|
|
 |
Impact of the graduated driver licensing program in Nova
Scotia
Report from the Traffic Injury Research Foundation
Discussion
The primary purpose of the present study was to determine the safety impact of the graduated
licensing program in Nova Scotia. A series of increasingly refined analyses, which controlled for
the influence of other explanatory variables, all showed that the implementation of graduated
licensing in Nova Scotia was associated with a significant reduction in collisions.
Table 8 summaries the results for drivers age 16. There was a substantial decrease in their
collisions and casualty collisions following the introduction of the graduated licensing program.
This was evident in the initial year of the program and across its first three years of operation.
Pre-post comparisons showed that the total collisions in 1995 – the first full year of the graduated
licensing program was in effect – were 24% lower than they were in 1993. Collisions in 1996 were
36% lower than in 1993. Time series analyses showed that collisions decreased by 37% during
the first three years of the program. Comparable decreases occurred in casualty crash ratios.
Improvements were also observed for all novice drivers not just those who are young. The collision
rate for all novice drivers dropped by 19.4%, from a rate of 1,418.9 per 10,000 learners in 1993 to
1,143.0 in 1995.
The weight of this converging evidence clearly demonstrates that the graduated licensing program
has proven effective.
A secondary purpose of this study was to gauge, to the extent possible, the effectiveness of the
Nova Scotia program, relative to other graduated licensing programs that have been evaluated. In
this regard, several of the analyses used in this study were similar to those used in previous
evaluations. This makes it possible to compare findings across studies.
To date, evaluations have been completed in the following six jurisdictions:
- Maryland McKnight et al. 1983
- California Hagge and Marsh 1988
- Oregon Jones 1994
- New Zealand Langley et al. 1996
- Ontario Tasca and Boase 1998
- Florida Ulmer et al. 1999
Strictly speaking, three of these programs – Maryland, California, and Oregon – were not
integrated graduated licensing systems but did contain some of the elements of a graduated
approach. And all three of these early initiatives have been shown to reduce crashes. The
magnitude of the reductions, however, have been modest compared to the findings for Nova Scotia
in the current study. For example, McKnight et al. (1993) evaluated the Maryland program using
time series and found no reduction in nighttime crashes but a 5% drop in daytime crashes for both
16- and 17-year-olds. Comparable reductions were found in California by Hagge and Marsh (1988)
using time series – a 5.3% reduction in the rate of total accidents among 15- to 17- year olds. The
37% reduction in collisions among drivers age 16 observed in Nova Scotia using time series was
considerably greater than the decreases reported for California and Maryland.
The evaluation of the program in Oregon had mixed results. Jones (1991) using pre-post
comparisons found a 16% reduction in crashes among male drivers age 16 to 17 in their first year
of driving but no significant change for females. By comparison, the current study of the Nova
Scotia program found larger reductions in collisions among all 16 year olds during the first full year
– a 24% decrease in collisions from 1993 to 1995.
The other jurisdictions – New Zealand, Ontario, Florida – introduced some version of a graduated
licensing system. The New Zealand program was the first truly integrated system and it was
evaluated by Langley et al. (1996) using time series. These authors reported that the introduction
of the graduated licensing program in 1987 resulted in a 7% decrease in casualty crashes among
15-19 year olds. By comparison, the time series analysis in the current study showed a 31%
decline in casualty collisions among 16-year-old drivers over the three years following the
implementation of the graduated licensing program in Nova Scotia.
Thus, Nova Scotia experienced decreases in the collisions of 16-year-old drivers that far exceeds
the reductions reported in previous evaluations of early programs in Maryland, California, Oregon
and New Zealand. However, these programs were introduced many years before the Nova Scotia
program, making a direct comparison of results somewhat tenuous.
A more appropriate comparison is with Ontario because it also implemented its program in 1994.
However, given that the features of the two programs are very different, it is reasonable to expect
that the magnitude of collision reductions might differ. For example, the Ontario program includes
a night curfew and a highway restriction in the first stage of the program, whereas the night curfew
in Nova Scotia applies to the second stage and there is no highway restriction. The learner stage
in Ontario is 12 months, compared to only six months in Nova Scotia; the second intermediate
stage in Ontario is 12 months, compared to 24 months in Nova Scotia.
A comparison of results from the Boase and Tasca (1998) evaluation of the Ontario program and
those reported previously for the current study of the Nova Scotia program initially suggest that
Ontario experienced greater reductions in collisions. For example, Boase and Tasca (1998) found
that the overall collision rate for novice drivers of all ages decreased by 31% after Ontario’s
graduated licensing program was implemented. By comparison, the present evaluation of the Nova
Scotia program found only a 19% reduction in the overall collision rate of novice drivers of all ages.
However, the analysis used by Boase and Tasca produces a liberal estimate of the impact of the
Ontario program. This is so because the 1995 post-program group in the Boase and Tasca study
included drivers issued a learner’s permit ("G1" in Ontario) but the 1993 pre-program group did not.
Given that learner’s are under supervision at all times and, therefore, have a relatively low crash
rate, excluding them from the pre-program group would artificially increase crash rates in the
period before the program came into effect, thereby, overestimating the subsequent decrease.
The present study overcame this problem by focussing on two equivalent groups of novices – i.e.,
novices obtaining learner licenses in 1993 before program implementation and those who became
learners in 1995 after program implementation.
Indeed, if the approach used by Boase and Tasca is applied to the Nova Scotia data the decrease
in collisions in Nova Scotia was actually greater than the one found in Ontario – a 38% reduction in
the overall collision rate among novice drivers in Nova Scotia, compared to only 31% in Ontario.
Finally, Florida implemented a graduated licensing program even more recently than the ones in
Nova Scotia and Ontario and it has been evaluated by Ulmer et al. (1999) using pre-post
comparisons with a control group – older drivers unaffected by the license change. The authors
found a significant 9% reduction in the casualty involvement ratio for 15- to 17-year-old drivers in
Florida. The greatest reduction (19%) occurred among 15 year olds, followed by 16 year olds
(11%) and 17 year olds (7%).
Applying similar procedures to a comparable time frame, the current analyses showed that the
casualty collision ratio among 16 year olds decreased by 34% in Nova Scotia during the first full
year of graduated licensing. This is substantially greater than the decrease recorded in Florida.
Although there was no significant decrease in collisions among 17-year-old drivers in Nova Scotia
in the year immediately following the implementation of the program, there was a 17% decrease in
the subsequent year when all 17 year olds would have been licensed under the new program. This
change is also more pronounced that the one reported in Florida.
In summary, the results of the current evaluation provide further compelling evidence of the
effectiveness of graduated licensing. Moreover, the results suggest that the Nova Scotia program
has been more effective than others that have been evaluated.
Two key issues remain for further research. The first is why the program works.
In particular, there is a need to know which specific features of the program
– i.e., the extended period of adult supervision, the night curfew, low BAC
requirement, driver education/training – contribute most to the observed safety
gains. A second issue concerns the persistence of the effects – whether the
safety benefits extend beyond the time when the novice driver is covered by
the graduated licensing program. Both these issues are currently under investigation
by the authors. |
|
