|
Impact of the graduated driver licensing program in Nova
Scotia
Report from the Traffic Injury Research Foundation
Method
The primary objective of the evaluation was to determine if any significant
changes occurred in collisions following the introduction of the graduated
licensing program. A secondary purpose was to gauge, to the extent possible,
the effectiveness of the Nova Scotia program, relative to others that have
been evaluated. The fact that other evaluations of graduated licensing programs
have used analytic procedures similar to the methods used in the current
investigation provides the basis for comparisons across studies.
To control for the influence of potentially confounding variables and to
provide the basis for comparisons with other previous studies, a series
of increasingly refined analyses were applied to the data.
Pre-post comparisons were used to assess changes in collisions among drivers
age 16 and 17 (the largest group covered by the program). Essentially, the
prevalence of collisions among 16 and 17-year-old drivers prior to the introduction
of the program was compared to the collision experience of these groups
after the program was implemented. Appropriate internal and external control
groups not exposed to the new program were used to account for the effects
of other factors that could be influencing the prevalence of collisions.
In addition, as a control for changes in population, the comparison data
for the target and control groups were based on crash rates – the number
of collisions per 10,000 population. This analytic approach was used recently
in the evaluation of the graduated licensing program in Florida (Ulmer et
al. 1999).
The next level of analysis used time series to rule out the possibility
that the observed short-term changes could be accounted for by longer-term
trends and not the program itself. Monthly collision data for the primary
target group (16-year-old drivers) were examined for an extended period
before the program was implemented (beginning in January 1, 1986), through
to December 1997, over three years following its implementation. To control
for the potential effects of other concurrent factors or events, various
internal and external comparisons were also made. A comparable approach
was used to evaluate early provisional programs implemented in Maryland
and California (McKnight et al. 1983; Hagge and Marsh 1988) as well as the
graduated licensing program implemented more recently in New Zealand (Langley
et al. 1996).
Taken together, the pre-post comparisons and time series analyses provide
a substantive evaluation of the impact of the Nova Scotia graduated licensing
program. However, although these analyses control for fluctuations in population,
they do not necessarily control for changes in licensing in the target group,
which could account for changes in the frequency of collisions. Moreover,
the primary analyses were, for various reasons, restricted to 16- and 17-year-old
drivers, so they reveal nothing about the impact of the program on older
novices.
To examine the effects of the program, independent of licensing changes,
as well as to examine its effects on all beginners, two groups of novice
drivers were selected. One group consisted of persons licensed for the first
time before October 1994 and the second consisted of persons licensed for
the first time after October 1994. This made it possible to compare the
collision rates of novice drivers of all ages under the current graduated
licensing program to those of novice drivers of all ages under the previous
licensing program. Any observed changes from the pre- to the post-program
periods could not be accounted for by changes in the number of licensed
drivers in the target group. A similar method was used in the evaluations
of the Oregon provisional license program and the Ontario graduated licensing
program (Jones 1994; Boase and Tasca 1998).
Dependent Variables
Control for Other Factors
Data Sources
|
|
