 to 54 tickets per 10,000 residents in the two waves (7 days and 12 days, respectively) that emphasized ticketing. Publicity included a combination of earned and paid media and, in the 1999 effort, feedback signs. No mea- sure of paid media was available for the third blitz, but the first two involved expenditures of 26¢ to 31¢ per resident. 2. The Albany and Greece, New York, demonstrations (Rood et al., 1987) achieved usage increases of 13 points (blitz) and 17 points (sustained), respectively, from baselines of about 50%.88 The blitz effort averaged about 40 citations per 10,000 residents (per week), comparable to those reported for Elmira, while the sustained effort resulted in only about one citation per 10,000 (per week) over a period of nearly 6 months. Over the 6-month period, the sustained effort achieved an increase that was slightly greater than the blitz program. In addition, after 4 months of follow-up, the sus- tained program was followed by a smaller postprogram decline than the blitz program. Similar findings were found in Texas (Mounce et al., 1990) and Illinois (Mortimer et al., 1990; Mortimer, 1992).89 3. The three North Carolina blitz programs conducted in High Point, Elizabeth City, and Haywood County were associ- ated with increases of 13 points, 10 points, and 38 points, respectively, from modest baselines of 65%, 69%, and 43%, respectively (Williams et al., 1994). The large 38-point increase in Haywood County was likely influenced by the low baseline rate and the fact that this was the first such pro- gram to be implemented in that county. Collectively, these three demonstrations showed that a combination of intense enforcement (20 to 80 tickets per 10,000 residents per week over 4 weeks) and paid media (about 8¢ per capita per week over nearly 5 weeks) with a hard enforcement message was associated with large and significant increases in usage from modest baselines (average = 60%). 4. Nighttime enforcement efforts have been shown to increase usage among late-night, high-risk occupants, including bar patrons. These studies are of increasing importance due to high rates of observed use in some states, most of which have much lower usage among crash victims (see “Discussion” section).90, 91 Local programs that have con- tributed to the understanding of the nighttime usage issue include the bar patron studies conducted in New York (Preusser et al., 1986) and Nova Scotia (Malenfant and Van Houten, 1988), as well as the early blitz programs con- ducted in Ottawa, Ontario (Grant, 1991) and Binghamton, New York (Wells et al., 1992). More recently, nighttime enforcement has been evaluated in Reading, Pennsylvania (Chaudhary et al., 2005). Collectively, these studies show that: a) nighttime belt use is lower than daytime use, particularly among late-night road users and those leaving drinking establishments; b) while day- time enforcement is nearly always associated with significant increases in daytime usage, it generally does not impact night- time SBU to the same extent;92 and c) nighttime enforce- ment, whether focused on safety belts alone or combined with alcohol-impaired driving programs, has been shown to impact nighttime usage, including usage among late-night road users and bar patrons. In some cases, there were increases in night- time usage while daytime usage remained unchanged. Again, these findings are particularly relevant to current situations where observed usage may be very high but where nighttime 38 88The increase associated with the blitz approach in Albany was meas- ured over a period of 6 months that included four waves of blitz enforce- ment. Thus, it is a longer term increase than that reported for many of the blitz efforts. 89Because these programs are 15+ years old and did not use paid media, their current relevance is limited. Still, the consistency of results, com- bined with the fact that blitzes have had diminishing returns in recent years, suggests further examination of the potential for sustained enforce- ment efforts. 90Of current interest are studies that either measured the impact of daytime enforcement on nighttime SBU or the impact of nighttime enforcement efforts (on daytime or nighttime usage). Statewide obser- vational surveys do not include nighttime usage, even though this is when many high-risk drivers and occupants (e.g., young persons, males, drinking drivers and passengers, and traffic law violators) are more prevalent on the roadways. There is reasonably consistent evidence that observed nighttime usage is modestly lower than daytime usage, and even lower very late at night (i.e., between 9 p.m. and 3 a.m.). There is an even greater gap between daytime and nighttime usage among occu- pants killed in crashes (i.e., FARS usage), particularly when comparing usage among those killed very late at night (see summary of nighttime usage and enforcement issues and results provided in Appendix C). 91Increasing usage among occupants involved in serious (potentially fatal) crashes will be essential to further reduce deaths and injuries in high-use states. However, this group constitutes but a small subpopu- lation of road users—day or night. Even with measures of impact on nighttime usage, one cannot determine the precise impact on the crash- involved population. However, if it can be demonstrated that usage is increased among nighttime drivers (and their passengers), particularly among those likely to have been drinking and/or those on the road very late at night, there is an increased likelihood that usage among the crash-involved occupants is also affected. 92A very recent statewide study conducted in Indiana and reported by Vivoda, Eby, St. Louis, and Kostyniuk (2007) reinforces the finding that daytime blitz enforcement may not increase usage among nighttime road users. In this study of a daytime CIOT mobilization, nighttime usage actually declined, which was apparently associated with the mobi- lization. Other studies have either found a smaller nighttime increase (compared with daytime) or no increase at all in nighttime usage.

usage and usage among fatally injured occupants remains much lower (see “Discussion” section). Statewide, Regional, and National Results and Trends Multijurisdictional HVE efforts have been shown to have a significant impact on SBU statewide, regionally, and nation- ally. Seventeen years of experience with such programs in- clude the 1991–92 Operation Buckledown/70% by ‘92 pro- gram, the 1993–94 North Carolina CIOT program, 6 years of OP-sTEP demonstrations implemented in 20 states as part of Campaign Safe and Sober, and 8 years of Operation ABC/CIOT mobilizations. These efforts, in conjunction with law upgrades, have not only demonstrated immediate increases in those juris- dictions where HVE has been implemented, they also have been accompanied by a 30-point nationwide increase in observed usage, from just under 50% in 1990 to just over 80% today. In addition, these programs have provided much information regarding the types and levels of enforcement and media asso- ciated with state, regional, and national implementations. The Benchmark Statewide STEP. Implemented in North Carolina as the CIOT program in 1993, it was part of a 5-year combined effort to increase SBU and decrease alcohol-impaired driving (the latter objective to be accomplished via a similar effort called “Booze It and Lose It”). The first two waves of the CIOT program were accompanied by an immediate 16-point increase in statewide usage, from 63% to 79% (Williams et al., 1996). This effort involved more than 6,000 checkpoints over 7 weeks (an average of 900 per week); nearly 60,000 citations (about 81 per 10,000 residents overall or about 12 per week); and $600,000 in paid media over 15 weeks (about 8¢ per res- ident overall or about 0.5¢ per week). Media conveyed a “hard” enforcement message—CIOT. Following this initial effort, ob- served usage in North Carolina remained at about 80% over the next 4 years of the CIOT program and has since increased (modestly) to a 2006 reported rate of 89% (NHTSA, 2007). Magnitude of Impact. Most multijurisdictional efforts have not achieved immediate increases of the magnitude found in North Carolina. The most often cited programs achieved immediate increases of 7 to 9 percentage points (e.g., 7 points in the 2000 South Carolina CIOT; an average of 9 points in the eight-state southeast regional CIOT implemented in 2001 (range: 4 to 20 points); and an average of 9 points in the 10-state fully implemented model programs implemented in 2002 (range: 3 to 19 points)). Further, there is some evidence of smaller gains associated with programs implemented since 2003. For example, there was a median 5-point gain in the six states participating in the 2005 Great Lakes RDP (range: 3 to 8 points), and there have been decreasing overall gains in the national CIOT mobilizations implemented since 2003. Program Intensity. It is likely that some of the declining impact associated with statewide and national programs is due to increasing baselines and repeated implementations. How- ever, there is also evidence of lesser intensity and differences in approach associated with recent implementations, compared with the benchmark effort. Few programs implemented since 2000 have used roadblocks or checkpoints to the extent that North Carolina did and the duration and intensity of recent enforcement efforts, as measured by citation rates, have generally not been as great as those of the benchmark (measured per wave). The latter included 7 weeks of enforcement (two succes- sive waves) with about 81 citations issued per 10,000 residents (overall). Most subsequent statewide programs have imple- mented 2 weeks of enforcement, resulting in about 21 citations issued per 10,000 residents (over the 2-week period). On the other hand, the intensity of media efforts in recent mobilizations has been similar to that of the benchmark (at the program or wave level). Expenditures have averaged about 6¢ to 12¢ per capita (over 2 weeks), compared with 8¢ per capita (over 7 weeks) in North Carolina. Thus, on a per-week basis, the intensity of recent media efforts has generally been greater (3¢ to 6¢ per capita) than that of the benchmark (0.5¢ per capita). Awareness. In the end, however, awareness of enforce- ment, which is thought to be the product of enforcement and media intensity and duration, has generally been lower in post-2000 mobilizations than in the benchmark. The combi- nation of media and enforcement activity in North Carolina resulted in an 89% awareness of checkpoints and an 85% awareness of the CIOT program. By comparison, awareness of enforcement averaged 59% in the 10 fully implemented states in 2002 and only 26% to 35% in the less fully implemented states. Nationwide, awareness of enforcement was only 40% to 41% following the 2003 and 2004 national CIOT mobiliza- tions, less than half that measured in North Carolina.93 This may reflect the lower levels of enforcement intensity and lesser use of checkpoints, in spite of equal or greater media intensity in the recent programs. Some Observations. The review of HVE implementations suggests that: a) it is more difficult to implement a statewide (or broader-based) program with the same intensity as a local 39 93Only the 2000 CIOT mobilization in South Carolina resulted in a rel- atively comparable awareness of enforcement (73% to 82%, depending on the measure used). Unfortunately, that program was followed by a political controversy that greatly reduced enforcement in subsequent months and years.

program; b) paid media appears to greatly facilitate the impact of statewide (or broader-based) programs (see Milano, McInturff, and Nichols, 2004); c) as with local programs, repeated waves of activity produce a ratcheting or saw blade effect, with initial increases followed by modest postprogram declines, followed by additional increases and modest post- program declines with each subsequent wave of activity (see Solomon et al., 1999); d) successive program implementations will likely result in progressively higher levels of usage and awareness (of both enforcement and safety belt messages) and smaller increments of change from these higher levels (see Solomon et al., 1999 and Nichols et al., 2007); and e) in current secondary states, a primary law upgrade will substantially increase the gains associated with HVE mobilizations. Longer-Term Impacts of HVE/STEP Efforts. Fifteen states participated in the state, regional, or model STEP (HVE) programs implemented in 2000, 2001, or 2002. Nearly all of these states also participated in subsequent national CIOT mobilizations. Looking only at these 15 states, there was an average 15-point increase in observed usage, from 1999 through 2005.94 The greatest increase was found among five states (Washington, Illinois, Indiana, Alabama, and Tennessee) that also upgraded their secondary laws during this period. These states experienced an average increase of 19 points from aver- age baselines of about 65%. One state with an existing primary law (Texas) experienced a 16-point increase, from a baseline of 74% usage. Increases among seven states that retained their secondary laws (West Virginia, Vermont, Florida, Mississippi, South Carolina, Kentucky, and Nevada)95 were lower, with an average increase of 14 points from an average baseline of 63%. Thus, most of these states continued periodic HVE activity over several years and initial increases, which averaged 8 to 9 per- centage points, were subsumed within longer-term increases averaging 15 points. In addition, states that upgraded their laws during this period nearly always experienced larger increases than states that retained their secondary laws.96 Impact of HVE on Usage Among Crash Victims. Changes in FARS use in these 15 states were somewhat smaller than increases in observed use and they emerged from much lower baselines. Six-year increases in use among crash victims aver- aged 7 points in the 15 participating states: 10 points in five states with new upgrades (average baseline of 36%); 9 points in six states with existing upgrades (average baseline: 36%); and 6 points in seven secondary law states (average baseline of 32%). Usage among victims did not increase appreciably in North Carolina during this period, which is presumably related to the fact that baseline FARS usage was already higher than in any of the participating states (50%), and there is some indication that recent mobilizations in North Carolina have not been of the same intensity as the benchmark program. Establishing Performance Measures. Much information has been gained over the past decade with regard to the charac- teristics of media and enforcement that have been associated with HVE, STEP, or mobilization efforts. In spite of these increases in knowledge, it has been difficult to quantify thresh- old levels of the intensity or duration of enforcement and media required for impact or to establish a quantitative relation- ship between intensity, duration, and magnitude of increase. This difficulty is due in part to the substantial variation in enforcement and media approaches, types of messaging, fine levels, baseline rates, etc. Summarized in this report are the characteristics of enforcement, media, and awareness that have been part of past program implementations. Efforts to use these data to develop performance measures are currently being explored. 40 94Appendix D provides more detailed data on usage rate trends (Observed and FARS) in these states. 95South Carolina (Dec. 2005), Mississippi (April 2006), and Kentucky (Jan. 2007) have since upgraded their secondary laws but not in time to impact 2005 usage rates. 96Low usage persists in South Carolina, which was one of the original CIOT states, but which suffered from political opposition to safety belt enforcement and only recently implemented a primary law (December 2005). Thus, the impact of this upgrade would not be apparent in this comparison.