ITE Journal April 2018 - 48
angles from crossroads to entrance ramps were verified to be
right angles. Based on methods from Wang et al., measurements
were conducted with measuring tools of Google Earth for each
study location and verified in field, including the turning radius
for left-turn drivers on crossroad, median type on crossroad, and
the median width between exit ramps and entrance ramps.12 In
addition, two lengths were measured, i.e. L1= the length from the
stop line of the left-turn lane(s) to the centerline of the median on
two-way ramps and L=the length of the entire intersection (the
distance between the stop lines in two opposite directions on the
crossroad). The stop line positioning through the intersection was
calculated by the following equation:12
S (%)= -1 * 100
Table 1. Analysis Results of Geometric Elements vs. WWD Crashes and ORs
WWD Crash Data
For WWD crashes in the selected sites, one crash database for
Illinois for a six-year period (2004-2009) was obtained from the
Highway Safety Information System (HSIS) of the Federal Highway
Administration (FHWA) that included crash coordinates and
wrong-way movement descriptions. Firstly, the route number,
roadway description, and roadway functional class of crashes
were used to verify the crashes were at interchange vicinities, with
street views and aerial photography Google Earth used as supplementary tools. Then, crash narratives were reviewed to confirm
that each crash resulted from WWD and to determine whether
the wrong-way entries occurred at interchange terminals. Thus,
wrong-way entry points for each crash were identified or estimated.
Based on the review of all hardcopy crash reports, 217 WWD
crashes were identified in total in the six-year period in Illinois,
fifteen of which occurred at signalized ramp terminals of parclo
interchanges, including confirmed and possible WWD entries.
The effects of the existing applications of geometric elements were
analyzed for the 44 study sites. The frequency of WWD crashes
experienced at study sites was calculated for each technique. The
most common condition for WWD entries for each technique is
stop line positionings on crossroads more than 60 percent, turning
radii larger than 100 feet (ft), the traversable median, and medians
10 ft or less (See Table 1 column 3).
Furthermore, the Peto odds ratio (OR) assessed the contribution of several geometric elements and WWD using a 95
percent confidence interval as the relative measure at statistical
significance level of 0.05. This calculation is based on the null
hypothesis that treatment has no effect on outcome. Under
the null hypothesis, the difference between the observed and
the expected would have zero difference and variance. The
advantage of the OR approach over direct calculations is that
Ap r i l 2018
i te j o urn al
it allows for zero results without generating infinity, which is
suitable for some categories of variables in the dataset that have
no events. The OR value reflects the impact of a specific category
with larger numbers reflecting greater contribution. The results
were presented in Table 1. It reveals that WWD occurrences
increase when the stop line positioning on the crossroad is
more than 60 percent, the turning radius is larger than 80 ft, the
median is traversable, and the medians width is within 20-30 ft
between ramps. At statistical significance level of 0.05 (depicted
as red text in the table), significant differences exist between
different categories of each variable.
More than 60%
50 ft. and less
51 to 60 ft.
61 to 70 ft.
71 to 80 ft.
81 to 90 ft.
91 to 100 ft.
More than 100 ft. 60.00
Type of Median Non-traversable
on Crossroads Traversable
10 ft. and less
10 to 20 ft.
21 to 30 ft.
31 to 40 ft.
41 to 50 ft.
51 to 60 ft.
More than 60 ft.
Note: there is no interchange terminal with median widths on ramps more than 60 ft., hence "NA".
Conclusions and Recommendations
This paper summarizes the existing applications of geometric
elements that have been proposed to deter wrong-way movement
at parclo interchange terminals. It analyzed the collected WWD
crash and geometric design data to quantify the relationships
between geometric elements and WWD at parclo interchange