4 – RESULTS
Our evaluation of Pueblo County’s Smart Roads program used alcohol-related crashes involving a driver of the of target age group (21- to 34-year-olds) as the measure of effectiveness. The relatively small size of Pueblo County (approximately 141,000) dictated the use of a before-and-after, test-and-comparison group evaluation design with surrogate measures of alcohol-related crashes. A larger population base would have allowed the use of the more desirable interrupted time series design with measured values of BACs in fatal crashes as a basis of the measures of effectiveness.
The test group of primary interest was Pueblo County, but because of the nature of the Pueblo County intervention (which had a strong public information component which could have spread to adjoining counties), we also considered a test group composed of Pueblo County plus eight other low-population surrounding counties. These other counties were Teller, Fremont, Custer, Huerfano, Las Animas, Otero, Crowley, and Lincoln counties. The comparison groups were composed of crash-involved, 21- to 34-year-old drivers in the non-test counties. For the test group containing only Pueblo, the comparison group was all Colorado counties except Pueblo. And for the test group containing Pueblo plus the above eight surrounding counties, the comparison group was all Colorado counties except Pueblo plus the above eight surrounding counties. The “before” period was 1998 and 1999, and the “after” period was 2000 and 2001.
Crash data was provided by the Colorado Department of Transportation, which maintains computerized records of police crash reports in the State. We used two types of crashes as surrogate measures of alcohol-related crashes: nighttime injury crashes (NI) and nighttime single-vehicle injury crashes (NSVI). The NSVI measure was applied only to Pueblo plus the eight adjoining counties because of the small number of NSVI crashes in Pueblo County alone. The results of evaluation are presented below in two sections, the first addresses Pueblo County alone as the test group, and the second addresses Pueblo County plus the eight adjoining counties as the test group.
Nighttime Injury Crashes
NI crashes in Pueblo County decreased from 38 in the before period to 23 in the after period, a 40 percent decrease. By comparison, NI crashes in the county’s comparison group increased 2 percent from 7,965 to 8,147. These changes were significant at a level of 0.045 (χ2=4.02, d.f.=1). Similar changes occurred for NI crashes as percentage of all crashes (Figure 4-1). For the test group, the percentage was 10.9 percent in the before period and 6.2 percent in the after period, a decrease of 43 percent (p=0.041). And for the comparison group, the percentage was 10.5 percent in the before period and 10.9 per-cent in the after period, an increase of 4 percent (p=0.0382).
Nighttime Single-Vehicle Injury Crashes
As noted above, these crashes were not analyzed because of too few such crashes in Pueblo County alone.
Figure 4-1: Nighttime Injury Crashes as a Percentage of All Crashes Involving Drivers Age 21 to 34, Test Counties (Pueblo) and Comparison Counties
[d]
PUEBLO COUNTY PLUS EIGHT ADJACENT COUNTIES
Nighttime Injury Crashes
In this test group, NI crashes decreased from 251 in the before period to 161 in the after period, a decrease of 39 percent (Figure 4-2). In this test group’s comparison counties, NI crashes increased 3.3 percent from 7,752 to 8,009. These changes were significant at the p<0.0001 level (χ2=22.13, d.f.=1). For NI crashes as a percentage of all age-21-to-34 crashes, the test group showed a decrease of 13.2 percent in the before period to 9.7 percent in the after period (p<0.0.0001), while the comparison group still had an in-crease of 10.5 percent to 10.9 percent (p=0.0042).
Nighttime Single-Vehicle Injury Crashes
In the test group, NSVI crashes decreased from 153 in the before period to 115 in the test group, a decrease of 24.8 percent (Figure 4-3). Again, there was an increase for the comparison group, this time from 2,401 to 2,496 (4%). These changes were statistically significant with p=0.01.
And while NSVI crashes as percentage of all crashes decreased again for the test group (Figure 4-3), the decrease was smaller than that for NI crashes, amounting to only 13.5 percent (from 8% to 6.9%). At the same time, the increase for the comparison group was slightly higher for NI crashes at 4.8 percent. Neither change was statistically significant.
Figure 4-2: Nighttime Injury Crashes Involving Drivers Age 21 to 34 as a Percentage of All Crashes, Test Counties (Pueblo Plus Eight Adjacent Counties) and Comparison Counties
[d]
Figure 4-3: Nighttime Single-Vehicle Injury Crashes Involving Drivers Age 21 to 34 as a Percentage of All Crashes, Test County (Pueblo Plus Eight Adjacent Counties) and
Comparison Counties
[d]