The following results address: (1) how much media and enforcement activity occurred as part of the RDP and CIOT; (2) changes in public awareness and perceptions; and (3) changes in observed seat belt use associated with each phase of the mobilization.
During CIOT, an average of about 4,000 television ads and just over 2,200 radio ads aired per State, totaling about 6,200 electronic media ads per State over the two-week period (about 7 ads per 10,000 residents). By this measure, the greatest saturation during CIOT was in Indiana (11.9 ads per 10,000), Wisconsin (11.5), and Minnesota (10.7), with lower levels in Ohio (4.3), Michigan (5.5), and Illinois (6.4). GRP data, available for only three States, showed and average of 517 GRPs (per market per week) in these States, well beyond the targeted minimum of 300-400. Table 8 summarizes various indices of paid media activity by State.Earned media was generated in every State, generally associated with press events, press releases or outreach activities. However, there was limited documentation of the number of media events held or news stories aired during the RDP. More complete data were provided for the CIOT phase, when at least 100 media events were conducted across the region, mostly as kick-off events. Ohio reported the most events (54). Other States reported 6-16 events per State. More than 500 television (TV) news stories and perhaps twice as many radio news stories were aired across the region.
During the CIOT phase, nearly 2,300 enforcement agencies participated in the GLR mobilization, representing an average of about 65 percent of all relevant agencies in each State. Illinois, Indiana, and Michigan conducted a total of 5,070 EZs (plus special and regular patrols). Minnesota, Ohio, and Wisconsin conducted special patrols. Nearly 120,000 citations were issued for seat belt and child restraint violations across the region (about 23 per 10,000 residents). Michigan had the highest citation rate (32 per 10,000 residents), followed by Illinois and Indiana (25), Minnesota (24), Wisconsin (20) and Ohio (15). Ohio reported 83 hours worked per 10,000 residents, followed by Wisconsin (59), Michigan (44), Indiana (23), Minnesota (16) and Illinois (11).11 Finally, Illinois implemented 2.3 enforcement zones per 10,000 people, followed by Indiana (2.2), and Michigan (1.8).
General Seat Belt Messages
Enforcement-Related Messages and Activity
Source(s) and Formats of Messages Received
Figure 2 shows RDP-related increases (w2-w1) in all three general awareness indices: buckle up (13 points); more than usual messages (26 points), and recognition of CIOT (14 points). Increases in these indices were significant in nearly all States. In addition, increases in awareness of special efforts by police to ticket were significant in all five States that provided data on this index (average increase = 19 points; p ≤ 0.05).14 There were smaller increases in the remaining indices: specific enforcement efforts (8 points), police writing more tickets (4 points); and risk of receiving a ticket (3 points). These latter changes, while consistent, generally did not reach significance during the RDP. 15
Following CIOT, all States reported increases regarding all indices in their targeted rural areas. This was the case for both general and enforcement-related messages. The largest average change (+22 points) was in awareness of special efforts by police to ticket. This is consistent with the fact that all States intensified enforcement and implemented their CIOT paid media efforts during this phase. Figures 3 and 4 show changes in awareness of general and enforcement-related messages, respectively. General message awareness tended to increase more during the RDP while awareness of enforcement-related messages tended to increase more during CIOT. Awareness of special police efforts to ticket increased in a nearly linear fashion throughout the mobilization.
Nearly every State experienced significant overall increases in every index (for which data were available). The only exception involved perceived risk of receiving a ticket, where 4 of 6 States reported a significant increase. Awareness of special efforts by police to ticket and of specific enforcement efforts (e.g., enforcement zones or road checks) increased more than perceptions of more tickets being issued or increased risk of receiving a ticket. 17
Overall, statewide changes were similar to those in targeted rural areas. Three waves of statewide and rural surveys in Michigan provided indices of change after each phase in this State, which did not intensify enforcement during the RDP.18 Figure 7 shows similar rural and statewide trends for two enforcement-related indices. Each increased more during CIOT than during the RDP.
Indiana conducted three waves of targeted and nontargeted rural telephone surveys. Figure 9 shows the results of these surveys for the three general seat belt indices. It shows RDP-related increases in targeted areas and CIOT-related increases in both areas.
Surveys in Illinois, Minnesota, and Wisconsin found television to be the dominant message source, followed by radio and newspapers. The prevalence of television increased during the RDP; then did not change much during CIOT. The prevalence of radio, on the other hand, did not change much during the RDP, but increased during CIOT. Rural trends are shown in Figure 11. Similar trends were found statewide.
Advertisements (ads) were part of paid and public service media efforts. News stories were generated by earned media efforts. Figure 12 shows that ads were reported as the source of information about four times as often as news stories.19 Similar trends were found statewide.
1. Statewide and Rural Area Changes
Seat belt use increased significantly in all GLR States. As Table 10 shows, there was a median overall increase of 4.8 percentage points (range: 3.2 to 7.7).23 Significant RDP-related increases (w2-w1) were found in Illinois, Minnesota, and Ohio, and significant CIOT-related increases (w3-w2) were found in all States except Ohio. Figure 13 shows trends for each State, differentiating between primary-law States (solid lines) and secondary-law States (dotted lines). The median increase (w3-w1) was about 5 points for both law types.
Another way to express these changes is in terms of percentage of nonusers converted to users. Figure 14 shows that the greatest change was in Michigan (33%); followed by Illinois (29%), Wisconsin (22%), Indiana and Minnesota (both 21%), and Ohio (13%). The median conversion rate was 29 percent in primary law States and 21 percent in secondary law States, due in part to smaller proportions of nonusers in primary States.
Table 11 shows that, following the RDP, the States with significant increases in usage were Illinois (3 points), Indiana (2.5 points), and Ohio (8 points), each of which intensified enforcement during this period. Following CIOT, five States reported significant increases in usage. They were Illinois (4 points), Indiana (7 points), Michigan (2 points), Ohio (4 points), and Wisconsin (8 points). Included among these States were the three RDP-enforcement States and the three States that employed enforcement zones and/or roadside checkpoints (Illinois, Indiana, and Michigan).
Figure 15 shows a median 7-point increase in usage in rural targeted areas (w3-w1), compared with a median 5-point statewide increase, possible evidence of additional impact in the rural areas. Perhaps more importantly, there were clear differences between States that intensified enforcement during the RDP and States that did not. As Figure 15 shows, there was a 9-point median increase in the targeted areas of the three RDP-enforcement States (w3-w1), compared with a 3-point increase in the nonenforcement States. Thus, the rural estimates of change in the three enforcement States were considerably greater than the statewide estimates, an even stronger indication that two waves of enforcement (RDP and CIOT) were associated with a greater impact on usage than one wave (CIOT only).
In Illinois and Indiana, as in Ohio, increases in rural targeted areas were greater than statewide increases (Illinois: 7 points rural versus 4.8 points statewide; Indiana: 9 points rural versus 4.9 points statewide). In addition, as Figure 17 shows, increases in rural targeted areas of Indiana were greater than increases in nontargeted areas (9 points versus 2.4 points). As would be expected, statewide usage in both States increased during CIOT, making rural and statewide trends more similar than in Ohio.25 Sufficient raw data were not available from these two States to conduct regression analyses.
In Michigan, a comparison of rural and statewide trends was possible only for the CIOT phase. Figure 18 shows that the statewide increase that was greater than the rural increase during that phase and regression analysis of the Michigan data found this difference in rate of increase to be significant (Wald = 8.45; df = 1; p = .004). Thus, the Michigan data support the expectation that there would be significant statewide increases during CIOT. They also suggest that the impact of CIOT was greater statewide than in rural areas.
In Minnesota, estimated changes in neither of two targeted area surveys reached statistical significance (during either phase). However, a 4.5 point statewide increase was significant (Χ2 = 55.3; df = 1; p < .001). Figure 20 shows the trends for the statewide survey and for both targeted-area surveys. Regression analyses, using data from the statewide survey, the southeast target area sample, and the nontarget area sample, found a significant aggregate increase for the three groups (Wald = 8.1; df = 1; p = 0.088) but there was no significant difference in the rates of increase for the three groups. As in Indiana, baseline usage for one group (the nontargeted sample) was significantly lower than baselines for the other groups, making it less useful as a control condition.
Figure 19 shows that, in Wisconsin, rural and statewide trends were similar, with possible declines during the RDP and significant increases during CIOT, statewide and in rural areas. Regression analyses, using statewide, targeted and nontargeted area data, found a significant upward shift in aggregate usage during CIOT (Wald = 7.5; df = 1; p = .006), but no evidence of different rates of increase.26
Table 12 provides median usage rates and changes for various sub-groups included in observational surveys. The last column indicates how many States contributed data regarding each sub-group. In general, this table shows that there were substantial differences in the usage rates of males versus females, younger versus older occupants, and occupants in pickups versus other vehicles. With regard to change, the largest changes occurred during CIOT. Changes within the various categories (age, sex, etc.) are similar. However, this table masks the considerable differences between when enforcement was present and when it was not. Those differences are shown in Table 13.
Following are additional highlights regarding rates and changes among the various sub-groups. Again, all data are from observation surveys conducted in rural targeted areas.
Role in Vehicle. About 80 percent of observations involved drivers. In nearly every State, there was little difference between driver and passenger use or change in use associated with the mobilization. Indiana provided the one exception in that driver use was substantially lower than that of passengers, before and after the mobilization.
Sex. About 55 percent of occupants surveyed were males and had substantially lower usage rates than females. Changes associated with the mobilization were similar for both groups. Here again, differences were greatest in Indiana, where usage among males was 19 points lower than among females. Usage increased significantly for both groups but males appeared to be most affected during CIOT. The smallest sex differences were in Michigan. Baseline usage among males was 8 points lower than among females and this difference declined to 2 points after CIOT. In Ohio, the sex gap declined from 13 points to 4 points, with much of the increase among males occurring during the RDP. In Wisconsin, where male usage was 12 points lower than female usage, CIOT-related increases were similar among both groups.
Age. On average, 22 percent of those observed were categorized as young adults (ages 16-29); 56 percent as adults (ages 30-64); and 22 percent as seniors (age 65+). Usage was consistently lower among the youngest occupants. In Ohio, usage among this group increased by 15 points and the gap between younger and older groups decreased by 4 points. These increases occurred during both phases of the mobilization.
Race. Only three States provided data by race and, in most cases, the samples were too small to make meaningful comparisons between blacks and whites. In Michigan, however, there were sufficient data to examine changes among blacks and there was evidence of a significant CIOT-related increase in usage among that group.
Vehicle Type. Usage by vehicle type was reported by all six States. Nearly half of those observed (46%) were in passenger cars; 25 percent were in pickup trucks; 16 percent were in SUVs; and 13 percent were in vans. At baseline, the highest rates were found among occupants of vans (81%), followed by cars and SUVs (76% and 74%, respectively) and pickup trucks (56%). The lowest rate among pickup occupants was found in Indiana (33% at baseline).28 The next lowest rates were in Minnesota, Ohio, and Wisconsin (about 55%).29 Significant RDP-related increases in pickups were found in Illinois (3 points) and Ohio (7 points).30 CIOT-related increases were found in Illinois, Indiana, Michigan, Ohio, and Wisconsin.
Table 14 shows States, RDP versus CIOT phases, and the occurrence of significant increases in usage for three target groups (occupants of pickup trucks, males, and young occupants). These target groups were examined because they consistently represent lower-use groups. All shaded cells represent conditions in which enforcement was present. Nonshaded areas represent conditions under which no enforcement was present. There were no instances where significant increases resulted when enforcement was absent. By comparison, there were significant increases in 70 to 80 percent of the situations in which enforcement was present. Two-by-two chi-square analyses found these differences in proportions to be significant in each case.
12 The estimates provided in this table were derived from RDP and CIOT reports submitted by the states. Some of the variation among the states likely results from differences in reporting criteria (e.g., with regard to enforcement hours). The (%) under “EZs” refers to the percent of hours dedicated to enforcement zones. The term “cites” refers to citations for seat belt violations (1st row) and child restraint violations (2nd row).
18 Michigan was not an RDP-enforcement State, but it did implement paid media during the RDP (ranking 4th in expenditures per capita and 2nd in ads per capita). During CIOT, it was a strong enforcement state, with 781 enforcement zones, and with a CIOT media campaign ranking 2nd in expenditures per capita.
20 All States conducted surveys in rural targeted areas. In Indiana and Minnesota surveys were conducted in rural nontargeted areas by the regional evaluation contractor; limited data regarding nontargeted rural areas was also extracted from surveys conducted in Wisconsin but this was a small sample and it involved a baseline that was significantly higher than that in the targeted rural areas.
24 Minnesota results are from the 28-site sub-sample of a statewide mini-survey. Results from this survey are shown because they covered a broader geographical area than the mini-survey conducted in southeast Minnesota. Also, while the estimated increase in Minnesota (+2.9) was larger than that in Michigan (+2.3), it did not reach statistical significance due to the much smaller number of observations in Minnesota.
26 Here again, the baseline rate of the control group was significantly different (in this case higher) than that of the other groups, making it less appropriate as a control condition. In addition, the Wisconsin control was small, consisting of only eight rural sites in a nontargeted media market.