UVM Theses and Dissertations
Format:
Print
Author:
Harootunian, Kristine
Dept./Program:
Civil and Environmental Engineering
Year:
2013
Degree:
MS
Abstract:
Drivers outside their country of residence are at a safety disadvantage when compared to native drivers. This research aimed to identify two things. The fIrst was whether out-of-state drivers in the United States were comparable to foreign drivers in the context oftraffic safety. This was accomplished using 2007-2011 crash data from Florida, Maine, Minnesota, and Nevada to model fault using logistic regression where factors between in-state and out-of-state drivers were compared. Next, the effect of distance from home with crash location was examined using geo-coded crash locations and home address for 30,675 crashes between 2007 and 2011 in the state of Vermont.
In the first analysis, regressions showed that out-of-state drivers had increased odds of fault, ranging from 18% to 93%, for a single-vehicle crash compared to in-state drivers in Maine, Minnesota, and Nevada. In Florida, there was no statistical difference between groups. Odds were higher for out-of-state drivers in two-vehicle in Maine (3%) and Nevada (19%). In Florida and Minnesota, there was no statistical difference between groups. Next, factors such as age, sex, driving conditions, and seasons were tested with multivariate variate regressions for in- and out-of-state groups separately and their odds ratios were compared between groups. For single-vehicle crashes, age, sex, road grade, surface condition, light conditions, and day of week were factors that increased at least one of the two groups' odds of fault in all states. Sex, surface condition, and light conditions increased the odds of fault for at least one of the groups in two vehicle crashes in all four states. Although there were several factors in each state that increased odds of fault for out-of-state drivers more so than in-state drivers, no factor consistently increased odds of fault this group across all four states.
The Vermont study showed that longer distances from home increased odds of fault for drivers in single-, and two-vehicle crashes. Limiting the database to only in-state drivers, distance as a main effect was not statistically significant. When incorporated as an interaction term, it had a significant impact for some variables. For single-vehicle crashes, increased distance from home paired with driving on grades, wet or icy surfaces, and at intersections increased a driver's odds of fault compared to when distance was held constant. This finding was true for the set of all drivers in single-vehicle crashes, as well as the subgroup of Vermont drivers. Interestingly, increased distance from home alleviated the effects of certain hazardous factors.
For all drivers, being a non-owner, driving on the weekend, or during the summer increased one's odds of fault (holding distance constant). As distance increased, however, the effects of these conditions for predicting fault decreased. This was also true for driving in differing weather conditions for two-vehicle crashes. Holding distance constant, drivers had higher odds of fault when driving on clear days compared to driving in inclement weather. Letting distance vary, a driver who was 100 miles from home and driving when there was freezing precipitation had the same odds of fault as a driver who was 10 miles from home and driving in good weather. The results add to a growing body of evidence that drivers away from home have increased risk and that in-depth consideration of cause and countermeasures is warranted.
In the first analysis, regressions showed that out-of-state drivers had increased odds of fault, ranging from 18% to 93%, for a single-vehicle crash compared to in-state drivers in Maine, Minnesota, and Nevada. In Florida, there was no statistical difference between groups. Odds were higher for out-of-state drivers in two-vehicle in Maine (3%) and Nevada (19%). In Florida and Minnesota, there was no statistical difference between groups. Next, factors such as age, sex, driving conditions, and seasons were tested with multivariate variate regressions for in- and out-of-state groups separately and their odds ratios were compared between groups. For single-vehicle crashes, age, sex, road grade, surface condition, light conditions, and day of week were factors that increased at least one of the two groups' odds of fault in all states. Sex, surface condition, and light conditions increased the odds of fault for at least one of the groups in two vehicle crashes in all four states. Although there were several factors in each state that increased odds of fault for out-of-state drivers more so than in-state drivers, no factor consistently increased odds of fault this group across all four states.
The Vermont study showed that longer distances from home increased odds of fault for drivers in single-, and two-vehicle crashes. Limiting the database to only in-state drivers, distance as a main effect was not statistically significant. When incorporated as an interaction term, it had a significant impact for some variables. For single-vehicle crashes, increased distance from home paired with driving on grades, wet or icy surfaces, and at intersections increased a driver's odds of fault compared to when distance was held constant. This finding was true for the set of all drivers in single-vehicle crashes, as well as the subgroup of Vermont drivers. Interestingly, increased distance from home alleviated the effects of certain hazardous factors.
For all drivers, being a non-owner, driving on the weekend, or during the summer increased one's odds of fault (holding distance constant). As distance increased, however, the effects of these conditions for predicting fault decreased. This was also true for driving in differing weather conditions for two-vehicle crashes. Holding distance constant, drivers had higher odds of fault when driving on clear days compared to driving in inclement weather. Letting distance vary, a driver who was 100 miles from home and driving when there was freezing precipitation had the same odds of fault as a driver who was 10 miles from home and driving in good weather. The results add to a growing body of evidence that drivers away from home have increased risk and that in-depth consideration of cause and countermeasures is warranted.