U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-02-089
Date: July 2002

Safety Effectiveness of Intersection Left- and Right-Turn Lanes

PDF Version (1.48 MB)

PDF files can be viewed with the Acrobat® Reader®

2. LITERATURE REVIEW ON SAFETY EFFECTS OF INTERSECTION DESIGN ELEMENTS

This section of the report presents the results of the literature review that was conducted as part of the research. This literature review covered all aspects of intersection safety. Based on this review, a decision was reached to focus the research on the safety effectiveness of left- and right-turn lanes at intersections. Therefore, these issues are addressed in greater detail than most other issues.

Overview

The scope of the literature review includes studies related to the safety effects of a wide variety of geometric design, traffic, and control elements of at-grade intersections. Although the research presented in this report focuses on the safety effectiveness of intersection left- and right-turn lanes, the initial scope of the research was not limited to this topic and could potentially have included the safety evaluation of any type of intersection design improvement. Therefore, this literature review is organized to emphasize studies related to the safety effectiveness of turn lanes, but it also includes a review of all geometric, traffic, and control elements that affect the safety of at-grade intersections.

The review identifies studies that address general intersection geometric design, traffic, and control issues with emphasis on studies that provide a quantitative estimate of the factor of interest. Some studies find a factor to be related to safety, but do not quantify the effect of that factor. With minor exceptions, the review does not address studies that investigated a factor but did not find it to be important or statistically significant. In such cases, it would be difficult without more detailed review of the study to judge whether the lack of an observed effect resulted from the true lack of a relationship of that factor to safety or from a limited sample size or poor study design. The review considers both studies that directly evaluated relationships between the factors of interest and safety and studies that summarized and synthesized past research. The latter were included to take advantage of the judgements made by previous reviewers.

Table 1 presents a list of the intersection features that are discussed in this review. The table is organized into three categories: intersection geometric design features, traffic control and operational features, and traffic characteristics. The specific topics that are most directly related to the safety effectiveness of turn lanes are listed first under each category.

The remainder of this section of the report presents the findings of the literature review of the specific topics identified in table 1. The findings are also presented in an extensive summary table in appendix A.

Table 1. Intersection Features Addressed in the Literature Review.
Intersection geometric design features
Traffic control and operational features
Traffic characteristics
Left-turn lanes

- offset left-turn lanes

Right-turn lanes

Channelization

- island design

Number of intersection legs (e.g., 3, 4, 5)

Intersection type (e.g., cross, T, Y, offset)

Roundabouts

Angle of intersection (e.g., skew)

Curb return radius

Sight distance

- intersection sight distance

- stopping sight distance

- sight distance to traffic control device

Approach width

Number of approach lanes

Median width and type

Vertical alignment on approaches

Horizontal alignment on approaches

Type of traffic control:

- uncontrolled

- YIELD-controlled

- STOP-controlled

- signal-controlled

- roundabouts

Turn prohibitions

Presence and type of crosswalks

Posted speed limit on approaches

Advance warning signs

Lighting

Average daily traffic (ADT)

- total entering ADT (all approaches)

- entering ADTs for major and minor approaches

Turning movements

Peak hour approach volumes

Vehicle mix/percent trucks

Distribution of total entering volume by hour of day

Distribution of approach volume by hour of day

Average approach speed

Volume of bicycle traffic

Volume of pedestrian traffic

Intersection Geometric Design Features

Left-Turn Lanes

Installation of left-turn lanes has been the focus of many research studies. Various safety-related impacts have been documented depending upon the type of intersection (signalized, unsignalized, four-leg, etc.) where the left-turn treatment was implemented, as well as the different types and/or severity of accidents. Parker(3) determined that the addition of left-turn lanes at rural intersections along two-lane highways can reduce the potential for passing-related accidents. On urban four-lane roadways, McCoy and Malone(4) found that installation of left-turn lanes reduced rear-end, sideswipe, and left-turn accidents. Foody and Richardson(5) found that accident rates decreased by 38 percent with the addition of a left-turn lane at signalized intersections and by 76 percent at unsignalized intersections. Gluck et al.(6) reported accident rate reductions ranging from 18 to 77 percent due to the installation of left-turn lanes, based on the review of work by the New Jersey Department of Transportation,(7,8) Griewe,(9) Agent,(10) Ben-Yakov and Craus,(11) Craus and Mahalel,(12) Tamburri and Hammer,(13) and Wilson et al.(14)

When implemented with additional safety measures, left-turn lanes have been very effective in increasing safety. Haler reported that left-turn channelization reduced accidents to varying degrees depending upon the intersection configuration.(15) Based on a synthesis of work by McFarlane,(16) Haler reported that the provision of left-turn lanes at unsignalized intersections, when combined, with installation of curbs or raised medians, reduced accidents by 70, 65, and 60 percent in urban, suburban, and rural areas, respectively. When the channelization was painted rather than raised, accidents decreased only by 15, 30, and 50 percent in urban, suburban, and rural areas, respectively. At signalized intersections, installation of left-turn channelization accompanied by a left-turn signal phase reduced accidents by 36 percent; however, without the left-turn phase, accidents decreased only by 15 percent.(16) At unsignalized intersections, findings of a California study indicate greater reductions in accidents with the use of a left-turn lane in a raised median than with painted left-turn lanes.(17) Similarly, Lacy(18) found that a left-turn lane, when coupled with several other safety improvements, reduced accident frequency by 35 percent and accident severity by 80 percent. Dale(19) found that installation of a traffic signal and left-turn channelization at intersections along rural two-lane highways reduced the total number of accidents by 20 percent, while the installation of a traffic signal without any channelization reduced the total number of accidents by only 6 percent.

Not all studies, however, have shown that left-turn lanes reduce accidents. Bauer and Harwood(20) found that left-turn lanes were associated with higher frequencies of both total multiple-vehicle accidents and fatal and injury multiple-vehicle accidents. However, this result was not advanced by the authors as a basis for policy because the directions of specific effects in predictive models often represent the surrogate effects of other variables, rather than the true effect of the variable of interest. At unsignalized intersections, McCoy and Malone determined there was a significant increase in right-angle accidents.(4) However, at unsignalized intersections on rural two-lane highways, McCoy et al.(21) found no significant difference in rear-end and left-turn accident rates between intersections with and without left-turn lanes. Poch and Mannering(22) also found some situations in which accidents of specific types increased with installation of left-turn lanes.

Several predictive models and accident modification factors have been developed that indicate left-turn lanes have a positive effect on safety. Maze et al.(23) developed a model that predicted a reduction in left-turn accident rate of 6 percent due to the installation of a left-turn lane with permitted signal phasing and a reduction of approximately 35 percent from installation of a left-turn lane with protected/permitted signal phasing. Vogt(24) developed a model for a four-leg rural intersection of a four-lane major road with STOP-controlled two-lane minor roads which yielded an accident reduction factor for total accidents of 38 percent due to the installation of a left-turn lane along the major road.

In another study, Harwood et al.(25) developed algorithms to predict the expected safety performance of rural two-lane highways. The prediction algorithms combined elements of historical accident data, predictions from statistical models, results of before-after studies, and expert judgements made by experienced engineers. As part of the research, an expert panel of safety researchers developed accident modification factors (AMFs) for specific geometric design and traffic control features. AMFs are used in the accident prediction algorithms to represent the effects of safety of the respective features. The base value of each AMF is 1.0. Any feature associated with a higher accident experience than the base condition has an AMF with a value greater than 1.0, and any feature associated with lower accident experience than the base condition has an AMF with a value less than 1.0.

In developing AMFs for the installation of left-turn lanes on the major-road approaches to intersections on two-lane rural highways, the expert panel conducted an extensive review of past research on the safety effectiveness of left-turn lanes, including most of the studies discussed above. The panel was charged with defining the safety effectiveness of intersection left-turn lanes based on the best study of this issue or based on results from a combination of studies. The panel concluded that there have been no well-designed before-after evaluations of intersection left-turn lanes and no single study that was considered more reliable than others. Therefore, the panel combined results from several studies and developed AMFs for left-turn lanes, which are presented in Table 2. The AMFs represent a judgement by the panel. The panel estimated that installation of a left-turn lane along one major approach reduces intersection-related accidents by 18 to 24 percent, depending upon the type of traffic control and the number of legs, and installation of left-turn lanes along both major approaches to a four-leg intersection reduces intersection-related accidents by 33 to 42 percent, depending upon the type of traffic control. These results are presented in table 2 in the form of AMFs, as defined above.

No research was found that quantifies the safety effectiveness of extending the length of existing left-turn lanes to eliminate traffic overflows into through travel lanes and to allow a greater proportion of vehicle deceleration to occur in the turn lane rather than in the through travel lanes.

Table 2. Accident Modification Factors for Installation of Left-Turn Lanes on the Major-Road Approaches to Intersection on Two-Lane Rural Highways(25).
Intersection type Intersection traffic control Number of major-road approaches on which left-turn lanes are installed
One approach Both approaches
Three-leg intersection

STOP signa

Traffic signal

0.78

0.85

Four-leg intersection STOP signa
Traffic signal

0.76

0.82

0.58

0.67

a STOP signs on minor-road approach(es)

An emerging issue in the design of left-turn channelization is the restriction in sight distance that opposing left-turn vehicles cause one another. As an indication of this safety problem, David and Norman(26) determined that for average daily traffic (ADT) volumes between 10,000 and 20,000 veh/day, four-leg intersections with opposing left-turn lanes had more accidents than those without. A potentially effective countermeasure for safety problems where opposing left-turn lanes are present is to eliminate the sight restrictions by offsetting the left-turn lanes. Harwood et al.(27) reviewed the safety performance of a limited set of tapered and parallel offset left-turn lanes and found no safety problems. Both McCoy et al.(28) and Joshua and Saka(29) developed procedures to compute the amount of offset required for clear sight lines. However, no evaluations of the accident reduction effectiveness of offset left-turn lanes have been found.

Table 3 summarizes the results of those studies that provided quantitative estimates of the effectiveness of installing left-turn lanes at intersections.

Right-Turn Lanes

Compared to left-turn lanes, very few studies have been conducted on the safety effectiveness of right-turn lanes. Bauer and Harwood(20) indicate that right-turn channelization resulted in a decrease in both total multiple-vehicle accidents and fatal and injury multiple-vehicle accidents. However, Vogt and Bared(30) modeled accidents for three-leg unsignalized intersections along rural two-lane highways, and based upon the prediction model, the presence of a right-turn lane increases intersection-related accidents by 27 percent.

The expert panel discussed above also developed estimates of the safety effectiveness of right-turn lanes; these AMFs are presented in table 4.(25) In their review of information, the expert panel did not find any well-designed before-after studies on the accident reduction effectiveness of right-turn lanes. Based on a review of the available studies, the expert panel estimated the presence of a right-turn lane along one approach to a rural STOP-controlled intersection reduces intersection-related accidents by 5 percent, and the

Reported LTL effectiveness (percent change in accident frequency)

Table 3. Summary of Research Results Concerning the Safety Effectiveness of Installing Left-Turn Lanes.
Source Total intersection accidents Left-turn accidents Conditions/comments
Harwood et al. [2000](25) -18 to- 24 two-lane highway; LTL on one major-road approach
-32 to -42 two-lane highway; LTLs on two major-road approaches
Vogt [1999](24) -38 LTL at four-leg rural intersection with four-lane major road and two-lane minor road
Maze et al [1994](23) 6 signalized intersection; LTL with permitted phasing
-35 signalized intersection; LTL with protected/permitted phasing
New Jersey Department of Transportation [1993](8) -35 to -51 LTL installation on Route 130 in

New Jersey

Griewe [1986](9) -58 -62 eight LTLs added by restriping
Agent [1983](10) -77 unsignalized intersection
-54 signalized intersection
Ben Yahov and Craus [1980](11)/Craus and Mahalel [1980](12) -38 LTL installation
McFarlane [1979](16) -70 LTL with curbed median; urban
-65 LTL with curbed median; suburban
-60 LTL with curbed median; rural
-15 LTL with painted median; urban
-30 LTL with painted median; suburban
-50 LTL with painted median; rural
-36 signalized intersection with LTL and exclusive phase
-15 signalized intersection with LTL but no exclusive phase
Foody and Richardson [1973](5) -38 signalized intersections
-76 unsignalized intersections
Dale [1973](19) -20 two-lane highway intersection; installation of signal with LTL
Lacy [1972](18) -35 installation of LTL with other improvements
Tamburri and Hammer [1968](13)/ Wilson et al [1967](14) -18 unsignalized intersection

 

Table 4. Accident Modification Factors for Installation of Right-Turn Lanes on the Major-Road Approaches to Intersection on Two-Lane Rural Highways.(25)
 Intersection traffic control Number of major-road approaches on which right-turn lanes are installed
One approach Both approaches
STOP signa 0.95 0.90
Traffic signal 0.975 0.95

a STOP signs on minor-road approach(es)

presence of a right-turn lane along both major approaches reduces intersection-related accidents by 10 percent. Similarly for rural signalized intersections, the expert panel estimated a reduction of 2.5 percent in total intersection-related accidents due to the presence of a right-turn lane along one major-road approach and 5 percent for right-turn lanes along both major-road approaches.

No research was found that quantifies the safety effectiveness of extending the length of existing right-turn lanes to eliminate traffic overflows into through travel lanes.

Table 5 summarizes the results of available studies on the safety effectiveness of right-turn lanes.

Reported LTL effectiveness
(percent change in accident frequency)

Table 5. Summary of Research Results Concerning the Safety Effectiveness of Installing Right- Turn Lanes.
Source Total intersection accidents Right-turn accidents Conditions/comments
Harwood et al. [2000](25) -5 two-lane highway; RTL on one major road approach to an unsignalized intersection
-10 two-lane highway; RTLs on two major-road approaches to an unsignalized intersection
-2.5 two-lane highway; RTL on one major-road approach to an unsignalized intersection
-5 two-lane highway; RTLs on two major-road approaches to signalized intersection
Vogt and Bared [1998](30) 27 based on multivariate modeling with Minnesota data

Channelization

Four functional objectives form the basis for channelization design concepts:(31,32)

  • Limiting the points of conflict.
  • Limiting the complexity of the conflict area.
  • Limiting the conflict frequency.
  • Limiting the conflict severity.

A variety of measures such as designation and arrangement of traffic lanes, traffic islands, median dividers, and various signs, signals, and markings may be used for channelization purposes. Studies on channelization by David and Norman,(26) Exnicios,(33) and Rowan and Williams(34) in general indicate that channelization improves safety. Exnicios(33) found reductions in accidents as high as 100 percent over a 26-month period. Haler(15) also reported that channelization can reduce accidents.

Channelizing islands are defined areas between traffic lanes that control vehicle movements and serve as refuge points for pedestrians.(35) Islands also provide suitable locations to place traffic control devices. Islands vary in both size and shape, as well as the type of surfacing material used.

Washington et al.(36) found that intersection approaches with raised medians had accident rates 40 percent lower than intersection approaches with flush medians. Forrestel(37) found that installation of a raised median island reduced the pedestrian accident rate by 11.5 percent. In another study, Templer(32) found that a raised median reduced the number of conflicts between pedestrians and vehicles, but the difference was not statistically significant.

Number of Intersection Legs

There is broad agreement in the literature that four-leg intersections experience more accidents than comparable three-leg intersections. This finding is logical because four-leg intersections have more conflict points than three-leg intersections and, therefore, present more opportunities for accidents to occur. Four studies have quantified this effect.

Bauer and Harwood(20) found that both rural and urban STOP-controlled intersections with four legs experienced approximately twice as many accidents as three-leg intersections. Specifically, rural four-leg STOP-controlled intersections experienced an average of 1.1 accidents per year, while three-leg intersections experienced 0.6 accidents per year. Urban four-leg STOP-controlled intersections experienced 2.2 accidents per year, while three-leg intersections experienced 1.3 accidents per year.

Predictive models developed by Harwood et al.(27) showed that typical divided highway intersections with four legs had about twice as many accidents as three-leg intersections for narrow medians and more than five times as many accidents as three-leg intersections for wide medians.

Hanna et al.(38) found that, in rural areas, four-leg intersections experience approximately 69 percent more accidents than T intersections. T intersections are three-leg intersections at which the legs meet at a right angle, while Y intersections are three-leg intersections where one or more of the legs are skewed. David and Norman(26) found that for STOP-controlled intersections in urban areas with total entering traffic volumes under 20,000 veh/day, the accident frequencies for three- and four-leg intersections were very similar; however, for intersections with total entering volumes over 20,000 veh/day, four-leg intersections experienced twice as many accidents as three-leg intersections.

Intersection Type

The review of intersection type focused on the differences between conventional and offset four-leg intersections and between T and Y three-leg intersections. Lau and May(39,40) found these differences to be statistically significant in modeling of injury accidents at both signalized and unsignalized intersections, but their classification and regression tree (CART) analysis results are difficult to interpret as a specific effect of these factors. Lau and May also modeled fatal and property-damage-only (PDO) accidents but, for the sake of simplicity, the discussions in this paper focus on the findings of injury accident modeling that are typical of the others.

Hanna et al.(38) found that, for three-leg intersections, Y intersections have accident rates approximately 50 percent higher than T intersections; for four-leg intersections, offset intersections had accident rates that were approximately 43 percent of the accident rate of conventional four-leg intersections. The effect observed by Hanna et al. is interesting. The operating experience of some highway agencies indicates that offset intersections can create operational and safety problems as through vehicles on the crossroad must turn onto and off of the major road rather than making a simple crossing maneuver. A number of projects have been constructed to realign the legs of offset intersections to convert them to conventional four-leg intersections. However, the results of Hanna et al. suggest the opposite—that offset intersections operate more safely than conventional four-leg intersections. This finding may indicate that, where there is little through traffic on the crossroad, two T intersections operate more safely than one conventional four-leg intersection.

Roundabouts

Roundabouts are a unique topic. They can be considered both an intersection geometric design feature and a form of intersection traffic control. Because roundabouts are classified as a form of intersection traffic control in Roundabouts: An Informational Guide, the safety effectiveness of roundabouts is discussed in the section on type of traffic control.(41)

Angle of Intersection

The angle between the legs of an intersection, particularly whether the legs intersect at a right or an oblique angle, has long been considered to affect the safety performance of the intersection. McCoy et al.(42) found that accidents at rural two-way STOP-controlled intersections increase with increasing skew angle; this result applies to both three-leg and four-leg intersections. In addition, the previously discussed difference in safety performance between three-leg T and Y intersections found by Hanna et al.(38) represents an effect of the angle of the intersection.

Harwood et al.(25) incorporated AMFs for intersection skew angle when they developed algorithms to predict the expected safety performance of rural two-lane highways. The AMFs for intersection skew angle were derived from statistical modeling and apply to total intersection-related accidents. Thus, the AMFs were formulated from data and do not represent judgements by the expert panel on the accident reduction effectiveness of this design feature. For a three-leg STOP-controlled intersection, the AMF was calculated as:

AMF = exp(0.0040 SKEW) (1)

For a four-leg STOP-controlled intersection, the AMF was calculated as:

AMF = exp(0.0054 SKEW) (2)

where:

SKEW = intersection skew angle (degrees), expressed as the absolute value of the difference between 90 degrees and the actual intersection angle.

Curb Return Radius

The curb return radius of an intersection controls the turning speed for vehicles making right turns. In addition, larger curb return radii make it possible for intersections to better accommodate right turns by large trucks. Haler(15) cited curb return radius as an important factor in safe intersection operations, but apparently no specific evaluations of the effect of curb return radius on safety have been conducted.

Sight Distance

Sight distance is the distance ahead or along an intersecting roadway that a driver can see from any location on the roadway system. Provision of adequate sight distance is fundamental to the design of roadways and intersections for safe operations. Three types of sight distance are particularly critical to the safe operation of at-grade intersections: intersection sight distance, stopping sight distance, and sight distance to traffic control devices.

Three studies have addressed the safety effects of intersection sight distance. David and Norman(26) found that within specific ADT levels the reduction in accident experience from a sight distance improvement was, in most cases, highest for intersection approaches whose initial sight distance was lowest. Hanna et al.(38) found that intersections with "poor" sight distance had an observed accident rate of 1.33 accidents per million entering vehicles, while intersections as a whole had an accident rate of 1.13 accidents per million entering vehicles. Mitchell(43) found that total intersection accidents were reduced by 67 percent when intersection sight obstructions were removed. Unfortunately, none of these studies were specific concerning the magnitude of the sight distance improvements made.

Fambro et al.(44) found that accident rates were high for intersections located on crest vertical curves with limited sight distance. The results of another recent study by Fambro et al.(45) are consistent with that finding.

No evaluations were found of the safety effects of limited sight distance to traffic control devices, such as STOP signs and signals.

The expert panel of safety researchers discussed earlier reviewed several sources of information to evaluate the effects of intersection sight distance on intersection-related accidents. The panel did not find any single evaluation to be the most credible. Therefore, the AMFs established by the panel represent the panel's best judgment on the safety effects of intersection sight distance. The AMFs are as follows for intersection sight distance at intersections with STOP control on the minor leg(s):

  • 1.05 if sight distance is limited in one quadrant of the intersection.
  • 1.10 if sight distance is limited in two quadrants of the intersection.
  • 1.15 if sight distance is limited in three quadrants of the intersection.
  • 1.20 if sight distance is limited in four quadrants of the intersection.

In applying these AMFs, sight distance in a quadrant of an intersection is considered limited if the available sight distance is less than the sight distance specified by AASHTO policy for a design speed of 20 km/h (12 mph) less than the major road-design speed and the sight distance restrictions are due to roadway alignment and/or terrain.

Approach Width

The width of an intersection approach includes the combined widths of the approach lanes and, in some cases, the width of the shoulder, as well. Studies by Bauer and Harwood,(20) Neuman,(31) and Lacy(18) found that increasing the approach width to an intersection reduces the accident rate along the approach. Bauer and Harwood(20) found that as lane width decreases on an intersection approach, accidents tended to increase. Similarly, Neuman(31) indicated that accidents may be reduced by widening the shoulder at intersections on narrow two-lane roadways. Widening of the shoulders may reduce accidents by providing space for collision-avoidance maneuvers and by providing better sight lines if sight distance is limited on the approach. Lacy(18) also found that widening the approaches, combined with other safety improvements, decreased accident frequency by 35 percent and accident severity by 80 percent. By contrast, David and Norman(26) did not find any evidence that incremental changes in lane or shoulder width near intersections affects accident rates.

Number of Approach Lanes

The number of lanes on an intersection approach is determined primarily by traffic demand and the desired level of service. Intuitively, one might assume that the number of accidents is proportional to the number of lanes (i.e., as the number of lanes increases so would the total number of accidents, since the potential number of conflicts would appear to increase). However, Bauer and Harwood(20) found that for unsignalized intersections in both rural and urban areas, the number of accidents tended to be higher on facilities with one approach lane and accidents tended to be lower at intersections with two or more approach lanes. The opposite appears to be the case for urban, four-leg, signalized intersections. David and Norman(26) also indicated that accident frequencies can be reduced for intersections with total entering volumes under 10,000 veh/day by adding through lanes. It should be noted that with a demand-related design parameter such as number of lanes, it is difficult to assess directly whether any observed safety effects are due to the number of lanes or to the traffic volume on the approach.

Median Type and Width

The width of a divided highway median influences the safety performance of intersections on that highway. Harwood et al.(27) found that accident frequencies at rural four-leg signalized intersections decrease as median width increases. In contrast, at both signalized and unsignalized intersections in urban and suburban areas, accident frequencies were found to increase with increasing median width. Similar results for rural divided highway intersections were found in an earlier Ohio study by Priest.(46) An Indiana study by Van Maren found no statistically significant relationship between median width and accident rates at divided highway intersections.(47)

Vertical Alignment

Crest and sag vertical curves are used to provide a smooth transition between roadway segments with different grades. From a safety standpoint, it is undesirable to locate intersections on steep grades or on crest vertical curves with limited sight distance. Steep upgrade approaches to intersections cause difficulty because vehicles accelerate more slowly, resulting in increased time during which the vehicle is exposed in the conflict area of the intersection. Steep downgrade approaches to intersections result in longer stopping distances, which may cause potential problems, as well. Surprisingly, however, Hanna et al.(38) found the accident rates for intersections with grades steeper than five percent to be lower than the average accident rate for all intersections. The average accident rates were 0.97 and 1.13 accidents per million entering vehicles for intersections with steep grades and for all intersections, respectively.

As discussed above, vertical curves cause potential problems at intersections where sight distance is limited. In particular, Fambro et al.(44) concluded that accident rates were high at intersections on crest vertical curves where sight distance was limited.(27)

Horizontal Alignment

From a safety standpoint, it is desirable for the alignment of intersecting roadways to be straight as practical. Horizontal curves on the approaches to intersections make it difficult for a driver to discern the proper path of travel and also affect a driver's visual perspective, since the driver's focus is directed tangentially to the travel path.(48) Horizontal curves also add complexity to the driving environment. Past research has shown that the distance from a horizontal curve to the nearest intersection is related to safety.(49) However, no studies were found which indicate that any specific threshold value for degree of curvature adversely affects safety on intersection approaches.

Traffic Control and Operational Features

Type of Traffic Control

A variety of different traffic control types are used for at-grade intersections including no control, YIELD-control, STOP-control, signal control, and roundabouts.

Poch and Mannering(22) indicated that intersections with no control on any of the approaches experience fewer total and angle accidents than intersections with other types of traffic control. However, this effect could have been observed solely because uncontrolled intersections typically have lower traffic volumes than other intersection types.

Hauer,(15) in a synthesis of past research, noted that conversion from no control to YIELD control reduced accidents by 44 to 52 percent in one study and by 23 to 63 percent in another.

Hall et al.(50) found that accidents can be reduced by 20 to 60 percent by proper use of YIELD signs. However, little additional benefit was found if the YIELD signs were replaced by STOP signs. Agent and Deen(51) found that at YIELD-controlled intersections, over half of the accidents were rear-end collisions, while angle collisions made up over half of the accidents at STOP-controlled intersections. Hanna et al.(38) found that accident rates at STOP-controlled intersections were lower than those at intersections having higher traffic flow.

No safety evaluations were found in the literature for intersections where flashing beacons were used in conjunction with STOP signs at either two-way or all-way STOP-controlled intersections.

Research by Hanna et al.(38) indicates that signalization of intersections that are currently unsignalized typically results in a slight increase in accident rate, a substantial increase in rear-end collisions, and a comparable decrease in angle collisions. Poch and Mannering(22) found that total and angle accidents for signal-controlled intersections were lower than for other traffic control types.

Maze et al.(23) developed predictive models which indicate that a protected left-turn signal phase without a left-turn lane has a positive effect on safety. A numerical example developed by the authors indicates an anticipated reduction in left-turn accidents of 50 percent from installation of a left-turn signal phase. David and Norman(26) found that in urban areas, multiphase traffic signals appear to have lower percentages of fatal and injury accidents than two-phase signals. King and Goldblatt(52) found that signalization leads to a reduction in angle collisions and an increase in rear-end collisions; their results also indicate that signalized intersections have higher accident rates, although this is often offset by reduced accident severity.

U.S. experience with roundabouts is rather limited, but interest has increased recently, partially due to the operational and safety benefits being reported in documents such as Roundabouts: An Informational Guide.(41) The Informational Guide indicates roundabouts may improve the safety of intersections by eliminating or altering conflict types, by reducing speed differentials at intersections, and by decreasing overall speeds into and through intersections. The Informational Guide summarizes the overall safety performance of roundabouts in various countries, including the U.S. After converting intersections with conventional traffic control to roundabouts, a reduction in accidents is reported of about 37 percent for all accidents and 51 percent for injury accidents. These values are consistent with experiences in the U.S. and internationally. Persaud et al.(53) found similar results after performing a before-after accident analysis following the conversion of twenty-three intersections from STOP-control and signal-control to roundabouts. Persaud et al. reported a 40 percent reduction in total accidents, an 80 percent reduction for all injury accidents, and about a 90 percent reduction of fatal and incapacitating injury accidents.

Turn Prohibitions

Research by Lau and May(39,40) found that left-turn prohibitions were a significant factor in predicting injury accidents at both signalized and unsignalized intersections. However, the results of this CART analysis are difficult to interpret in order to obtain an explicit estimate of this effect.

Presence and Types of Crosswalk

The purpose of marked crosswalks is to guide pedestrians across a busy roadway, as well as to increase drivers' awareness of pedestrians. Some intersections provide designated crosswalks for pedestrians, while others do not. Research results provide conflicting conclusions as to whether the provision of marked crosswalks actually improves safety for pedestrians. Several studies have concluded that marked crosswalks decrease accident rates, in some cases by as much as 50 percent.(15,54) On the other hand, perhaps the best-known study on crosswalks, conducted by Herms(55) in 1970, concluded that approximately twice as many pedestrian accidents occurred in marked crosswalks as in unmarked crosswalks. Another study found that pedestrian accidents increased by 86 percent after crosswalks were marked.(15) As Herms pointed out, the increase in accident rates resulting from marked crosswalks may "not be due to the crosswalk being marked as much as it is a reflection on the pedestrian's attitude and behavior when using the marked crosswalk." Other factors which may affect the safety of marked crosswalks include visibility, intersection type, and signal timing.

Although crosswalks typically affect pedestrian safety, it is also important to note that vehicular accident rates may also be affected. Hauer(15) noted that rear-end collisions increase after crosswalks are marked. Thus, the need for crosswalks should be examined from the standpoint of both pedestrian safety and vehicular safety.

Posted Speed Limit

It is rational to assume that the likelihood and severity of accidents on an intersection approach increases as the posted speed limit on the approach increases. Higher posted speed limits are generally associated with higher approach speeds, which require longer distances to bring an approaching vehicle to a complete stop. Therefore, drivers must react more quickly to potential conflicts encountered at the intersections. However, no studies were found that quantify the extent to which accidents increase or decrease with changes in posted speed limits or operating speeds on intersection approaches.

Advance Warning Signs

Advance warning signs are intended to increase a driver's awareness of upcoming traffic situations. Studies of specific types of advance warning signs provide varied results. Gattis and Iqbal(56) found that most drivers do not abide by the "Do Not Block Intersection" sign. Washington(36) found that accident rates increased for approaches to skewed intersections where advance warning signs were provided. Pant and Huang(57) found the "Prepare To Stop When Flashing" sign raised conflict rates by 15 percent on curved approaches but had no influence on conflict rates on tangent approaches. Pant and Huang also noted that the flashing symbol "Signal Ahead" sign had no impact on traffic conflict rates.

Research has also shown positive effects for certain supplements to advance warning signs. Washington(36) found that advance warning signs with flashers (AWFs) can reduce approach accident rates at high-speed signalized intersections by as much as 50 percent. He also concluded that right-angle accidents were reduced when route markers and/or advance warning signs were present. Klugman(58) found that total accident rates decreased from 1.22 to 1.09 accidents per million entering vehicles at AWF-equipped intersections; right-angle and rear-end accident rates also decreased from 0.68 to 0.63 accidents per million entering vehicles. In other related work, Styles(59) concluded that the "Red Signal Ahead" warning sign reduced right-angle accident rates by 42 percent on intersection approaches with crest vertical curves and reduced the total accident rate on intersection approaches with horizontal curves and tangent alignments by 14 and 41 percent, respectively. In a separate study, Styles(60) found that flashing red strobe lights are also effective in reducing right-angle accidents.

It is important to note that many of the studies related to advance warning signs stress the importance of factors such as approach alignment, type of sign, and type of accident as influencing the accident reduction effectiveness for such devices.

Lighting

Intersection lighting is potentially effective as a countermeasure to reduce nighttime accidents. Bauer and Harwood(20) found that rural, four-leg, STOP-controlled intersections that were lighted would be expected to experience 21 percent fewer fatal and injury accidents than unlighted intersections. However, for other intersection types, no similar effect was observed and, in some cases, an opposite effect that may represent a surrogate effect of some other variable was observed. It is important to note that this study evaluated total accidents (daytime plus nighttime), rather than nighttime accidents alone.

Box(61) found that improved lighting reduced the proportion of pedestrian/bicycle, fixed-object, sideswipe, and other accidents that occurred at night on a 4.5-km (2.8-mi) section of a suburban arterial in Illinois. Only nighttime head-on accidents increased as a proportion of total (daytime plus nighttime) accidents.

An extensive study in Los Angeles found no statistically significant reduction in nighttime accidents due to lighting improvements at intersections. Statistically significant reductions in nighttime accidents were found for a few intersections.(62)

Traffic Characteristics

Average Daily Traffic (ADT) Volume

Many studies have found approach traffic volumes to have a strong relationship to intersection accidents. A number of studies have used the total entering ADT as an exposure measure in determining intersection accident rates. Bauer and Harwood(20) found better results in accident prediction modeling when the major-road and crossroad ADTs were treated as separate independent variables than when they were combined as a product or a sum. Lau and May(39,40) represented the relative traffic volumes on the intersecting roadways by the ratio of the crossroad volume to the total entering ADT, expressed as a percentage.

Turning Movements

Hauer et al.(63) developed relationships between accident frequency for specific accident types (e.g., left-turn accidents) and the turning movement volumes most specifically related to that accident type.

Other Traffic Characteristics

No studies were found relating the following traffic flow measures to accidents:

  • Peak hour approach volumes.
  • Vehicle mix/percent trucks.
  • Distribution of total entering volume by hour of the day.
  • Distribution of approach volume by hour of the day.
  • Average approach speed.
  • Volume of bicycle traffic.
  • Volume of pedestrian traffic.

Summary

The scope of this literature review covers the safety effectiveness of general intersection geometric design features, traffic control elements, and traffic characteristics, focusing on studies that provide a quantitative estimate of the factor of interest. Based upon the review, it is evident that many design features have the capability to improve the safety of an at-grade intersection. It is also evident by the quantity of the studies related to left- and right-turn lanes that there is considerable interest in quantifying their safety effectiveness. This interest has been stimulated by the number of highway agencies that have installed turn lanes and by the results of previous studies that give a strong indication that installation of left- or right-turn lanes improves the safety of at-grade intersections.

Based on these considerations, representatives of the state highway agencies in this pooled-fund study decided to focus this research on the evaluation of the safety effectiveness of intersection left- and right-turn lanes.

 

Previous | Table of Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101