Methods

A. Program Description

The RDP was implemented just prior to the May 2005 Click It or Ticket mobilization. The scheduling of media, enforcement, and survey activity for both the RDP and CIOT phases of the mobilization is shown in Figure 1.

Figure 1
Media, Enforcement and Evaluation Activity Schedule
For the RPD and CIOT Phases of the May 2005 Mobilization


milestone chart

1. Media

Developing the State media plans required: (1) selecting target markets and counties within a State; 2) developing media placement plans; and (2) developing media concepts and materials to be used in the execution of the media plan.

Target Markets. Because funding was not available to cover the entire rural population in any State, only selected media markets or, in some cases, specific counties within these markets were targeted. The Tombras Group provided the States with maps of media markets and estimates of the costs of placing ads in each of them. In consultation with Tombras and PRG, the States then made the final selection of counties. Factors which entered into this decision included population density, number of fatalities, and the estimated cost of placing ads in the various markets (or counties within markets). The regional map that follows shows the areas across the region that were targeted for paid media during the RDP. Table 1 provides additional information for each State including: media markets involved, number of counties, population of the targeted area, and percent of total State population.

Figure 2

map

 

Table 1
Media Markets, Counties, and Population in RDP Rural Targeted Areas

State

Media Markets

# of Counties

Population

% of State Population

IL

5 markets: Peoria, Champaign, Rockford, Davenport, and St. Louis (MO)

42

2.88 million

23%

IN

5 markets: Cincinnati, Evansville, Indianapolis, Louisville, Terre Haute

27

0.66 million

11%

MI

5 markets: Flint/Saginaw/Bay City, Traverse City/Cadillac, Marquette, Lansing, and Alpena

56

2.78 million

28%

MN

4 markets: Rochester, Duluth, Minneapolis/ St. Paul, and La Crosse (WI)

54

1.67 million

33%

OH

7 markets: Toledo, Dayton, Columbus, Fort Wayne, Lima, Zanesville, and Charleston (WV)

15

0.66 million

6%

WI

4 markets: Duluth/Superior, Wausau/Rhinelander, La Crosse/Eau Claire, and Madison

40

1.86 million

34%

GLR

30 markets

234 counties

10.52 million

20.6%


Media Placement Plan. After the States identified their markets and resources, Tombras developed media placement plans for each of them. These plans covered audience segment profiles, geography, television and radio strategies, and dollar allocation by medium and by market. Each State provided feedback to Tombras, which then finalized the plan. Several States selected Tombras to execute this plan, while other States gave the plan to their own media contractors who executed it, either in modified or unmodified form. Section 157 (Title 23 U.S. Code) grant awards to the States provided for nearly all the funding for paid media in the States.

Concept Development. Based on problem identification information provided by PRG, 1 the Tombras Group developed three media concepts, focus-tested these concepts in four States, and provided the findings to the States for discussion. Following these discussions, the States selected a single concept to be used in the paid media effort. It was titled “The Friendly Cop.” In this ad, a local police officer pulls over a driver for not buckling up. After issuing a ticket for the seat belt violation, the officer tells the young man that he will see him later at the ball game. The violator looks on in amazement, clearly surprised that he has received a ticket from this officer whom he knows. The “Friendly Cop” concept addressed the State’s objective of increasing young male drivers’ perception of being caught for a seat belt violation. More focus group respondents, which consisted primarily of young men, identified the main message as “They’re getting serious about writing tickets for seat belts,” or “They’re STEPping up enforcement of seat belt laws,” or “No exceptions.”

Production and Placement. Tombras produced two 30-second television ads, one for primary law States and one for secondary law States. The agency also produced a 30-second radio script based upon the same script. Complementary outdoor billboard art, posters, and print ads (three versions) were developed using a complementary “No exceptions/No excuses” theme that was to be delivered by law enforcement officers. The “No Exceptions” poster can be seen in Appendix A.

Ads were purchased for a two-week period (May 2-15). The media contractors purchased space at the times of day and on the formats and shows that had the greatest opportunity to be viewed or heard by young men age 18 to 34. The greatest emphasis was on the youngest males, under 25. These contractors placed advertising within the specified parameters and they obtained “bonus” or “added value” spots to be played at similar times. Due to common demographics in the rural areas of the GLR States, they purchased similar programs and similar time parts in each State, targeting broadcast and cable television programming such as: Saturday Night Live, Mad TV, Fear Factor, WWF Smackdown, NASCAR, Extreme Sports, King of the Hill and Everybody Loves Raymond. The radio strategy focused on Alternative, Country, Top 40 and Rock, time parts including afternoon and evening drive time, some morning drive time, and weekend days. Funds spent on paid media were monitored in each State. Gross rating points (GRPs) and number of ads purchased were also monitored, usually only where the NHTSA contractor made the media purchase2 .

Earned Media. Although paid media was the major media activity, the RDP also included an earned (news) media effort. NHTSA asked each Highway Safety Office to contact enforcement and rural community partners to enlist their aid in getting news media attention for the rural crash problem and for the ongoing enforcement effort. After soliciting input from States with regard to the type of material that would be most useful, Tombras developed two versions of a media planner to aid the States in this activity. One planner was designed for enforcement agencies and one was designed for rural community organizations. The planners contained a news release, an op/ed article, a letter to the editor, a fact sheet with talking points, a drop-in newsletter article, and a PowerPoint presentation. The State Highway Safety Offices delivered these planners to the appropriate organizations, either in person or via mail. Law enforcement liaisons (LELs) frequently delivered the material to enforcement agencies as part of scheduled visits. One planner is provided in Appendix B. Both are available on the RDP Web site at http://www.greatlakesproject.org/.

2. Enforcement

Illinois, Indiana, and Ohio intensified enforcement of seat belt laws during the RDP. Illinois and Indiana conducted enforcement zones (EZs), which are similar to checkpoints or safety checks, except that vehicles passing through a zone are stopped only if an occupant covered by the seat belt law is observed to be unbuckled. Ohio conducted other types of enforcement. Rural enforcement grants were provided to State and local agencies in Illinois. Similarly, 11 rural agencies in Indiana received grants to conduct EZs and three received funds to conduct special patrols. The Indiana State Police (ISP) agreed to conduct a minimum of 10 EZs in each of 13 targeted counties. All enforcement efforts were supported by a combination of overtime and incentive grants. In Ohio, LELs asked local police departments and county sheriffs to conduct rural enforcement efforts during the RDP. The Ohio State Highway Patrol conducted seat belt enforcement efforts in the 15 targeted counties. Wherever available, information was gathered by the States relative to the level and type of enforcement activity being conducted (e.g., number of agencies, citations, hours worked, overtime hours, etc.) and this information was provided to the central evaluator.

B. Evaluation

Each State’s evaluators designed, implemented, and analyzed the results of observational, telephone, and (in three States) motorist surveys. Summary results and, in some cases, raw survey data were provided to PRG for the regional evaluation. Baseline surveys were conducted prior to the start of the RDP program (Wave 1); just prior to the start of CIOT paid media (Wave 2); and after the completion of CIOT enforcement (Wave 3).

1. Measuring Changes in Awareness
Telephone and motorist surveys3 were used to measure changes in awareness of general seat belt messages and of enforcement-related messages and activity. Table 2 provides a summary of key characteristics of telephone surveys conducted in the States.

Table 2.
Number and Size of State Telephone Surveys
Statewide, Rural-Targeted, and Nontargeted Areas4

State

Sample Frame

Wave 1 Pre-RDP

Wave 2 Pre-CIOT

Wave 3 Post-CIOT

Estimated. Error (+/-)

IL

Statewide

n = 580

 

n = 563

4.2 %

 

Rural Targeted

n = 222

n = 257

n = 225

6.7%

 

 

 

 

 

 

IN

Statewide

n = 1520

 

n = 1520

2.6%

 

Rural Targeted

n = 364

n = 319

n = 385

5.6%

 

Non-targeted

n = 330

n = 318

n = 335

5.6%

 

 

 

 

 

 

MI

Statewide

n = 400

n = 400

n = 400

4.9%

 

Rural Targeted

n = 150

n = 150

n = 150

8.2%

 

 

 

 

 

 

MN

Statewide

n = 768

 

n = 768

3.6%

 

Rural Targeted

n = 352

n = 202

n = 352

7.1%

 

 

 

 

 

 

OH

Statewide

n = 907

 

N = 927

3.3%

 

Rural Targeted

n = 300

 

N = 331

5.8%

 

 

 

 

 

 

WI

Statewide

n = 411

 

N = 411

4.9%

 

Rural Targeted

n = 252

n = 239

N = 238

6.5%

In addition to two waves of Statewide surveys (w1 and w3), most States conducted three waves of rural surveys (w1, w2, and w3). The rural surveys were usually over-samples conducted as part of the statewide polling. They provided an opportunity to measure change from baseline to post-RDP (w2-w1) and from post-RDP to post-CIOT (w3-w2). They also permitted a comparison of overall change (w3w1), statewide and in rural areas. All three waves of rural surveys were implemented in five States. Ohio conducted only two waves of rural surveys (w1 and w3) and Indiana conducted three waves of surveys in both targeted and nontargeted rural areas. All these surveys were random-digit-dial (RDD) surveys conducted either by commercial polling firms (as in Michigan and Minnesota) or by university research departments (as in Illinois, Indiana, Ohio, and Wisconsin). Each State used a modified version of a survey instrument developed by NHTSA.5 The number of respondents in the statewide samples ranged from 400 in Michigan and Wisconsin to 900 in Ohio. The number of respondents in the rural surveys ranged from 150 in Michigan to 300 in Ohio.6

In three States, awareness surveys were also conducted at Department of Motor Vehicles or Bureau of Motor Vehicles licensing centers. These surveys provided for additional comparisons of changes in rural and statewide samples (in Illinois and Indiana) and in targeted and nontargeted rural samples (in Indiana and Wisconsin). They involved one-page, paper-and-pencil questionnaires, with questions similar to those included in telephone surveys. Surveys were conducted at 76 centers in the three States; 31 centers were in targeted rural counties and 14 were in nontargeted counties. In nearly all cases, these centers were in counties where rural observational surveys were conducted. Table 3 shows the sample sizes of DMV/BMV surveys conducted in the three States.

Table 3.
Number and Size of State BMV/DMV Motorist Surveys
Statewide, Rural-Targeted, and Control Samples

State

Sample Frame

Wave 1 Pre-RDP

Wave 2 Pre-CIOT

Wave 3 Post-CIOT

IL

Statewide

N = 923

n = 701

n = 857

 

Rural Targeted

N = 544

n = 211

n = 103

IN

Statewide

N = 2011

n = 1895

n = 1628

 

Rural Targeted

N = 1376

n = 1086

n = 915

 

Non-targeted

N = 877

n = 838

n = 703

WI

Statewide

n/a

n/a

n/a

 

Rural Targeted

N = 481

n = 530

n = 540

 

Non-Targeted

N = 97

n = 95

n = 99


2. Measuring Changes in Seat Belt Usage
Changes in seat belt usage were measured by means of observational surveys. These efforts, summarized in Table 4, consisted of statewide and rural-targeted surveys in all six States and rural, nontargeted surveys in Indiana and Minnesota. Statewide surveys were of two types: full surveys and mini-surveys. Full statewide surveys met the requirements established for statewide observational surveys under Section 157 (U.S. Code 23).7 These surveys ranged in size from 117 sites in Indiana to 265 sites in Ohio. One full survey (at w3) was conducted in Illinois, Indiana, Minnesota, and Wisconsin. In Michigan, a full survey was not conducted as part of this evaluation. Instead, two mini-surveys each taken at 192 sites were used to measure statewide change (at w2 and w3). Similarly, while Illinois conducted a full statewide survey (at w3), it also conducted a 50-site mini-survey at that time. The mini-survey was used for this evaluation as it was most comparable to previous surveys (at w1 and w2). Ohio conducted three full (265-site) surveys (w1, w2, and w3).

Mini-surveys were used to measure statewide baseline and post-RDP usage in most States (at w1and w2). These surveys were smaller than the full surveys, with the number of sites ranging from 50 in Illinois to 192 in Michigan. They could be completed in a few days, rather than a few weeks, making them more suitable for use at several stages of a one-month program. Mini-survey sites were nearly always selected from sites in the full survey and the same procedures were followed in conducting both types of surveys.

Mini-surveys were also used to measure change in rural-targeted areas in all six States. These surveys ranged in size from 25 sites in Illinois to 60 sites in Michigan.8 Mini-surveys were also conducted in nontargeted rural areas in Indiana and Minnesota.9 The nontargeted-area surveys in Indiana and Minnesota each consisted of 30 sites. They were conducted by the regional evaluation contractor. In these cases, the State evaluators provided PRG with a list of sites, along with written procedures for conducting the surveys.

Changes in awareness and observed seat belt usage in the rural targeted areas constituted the primary measure of impact of the RDP. These changes were measured post-RDP (w2-w1), post-CIOT (w3-w2), and overall (w3-w1). The significance of changes was tested by means of chi-square analyses. Where complete data was available, logistic regression analyses were also applied to statewide and rural data to examine differences in trends among the various samples.

Table 4.
Observational Surveys Conducted
Type of Survey and Number of Sites
Statewide, Rural-Targeted, and Nontargeted Areas

State

Sample Frame

w1 Pre-RDP

w2 Pre-CIOT

w3 Post-CIOT

 

State

Sample Frame

w1 Pre-RDP

w2 Pre-CIOT

w3 Post-CIOT

IL

Statewide

mini 50 site

mini 50 site

mini/full* 50 site 258 site

MN

Statewide

mini 84 site

mini 84 site

full 240 site

Rural Targeted

mini 25
site

mini 25
site

mini 25
site

Rural Targeted 1 2

mini 28
site 30 site

mini (2)
28 site 30 site

mini (2)
28 site 30 site

Non-targeted

n/a

n/a

n/a

Non-targeted

mini 30 site

mini 30 site

mini 30 site

IN

Statewide

Mini 75 site

mini 75 site

full 117 site

OH

Statewide

full 265 site

full 265 site

full 265 site

Rural Targeted

mini 30 site

mini 30 site

mini 30 site

Rural Targeted

mini 49 site

mini 49 site

mini 49 site

Non-targeted

Mini 30 site

mini 30 site

mini 30 site

Non-targeted

n/a

n/a

n/a

MI

Statewide

--

mini 192 site

mini 192 site

WI

Statewide

mini 56 site

mini 56 site

full 240 site

Rural Targeted

mini 60 site

mini 60 site

mini 60 site

Rural Targeted

mini 32 site

mini 32 site

mini 32 site

Non-targeted

n/a

n/a

n/a

Non-targeted

8 site

8 site

8 site

* Illinois conducted a mini and a full statewide survey at w3 (post-CIOT)

 


1 The Tombras Group used the demographic profile provided by PRG. Key characteristics included:

  1. Passenger cars account for more rural deaths than any other vehicle type with pickup trucks second.
  2. Young drivers (16 to 24) have the highest number of rural deaths per age and year but occupants in the mid range (25 to 54) account for a high number of deaths as well.
  3. Young drivers have the greatest number of unrestrained deaths as occupants of passenger cars.
  4. Males and drivers account for 2 ½ times as many unrestrained deaths as females and passengers.

2 Gross rating points (GRPs) are a measure of exposure. They are measured per market, per week of advertising. A GRP of 300 was considered to be a “strong” media effort in this program.

3 Motorist surveys are surveys conducted at driver license centers of State Departments of Motor Vehicles or Bureaus of Motor Vehicles. The surveys consist of a one-page questionnaire regarding knowledge, attitudes, and perceptions regarding (in this case) seat belt use, media, and enforcement activities, usually printed in English on one side and in Hispanic on the other side. Most often, a contractor working for the state Office of Highway Safety offers these surveys to drivers visiting licensing offices and asks these drivers to complete the surveys while waiting for photos to be taken.

4 The error estimate (for each row) reflects the expected sampling error of a simple random sample (for the smallest sample size in each row), at the 95% confidence level, when p = q = 0.5.

5 The occupant protection portion of this survey (as adapted by Minnesota) is shown in Appendix C

6 In situations where a rural over-sample is added to a statewide sample, the resulting sample size is larger than just the over-sample. A certain percentage of the statewide sample itself involves rural respondents.

7 These requirements were established as part of Section 157 of the Transportation Equity Act for the 21st Century (TEA-21)
and are found in Section 157, 23 U.S. Code. A copy of these guidelines can be found in
Appendix E.

8 Most of the mini-surveys consisted of sub-samples of rural sites included in the full statewide survey and within the
boundaries of the targeted markets. In at least one case, a totally independent, probability-based survey was designed to
measure seat belt use in the rural targeted areas.

9 In Wisconsin, sites from one nontargeted media market were examined as well.