Chapter 2
Understanding Red-Light Running

This chapter is provided to better understand the problem of red-light running and its characteristics through review of past research. Investigation of the red-light running problem includes:

• Definitions;
• Understanding the frequency;
• Understanding safety implications;
• Discussing the relationship between a red-light runner or intersection characteristics and the likelihood of red-light running;
• Understanding the stop-go decision process; and
• Relating causes of red-light running to engineering, education, or enforcement countermeasures.

RED-LIGHT RUNNING DEFINED

Simply stated, red-light running is entering, and proceeding through, a signalized intersection after the signal has turned red. According to the Uniform Vehicle Code (UVC) (3), a motorist "...facing a steady circular red signal shall stop at a clearly marked stop line, but if none, before entering the crosswalk on the near side of the intersection, or if none, then before entering the intersection and shall remain standing until an indication to proceed is shown..." (section §11-202). An intersection is defined in the UVC as "... the area embraced within the prolongation or connection of the lateral curb lines, or if none, then the lateral boundary lines of the roadways of two highways which join one another at, or approximately at right angles, or the area within which vehicles traveling upon different highways joining at any other angle may come in conflict" (section §1-132). From an enforcement perspective, if the motorist stops past the stop line or into the crosswalk, or even slightly into the physical intersection, a citation would likely not be given. The motorist usually has to pass through the intersection to be cited for running a red light.

The law as stated in the UVC is considered a permissive yellow law, meaning that the driver can enter the intersection during the entire yellow interval and be in the intersection during the red indication as long as he/she entered the intersection during the yellow interval. As of 1992, permissive yellow rules were followed by at least half of the states (4). However, in other states there are two types of restrictive yellow laws that apply, namely:

• Vehicles can neither enter the intersection nor be in the intersection on red; or
• Vehicles must stop upon receiving the yellow indication, unless it is not possible to do so safely.

In those states where the yellow phase is considered restrictive, it is possible that an officer might stop the driver to discuss the law and to take appropriate action as required. Doing such, however, includes subjectivity of the officer.

The slight differences mentioned above will surface when developing a plan to address red-light running. For instance, if an automated-enforcement plan is implemented, then the area defining the intersection will affect the placement of the detector loops and the law regarding the yellow phase may play into the decision of a grace period. If a public information and education campaign is conducted, then the public should be educated regarding the use of the yellow phase.

RED-LIGHT RUNNING VIOLATION FREQUENCY

Red-light running can be considered a "big" problem with respect to the number of violations that occur. A two-hour traditional enforcement effort at a highvolume intersection in Raleigh, NC resulted in 36 tickets, which is a rate of 18 violations per hour or an average of one violation about every 3.5 minutes (min.) (5). A study conducted over several months at a busy intersection (30,000 vehicles per day) in Arlington, VA revealed violation rates of one red-light runner every 12 min. and during the morning peak hour, a higher rate of one violation every 5 min. A lower volume intersection (14,000 vehicles per day), also in Arlington, had an average of 1.3 violations per hour and 3.4 in the evening peak hour (6).

The Center for Transportation Research and Education (CTRE) at Iowa State University completed a study in December 2000 that "examined the frequency and effects of red-light running at intersections within selected communities and estimated the overall scope of this practice in the State of Iowa" (7). Table 2-1 summarizes the occurrence of violations that were presented in the final report for the study. Violations were counted by videotaping the intersection and then analyzing the tape. As seen in the table, the violation rate for the various intersections covers a wide range of values ranging from 0.45 violations per 1,000 entering vehicles to 6.08 (discounting the outlier value of 38.50 for Intersection 1 in Dubuque).

Table 2-1
Summary of Violation Data for Iowa Cities

SAFETY IMPACTS OF RED-LIGHT RUNNING

Safety is the biggest concern associated with red-light running. Safety is often measured by the number and severity of the crashes occurring. Numerous statistics have been published quantifying the problem in a specific city, state, or across the country. Some of the facts regarding the safety impacts of red-light running are presented below.

• Highway Safety Information System (HSIS) (a multi-state crash database maintained by FHWA) data from four states show that red-light running crashes account for 16 to 20 percent of the total crashes at urban signalized intersections (8).
  
  
• In a study of police-reported crashes for four urban areas, "ran traffic control" was the single most common type of crash accounting for 22 percent of urban crashes and 27 percent of all injury crashes (9). Of these crashes, 24 percent are the result of red-light running, meaning that about 5 percent of urban crashes are the result of red-light runners (10).
  
  
• Overall, 56 percent of Americans admit to running red lights. Yet 96 percent of drivers fear a red-light runner will hit them when they enter an intersection (11).
  
  
• One in three people claim they personally know someone who has been injured or killed in a redlight running crash-similar to the percentage of people who know someone who was killed or injured by a drunk driver (11).
  
  
• According to a survey conducted for FHWA, approximately 21 percent of respondents said they felt that drunken driving incidents are decreasing, but only 6 percent felt that red-light running incidents were decreasing (11).

These statistics show that red-light running has specifically impacted the safety of signalized intersections and is considered a very dangerous act by the motoring public.

National/Multi-State Data

Retting et al. (9) used two national databases to quantify the occurrence of red-light running crashes, as well as to summarize the characteristics of red-light runners. The databases include the Fatal Analysis Reporting System (FARS), which includes virtually all U.S. police-reported crashes involving a fatality, and the General Estimates System (GES), which is based on a nationally representative probability sample of crashes with a varying degree of injury and property damage. As reported between 1992 and 1996, FARS data indicated that 3,753 crashes could be attributed to redlight running, resulting in 4,238 fatalities. Also, 97 percent of the 3,753 fatal crashes from 1992 through 1996 involved two or more vehicles. The remaining 3 percent involved pedestrians or bicycles. The FARS statistics quoted in Retting's report were updated to show trends in red-light running crashes after 1996. A definition of a red-light running crash that best duplicates Retting's results isolates such crashes as those where a vehicle was proceeding straight through the intersection, a driver factor as failure to obey traffic control and at a signalized non-interchange intersection. Retting's research also used GES to look at the impact of red-light running crashes beyond fatal crashes. The results from the GES system indicate a total of 257,849 red-light running crashes during 1996, which is approximately 4 percent of the estimated total number of police-reported crashes. Additional conclusions made by Retting regarding these red-light running crashes are summarized below (9):

• Red-light running crashes were more likely than other crashes to produce some degree of injury (47 percent versus 33 percent);
• Red-light running crashes were more likely to occur on urban roads than other fatal crashes; and
• Red-light running crashes were somewhat more likely to occur during the day.

For the purposes of this report, the statistics using GES were updated using the database to reproduce the injury distribution for 1999's crash data. A red-light running crash was defined as one that took place at an intersection (either at an interchange or non-interchange location) controlled by a traffic signal and involved either (1) a driver who was charged with the violation "running a traffic signal or stop sign" or (2) the accident type was either a "changing traffic-way, turn into path, turn into opposite direction crash" or an "intersecting path, straight path" crash. Using the GES weights, there were an estimated 252,506 red-light running crashes in 1999. The severity distribution for these crashes is shown in Figure 2-1.

The data obtained from both the FARS and GES databases highlight the danger of red-light running. Although the number of fatal red-light running crashes seemed to peak in 1996 and has since decreased slowly, red-light running crashes still account for almost 40 percent of fatal signalized intersection crashes.

Although the magnitude of the statistics presented here is useful in portraying the size and injury severity of red-light running crashes, caution should be taken with regard to the specific numbers reported. It is difficult to create a definition for a red-light running crash, with the available database variables, that catches all true redlight running crashes and does not catch other crash types.

State/Local Data

The impacts of red-light running crashes are also reported on a state level and in smaller jurisdictions. For example, in December 2000, CTRE at Iowa State University completed a study of crashes and associated costs resulting from red-light running (7). Table 2-2 shows the crash frequencies by severity type and costs for a 3-year period in seven cities in Iowa, as well as for the entire state.

In 1999, Oregon legislation approved a bill that allowed six cities to conduct a 2-year demonstration project of photo red-light enforcement. The City of Beaverton was the first Oregon city to enact the program, with cameras activated in January 2001. This action was prompted, in part, by the fact that in the 3-year period of 1997 to 1999, crashes caused by red-light running had increased by 20 percent over the prior 3-year period of 1994 to 1996, and injury crashes related to red-light running increased by 82 percent for the same periods (12).

Table 2-2
Summary of Costs Resulting from Ran Traffic-Signal Crashes in Iowa (1996-1998)

CRASH TYPES RELATED TO RED-LIGHT RUNNING

While most red-light running crashes involve at least two vehicles, crashes involving a single vehicle and an alternative transportation mode (pedestrian or bicyclist) can occur. A single vehicle, hit fixed object crash could occur when either the running-the-red violator or the opposing legal driver takes evasive action to avoid the other and crashes into an object, e.g. a signal pole. Also, a running-the-red violator can hit a pedestrian or bicyclist who is legally in the intersection. The two most prominent crash types involving multiple vehicles are the angle- and turning-crash types. The angle crash is typically the offending motorist hitting or being hit by a vehicle legally in the intersection from the adjacent approach. A i can occur when a left-turning vehicle collides with an on-coming vehicle from the opposite direction; either vehicle may be the red-light violator.

Past research studies that have evaluated the effect of cameras or other programs on red-light running focused on three crash types. This includes the two mentioned above (angle and turning) and also rear-end collisions. Rear-end collisions are not the result of red-light running but rather the result of vehicles stopping for a signal at an intersection while others behind them do not. Some studies have noted a decrease in angle and turning crashes but an increase in rear-end crashes as a result of concentrated enforcement of red-light running. Figure 2-2 displays the three crash types.

In developing a red-light camera effectiveness study, Persaud and Council (13) noted that the following crash types could be possible target crashes for a red-light study:

1. Right-angle (side impact) crashes;
2. Left turn (two vehicles turning);
3. Left turn (one vehicle oncoming);
4. Rear end (straight ahead);
5. Rear end (while turning); and
6. Other crashes, specifically identified as red-light running.

However, with these crash types, there are numerous situations where a crash could have occurred that was unrelated to red-light running.

When using an accident database to highlight red-light running crashes, defining such a crash using database variables requires particular and detailed thought. However, a jurisdiction that is looking at a specific intersection and the problem of red-light running should feel comfortable in investigating the angle, turn and rear-end crashes to monitor red-light running problems.

DRIVER CHARACTERISTICS

As mentioned in the Introduction, red-light running is a complex problem. Causal factors range from driver- to intersection-related and there is also an element of driver psychology and sociology involved in the action of violating a signal. The likelihood of committing a violation varies from day to day, intersection to intersection and from person to person. A few studies have been conducted to identify driver characteristics of red-light runners. These studies have used a variety of methods including focus groups, field data collection and observation and crash databases. These studies provide valuable information to address red-light running.

Red-Light Violators

The Department of Public Health (DPH) in San Francisco, CA has been very involved in the city's redlight running program. The agency has developed a "Stop Red-Light Running" campaign that highlights the issue of red-light running to both the public and the media through bumper stickers, billboards and press conferences. The DPH also has conducted focus groups to better understand the psychology of red-light runners and hence, target campaign messages appropriately (14).

In 1998, the DPH conducted focus groups that divided red-light runners into two groups, aggressive drivers and distracted drivers. The information used in this study identified the average red-light runner in San Francisco as a male older than 40 years of age. This information was used to better focus public education efforts (14).

An additional set of focus groups was held in June 2001, with plans to use the results in another media campaign. Three different groups were developed:

• Group One-Violators who live outside of San Francisco but regularly drive on San Francisco's streets and have engaged in at least four traffic infractions and have at least two tickets in the last year.
• Group Two-Violators who live in San Francisco and regularly drive on San Francisco's streets and have engaged in at least four traffic infractions and have at least two tickets in the last year.
• Group Three-Non-violators and pedestrians/bicyclists who live in San Francisco and have engaged in traffic infractions no more than two times and have not received any tickets in the last year.

Infractions, as defined in this study, include running a red light, speeding, running a stop sign and running through a pedestrian crosswalk without stopping when someone was present (15). The findings indicate both differences and similarities among the three groups. For instance, among those participants who lived in San Francisco, there was a difference regarding driver courtesy between the violators and non-violators. Violators did not want to be "taken advantage of" while driving as opposed to the non-violators who had a more courteous attitude. On the other hand, groups felt similar regarding red-light running. Participants spoke of running red lights because they felt the person behind them was going to run it and they noted they were in a hurry and would do anything (including running red lights) to get to their destination more quickly.

Another study, involving field data collection, was conducted to profile red-light violators (16). The study compared characteristics of drivers that run red lights with a group of drivers that had the opportunity to run a red light, but did not. Field observation, cameras and driver records were used to record characteristics of 462 violators and 911 compliers at one particular intersection in Arlington County, VA. Analysis of the field data and a comparison between violators and compliers indicate the following:

• Violators were less likely to drive vehicles manufactured after 1991 and drive to large cars;
• Violators were less likely than compliers to wear safety belts;
• Violators were generally younger, however, there was no difference in gender distribution between the compliers and violators; and
• Violators had significantly more tickets for moving violations and were three times more likely to have multiple speeding convictions.

During the summer of 1999, the Social Science Research Center (SSRC) of Old Dominion University conducted a telephone interview to learn more about red-light runners and driving behavior. The survey was sponsored by DaimlerChrysler Corporation, the American Trauma Society and FHWA in order to gain data for the "Stop Red-Light Running" Program. The survey focused on what drivers reported to be their redlight running behaviors as opposed to their beliefs. The researchers acknowledge that self-reported data is only a proxy for actual driver behaviors as respondents seek to present themselves as best possible and may stretch the truth. However, the survey results reveal interesting trends (17).

Overall, 55.8 percent of the respondents reported running red lights. General characteristics of red-light runners include:

• Younger drivers;
• Persons without children;
• Driving alone (the presence of passengers significantly decreased the likelihood of running a red light, especially child passengers);
• Employed in jobs requiring less education or unemployed;
• Rushing to work or school in the morning weekday hours;
• Driving more than two miles from home; and
• More likely to have been ticketed for red-light running.

In the same survey, drivers were also asked of their response to the following scenario:

You are late for work, school, or an appointment and have been stopped by several red lights in a row. You are approaching another intersection that has had a yellow light for several seconds, but you know it is about to turn red. Which of the following would you likely do?
A.Slow down and prepare to stop at the red light.
B.Speed up to beat the red light.

Seventy-one percent of the respondents said they would slow down and stop while 29 percent indicated they would speed up. Of those that would speed up, they were asked why. The majority of the responses (69 percent) were due to being in a rush and to save time. Only 12 percent reported being frustrated. Additionally, the main source of this frustration is discourteous drivers and not congestion. This was highlighted as a major finding to the survey "given the general assumptions among safety experts that congestion is a leading and perhaps most important factor in predicting risky driving actions such as red-light running or aggressive driving."

The final survey questions dealt with the problem and danger of red-light running. About 80 percent of respondents believe red-light running is a problem and 99 percent believe it is dangerous. When asked: "Out of every 10 red-light runners, how many do so intentionally?" the mean response was more than five. However, respondents believe that less than two out of 10 would be stopped or ticketed by police. The report summarizes that "drivers believe red-light running was often a choice with few legal consequences."

Bonneson (39) reports that in examining 10,018 signal cycles in 6 hours, 586 vehicles entered the intersection after the indication turned red. Of these, 84 were heavy vehicles. Overall, 0.86 percent of all heavy vehicles violated the red indication as compared to 0.38 percent of passenger vehicles running the red. Bonneson concludes that heavy vehicle drivers are twice as likely to run red lights as are passenger drivers.

Red-Light Related Crashes

Retting et al. used crash databases to investigate the occurrence of red-light running crashes nationwide and to investigate the characteristics of red-light runners (9). In order to compare driver characteristics, a subset of the red-light crashes were selected. These crashes were those involving two vehicles (as opposed to involvement of a pedestrian or bicyclist), both of which were proceeding straight through the intersection and only one was charged with red-light running. The characteristics of the violators and non-violators in the same crashes were then compared. Some results of the comparison are highlighted below.

From FARS:

• Red-light violators were more likely than nonviolators to be younger than age 30 and slightly more likely to be male;
• The driver who violated the red light was more likely to be fatally injured than the non-violator (40 percent vs. 34 percent) in the same crash;
• Police were far more likely to report alcohol consumption for the red-light violators (34 percent vs. 4 percent) than the non-violators;
• In non-alcohol crashes, red-light running drivers and their passengers were more likely to be killed (58 percent vs. 40 percent). However, when the red-light runner was affected by alcohol, the non red- light running driver and passengers were more likely to be fatally injured (55 percent vs. 41 percent); and
• Red-light runners were more likely to have been driving with a suspended, revoked, or invalid driver's license and were also more likely to have prior driving while intoxicated convictions and two or more moving violations than the non-violators.

From GES:

• Red-light runners were more likely to be younger than age 30 and slightly more likely to be male than non-violators; and
• Red-light drivers were slightly more likely to have been drinking alcohol than the non-violators (5 percent vs. 1 percent). The difference was more dramatic for nighttime crashes-12 percent and 1 percent for red-light runners and non-red-light runners, respectively.

INTERSECTION CHARACTERISTICS

Bonneson et al. (18, 39) reviewed many past studies regarding various intersection characteristics as they relate to red-light running. Three intersection characteristics were highlighted as exposure factors including flow rate, number of signal cycles and phase termination by max-out. Field studies support the logical conclusion that as more vehicles are exposed to the potential of red-light running, the violation rate increases. The findings from that report are summarized below.

• Flow rate or volume: Every vehicle approaching the intersection at the onset of the yellow is exposed to the potential of red-light running. A decision must be made to stop or proceed through the intersection. As the number of approaching vehicles increases, the number of red-light runners will likely increase.
• Number of signal cycles: The more times the yellow phase is displayed, the more potential for red-light running. Hence, researchers should report that violation rates normalized by the number of signal cycles.
• Phase termination by max-out: Actuated signal systems operate using green extension time as long as the approach is occupied. However, the green may reach its maximum limit and "max-out" forcing the green phase to end regardless of whether the approach is occupied. Conversely, the signal may "gap-out" because the approach has been unoccupied for a set period of time. There is greater potential for red-light running as the frequency of max-out increases.

Bonneson cites other intersection characteristics and driver behaviors that are considered contributory factors. These include the following:

• Actuated control and coordination: This factor has to do with driver expectancy. In actuated control systems, vehicles often travel in platoons through several interconnected signals. Drivers expect the signal to remain green until they pass through the intersection. As a result, drivers expect the yellow to be long enough for them to make it through the intersection so they can stay with the platoon. This may result in violations.
• Approach grade: Drivers on downgrades are less likely to stop (at a given travel time from the stop line) than drivers on level or upgrade approaches.
• Yellow-interval duration: Long yellow intervals can violate driver expectancy, as drivers that stop are not "rewarded" with the red signal. In contrast, yellow intervals shorter than ITE-suggested values (19) have caught drivers off guard and resulted in a high number of red-light violations. (This was not discussed in the Bonneson report.)
• Headway: Drivers that follow closely (headway of less than 2 sec.) are more likely to run a red light. When a driver leaves a small headway, the following car is "drawn" into the intersection as more attention is given to the leading vehicle as opposed to the environment and traffic signal.

A similar study investigating intersection characteristics was conducted using accident and intersection data for California, available in HSIS (8). The study investigated select intersection characteristics and the relationship to red-light running crashes by developing mathematical models. The main intersection variables of interest include the number of cross-street lanes (surrogate for intersection width), average daily traffic (ADT) and traffic-control type.

Separate models were developed for the "mainline as entering street" and the "cross-street as entering street." Therefore, a crash where the violating vehicle entered the intersection from the lower-volume road was modeled using "cross-street as entering street" and the mainline road is considered the crossing street. The analysis used a total of 4,709 two-vehicle red-light running crashes for a 4-year period. The findings are summarized in the points below and in Table 2-3.

• Effect of cross-street lanes: For the "cross-street as entering street" model, there is a 7 percent increase in red-light running crashes for each one-lane increase on the mainline when controlling for signal operation type, opposite street ADT and leftturn channelization. The results are different for the "mainline as entering street" where the number of cross-street lanes had no effect on the number of red-light running crashes.
• Effect of ADT: The number of crashes could be affected by both an increase of traffic on the entering street (more possibility of red-light runners) and an increase of traffic on the crossing street (the greater the possibility of hitting another vehicle). For the "mainline as entering street," an increase in entering street ADT as well as an increase in crossing street ADT both resulted in an increase in red-light running crashes. For the "crossstreet as entering street," there was an increase in crashes with an increase in the entering street volume, but there was no increase in crashes with an increase in the crossing-street (or mainline) volume.
• Effect of traffic control: For both models, fully actuated signals tend to have more crashes per approaching street than approaches with semiactuated or pre-timed signals. The models indicate a 35 to 39 percent greater number of red-light running crashes at fully actuated signals as compared to pre-timed signals.

Table 2-3
Summary of Intersection Characteristics on Likelihood of Red-Light Running Crash

Those variables shown to increase the likelihood of a red-light running crash are not negotiable design features. For example, in order to reduce the likelihood of a red-light running crash, one cannot reduce the number of lanes on the cross-street from four lanes to two lanes, if the volume requires four lanes. The study results are helpful, however, in identifying intersections that may have a high number of red-light running crashes and where careful engineering evaluations and enforcement may be necessary.

DRIVERS' STOP-GO DECISION

Bonneson also discussed the factors that affect the driver's decision to stop or proceed through the intersection upon seeing the onset of the yellow (18). Based on earlier work by Van der Horst (20), there are three main components of the decision process: driver behavior (expectancy and knowledge of operation of the intersection), estimated consequences of not stopping and estimated consequences of stopping. Table 2-4 summarizes these factors.

Table 2-4
Factors Affecting the Stop-Go Decision

What if the driver makes his decision to proceed through the intersection based on the factors above, but ends up running the red light? Bonneson divides redlight runners into two categories. The first is the intentional violator who, based on his/her judgment, knows they will violate the signal, yet he/she proceeds through the intersection. This type of driver is often frustrated due to long signal delays and perceives little risk by proceeding through the intersection. The second type of driver is the unintentional driver who is incapable of stopping or who has been inattentive while approaching the intersection. This may occur as a result of poor judgment by the driver or a deficiency in the design of the intersection. Bonneson further indicates that intentional red-light runners are most affected by enforcement countermeasures while unintentional redlight runners are most affected by engineering countermeasures.

Milazzo et al. (10) constructs similar distinctions in redlight running driver types. Table 2-5 describes four different driver types. As pointed out in the report, "note that any driver can assume any of these roles depending on the situation, the driver's current mindset and chance itself."

Such distinctions in driver types highlight the need for different types of countermeasures. It is difficult to determine the percentage of crashes as a result of a specific red-light running driver type since each driver is capable of acting like any of the driver types above depending on the current situation. However, engineering, enforcement and education countermeasures are plausible solutions to address the differences.

Table 2-5
Driver Population Types

CAUSAL FACTORS AND POTENTIAL COUNTERMEASURES

As part of a FHWAstudy (21) on the feasibility of using advanced technologies to prevent crashes at intersections, the researchers reviewed the police reports of 306 crashes that occurred at 31 signalized intersections located in three states. Traffic-signal violation was established as a contributing factor and the reason for the violation was provided in 139 of the crashes. The distribution of the reported predominant causes is as follows:

• 40 percent did not see the signal or its indication;
• 25 percent tried to beat the yellow-signal indication;
• 12 percent mistook the signal indication and reported they had a green-signal indication;
• 8 percent intentionally violated the signal;
• 6 percent were unable to bring their vehicle to a stop in time due to vehicle defects or environmental conditions;
• 4 percent followed another vehicle into the intersection and did not look at the signal indication;
• 3 percent were confused by another signal at the intersection or at a closely spaced intersection; and
• 2 percent were varied in their cause.

Care should be taken in interpreting this information because it is self-reported, cannot be independently verified and is based on a small sample. If these causes are statements from the driver, it is safe to say that there will be few who will not admit that they "intentionally violated the signal." Nonetheless, from these examples, countermeasures can be identified that would address one or more of the causes.

Countermeasures can be engineering, education, or enforcement actions. Different types of measures may be more appropriate to address the variety of causes listed above. Table 2-6 correlates the causes discussed above to the appropriate category of countermeasure. A check mark () signifies that the countermeasure type is likely to address the cause, while a bullet mark (.) signifies that the countermeasure type could possibly address the cause.

Table 2-6
Possible Causes and Appropriate Countermeasures for Red-Light Running

Possible Causes of Red-Light Running Engineering Enforcement Education 1. Did not see signal     2. Tried to beat yellow 3. Reported they had geen     4. Intentional violation 5. Unable to stop vehicle     Followed another vehicle   Confused by signal     : Likely countermeasure : Possible countermeasure

SUMMARY

Based on what we know of the extent and nature of the problem of red-light running, the three "E"s (engineering, enforcement and education) must be considered as a part of an effective solution. The three "E"s are sometimes considered separately; however, an effective program uses all types of solutions to target the problem. Each "E" addresses different deficiencies contributing to red-light running, whether that of the driver, vehicle, or intersection. However, as seen from Table 2-6, engineering countermeasures would appear to be those that would address the majority of causes of red-light running. The engineering countermeasures that aim to reduce red-light running or mitigate its consequences are the focus of this document and are discussed in detail in the next chapter.