Printer Friendly

An investigation of passing accidents on two-lane, two-way roads.


The passing maneuver on two-lane highways is one of the most demanding and hazardous operations performed by motorists. Its danger lies in the fact that a passing vehicle must occupy an opposing lane of traffic to complete the maneuver, and it is not uncommon for drivers at some point in time to want to pass a slower moving vehicle on a two-lane highway. Two-lane highways comprise the majority of the roadway mileage in the United States. On the Federal-aid Highway System alone, nearly 960,000 km (600,000 mi) are classified as two-lane rural highways, and an additional 193,000 km (120,000 mi) are classified as two-lane urban highways. (1) [1]

A recent study undertaken by the Federal Highway Administration (FHWA) analyzed the issues surrounding passing on two-lane rural roads. The goals of this research were to:

* Determine if there is a safety problem in passing zones on two-lane, two-way highways.

* Determine the magnitude of the problem and associated risk.

* Identify causal and/or contributory factors to the problem.


Since the 1920's, much has been written about passing problems on two-lane rural roads. (2, 3, 4) A critical issue that was identified in the past research is the difference between the American Association of State Highway and Transportation Officials (AASHTO) passing sight distance design criteria for two-lane roads found in the manual A Policy on Geometric Design of Highway and Secondary Streets and the marking criteria for no-passing zones on the same roads found in the Manual on Uniform Traffic Control Devices for Streets and Highways (MUTCD). (5, 6) These criteria are presented in tables 1 and 2, respectively.

In the field, passing zones are not specifically marked; instead, they exist where no-passing zones are not warranted. The warrants for no-passing zones are established by the MUTCD, and the MUTCD values are clearly different from the AASHTO values, which are used in the design of two-lane roads and not for traffic control. The net result was that two different sets of "minimum passing sight distances" were developed. However, the FHWA and others recognized that the values in the MUTCD are not minimum passing sight distances, but rather they are minimum sight distance requirements. Furthermore, it was recognized that these two sets of values serve two different functions. (4)

Many past studies investigated the validity and practicability of these two sets of requirements. The general consensus was that the MUTCD values were too low. The AASHTO requirements were also questioned by some researchers as too conservative; others concluded that they were inadequate for car-passing-truck and truck-passing-truck scenarios.

Other issues were overlooked or not addressed by past studies on this topic. For example, there was no complete investigation of accidents relating to passing zone pavement markings to determine if there is a passing zone safety problem, nor was the magnitude of the problem documented. Also, only a few investigations were conducted to determine and analyze the casual and/or contributory factors of passing accidents.

Assessment of previous studies identified the need to do a detailed analysis of passing accidents with respect to the above-described factors to determine if there are deficiencies in current practices for marking passing zones as specified in the MUTCD. In other words, the intent of this study was not to compare the AASHTO values to the MUTCD values but rather to assess the level of safety of the two-lane roads with respect to the vehicles passing each other as defined by the MUTCD. If deficiencies were identified, then steps would be taken to correct the problem. If it was determined that current marking practices do not create unsafe driving conditions, then other concerns, including those related to the different sets of sight distances, could be resolved.

Research Scope

The scope of this research project consisted of the study of the accident experience of two-way, two-lane highways with passing zones marked in accordance with MUTCD values. To fulfill study objectives, data were needed on passing zones that included a large number of accident, traffic, roadway, environmental, and vehicle factors. Information on passing zone pavement markings and their relationship to the accident data was also required. The only viable sources for these types of data are State files. Since the FHWA Highway Safety Information Systems (HSIS) has been developed using State data, it was selected for use in this project.

The HSIS is a roadway-based system that provides quality data on a large number of accidents, vehicles, drivers, roadway segments, and other roadside inventories for five States. Six years of data from 1985 through 1990 are currently available for each State. At the time that this research was conducted, data were available for the years 1985 through 1989. The five States were comparatively reviewed to determine the amount, nature, availability, and workability of variables and data points in order to select the State(s) to be used in this study.

Table 3 presents a summary of the accident data for the State that was eventually chosen. One of the key factors in the State selection process was the availability of data pertaining to the location of passing and no-passing zones. Since this information is not generally included on the accident report form, and because it was not feasible to collect these data in the field, it was imperative that the HSIS State chosen provide some means to stratify the passing accidents into the respective zones in which they occurred. The selected State also had videodisc photologs that could provide the visual images of the accident sites and the location of the passing zones. On the videodisc photolog, pictures of the roadway are taken every 0.016 km (0.01 mi) from the driver's perspective. Using a combination of a videodisc player and personal computer, the images can be referenced easily according to location, and they can be quickly retrieved.

Research Methodology

In 1991, the FHWA funded a research effort to examine the key issues relating to passing zone design and markings and to develop study designs to address these issues. The contractor developed a plan to address the highest priority issue: Is there a safety problem associated with the current minimum sight distance requirements found in the MUTCD? If such a problem existed, then the magnitude of it was to be determined and. to the extent possible, the causes of the safety problem were to be identified.

A two-stage approach was used in this effort. Stage one involved the preparation of the data base, the determination of the most appropriate definition of a passing accident based on the variables in the data base, determination of the magnitude of the passing accident problem, and a comparative analysis of passing and nonpassing accidents. Stage two involved the creation of the passing zone inventory and analysis files and an analysis of the passing accidents with respect to the zones in which they occurred.

For this research effort, a passing accident was defined as one that either directly or indirectly involved a passing maneuver. In most cases, it was expected that the passing vehicle was directly involved in the accident. However, in some cases, the passing vehicle could have contributed to the accident but was not directly involved in the accident.

Since the accuracy and precision of selecting the correctly defined passing accidents determined the project's ultimate results, this was the most crucial task of the project. There is no one variable within the data base that identifies an accident as involving a passing maneuver. In fact, there are 37 variables in the HSIS accident file and 17 variables in the vehicle file that could be used in defining a passing accident. From these, two accident variables and three vehicle variables were selected:

* Accident variables:

Accident diagram = sideswipe--passing.

Vehicle movement = sideswipe--passing not at an intersection and at an intersection.

* Vehicle variables:

First contributing factor = improper overtaking or passing.

Second contributing factor = improper overtaking or passing.

Action prior to accident = vehicle--overtaking or passing.

A series of univariate and cross-tabulation tables were developed from these and other key related variables. This data manipulation gave some perspective to the selection of the most appropriate combination of variables that best fit the proposed definition. It was also determined at this point that the data should be further screened to drop the following observations from the final file:

* All passing accidents involving pedestrians and bicyclists.

* All "passing accidents" occurring at intersections or driveways from which any turning movements were possible.

Next, a micro-analysis of a random sample of 200 hard copy reports of these passing accident reports was undertaken. This analysis consisted of a manual review of the narrative and descriptive sections of this sample to confirm that the accidents extracted from the data base fit the definition of a passing accident. During this assessment, special attention was given to the identification of the factors that contributed to the accidents in order to get a better insight into the kinds of factors--human, vehicular, roadway, traffic, or environmental--that needed to be examined more closely.

This review revealed that the passing accidents that were coded "passing" from the accident diagram and/or the vehicle movement variables did not fit the desired definition of a passing accident. The term "passing" was used in these cases to refer to situations where two vehicles traveling in opposite directions collided in passing. In other words, none of the vehicles involved in these accidents was engaged in a passing maneuver. In the HSIS data base, this distinction between these two types of passing accidents was not possible. As a result of this finding, the passing accidents coded by the accident diagram and vehicle movement variables were deleted from the final passing accident file.

Passing Accident Analysis

Magnitude of passing accidents. To assess the magnitude of passing accidents, aggregate statistics were compiled on the following frequencies:

* All accidents.

* Fatal and incapacitating-injury accidents.

* All passing accidents.

* Fatal and incapacitating-injury passing accidents.

Rates, ratios, and percentages were developed on the four groups of accidents listed above. From these statistics, the indicators that were used to determine the magnitude of the problem were as follows:

[Mathematical Expression Omitted]

Comparative analysis of the passing and nonpassing accidents. The second step of the passing accident analysis consisted of a comparative analysis between the passing accidents and all other accidents that occurred on two-lane U.S. and State rural roads. These roads were selected because the videodisc photolog coverage was limited to these two route systems. Even though these two route systems make up only 27.5 percent of all the State's rural roads, they account for approximately 69 percent of the passing accidents.

The comparison was conducted on the passing and nonpassing groups of accidents with respect to a list of variables that were selected based on findings from the literature review, the results of the hard-copy analysis, the experience of the highway officials consulted in this project, and the availability of the variables in the data base. The comparative analysis was done on the two data sets, taking one variable at a time, as listed in table 4.

Passing Zone Accident Analysis

Passing zone inventory file preparation. The second stage of the research involved the analysis of the passing accidents with respect to the zones in which they occurred. First, the accidents had to be stratified into those that occurred in passing zones and those that occurred in no-passing zones. This delineation was done by reviewing the visual images of the passing accident sites on the HSIS Photolog Laser Videodisc Retrieval System.

The data extracted from the photologs were recorded on log sheets specially created for this purpose. The images on the photologs were taken between 1988 and 1990. Even though the accidents being analyzed occurred during the period 1985 through 1989, it is assumed that there were no changes to the pavement markings on these roads between the time that the accidents occurred and the time the roads were photologged.

Comparative analysis of passing zone and no-passing zone accidents. As in stage one, a comparative analysis was conducted on the two groups of passing accidents--those occurring in no-passing zones and those occurring in passing zones. The comparison was conducted for each variable as listed in table 4. The research approach used in the preparation of the data files for these analyses is summarized in figure 1.


Magnitude of Passing Accidents

Table 5 presents the results of the aggregate statistics developed to determine the magnitude of the passing accidents. From this table, the following were observed:

* A slightly higher percentage of passing accidents occurred on two-lane rural roads than on all two-lane roads.

* The percentage of fatal and incapacitating passing accidents on two-lane rural roads was also higher than that on all two-lane roads.

* The number of passing accidents that occurred between 1985 and 1989 was very small--less than 1 percent of all two-lane road accidents and just over 1 percent of all two-lane rural road accidents. However, passing accidents represent just under 5 percent of all the nonturning motor vehicle accidents that occurred on both two-lane roads and two-lane rural roads during the 1985 through 1989 period.

Comparative Analysis of Passing and Nonpassing Accidents

Table 6 presents a breakdown of the accidents that occurred on U.S. and State rural roads that were used in the comparative analysis of these two accident groups. For this analysis, passing accidents excluded those involving turning movement, pedestrians, and nonmotorized vehicles.

Several variables showed marked differences between the passing and nonpassing data sets. These variables are presented in table 7 and are examined more fully in figures 2 through 4.

Descriptive characteristics. The variables in this group present descriptive characteristics of accidents that occurred; they do not represent causal or contributory factors. Figure 2 highlights key differences and similarities for passing and nonpassing accidents among these descriptive variables. Analysis of all descriptive characteristics yielded the following observations:

* Over half of both passing and nonpassing accidents involved collisions with another motor vehicle. Vehicles involved in passing accidents were about 7 percent more likely to overturn than were those in nonpassing accidents.

* Occupants of vehicles involved in passing accidents had a slightly greater likelihood of sustaining injuries and fatalities than those involved in nonpassing accidents.

Environmental conditions. Figure 3 illustrates those environmental factors that showed the most marked difference between the two data sets. These factors are discussed below:

* Nonpassing accidents were more likely to occur under non-ideal environmental conditions--e.g., dark lighting or inclement weather. This observation might indicate that motorists attempt fewer passing maneuvers under non-ideal environmental conditions when the effective passing sight distances might be significantly reduced.

* Partial or total obscurement of pavement markings by snow or ice presents more of a potential driving hazard to the passing motorist than to the nonpassing motorist. This finding might be an indication of the level of motorist compliance with, or their dependency on, existing pavement markings. The obscurement of pavement markings seems to add a hazardous element to vulnerable driving maneuvers, such as those involved in passing. Although this finding might seem to contradict the previous observation, it should be noted that obscured pavement markings can easily exist under daylight or non-inclement weather conditions.

Human characteristics. Figure 4 illustrates those human factors that showed the most marked difference between the two data sets. The findings from the review of these factors follow:

* Driver age did not appear to affect passing accidents. Even though experience seems to improve the ability of drivers to make better driving decisions, it is apparent that for senior drivers other physical impairments may affect their driving abilities. However, because exposure data by age were not available, no further investigation into the magnitude of the older driver problem could be conducted.

* Overall, the number of male drivers involved in both passing and nonpassing accidents was twice that of females. There was a slightly higher proportion of males involved in passing accidents than in nonpassing accidents.

Geometric/roadway characteristics and conditions. Some of the variables analyzed from this group included roadway characteristics, posted speed, shoulder type and width, and pavement type and width. The results of the comparative analysis indicated that there were marked differences for these variables between the two groups. However, due to the direct correlation between roadway design and the type of accidents that would occur, it is difficult to examine and make inferences about these variables unless exposure data for each variable were available.

Comparative Analysis of Passing Accidents in Passing Zones and No-passing Zones

Table 7 presents the passing accidents that are categorized by zone type. Following is a discussion of the comparative analysis conducted on these two groups of passing accidents:

Table 8 lists the variables that showed the most marked differences between the two groups of passing accidents.

Descriptive characteristics. Figure 5 highlights findings from the analysis of descriptive characteristic variables. From this analysis, the following observations were apparent:

* There was a greater tendency for sideswipe-in-passing accidents to occur in no-passing zones, while the tendency for head-on and rear-end collisions was noticeably higher in passing zones.

* Occupants involved in passing zone accidents were more likely to sustain injuries and fatalities than those involved in no-passing zone accidents.

Environmental conditions. Figure 6 depicts the key environmental conditions that differed between the two sets of data. These findings were observed regarding environmental conditions:

* Accidents in no-passing zones were more likely to occur under dark lighting conditions. This finding seems to indicate that the adverse effect of limited sight distance is further compounded under limited lighting conditions for the passing motorist in no-passing zones.

* Passing zone accidents were more likely to occur under adverse weather conditions than were accidents in no-passing zones. This finding may indicate that drivers passing in a no-passing zone are more hesitant to execute the passing maneuver under adverse weather conditions.

* The higher portion of passing zone accidents that occurred when pavement markings were obscured by snow or ice may reflect the fact that motorists had to rely on their own judgment of effective sight distance in determining whether to attempt passing maneuvers. That is, even under these hazardous conditions, drivers still feel the need to pass other vehicles. It is expected that drivers recognize the added hazards and will not execute a pass if a certain level of risk is exceeded. When accidents occur, it is evident that some drivers are not capable of assessing the level of risk on their own and that, at a minimum, they require the guidance provided by the pavement markings and signs.

Human characteristics. Driver age, presented in figure 7, was the only human characteristic variable that showed a marked difference between the two data sets. From a review of this variable, the following was apparent:

* Driver age has no relationship to accident zone.

* Of the drivers involved in no-passing zone accidents, 75 percent were under the age of 40; 63 percent of the drivers involved in accidents in passing zones were also under 40. This finding might be a reflection of the higher risk-taking behavior and lack of experience on the part of younger drivers.

Geometric/roadway characteristics and conditions. Some of the variables analyzed from this group included terrain, posted speed, shoulder type and width, and pavement type and width. The results of the comparative analysis indicated that these variables showed marked differences between the two groups. This finding is expected, because these geometric/roadway variables are in direct correlation to the type of passing zone in which the accidents occur. Definite inferences regarding these variables cannot be made without exposure data for each variable.

Traffic conditions. Figure 8 shows average daily traffic. Accidents in no-passing zones were more likely to occur than accidents in passing zones when daily traffic volumes were greater than or equal to 2,000 vehicles. Higher volumes indicate that motorists will have a higher urge to pass but fewer opportunities to do so. Consequently, they may be more likely to attempt a pass when they perceive an acceptable gap regardless of pavement markings. In no-passing zones, when the availability of these gaps are few and far between, increased driver frustration may cause motorists to attempt passing maneuvers even when sight distances are inadequate.


Magnitude of Passing Accidents

From the aggregate statistics developed to determine the magnitude of passing accidents, it can be concluded that there is not a significant problem in this area. In fact, the pattern of occurrence of passing accidents seems to be extremely random. Thus, based on the findings of this project, the current MUTCD minimum sight distance requirements do not appear to pose a driving hazard to motorists. Consequently, there is no immediate need to change these values.

Comparative Analysis of Passing and Nonpassing Accidents

From the results of this analysis, the following can be concluded:

* Most accidents involve property damage only, but when there are injuries, passing accidents seem to be slightly more severe than nonpassing accidents.

* Most accidents, both passing and nonpassing, occurred under favorable driving conditions-- dry roadway surfaces, straight and level roads with shoulders, and clear weather.

* When pavement markings cannot be seen because of partial or total obscuring of the markings by snow or ice, motorists are forced to rely on their own judgment in determining passing sight distances. The findings of this study show that motorists have difficulty in making such judgments.

Comparative Analysis of Passing Accidents Between Passing and No-passing Zones

From the results of this analysis, the following can be concluded:

* Most of these accidents involve property damage only; however, when an injury occurs, passing zone accidents seem to be more severe than no-passing zone accidents.

* No geometric, environmental, or traffic factor was confidently identified as a primary cause of accidents in these two types of zones. Further, no definite conclusions can be drawn relating these variables to the accident experience unless exposure data are available for the analyses. Some factors seem to contribute to a degree, however. The data indicate that percentages of accidents were lower in no-passing zones for such variables as adverse weather and obscured road surface conditions. This finding is probably an indication that passing in a no-passing zone is a dangerous maneuver even under good driving conditions, and fewer of these maneuvers are attempted under poor driving conditions.

* Most of the drivers involved in passing zone accidents were under the age of 40.

* Increased driver frustration, brought on by the availability of fewer passing opportunities on high volume roads, appears to lead motorists to attempt passing maneuvers in no-passing zones even when sight distances were inadequate.


The following recommendations are based on the findings of the comparative analyses:

* Besides pavement markings, other traffic control devices such as the No-passing pennant should be installed more frequently to further demarcate a passing zone from a no-passing zone or a change between the two--specially where existing pavement markings are obscured.

* Since drivers' inabilities seem to contribute to the occurrence of many accidents, workshops should be conducted to further educate all drivers--especially those under the age of 40--about maneuvers in passing zones.

* Passing lanes should be provided on two-lane rural roads that have high average daily traffic volumes and high posted speeds.


(1) Highway Statistics 1987. Publication No. FHWA-PL-88-088, Federal Highway Administration, Washington, DC, 1988.

(2) Graeme D. Weaver and Donald L. Woods. Passing and No-Passing Zones: Signs, Markings, and Warrants-Final Report. Publication No. FHWA-RD-77-005, Federal Highway Administration, Washington, DC, September 1978.

(3) Anton Huber, Donald R. Hatcher, and Graeme D. Weaver. Passing and No-Passing Zones: Signs, Markings, and Warrants--Accident Analysis. Unpublished report, Federal Highway Administration, Washington, DC, September 1978.

(4) Warren E. Hughes, Sarath Joshua, and Hugh W. McGee. Study Design for Passing Sight Distance Requirements. Publication No. FHWA-RD-91-078, Federal Highway Administration, Washington, DC, September 1991.

(5) A Policy on Geometric Design of Highways and Streets. American Association of State Highway and Transportation Officials, Washington, DC, 1990.

(6) Manual of Uniform Traffic Control Devices. Federal Highway Administration, Washington, DC, 1988.

Henrietta B. Alexander is a 1991 Federal Highway Administration (FHWA) Graduate Research Fellow who conducted this research project within the Information and Behavioral Systems Division of the Office of Safety and Traffic Operations Research and Development. Ms. Alexander is a graduate student from Howard University in Washington, DC, and holds a bachelor of science degree in civil engineering.

Paul A. Pisano is a research highway engineer in the Information and Behavioral Systems Division of the Office of Safety and Traffic Operations Research and Development, He has worked at the Turner-Fairbank Highway Research Center for 3 years and has been with the FHWA for 7 years. His work focuses on highway safety, with a special emphasis on the relationship between roadway features and accidents. Mr. Pisano has both a bachelor's degree and a master's degree in civil engineering from the University of Maryland.

[1] Italic numbers in parentheses identify references on page 60.
Table 1.--AASHTO minimum passing sight distances
 for design of two-lane highways
Design Speed (mi/h) Distance (ft)
 20 800
 30 1,100
 40 1,500
 50 1,800
 60 2,100
 65 2,300
 70 2,500
 Source: AASHTO Policy, 1990 Edition
Table 2.--MUTCD's minimum passing sight
 distances for making no-passing zones
85 percentile Speed or Minimum Passing Sight
 Posted Speed Limit Distance (ft)
Whichever is Higher (mi/h)
 25 400
 30 500
 35 550
 40 600
 45 700
 50 800
 55 900
 60 1,000
 65 1,100
 70 1,200
 Source: MUTCD for Streets and Highways, 1988 Edition
Table 3.--Accident summary of the State data
 for 1985-89
 General Summary
* Total number of accidents 375,690
 Summary for Two-lane Roads
* Two-lane road miles 44,690.5
* Annual million vehicle miles (MVM)
 of travel 17,603.9
* Accidents on two-lane roads
 for period 1985-89 198,185
* Total MVM of travel
 for period 1985-89 88,019.8
- Annual accident rate per 100 MVM
 on two-lane roads: 225
 Summary for Two-lane, Rural Roads
* Two-lane, rural road miles 34,216.9
* Annual MVM of travel 10,749.4
* Accidents on two-lane, rural roads
 for period 1985-89 69,104
* Total MVM of travel
 for period 1985-89 53,746.9
- Annual accident rate per 100 MVM
 on two-lane roads: 129
Table 6.--Accidents on U.S. and State rural roads
 Accidents Vehicles
All accidents 38,080 61,445
Passing accidents 683 (1.8%) 1,174 (1.9%)
Nonpassing accidents 37,397 (98.2%) 60,271 (98.1%)
Table 7.--Passing accidents on U.S. and State
 State rural roads
 Accidents Vehicles
All passing accidents 683 1,174
Passing zone accidents 617 (90.3%) 1,056 (89.9%)
No-passing zone
accidents 66 (9.7%) 118 (10.1%)
Table 8.--Bariables that differed between the
 passing and no-passing zone accident groups
Variable Group Distinguishing Variables
General Descriptive Accident type, Number of
 injuries, Accident Severity,
 Number of Vehicles Involved
Environmental Lighting Conditions, Road Surface
 Conditions, Weather
Human Age of Driver
Traffic Average Daily Traffic

[Tabular Data Omitted]
COPYRIGHT 1992 Superintendent of Documents
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Federal Highway Administration research
Author:Alexander, Henrietta B.; Pisano, Paul A.
Publication:Public Roads
Date:Sep 1, 1992
Previous Article:The FHWA test road: construction and instrumentation.
Next Article:Investigation of passing accidents using the HSIS data base.

Related Articles
Investigation of passing accidents using the HSIS data base.
Accident rates using HSIS.
Be ALERT for efficiency and safety.
International Cooperation to Prevent Collisions at Intersections.
Reducing points of conflict: FHWA targets intersection safety.
Making trails.
Preventing roadway departures.
Safety scans--a successful two-way street.
Four die in one-vehicle crash on Highway 58.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters