ITE Journal - June 2021 - 35

L

owering crash severity is a critical priority for transportation professionals and
all those in the transportation community who are actively adopting Vision Zero.
With the goal of reaching zero fatality and serious injury crashes, the need to better
understand the factors related to severe cases becomes essential. Crash severity is

highly shaped by the crash type, or manner of collision.1-3 For example, the odds of being fatally
injured in a crash greatly increases for head-on and rollover crashes.1 Severity has been found to
be highest for head-on collisions, followed by angle crashes, and lower for rear end and sideswipe
crashes.4 Given this relationship between crash type and severity, it is critical that transportation
professionals understand the likelihood of specific crash types to occur at various traffic control
devices. A large-scale study in the United States revealing this relationship has yet to exist in
literature. This information will guide future decision-making, guiding transportation practitioners to implement the most critical countermeasures to achieve the highest level of safety.
This paper aims to investigate the probability of various crash types
at different traffic control devices. The results allow transportation
professionals to create informed decisions in their crash countermeasure selection process at various traffic control devices.

Procedure and Data Analysis
Data used in this study were obtained from the Federal Highway
Administration (FHWA) Highway Safety Information System
(HSIS) database. HSIS is a multistate database that contains
police-reported crash, roadway, and volume data. For this study,
data from the states of North Carolina and Ohio from 2010 to 2015
were obtained.
These two states were chosen as they both had current data from
the chosen study variables, unlike other states, who were missing
key contributing variables. Ohio and North Carolina have similar
percentages of rural versus urban roadway miles, rural vehiclemiles travelled, and deaths per vehicle-miles travelled compared
5, 6
to the greater United States area. Given this, North Carolina and
Ohio were deemed as reasonable states to use in reference to the rest
of the United States.
For this study, speed limit, traffic control device, crash type,
and weather condition data were all obtained from HSIS for the
two states. With this data, simplistic models could be created to
determine the presence of relationships. These simpler models are

not meant to detract from other variables, such as vehicle type,
sex, and age. This study rather presents a larger-scale, particular
perspective on crash types and traffic control with the recognition
that on a smaller-scale, with complementing studies, and with
analysis on a location-specific basis, these additional factors would
be considered.
All data rows (or crash occurrences) with missing values and
those with a speed limit of 0 miles per hour (mph) were omitted
from the dataset. To establish consistency between Ohio and North
Carolina coded data, each variable was recoded to match one
another. In this process, some traffic control types or crash types
were recoded into " Other " categories, as they were unmatchable
to the other state's data or not the focus of this study. After this
recoding, more than 1.8 million crashes remained for model
development. A graphical representation of this data is presented in
Figure 1. Table 1 presents a summary of crash types. It is noted that
Ohio law has a 20 mph [32.2 kilometers per hour (km/hr)] speed
limit in school zones during certain hours. In North Carolina,
speed limits in school zones are designated at the local level and
cannot be lower than 20 mph. School zones in each state are defined
differently, but are generally road sections that border school
grounds or pass a school fronting.
Analysis was completed using multinomial logit regression
(MLR) with relative risk ratios (RRRs). MLR with RRRs
w w w .i t e.or g

J u ne 2021

35


http://www.ite.org

ITE Journal - June 2021

Table of Contents for the Digital Edition of ITE Journal - June 2021

ITE Journal - June 2021 - 1
ITE Journal - June 2021 - 2
ITE Journal - June 2021 - 3
ITE Journal - June 2021 - 4
ITE Journal - June 2021 - 5
ITE Journal - June 2021 - 6
ITE Journal - June 2021 - 7
ITE Journal - June 2021 - 8
ITE Journal - June 2021 - 9
ITE Journal - June 2021 - 10
ITE Journal - June 2021 - 11
ITE Journal - June 2021 - 12
ITE Journal - June 2021 - 13
ITE Journal - June 2021 - 14
ITE Journal - June 2021 - 15
ITE Journal - June 2021 - 16
ITE Journal - June 2021 - 17
ITE Journal - June 2021 - 18
ITE Journal - June 2021 - 19
ITE Journal - June 2021 - 20
ITE Journal - June 2021 - 21
ITE Journal - June 2021 - 22
ITE Journal - June 2021 - 23
ITE Journal - June 2021 - 24
ITE Journal - June 2021 - 25
ITE Journal - June 2021 - 26
ITE Journal - June 2021 - 27
ITE Journal - June 2021 - 28
ITE Journal - June 2021 - 29
ITE Journal - June 2021 - 30
ITE Journal - June 2021 - 31
ITE Journal - June 2021 - 32
ITE Journal - June 2021 - 33
ITE Journal - June 2021 - 34
ITE Journal - June 2021 - 35
ITE Journal - June 2021 - 36
ITE Journal - June 2021 - 37
ITE Journal - June 2021 - 38
ITE Journal - June 2021 - 39
ITE Journal - June 2021 - 40
ITE Journal - June 2021 - 41
ITE Journal - June 2021 - 42
ITE Journal - June 2021 - 43
ITE Journal - June 2021 - 44
ITE Journal - June 2021 - 45
ITE Journal - June 2021 - 46
ITE Journal - June 2021 - 47
ITE Journal - June 2021 - 48
ITE Journal - June 2021 - 49
ITE Journal - June 2021 - 50
ITE Journal - June 2021 - 51
ITE Journal - June 2021 - 52
https://www.nxtbookmedia.com