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
Back to Publication List        
Publication Number:  FHWA-HRT-13-048    Date:  October 2013
Publication Number: FHWA-HRT-13-048
Date: October 2013

 

Driver Expectations When Navigating Complex Interchanges

Chapter 2. Task 2-Literature Review

The objective of task 2 was to conduct a literature review of prior work on driver navigation problems and driver expectations at interchanges and intersections and identify the key issues. This chapter provides a summary of the findings of this literature review. Task 2 was conducted to identify relevant methods for investigating driver expectations for future project tasks. It was also used to gain an understanding of the background literature on driver expectations.

This chapter contains three technical sections:

Task 2 Methods

This section describes the methods used to conduct the literature search and summarizes the resulting source information.

Literature Searches

The literature search began with defining keywords to find sources containing information regarding driver expectations and behavior at interchanges. As a part of the Human Factors Guidelines project, a number of terms were found that appear to be the most efficient at distinguishing research related to human factors and safety.(7) The four terms identified include human factors, driver behavior, driver behaviour, and driver performance. These terms were used in the initial literature search to narrow the results to those relevant to the driver. Note that much of the existing human factors research related to interchanges involves signage. Given that this topic was the focus of task 6-and comprehensively reviewed as part of that project-the researchers limited their inclusion of signage research to those studies that directly addressed driver expectations.

All of the searches were performed in the Transportation Research Information Services database. An initial search of the keywords "driver expectations" and "interchange" did not produce any relevant documents. As a result, the search terms were broadened, and a search was conducted using general terms to look for driver-relevant information at interchanges. Following this search, another search was performed to narrow the results to driver expectations and driver information needs at interchanges. Table 1 shows the search keywords that were used and the number of results that were obtained.

Table 1. Keywords used and number of records returned for literature searches.

Search Number

Keywords

Number of Records

0

"driver expectations" AND "interchange"

0

1

"interchange" AND ("human factors" OR "driver behavior" OR "driver behaviour" OR "human behavior" OR "human behavior" OR "driver performance" OR "human performance")

241

2

"interchange" AND driver AND ("confusion" OR "expectations" OR "expectancies" OR "information")

93 (50 records not included in search #1)

After conducting the searches, the researchers broadened the scope and reviewed some sources regarding driver expectations without focusing on interchanges. An initial search was performed using the PsycInfo database; however, no relevant results were found. General references containing information about driver expectations were found through references provided by existing sources and general Web searches.

Upon obtaining the search results, the titles and abstracts were thoroughly inspected, and articles that were clearly not related to the topic were removed from the lists. The remaining relevant documents were ordered or downloaded from the Internet for closer review.

Finally, while conducting the literature review, any secondary research sources cited in reviewed documents that seemed relevant to driver expectations were reviewed as well.

Structured Literature Review Forms

The general approach used to summarize the literature was based on literature syntheses that were conducted in past projects that involved using structured review forms to identify key information from research sources. (See references 8-11.) In particular, these structured review forms typically contain fields to capture information about dependent and independent variables, useful methodological approaches, and summaries of key findings and conclusions. A concise, structured review form was developed and is shown in figure 2.

This figure shows a structured review form, which has several labeled boxes to fill in information. The labels include document, method/study type, configuration, driver expectation elements, general approach, methods, independent variables, dependent variables, study findings and conclusions, other information, and caveats/comments.
Figure 2. Illustration. Structured review form.

The top two rows of the form include space to fill in general descriptive information about the source such as the full document reference and the study type. The third row contains two fields that can be used to classify the reviews. "Configuration" refers to the primary geometry that was examined in the study, while "Driver Expectation Elements" refers to any elements of the roadway environment that have been shown to potentially violate driver expectations (e.g., ramp spacing, guide signing/markings, route continuity, etc.). This field also provides an indication of the applicable engineering elements of the research (i.e., signs, markings, or geometry). The fourth line labeled "General Approach" provides a concise overview of the purpose of the study, while the fifth line labeled "Methods" describes the study methods in appropriate detail. Since one of the subgoals of this task was to examine the available methods for gathering expectation data from drivers, some of the reviews focused on the methodology used without providing findings or conclusions. In these cases, the methods are contained in the "Study Findings and Conclusions" section of the form. Another sub-goal of this task was to formulate and quantify the independent and dependent variables that can provide useful information about driver expectations. Therefore, the independent and dependent variables each have a corresponding section on the form. The remainder of the form is devoted to the findings, conclusions, and other information. At the bottom, there is a place for describing any comments related to the applicability of the particular source or caveats regarding the research method.

Task 2 Results

This section presents a summary of the results that were found in the sources examined during the literature review. These results are presented in four sections. The first section includes definitions of driver expectations from the research sources. The second section includes a summary of how empirical findings relate to specific interchange elements. The third section lists the independent and dependent variables used to measure expectations in the sources reviewed during the literature search. Finally, the fourth section provides a summary of how findings related to driver expectations have been incorporated in existing design guidance.

Definition of Driver Expectations

One of the objectives of the literature review was to try to provide a clear definition of the term "driver expectations," and to describe the implication that these expectations have for driver behavior, performance, and decisionmaking.

Based on the literature review, it was clear that the concept of driver expectations has not been thoroughly defined with specific regard to interchanges and common driver maneuvers at interchanges. However, this concept has received consideration with regard to broader driving tasks, and there are several existing definitions of driver expectations.(12,3,13) The majority of the definitions typically include some variation of the comprehensive definition and are provided by Lunenfeld and Alexander.(3) They define expectancy as "a driver's readiness to respond to situations, events, and information in predictable and successful ways."(pg. 153)(3) Attributes of driver expectations are as follows:(3)

Other researchers have added other useful aspects to this definition. These include elements such as the formation of expectations and the impact of expectations at different levels of the driving task. Expectations are formed from the following two processes:(14,15)

The following list includes various forms of expectations that are required while driving:(14)

For the most part, the items apply reasonably well to interchange driving situations. Together, these elements provide a broad view of how driver behavior may be affected by driver expectations at interchanges. This formulation of driver expectations was used in this project.

Driver Expectations Related to Roadway Elements

This section summarizes empirical findings about driver expectations related to specific interchange elements. The findings are organized by type of performance measure rather than by interchange element because of certain limitations regarding the interpretation of results.

In general, few on-road/simulator study results were found that directly address driver expectations regarding a specific roadway element. Consequently, researchers can only indirectly make inferences about driver expectations based on driver behavior, such as the number of missed exits or decision sight distances. In particular, differences in driver performance associated with these factors can indicate that the driving situation has matched or contradicted the driver's expectations of the situation. For example, a sign that systematically causes drivers to make required lane changes at the last possible second might contribute to drivers having incorrect expectations about which lane they should occupy. However, the same performance differences can also reflect problems with sign comprehension, reading time, or personal preference. The same late lane changes could also occur if it takes drivers longer to read and comprehend the sign, which likely has little to do with incorrect driver expectations.

Fisher et al. acknowledged this problem with interpreting performance results in terms of driver expectations. Specifically, they noticed that the driver's motives behind lane changes are not well understood.(16) For example, drivers may change lanes toward the mainline to get out of the way of exiting drivers, or they may prefer not to use an option lane to ensure that they do not miss their exit. In fact, the authors found that drivers change lanes as a factor of conservative behavior rather than a misunderstanding of the meaning of the sign.(16) They suggested that until drivers gain experience with a particular configuration, the effect of modifying the signage would be difficult to evaluate.

Keeping this caveat in mind, there is still information about driver expectations that can be implied from existing research. One study tested signage for two-lane exits with option lanes in which drivers were asked to take the lane necessary to reach their destination.(17) They found that the proportion of unnecessary lane changes (generally lane changes out of the option lane) was significantly higher when the destination was an exit rather than the mainline. This suggests that drivers expect to need to be in the lane closest to the exiting direction to be able to exit. However, as previously stated, drivers may have other unknown motives for their lane changes.

Some research sources used driver confidence as an indirect measure of expectation.(17,2,18) The basic logic is that drivers should be more confident in a decision if it matches their expectations. In the signage study performed by Chrysler et al., participants were asked which lane(s) would take them to their destination and how confident they were in their decisions.(2) In general, the signs that caused the highest percentage of correct lane choices also were assigned the highest confidence scores by participants. Some of the signs that were tested, however, received high confidence scores for one of the traffic movements (e.g., the through traffic condition) but not the alternative traffic movement (e.g., the exiting traffic condition), which suggests that drivers form expectations for individual traffic movements rather than for all information presented on the sign as a whole. In another study of central business district signage performed by McNees, the message that drivers expected to see was often the same as the message that drivers preferred to see.(19)

A common theme throughout the reviewed sources was that violated expectations cause longer response times and more incorrect responses. In one study examining lane drop markings, the installation of markings led to a significant decrease in lane changes, a decrease in erratic maneuvers, and earlier lane changes overall, which could be interpreted as the sign improving the drivers "readiness" to respond and make decisions.(20) In another study of lane drop signage, it was found that 58 percent of participants expected a lane drop to occur even when no lane drop panel was present.(18) This result could be due to the fact that lane drops were the focus of the study, or it could indicate a true trend in terms of driver expectations at freeway exits.

Driver expectations also impact driver information acquisition. Borowsky, Shinar, and Parmet studied the eye gaze of drivers while they viewed "No Right Turn" and "No Left Turn" signs in expected and unexpected locations.(21) They found that experienced drivers identified the signs more accurately than inexperienced drivers when the signs were located in the expected location. However, when the sign location did not match the expectations of the experienced drivers, their identification accuracy was worse than the inexperienced drivers. Experienced drivers seem to rely on their expectations and may be less likely to notice information in unexpected locations. "It was concluded that drivers should be warned when signs are placed in unexpected locations.

The empirical finding regarding driver expectations are summarized as follows:

As discussed at the beginning of this section, another important finding that must be acknowledged is that driver expectations are sometimes difficult to disentangle from driver motivations, performance factors (e.g., reading distance), preferences, and familiarity, among other factors. This suggests that the types of approaches may be less suitable for the current project than approaches that ask drivers directly about their expectations about interchange driving.

Variables

This section provides lists of all of the independent and dependent variables that were used to measure driver expectations in the sources reviewed during the literature search. The variables generally relate to signage, markings, or geometry. These variables were considered in project activities related to identifying candidate empirical methods for task 7 data collection.

Independent Variables

Table 2 lists the dependent variables that were used in the source documents. These variables are categorized by variable type (i.e., sign-related, markings-related, or geometry-related variable).

Overall, these variables mainly relate to the specifics of the driving environment as presented to the participants. Variables of this nature include the type of stimulus presented (e.g., signs and markings), qualities of that stimulus, and the geometry of the scenario. Other miscellaneous variables include those related to the participants (e.g., experience level) and the scenario in general (e.g., traffic volumes).

Table 2. Independent variables found during the literature search categorized by variable type.

Variable

Reference

Variables Related to Signs

Destination (mainline or exit)

16 and 17

Driver experience level

21

Original lane position

16 and 17

Sign location (right or left side of the road)

21

Sign location (with reference to the interchange)

19

Sign message (words, symbols, plaques, etc.)

18, 19, and 21

Sign position (over left or right lane)

18

Sign type (conventional, diagrammatic, etc.)

2, 17, 18 and 22-25

Variables Related to Markings

Type of pavement markings

19, 23, and 25

Variables Related to Geometry

Exit direction

24

Interchange geometry, including exit geometries

2 and 26

Interchange spacing

27

Number of exiting lanes

24

Overall scenario picture

15

Presence of option lanes

24

Side of lane reduction (left or right)

25

Traffic volumes

26 and 27

Dependent Variables

Table 3 lists the dependent variables that were used in the source documents. These variables are categorized by the main focus of the study (i.e., evaluate signage, markings, or geometric aspects).

Overall, the variables relate to three types of measurements used to record driver behavior and expectations. The three types of measurements include the following:

Table 3. Dependent variables found during the literature search categorized by study type.

Variable

Reference

Sign Evaluation Studies

Choice/response correctness

2, 18, 23, and 24

Decision sight distance

24

Lane change distance (first and final)

17 and 22

Number of fixations on the target

21

Number of missed exits

16 and 17

Number of switchbacks (lane change and then return to the initial lane)

22

Number of unnecessary lane changes

16 and 17

Response time

18 and 19

Sign expectation (which sign drivers expect)

19

Sign identification accuracy

21

Sign preference (which sign drivers prefer)

19

Subjective certainty of choice correctness

2, 17, 18

Subjective difficulty of sign comprehension

17

Markings Evaluation Studies

Number and location of lane changes

20 and 25

Number, location, and type of erratic maneuvers

20

Traffic volumes

20

Geometric Evaluation Studies

Crash frequency

27

Groupings of "if, then" statements (card sort)*

15

Labels for card groups (card sort)*

15

Lane change location

28

Lane position

26

Number of lanes crossed during a lane change

28

Responses to expectation questions (what participants expect and why they expect that)

15

Speed

26

*"Card sort" refers to an experimental methodology where participants sort cards printed with statements into piles of similar statements and then name each pile.

Applications of Expectations Research

Several of the sources that provided information about expectations were based on literature reviews, analytical reviews, observational data, or expert judgment. The findings from these reports often involved design principles or guidance that was relevant to the overall objectives of the project, including developing recommendations for navigation signage to aid complex interchange design. Some of the design guidance identified applied broadly to roadway design, general interchange design, and specific roadway elements. The following list provides a compilation of the most relevant design information categorized under broad themes that were seen across sources.

The principles refer to driver expectations as well as the related categories of design consistency and interchange design:

1. Provide adequate forward sight distance.

2. Provide transition cues.

3.Minimize attention dividing conditions.

4. Provide navigation information to address all of the driver information needs.

5. Maintain compatibility between the interchange and the visual cues.

6. Design to accommodate the drivers' expectations and abilities.

7. Warn drivers of situations which may violate their expectations.

8. Allow drivers to recover after making an error.

9. Design for simplicity.

10. Design for consistency and predictability.

Task 2 Conclusions

Overall, the amount of research that was directly related to driver expectations was relatively limited. However, enough information was obtained to support the development of other tasks in this project by broadening the scope of research examined to include driver expectations and comprehension of roadway elements. The following sections discuss the information provided by the literature review related to priority research gaps, relevant criterion for choosing interchanges, and potential scenarios for future tasks.

Priority Research Gaps

From the examination of the source documents, it is clear that there are several issues that have not been fully resolved by the current research on driver expectations at interchanges. Two of the most important and apparent research gaps are described in the following subsections.

Isolation of Driver Expectations

From this literature review, it is evident that research results about driver expectations are intertwined with results on driver performance and behavior. This finding makes it difficult to draw direct conclusions about driver expectations unless the researchers explicitly asked the participants about their expectations. In general, this also suggests that several types of empirical approaches used in previous research were less suitable for this project than approaches that directly ask drivers about their expectations of interchange driving. In particular, for other project tasks, it was necessary to be clear with participants that the data being collected were directly related to their expectations about interchange driving.

Driver Expectations Related to Specific Roadway Elements

Another important research gap identified in the current research literature is that there is little research that specifically identifies driver expectations for specific interchange elements1. A tentative list of expectations was compiled in the results section of this chapter; however, those represented tentative findings had alternative explanations that did not necessarily involve driver expectations. What is needed is a more systematic approach to cataloging driver expectations at different points during interchange driving.

Useful Metrics for Selecting Interchanges

In addition to addressing topics already discussed in this chapter, the literature review also identified useful metrics for selecting the interchanges that were examined in other project tasks.

A key metric identified in some studies was interchange complexity. More specifically, in one study, ratings of complexity were assigned to a number of intersection scenarios to select those that would provide information about the complexity of interaction situations.(15) The six dimensions of complexity used in the study included number of interaction partners, number of types of interaction partners, number of intersection branches, the type of intersection (signalized or priority), the number of pieces of static information present, and which partner has the right of way. Some of these are applicable to interchanges. This topic was also examined in task 6, which provided a list of interchange features that appeared to be related to interchange complexity.(33)

Another available metric is the driver information load as presented by Lerner et al.(34) This metric focuses mainly on the information present in the driving environment in terms of signs, and it also incorporates the roadway demand. This metric could be used to determine the load for the driver at an interchange. One disadvantage is that this method does not have an all-inclusive way to account for elements that violate driver expectations.

A final metric is the use of the existing design principles and expectation checklist items that were found in the research. Candidate interchanges could be examined using these checklist items to determine how many principles are violated in a particular situation.

Potential Scenarios for Focus Groups and Task Analyses

The literature review identified a variety of geometries and specific interchange elements associated with driver expectations (referred to as "expectation elements") that are useful to consider when developing scenarios for examining driver expectations. These components represent issues related to challenging situations at interchanges and include the following:

Expectation Elements

Elements of interchange navigation where expectation-related issues may exist include the following:

This list represents a core set of scenario elements that were considered during the development of interchange driving scenarios for other project tasks.

Overall, the amount of information from existing research sources that was directly related to driver expectations was limited. However, by using relevant information from other research domains, it was still possible to find information about methods, variables, and other results that were useful for developing and conducting several of the remaining project tasks.


1The exception to this may be sign information, but this was covered in less detail in the current report since it is the focus of task 6.

 

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