This article was downloaded by: [Texas A & M International University]

On: 21 August 2015, At: 09:53

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place,
London, SW1P 1WG

Ergonomics
Publication details, including instructions for authors and subscription information:
[Link]

Using the Event Analysis of Systemic Teamwork

(EAST) to explore conflicts between different road
user groups when making right hand turns at urban
intersections
a b c d
Paul M. Salmon , Michael G. Lenne , Guy H. Walker , Neville A. Stanton & Ashleigh
e
Filtness
a
Faculty of Arts and Business, University of the Sunshine Coast Accident Research (USCAR),
Maroochydore, Australia
b
Accident Research Centre, Monash University, Victoria, Australia
c
Click for updates School of the Built Environment, Heriot-Watt University, Edinburgh, UK
d
Transportation Research Group, University of Southampton, Southampton, UK
e
CARRS-Q, Queensland University of Technology, Brisbane, Australia
Published online: 10 Sep 2014.

To cite this article: Paul M. Salmon, Michael G. Lenne, Guy H. Walker, Neville A. Stanton & Ashleigh Filtness (2014) Using the
Event Analysis of Systemic Teamwork (EAST) to explore conflicts between different road user groups when making right hand
turns at urban intersections, Ergonomics, 57:11, 1628-1642, DOI: 10.1080/00140139.2014.945491

To link to this article: [Link]

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
[Link]/page/terms-and-conditions
 Ergonomics, 2014
Vol. 57, No. 11, 1628–1642, [Link]

Using the Event Analysis of Systemic Teamwork (EAST) to explore conflicts between different
road user groups when making right hand turns at urban intersections
Paul M. Salmona*, Michael G. Lenneb, Guy H. Walkerc, Neville A. Stantond and Ashleigh Filtnesse
a
Faculty of Arts and Business, University of the Sunshine Coast Accident Research (USCAR), Maroochydore, Australia; bAccident
Research Centre, Monash University, Victoria, Australia; cSchool of the Built Environment, Heriot-Watt University, Edinburgh, UK;
d
Transportation Research Group, University of Southampton, Southampton, UK; eCARRS-Q, Queensland University of Technology,
Brisbane, Australia
(Received 5 August 2013; accepted 2 July 2014)

Collisions between different types of road users at intersections form a substantial component of the road toll. This paper
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

presents an analysis of driver, cyclist, motorcyclist and pedestrian behaviour at intersections that involved the application of
an integrated suite of ergonomics methods, the Event Analysis of Systemic Teamwork (EAST) framework, to on-road study
data. EAST was used to analyse behaviour at three intersections using data derived from an on-road study of driver, cyclist,
motorcyclist and pedestrian behaviour. The analysis shows the differences in behaviour and cognition across the different
road user groups and pinpoints instances where this may be creating conflicts between different road users. The role of
intersection design in creating these differences in behaviour and resulting conflicts is discussed. It is concluded that
currently intersections are not designed in a way that supports behaviour across the four forms of road user studied.
Interventions designed to improve intersection safety are discussed.

Practitioner Summary: Intersection safety currently represents a key road safety issue worldwide. This paper presents a
novel application of a framework of ergonomics methods for studying differences in road user behaviour at intersections.
The findings support development of interventions that consider all road users as opposed to one group in isolation.
Keywords: intersections; road safety; EAST; systems analysis; drivers; cyclists; motorcyclists; pedestrians

Introduction
Collisions at intersections represent a major road safety issue. In Australia, for example, the majority of urban crashes and a
substantial proportion of rural crashes occur at intersections (McLean et al. 2010). This problem is not limited to one
particular road user group, with most forms of road users having a high level of involvement in crashes at intersection. In the
Australian state of Victoria, for example, during 2012 more than half of all car, bicycle, motorcycle and pedestrian crashes
occurred at intersections (VicRoads 2013).
The prevalent approach to understanding and preventing collisions at intersections has tended to be road user centric
and component driven, focussing on specific road user types (e.g. drivers) and/or fixing a component of the problem (e.g.
making motorcycles more conspicuous; Gershon and Shinar 2013). The fundamental limitation of this approach is that the
unit of analysis is only one road user group, or one component of the problem (e.g. motorcyclist conspicuity). Recently,
researchers have made the case for the ‘systems’ approach when considering road user behaviour (Cornelissen et al. 2013;
Larsson, Dekker and Tingvall 2010; Salmon, McClure and Stanton 2012). Although the systems approach has been
discussed in the context of road traffic crashes (Wagenaar and Reason, 1990), very few road safety studies go beyond a
focus on one or two components. Previous studies of behaviour at intersections, for example, have largely focussed on
individual road users (e.g. drivers, cyclists, motorcyclists or pedestrians alone) or factors such as driving errors (Gstalter and
Fastenmeier 2010; Sandin 2009), pedestrian behaviours (King, Soole and Ghafourian 2009) and engineering (Highways
Agency 2012). The consequence is that other factors influencing behaviour are often not considered when developing
interventions, potentially diminishing their effectiveness (Salmon, McClure and Stanton 2012). For this reason, the road
user-centric, component-driven approach to road safety has been criticised for failing to take into account the inherent
complexity in transportation systems (Larsson, Dekker and Tingvall 2010; Salmon, McClure and Stanton 2012).
It is argued here that the long-standing problem of collisions between road users at intersections cannot be solved
through focussing on single road user groups and not considering other components within the intersection system. The
analysis presented in this paper moves towards systems thinking in road transport by considering the intersection ‘system’,
comprising different road users (e.g. drivers, riders, cyclists and pedestrians), vehicles and the road environment. Using data

*Corresponding author. Email: psalmon@[Link]

q 2014 Taylor & Francis

 Ergonomics 1629

derived from a recent on-road study, the Event Analysis of Systemic Teamwork framework (EAST; Stanton et al. 2013) was
used to evaluate behaviour at three major signalised intersections. Specifically, task, social and situation awareness
networks were used to identify the tasks performed by the different road user groups when negotiating the intersections, to
determine which parts of the intersection environment the road users interacted with and to examine differences in situation
awareness between the different road users when moving through the intersections. The aim was to identify differences in
behaviour across different forms of road user when negotiating the three intersections, to identify factors that may play a
role in conflicts between different road users at intersections, and to gather information to inform the development of new
intersection designs that are appropriate for all road users.

Heading EAST: the Event Analysis of Systemic Teamwork

EAST (Stanton et al. 2013; Walker et al. 2006, 2010) provides an integrated suite of ergonomics methods for analysing system
performance. To date, the approach has been applied for this purpose in a range of domains, including land and naval warfare
(Stanton 2014), aviation (Stewart et al. 2008), air traffic control (Walker et al. 2010), railway maintenance (Walker et al. 2006)
and the emergency services (Houghton et al. 2006). Underpinning the framework is the notion that system performance can be
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

meaningfully described via a ‘network of networks’ approach in which three interlinked network-based representations are
used to describe and analyse activity. Task networks are used to provide a summary of the goals and subsequent tasks being
performed within a system. Social networks are used to analyse the organisation of the system and the communications taking
place between ‘agents’ (both human, such as drivers, cyclists, and non-human, such as traffic lights, vehicles, pedestrian
crossing warnings). Situation awareness networks describe situation awareness in terms of concepts and the relationships
between concepts and show how situation awareness is different across different components of the system. The EAST
‘network of networks’ approach is represented in Figure 1.

Intersection case study

The aim of this study was to investigate the differences in behaviour of different road users when negotiating three
intersections, and to identify factors that may play a role in conflicts between different road users at intersections. The data
were derived from an on-road study which covered a pre-defined 15- km urban route in the south-eastern suburbs of
Melbourne, Australia. The analysis focussed on three major signalised intersections located along the route (see Figure 2).
Each intersection required a right turn against conflicting traffic in order to pass through the intersection and remain on the
study route. All three intersections were fully signalised with traffic lights and turning arrows.

TASK
NETWORK

Knowledge required during Command and operational

task performance structure required

KNOWLEDGE SOCIAL
NETWORK NETWORK
Distribution and
communication of
information underpinning
knowledge

Figure 1. Network of networks approach to analysing complex sociotechnical system behaviour. The figure shows example
representations of each network, including task analysis (task network), social network analysis (social network) and situation awareness
network (information network) representations (Stanton et al. 2013).
 1630 P.M. Salmon et al.
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

Figure 2. The three intersections studied (white arrow denotes route though intersection).

Methodology
Design
The analysis presented in this paper is based on data derived from an on-road study investigating the influence of road
design on the behaviour of different forms of road user. The study used an on-road study paradigm whereby participants
negotiated a pre-defined 15- km urban route (on separate occasions but with similar traffic conditions). Drivers drove an
instrumented test vehicle, whilst motorcyclists and cyclists completed the route using their own motorcycle or bicycle
 Ergonomics 1631

which was equipped with video and audio recording equipment. Pedestrians negotiated the three intersections on foot whilst
wearing video recording glasses and a microphone connected to Dictaphone. All participants provided concurrent verbal
(think-aloud) protocols as they negotiated the route.

Participants
A total of 78 participants (52 male, 26 female) aged 21– 64 years (mean ¼ 35.81, SD ¼ 13.03) took part in the study. The
sample comprised 20 car drivers, 18 motorcyclists, 20 cyclists and 20 pedestrians. An overview of the participants in each
group, including a summary of their previous exposure to the intersections involved in the study, is presented in Table 1.
Although a gender balance was achieved for the driver and pedestrian participant groups, this was not the case for
cyclists and motorcyclists (see study limitations in Discussion section). All participants were recruited through a weekly on-
line university newsletter and were compensated for their time. Prior to commencing the study, ethics approval was
formally granted by the Monash Human Ethics Committee. Participants were reimbursed $35 for their time.
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

Materials
A demographic questionnaire was completed using pen and paper. A desktop driving simulator was used to provide verbal
protocol training. A 15- km urban route, located in the south-eastern suburbs of Melbourne, was used for the on-road study
component. The analysis presented in this paper focusses on the three major intersections presented in Figure 2. Drivers
drove the route in a 2004 Holden Calais sedan equipped to collect various vehicle-, driving scene- and driver-related data. A
Dictaphone was used to record participant’s verbal protocols. Each motorcyclist participant’s motorcycle was fitted with an
Oregon Scientific ATC9K portable camera, which, depending on motorcycle model, was fixed either to the handlebars or
front headlight assembly. The ATC9K camera records the visual scene, speed and distance travelled. A microphone was
fitted inside each motorcyclist participant’s helmet and attached to the Dictaphone to record their verbal protocols. For
cyclist participant’s, the ATC9K portable camera was fitted to their helmet and they wore Imging HD video cycling glasses.
Pedestrians negotiated the intersections on foot whilst wearing Imging HD video sunglasses and a microphone linked to a
Dictaphone. Motorcyclist, cyclist and pedestrian participants also carried a Garmin GPS unit to record aspects such as
speed, distance travelled and route taken. All verbal protocols were transcribed using Microsoft Word. During data analysis,
the LeximancerTM content analysis software was used to construct situation awareness networks.

Procedure (data collection)

Individual participants negotiated the intersections on separate occasions. In order to control for traffic conditions, all trials
took place at the same pre-defined times on weekdays (10 am or 2 pm, Monday to Friday). Participants first completed an
informed consent form and demographic questionnaire and were then briefed on the research and its aims (expressed
generally as a study of driver behaviour). Following this, they were given training in providing verbal protocols which
included instruction from an experimenter followed by a desktop driving simulator task where they were asked to complete
the drive whilst providing a verbal protocol. The instructions gave an overview of the verbal protocol method and included
examples of what should and should not be verbalised during the study. The driving simulator task involved participants
driving along a road through a town whilst providing a concurrent verbal protocol. The scenario included other traffic,
pedestrians, traffic lights and intersections. The experimenter monitored the drive and provided feedback to the participant
regarding the quality of their verbal protocol. The participants were deemed to reach an acceptable level of proficiency once
they were able to accurately verbalise the major elements of the driving scenario and appropriate features of their own
cognitive processes (e.g. ‘I am checking the traffic lights’, ‘I can see that the pedestrian is waiting to cross up ahead’).

Table 1. Participant demographics.

Hours using respective Number of participants
Mean age mode of transport with previous experience
Road user group (SD); year Gender per week (hour) of each intersection
Drivers 34.9 (12.53) 10 males, 10 females 11.5 19 19 19
Cyclists 32.4 (10.42) 15 males, 5 females 6.9 16 14 17
Motorcyclists 45.5 (12.87) 17 males, 1 female 7.0 16 15 15
Pedestrians 30.5 (11.86) 10 males, 10 females 8.9 9 8 12
 1632 P.M. Salmon et al.

Participants were then shown the study route and were given time to memorise it. When comfortable with the verbal
protocol technique and route, participants were taken to their vehicle and asked to prepare themselves for the test. They
were then given a demonstration of the video and audio recording equipment, which was also set to record at this point.
Following this, the experimenter instructed the participant to begin the study route. For the drivers, an experimenter was
located in the vehicle and provided route directions if necessary. For the motorcyclists and cyclists, an experimenter
followed behind (in a car for the motorcyclists, on a bicycle for the cyclists) ready to intervene if the participants strayed off
route. Pedestrians were taken by car to the first intersection and instructed to negotiate the intersection and walk to a set
point following the intersection. Once the participant reached this point, they were picked up by the experimenter and
driven to the next intersection. This process was repeated until all three intersections had been negotiated. Route completion
times varied across participants; however, drivers took approximately 30 minutes to complete the route, motorcyclists took
approximately 25 minutes and cyclists and pedestrians took approximately 45 minutes.

Procedure (data analysis)

Downloaded by [Texas A & M International University] at 09:53 21 August 2015

Participants’ verbal protocols were transcribed verbatim using Microsoft Word. For data reduction purposes, extracts of
each verbal transcript for each intersection were taken from the overall transcripts. The extracts were taken by identifying
set features in the environment denoting the beginning and end of each intersection (e.g. road markings, signage) and then
viewing each participant’s route video and marking the start and end points. The participants’ verbalisations recorded
within the start and end points were then extracted for further analysis via EAST. The EAST analysis involved constructing
the three network representations (task, social and situation awareness networks) for each participant at each intersection.
The term network is used as each of the three forms of analysis comprises a network of ‘nodes’ and connections between
them. The term ‘node’ refers to the linked elements within the networks. Within the task network, the nodes represent
tasks undertaken when passing through the intersections (e.g. operate vehicle, maintain situation awareness). Within the
social network, the nodes represent human and non-human agents that are connected through interactions such as visual
checks (e.g. driver, traffic light, road markings). Within the situation awareness network, the nodes represent pieces of
information relevant to situation awareness (e.g. ‘speed’, ‘traffic’, ‘position’). Constructing the three forms of network
occurred as follows:
1. Task networks: Three task networks were constructed for each road user group, representing their behaviour as they
negotiated each intersection. The task networks were built based on reviewing the video and verbal protocol
transcripts and identifying how each participant negotiated the intersections. This led to the construction of generic
task networks for each road user group incorporating the range of participant behaviours identified. The task network
thus shows a generic description of the tasks that each road user group undertook as they negotiated the three
intersections.
2. Social networks: In the present analysis, the social networks represent each participant’s interaction with the
intersection ‘system’ comprising other road users and the road infrastructure (e.g. traffic lights, signage).
Construction of the social networks involved reviewing the verbal transcripts and recording all instances in which
the participant in question verbalised an interaction with part of the intersection system (e.g. in a driver transcript the
verbalisation ‘I’m just checking the traffic lights ahead’ was coded as one interaction between the driver and the
traffic light, and the verbalisation ‘I’ve just noticed a car pulling up behind me’ was coded as one interaction between
the driver and traffic behind). The frequency of interactions between the road users and the other agents within the
intersection were calculated for each road user group and represented using standard social network diagrams. For
example, if a participant’s verbal transcript included seven references to ‘checking the traffic lights’, the resulting
social network would show seven interactions between the driver and the traffic lights.
3. Situation awareness networks: ‘Situation awareness’ networks show the information being used by participants and
the relationships between different pieces of information. These were constructed for each participant group at each
intersection through analysis of participants’ verbal transcripts using the LeximancerTM content analysis tool.
LeximancerTM uses text representations of natural language to interrogate verbal transcripts and identify themes,
concepts and the relationships between them. The software does this by using algorithms linked to an in-built
thesaurus and by focussing on features within the verbal transcripts such as word proximity, quantity and salience.
LeximancerTM thus automates the content analysis procedure by processing verbal transcript data through five
stages: conversion of raw text data, concept identification, thesaurus learning, concept location and mapping (i.e.
creation of network). The output is a network representing concepts derived from the verbal transcript and the
relationships between them reflected within the verbalisations. The software has previously been used for situation
awareness network construction and analysis (Salmon, Young and Cornelissen 2013; Walker, Stanton and Salmon
 Ergonomics 1633

2011) and is important to this kind of analysis, since it provides a reliable, repeatable process for constructing the
networks.
Each form of network was analysed using selected network metrics that have previously been used in EAST analyses
(e.g. Stanton 2014). Although predominately used for the analyses of social networks, Stanton (2014) demonstrates the
utility of applying selected network analysis metrics to the task, social and situation awareness networks generated through
EAST analyses. In this study, the following network analysis metrics were applied to each network:
1. Network density: Network density represents the level of interconnectivity of the network in terms of relations
between nodes. Density is expressed as a value between 0 and 1, with 0 representing a network with no connections
between nodes, and 1 representing a network in which every node is connected to every other concept (Kakimoto
et al. 2006, cited in Walker, Stanton and Salmon 2011). In the present analysis density was used as a way of
calculating the level of connectivity within each network. The formula for network density is presented below.

Network density ¼ nðn2e

2 1Þ
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

where :
ð1Þ
e ¼ number of links in network
n ¼ number of information elements in network

2. Sociometric status (individual nodes): Sociometric status provides a measure of how ‘busy’ a node is relative to the
total number of nodes within the network under analysis (Houghton et al. 2006). In the present analysis nodes with
sociometric status values greater than the mean sociometric status value plus one standard deviation are taken to be
‘key’ (i.e. most connected) nodes within each network. The formula for sociometric status is presented below.

1 X
g
Status ¼ ðxji þ xij Þ ð2Þ
g 2 1 j¼1

Results
Task networks
The overall task of passing through the networks was labelled ‘Negotiate intersection’. The ‘Negotiate intersection’ task
networks provide a summary of the main goals and tasks involved in negotiating signalised intersections for each road user
group. The task network for the drivers is presented in Figure 3.
The task network presented in Figure 3 shows at a high level the important sub-tasks (boxes in Figure 3) undertaken by
the drivers when negotiating the three intersections during the on-road study. The relationships between the sub-tasks are
represented by the links between boxes in Figure 3. For example, links are added if tasks are undertaken together (e.g.
‘operate vehicle’ and ‘maintain situation awareness’ are undertaken together and continuously). Similarly, sub-tasks are
linked if they are undertaken in sequence (e.g. ‘approach intersection’ is linked to ‘select and move into appropriate lane’ as
it occurs before selecting and moving into an appropriate lane).
The differences in task networks across the four road user groups were identified by examining the sub-tasks in each task
network. A summary of the task network sub-tasks for the four road user groups is presented in Table 2. Within Table 2 the
sub-tasks are presented in the left hand side column and shading is used to denote whether each road user group undertook
the sub-tasks when negotiating the intersections during the study.
As shown in Table 2, when negotiating intersections there are important differences in the tasks undertaken by the four
road user groups studied. The only tasks common across the four road user groups are ‘approaching the intersection’,
maintaining situation awareness’, ‘checking the intersection and determining a route through the intersection’, ‘proceeding
through the intersection’ and ‘leaving the intersection area’. It is notable, however, that the activities required for these tasks
are different across the road user groups. For example, ‘maintaining situation awareness’ entails maintaining awareness of
different aspects of the intersection depending on road user type (e.g. cyclists and motorcyclists required awareness of
debris on the road whereas drivers did not).
Important differences across the task networks are present. These include the task of filtering up through the traffic
queue undertaken by both cyclists and motorcyclists in pursuit of the goal to get to the front of the traffic queue, the
requirement for cyclists and motorcyclists to undertake activities in order to make themselves visible to other road users
(e.g. wiggling, flashing lights, eye-balling drivers, revving engine) and the requirement for cyclists and motorcyclists to
check the road surface for debris/hazards/obstacles and to avoid debris on the road as they turn through the intersection.
 1634 P.M. Salmon et al.

Check intersection and determine

Approach intersection
route through intersection

Select and move into appropriate lane

Maintain
situation
awareness
Take place in traffic queue/ Maintain safe
Operate (e.g. other Follow traffic
at front of traffic queue position on
vehicle traffic, own
road
position,
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

lights, speed
limits)
Check lights and wait for green light/turning arrow

Proceed through intersection

Figure 3. Driver intersection task network.

One key issue highlighted by the task networks is the difference in the level of flexibility afforded to each form of road user
by the intersections studied. As shown in Table 2 by the high number of shaded tasks, cyclists have a significant level of
flexibility in that they can proceed through the intersection in a variety of different ways. They can either turn right on the road

Table 2. Summary of sub-tasks within task networks.

Task Drivers Cyclists Motorcyclists Pedestrians

Operate vehicle
Walk/stop/stand
Maintain situation awareness
Make self visible to other road users
Maintain safe position on road
Avoid surface obstacles
Approach intersection
Check intersection and determine route through
Select and move into appropriate lane
Follow traffic
Take place in traffic queue/at front of queue
Filter to front of traffic queue
Check lights and wait for green light/turning arrow
Proceed forward into adjacent traffic flow and conduct hook turn
Activate pedestrian crossing
Wait for green man
Cross road using pedestrian crossing
Cross road without using pedestrian crossing
Proceed through intersection
Leave intersection area
 Ergonomics 1635

Table 3. Task network analysis.

Road user Density Sociometric status key tasks
Drivers 0.62 Operate vehicle
Maintain situation awareness
Maintain safe position on road
Cyclists 0.69 Operate bicycle
Maintain situation awareness
Make self visible to other road users
Maintain safe position on road
Motorcyclists 0.81 Operate motorcycle
Maintain situation awareness
Make self visible to other road users
Maintain safe position on road
Pedestrians 0.63 Walk/stop/stand
Maintain situation awareness
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

within the flow of traffic, turn right via the pedestrian crossings and along the footpath or turn right using a ‘hook’ turn whereby
they proceed straight on through the intersection, join the traffic queue to the left hand side and then wait for a green light and
proceed straight through the intersection (achieving the originally desired right hand turn). For pedestrians, a degree of
flexibility is extant through the ability to either cross via the pedestrian crossing or the road itself. Motorcyclists’ behaviour is
more restricted; however, they still have some flexibility in that they can choose to use the stay in the normal traffic queue or to
filter up between the traffic to the front of the queue. Of the four road user groups, drivers are the most restricted in behaviour
since they can only select one of two turning lanes and progress through the intersection.
The density and sociometric status calculations for the task networks are presented in Table 3.
Table 3 shows that the driver, cyclist and pedestrian task networks had similar levels of density; however, the
motorcyclist task network is the densest (0.81). This indicates that, of the four road users, motorcyclists have a more
connected set of tasks when negotiating intersections. The key tasks, as identified through the sociometric status metric,
demonstrate the importance of ‘operate vehicle’ and ‘maintain situation awareness’ tasks for all road users, and also shows
the importance of the ‘make self visible’ and ‘maintain safe position on road’ for the cyclists and motorcyclists.

Social networks
The social networks depict each road user groups’ interaction with the intersections studied. An interaction in this case
is defined as any instance where participants referred to interacting with a part of the intersection within their verbalisations
(e.g. ‘I am checking the traffic lights’, ‘I can see the road markings’). For example purposes, the social networks for
intersection 1 are presented in Figure 4. This shows the total number of reported interactions with different parts of the
intersection for each group based on their verbal protocol transcripts. The thickness of the arrows represents the frequency
of interactions, with the thicker arrows representing a higher frequency across each road user group.
Density was calculated for each network and the sociometric status metric was used to identify the most connected
nodes within each network (outside of the road users themselves). These results are presented in Table 4.
Table 4 shows that the pedestrian group were the most connected in terms of their interactions with the intersection
environments. They interacted with the greatest number of intersection agents (e.g. lights, button, traffic), had the highest
number of interactions and achieved the highest sociometric status values across all three intersections. The drivers on the
other hand were the least connected of the four groups, with smaller networks, the lowest number of interactions at two of
the three intersections and the smallest sociometric status values across the three intersections.
The social networks also reveal important differences in how the different road user groups interact with intersection
systems. In this study the drivers interacted less with certain parts of the intersection than the other road users. For example,
across the three intersections the most frequent interaction for drivers was observing the oncoming traffic, traffic
lights ahead and the traffic in front. Although cyclists and motorcyclists had frequent interactions with these parts of
the intersections, they tended to have more interactions with these components and also other interactions were prominent,
such as checking for traffic behind and observing the traffic generally (i.e. travelling across the intersection). The low
number of driver interactions focussing on the traffic generally, the traffic behind and adjacent traffic is a concern,
particularly when coupled with the task network finding that the other road user groups may be unpredictable in and around
the intersection.
 1636 P.M. Salmon et al.

Oncoming Traffic
Total interactions = 122 traffic lights
Oncoming Traffic
Road traffic lights
Total interactions = 204
Ped name
26
crossing (straight
44
ahead) 21
Traffic
Traffic Lane Parked cars
2 82 (service
1 Speed
lane)
limit 11 Service
4 Road 9 signage Traffic lane
5 lights 2
surface
Road 17

44
44
2 surface

44

44
Traffic 1
Green/ P.
lights Traffic in Red man
front Highway
Traffic in

44
7 9
front 9
6 44
44
Left turn

44
11 Ped
arrow 44 Turning
1 crossing
44

traffic
4

44
21
4
2
Pedestrians
Driver 1 Camera
6 Footpath
Cyclist
4
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

2
Pedestrians
6

44
14

Traffic
behind
Traffic
behind

Traffic

Traffic Total interactions = 423

Total interactions = 141 Oncoming
traffic lights

Lights

15 Footpath
Traffic
Parked
180
56 cars
133
12 11
Traffic Speed
lights 1 limit Bin 44 Road
Road signage

44
3 surface 2
44

3
44
44

P. Driveway 29 Crossing
Highway 44
44

Traffic in
9 8 44 10
front 1 44

Service
44

Ped 44 Pedestrian 44 Button

lane
crossing 12 15
44

2 17
1
Police
Cars P.
4
Signage (driveway) Highway
1 1
Rider Camera
44

44
4 1 Horizon
Island 1
10 1 Camera
Pedestrians Cyclist 19
2
44

10
1
Peds
Road 2 Gardener
surface 2
Traffic
behind
Shops
Car yard

Figure 4. Example driver and cyclist social networks for intersection 1.

Unsurprisingly, pedestrians had various interactions related to using the pedestrian crossing infrastructure, such as
pressing the ‘button’. Interestingly, cyclists also had a number of interactions with pedestrian crossing infrastructure, which
reflects the decision by selected cycling participants to cross the intersections using the footpath and pedestrian crossing.
Other notable interactions specific to road user groups were motorcyclists and cyclists checking of the ‘road surface’ for
debris and oil, cyclists and motorcyclists monitoring of traffic adjacent to them, motorcyclists checking of the traffic lights
facing other traffic (to work out when other traffic would be stopping and going) and cyclists monitoring of pedestrian
behaviours in and around the intersection.
Situation awareness networks
LeximancerTM was used to construct overall driver, cyclist, motorcyclist and pedestrian situation awareness networks for
each intersection. The situation awareness networks provide a representation of each road user groups’ awareness when
negotiating the intersections. For example, Figure 5 depicts each road user groups’ situation awareness network, overlaid on
one another, for intersection 1. Within Figure 5 the nodes and links are shaded to depict each road user group’s situation
awareness. Figure 5 therefore shows how situation awareness differed across the road user groups whilst negotiating
intersection 1, both in terms of the concepts underpinning situation awareness (i.e. nodes in the network) and also in the way in
 Ergonomics 1637

Table 4. Summary of social network analysis for each intersection.

Sociometric
Road user Number of Mean number status of
group nodes of interactions Density road user Nodes with highest sociometric status values
Intersection 1
Drivers 12 6.1 0.083 11.09 Lights, oncoming traffic, traffic (in front), lane, pedestrians,
traffic (behind)
Cyclists 13 10.3 0.08 17.17 Lights, oncoming traffic, traffic (in front), service lane,
traffic (behind)
Motorcyclists 14 7.8 0.07 11.31 Lights, traffic (in front), oncoming traffic, traffic (general),
traffic (behind), pedestrians, parked cars, road, camera,
cyclist
Pedestrians 20 21.1 0.05 22.78 Traffic, lights, road, pedestrians, service lane
Intersection 2
Drivers 10 6.3 0.1 14.11 Lights, traffic (in front), traffic (general), traffic (behind),
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

sign
Cyclists 20 10.4 0.05 10.95 Lights, traffic (general), service lane, traffic (in front), road
Motorcyclists 13 6.4 0.08 9.66 Lights, traffic (in front), road, traffic (general), traffic
(behind)
Pedestrians 28 19.4 0.036 14.37 Traffic (general), lights, road, intersection, button
Intersection 3
Drivers 14 7.6 0.071 11.69 Lights, traffic (general), traffic (in front), intersection, lane
Cyclists 20 12.1 0.05 12.78 Lights, traffic (general), traffic (behind), lane, traffic (in
front)
Motorcyclists 18 6.3 0.053 6.28 Lights, traffic (general), lane, road, intersection
Pedestrians 30 38.35 0.033 26.44 Traffic (general), lights, road, pedestrians, intersection

which the concepts were linked together (i.e. links between the nodes in the network). Moreover, the network demonstrates
that, even when the different road users were using the same concept, they were doing so in conjunction with other different
concepts. This pattern is repeated over the other two intersections studied. The multi-road user situation awareness networks
therefore confirm that driver, cyclist, motorcyclist and pedestrian situation awareness was different when negotiating the
three intersections studied.
A summary of the situation awareness network analysis is presented in Table 5. This includes the number of concepts
within each groups’ situation awareness networks, density and sociometric status scores and the key concepts underpinning
situation awareness for each road user group.
Table 5 highlights some important differences in situation awareness across the four road user groups. Pedestrian
situation awareness networks comprised more concepts at each intersection, whereas the driver networks were the smallest.
The driver situation awareness networks were the most dense at intersections 1 and 2, whereas the motorcyclists were most
dense at intersection 3. The same pattern was found for the mean sociometric status values, which indicates that the driver
networks were most connected in terms of links between concepts at intersections 1 and 2, whereas the motorcyclists were
most connected at intersection 3.
The key concepts underpinning each road user groups’ situation awareness also reveal important differences. At
intersection 1, the key concepts are broadly similar across all four road user groups, with situation awareness predominantly
underpinned by the lights, their status (red or green) and the traffic. Notable differences at intersection 1 include a focus of
cyclists on the ‘intersection’ itself, and a focus of motorcyclists on the ‘lane’ within which to progress through the
intersection. At intersection 2 more differences in key concepts were found. Again, the focus of the drivers is on the lights;
however, the cyclists also focussed on the ‘service lane’, which ‘lane’ to use, ‘turning’ traffic and the ‘road’ itself. The
motorcyclists’ key concepts include ‘looking’, ‘behind’, and again the ‘turning’ traffic and the ‘lane’ to use. At intersection
3 the key concepts are again different. The drivers’ key concepts now include distinct areas around the car, such as ‘front’
and ‘behind, and also the ‘lanes’. The cyclists’ key concepts also include ‘ahead’ and ‘behind’, ‘wait’ and ‘crossing’. The
motorcyclists’ key concepts relate to the lights (i.e. green) and the ‘lane’ within which to progress through the intersection.
Finally, the pedestrians have a larger set of key concepts at intersection 3, with key concepts covering the ‘road’, the traffic
and its behaviour (e.g. ‘cars’, ‘traffic’, ‘coming’) and the lights and their status.
The differences in the situation awareness networks were explored further by examining the ‘common’ concepts (i.e.
those found in all road user groups situation awareness networks) along with the concepts unique to each road user group at
each intersection. Figure 6 shows that the concepts common across all four road user groups at the three intersections were
 1638 P.M. Salmon et al.

= Cyclists = Drivers = Motorcyclists = Pedestrians

Checking
Orange
Lane
Service

Busy
Moment
Hand
Round Middle
Fairly
Car
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

Road
Looking Walking
I’ll
Check Able
Turning
Intersection

Walk
Looks
Cross
Princes Safe
Highway
Cars
Coming Stopping Proceed
Truck Sure

Green Unclear
Button

Traffic
Red

Front
Change Light Arrow

Man

Crossing Lights
Pedestrian
Slowly Wait

Amber

Behind
Straight
Oncoming
Waiting

Stay Ahead
People

Figure 5. Intersection 1 situation awareness network showing driver, cyclist, motorcyclist and pedestrian situation awareness networks
mapped onto each other.
 Ergonomics 1639

Table 5. Summary of situation awareness network analysis for each intersection.

Number of Mean sociometric
Road user group concepts Density status Key concepts
Intersection 1
Drivers 26 0.075 0.151 Green, oncoming, traffic
Cyclists 28 0.061 0.122 Coming, cars, intersection, red, traffic
Motorcyclists 28 0.066 0.132 Lane, traffic, cars, arrow
Pedestrians 29 0.069 0.138 Green, red
Intersection 2
Drivers 27 0.074 0.148 Lights, light, lane
Cyclists 29 0.069 0.138 Cars, green, road, take, service, lane, turning
Motorcyclists 28 0.071 0.143 Looking, behind, green, cars, turning, lane
Pedestrians 31 0.064 0.129 Lights, light, green, turning, coming, road, crossing
Intersection 3
Drivers 27 0.074 0.148 Front, red, Dandenong Road, making, behind, lanes
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

Cyclists 28 0.071 0.143 Turning, stay, behind, ahead, wait, lane, lights, crossing
Motorcyclists 26 0.075 0.151 Green, lane
Pedestrians 34 0.055 0.110 Road, cars, coming, cross, wait, traffic, lights, green, red, light

mainly related to the cars, traffic, the road, the lights and their status (e.g. green), the intersection and the act of turning. This
reflects a high focus of all road users on cars, the traffic lights and the intersection environment itself.
Important differences are shown in Figure 6. At intersections 1 and 2, the cyclist networks include the concepts
‘service’, ‘stay’ and ‘route’, all of which reflect the key decision that they face regarding whether or not to use the service
lane on approach to the intersection and then cross via the footpath and pedestrian crossing, or to stay on the road and go
through the intersection in the normal traffic flow. Prior to the intersection the cyclists also decide whether they will leave
the road and get back into the service lane once they have passed through the intersection. At intersection 3, the cyclists’
network included the ‘hook’ concept, which refers to their decision regarding whether or not to use a hook turn in order to
turn right at the intersection. Again this reflects a key decision whereby cyclists try to work out whether it is safe enough to
pass through the intersection on the road within the flow of traffic or whether they need to perform a hook turn to avoid
conflict with other traffic also turning right.
For the motorcyclists, the unique concepts relate primarily to the selection of the left or right hand lane to negotiate the
intersection (e.g. ‘hand’, ‘left hand’, ‘merging’), the motorcycle itself (e.g. ‘bike’, ‘gear’) and the ‘line’ that they should take
through the intersection. ‘Stopping’ relates to the motorcyclists’ braking behaviour and also to them checking that other
traffic approaching from behind are slowing down when the traffic lights are red. For the pedestrians, the unique concepts
were primarily related to the physical acts of walking (e.g. ‘walk/walking’), crossing the road (e.g. ‘cross/crossing’) and the
crossing infrastructure (e.g. ‘button’, green ‘man’). Only the pedestrian networks included the concepts ‘check/checking’
and ‘look/looking’, which indicate that the other road users placed less emphasis on checking other traffic and the road
environment when negotiating the intersections.

Discussion
The aim of this paper was to examine the differences in the behaviour of drivers, cyclists, motorcyclists and pedestrians
when negotiating three intersections in the south-eastern suburbs of Melbourne, Australia, through a novel application of
the EAST methodology. Following this, the aim was to use the findings to identify some of the sources of conflicts between
different road users at intersections.

Differences across road users

The analysis demonstrates that, even when faced with the same intersection and right hand turn task, the goals and tasks,
interactions and knowledge used by the different road users was different. This was exemplified through differences in the
task, social and situation awareness networks for each road user group. It is concluded that, when negotiating intersections,
different forms of road user perform different tasks, interact differently with the intersection environment and have a
different understanding of the situation. An important question to ask then is whether these differences create conflicts
between road users or whether they represent different but compatible behaviours?
 1640 P.M. Salmon et al.

Intersection 1
Common across all road Unique to Unique to Unique to
Unique to pedestrians
user groups drivers cyclists motorcyclists

Car(s) Road Walk/ Check/

Safe Service Stopping
Walking Checking

Light(s) Green I’ll Road name Fairly Button Ahead

Cross/
Intersection Traffic Slowly Moment Man
Crossing

Wait/
Coming Stay Round Able Straight
Waiting

Intersection 2
Common across all road Unique to Unique to Unique to
Unique to pedestrians
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

user groups drivers cyclists motorcyclists

Lane/
Car(s) Pull Service Hand Turned Anyway
Lanes

Light(s) Green Time Left Hand Button Man

Intersection Sure Ready Bike Seems

Red Straight Stay Gear

Turning Take Line

Route

Intersection 3
Common across all road Unique to Unique to Unique to
Unique to pedestrians
user groups drivers cyclists motorcyclists

Car(s) Coming Change/ Walk/

Hook I’ll Middle
Changing Walking

Light(s) Green Right hand Doing Hand Check/

Able
Checking

Intersection Red Forward Moving Clear Flashing

Turning Making Gear Look/

Button
Looking

Notice Merging Man

Route Assume

Stopping

Figure 6. Common and unique concepts across road user groups at each intersection.

Unfortunately, the EAST analysis suggests that compatibility between different forms of road user is under threat in
some instances. For example, the analysis demonstrates that the limited latitude for behaviour for drivers at intersections is
in fact in conflict with the highly flexible and thus unpredictable behaviour of cyclists, motorcyclists and pedestrians. In the
case of cyclists, for example, the analysis highlights their high level of flexibility since they can proceed through the
intersection in multiple ways (i.e. hook turn, in normal flow of traffic, filtering, on the footpath and via pedestrian crossings).
The analysis also shows that, as a corollary of this flexibility, cyclists’ behaviour and situation awareness are focussed on
 Ergonomics 1641

working out the safest way through the intersection, and that their behaviour is generally unpredictable since their chosen
route through can change on a moment-by-moment basis. Examination of the drivers’ social and situation awareness
networks, however, shows that they may not be expecting cyclists, and indeed motorcyclists, to make major manoeuvres in
close proximity to the intersection. Consequently, they do not appear to behave or think in a manner that is compatible with
other cyclists and motorcyclists. The driver situation awareness networks embodied a high focus on cars, the traffic lights,
and the act of turning right and an absence of ‘check/checking’, ‘look/looking’ and left or right ‘hand’ side concepts. The
social networks showed that, compared to the other road users, drivers had a limited interaction with the intersection that
mainly involved the oncoming traffic, traffic lights and traffic directly in front of the vehicle. Drivers also had only limited
interactions with the areas adjacent to, and behind their own vehicle. The conclusion from this is that driver behaviour at the
three intersections studied may not be compatible with cyclist and motorcyclist behaviour.
This ‘flexibility problem’ as we will call it is likely brought about by the fact that it represents flexibility not purposely
designed into the system, but flexibility derived from emergence. For example, cyclists’ range of possible behaviours at
intersections is brought about by the lack of support for, and high risk associated with, turning right within the normal flow of
traffic. The risk level for cyclists is elevated once they enter the intersection, hence their desire to seek a safer path through it. As a
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

result, cyclists and motorcyclists have developed new ways of taking a safer path through the intersection. The physical design of
the three intersections is therefore creating this flexibility. Interestingly, the design of the intersections and road environments
studied also acts to increase the potential of a conflict between the cyclist and the driver. There is currently no signage warning
drivers of the presence of cyclists or motorcyclists or the likelihood that they will be manoeuvring in and around the intersection.
The speed limit on approach to the intersections was 80 km/h (far exceeding the capability of cyclists). Also, where present the
cycling lanes preceding the intersections end a significant distance before the intersection itself, and there are no dedicated cycle
or motorcycle lanes proceeding through the intersections. In addition, the road rules prohibit cyclists from cycling on footpaths
and motorcyclists from filtering up the traffic queue. The flexibility brought about by the design of the intersections is
problematic because the intersections do not prepare other road users for the resulting emergent behaviours.

System redesign
To solve the flexibility problem, it would appear that the key is to ensure that different road users have some degree of
flexibility, but also to ensure that the way in which the intersection is designed supports this flexibility, communicates this
flexibility to other road users and ensures that the resulting behaviours are compatible. It is appropriate to maintain the level
of flexibility for cyclists; however, other road users should be made aware of it, and measures should be taken to support it.
For example, avenues for making other road users (pedestrians, drivers) more expectant of cyclists and their range of
behaviours will be fruitful. The use of cross mode training has previously been raised as a way of developing anticipatory
schema that supports perception of other road users (Magazzù, Comelli and Marinoni 2006; Walker, Stanton and Salmon
2011). Cross mode training incorporating both cyclist and motorcyclist training for drivers is likely to increase their
expectancy levels regarding cyclist and motorcyclist behaviours at intersections. There is also a clear role for road design,
with dedicated cycling lanes (on the road and on the footpath) and signage warning drivers of the presence of cyclists and
the likelihood that they will make major manoeuvres in close proximity to the intersection.
As a first test application of EAST in a road user behaviour context, this research did have some limitations. First, road
users negotiated the same intersections under similar traffic conditions, but not at the same time as one another. In addition,
they were followed by an experimenter throughout the study. Analysis of behaviour at intersections when road users are
interacting together at the same intersection at the same time whilst not being followed would provide more valid data.
Second, the social networks were based on the verbal protocols describing interactions with parts of the intersection rather
than eye tracking data showing actual eye fixations on objects in the road environment. Future studies in which eye tracking
data are used to construct the social networks will be useful. Third, interpretation of the networks produced is based on the
authors’ subjective judgement. This study attempted to remove some of this subjectivity through the use of network analysis
metrics; however, further investigation should be undertaken to explore making the approach more objective. Fourth,
although EAST has been applied in many areas there is a lack of validation evidence associated with the framework. It is
recommended that studies examining the reliability and validity of the EAST framework be undertaken. Fifth and finally,
this study focussed on large signalised intersections only. Further research focussing on multiple road users at other forms of
intersection (e.g. smaller, non-signalised) is required.

Conclusion
From the analysis presented it is concluded that there was a significant level of disharmony between road users in the
intersections studied. Differences were found in the way in which road users act, think and interact with the environment;
 1642 P.M. Salmon et al.

however, it appears that intersections are not designed with these differences in mind. Moreover, the analysis suggests that
the way in which the intersections are currently designed may be restricting compatibility between different road user
groups. It is argued that a failure to consider the overall intersection system during intersection design, including different
road users’ physical and cognitive behaviours and their differences, is facilitating the problem of intersection collisions.

Acknowledgement
This research was funded through the Australian Research Council Discovery Scheme [grant number DP120100199]. Professor Salmon’s
contribution was funded initially through an Australian National Health and Medical Research Council (NHMRC) postdoctoral training
fellowship followed by an Australian Research Council Future Fellowship [grant number FT140100681].

References
Cornelissen, M., P. M. Salmon, R. McClure, and N. A. Stanton. 2013. “Using Cognitive Work Analysis and the Strategies Analysis
Diagram to Understand Variability in Road User Behaviour at Intersections.” Ergonomics 56 (5): 764– 780.
Gershon, P., and D. Shinar. 2013. “Increasing Motorcycles Attention and Search Conspicuity by Using Alternating-Blinking Lights
Downloaded by [Texas A & M International University] at 09:53 21 August 2015

System (ABLS).” Accident Analysis & Prevention 50: 801– 810.

Gstalter, H., and W. Fastenmeier. 2010. “Reliability of Drivers in Urban Intersections.” Accident Analysis & Prevention 42 (1): 225– 234.
Highways Agency. 2012. Design Manual for Roads and Bridges. London: Highways Agency.
Houghton, R. J., C. Baber, R. McMaster, N. A. Stanton, P. M. Salmon, R. Stewart, and G. H. Walker. 2006. “Command and Control in
Emergency Services Operations: A Social Network Analysis.” Ergonomics 49 (12 – 13): 1204– 1225.
King, M. J., D. Soole, and A. Ghafourian. 2009. “Illegal Pedestrian Crossing at Signalised Intersections: Incidence and Relative Risk.”
Accident Analysis & Prevention 41 (3): 485– 490.
Larsson, P., S. W. A. Dekker, and C. Tingvall. 2010. “The Need for a Systems Theory Approach to Road Safety.” Safety Science 48 (9):
1167– 1174.
Magazzù, D., M. Comelli, and A. Marinoni. 2006. “Are Car Drivers Holding a Motorcycle License Less Responsible for Motorcycle-Car
Crash Occurrence? A Non-parametric Approach.” Accident Analysis & Prevention 38 (2): 365– 370.
McLean, J., P. Croft, N. Elazar, and P. Roper. 2010. Safe Intersection Approach Treatments and Safer Speeds through Intersections: Final
Report, Phase 1, AP– R363/10. Sydney: Austroads.
Salmon, P. M., R. McClure, and N. A. Stanton. 2012. “Road Transport in Drift? Applying Contemporary Systems Thinking to Road
Safety.” Safety Science 50 (9): 1829– 1838.
Salmon, P. M., N. A. Stanton, G. H. Walker, D. P. Jenkins, C. Baber, and R. McMaster. 2008. “Representing Situation Awareness in
Collaborative Systems: A Case Study in the Energy Distribution Domain.” Ergonomics 51 (3): 367– 384.
Salmon, P. M., K. L. Young, and M. Cornelissen. 2013. “Compatible Cognition amongst Road Users: The Compatibility of Driver,
Motorcyclist, and Cyclist Situation Awareness.” Safety Science 56: 6– 17.
Sandin, J. 2009. “An Analysis of Common Patterns in Aggregated Causation Charts from Intersection Crashes.” Accident Analysis &
Prevention 41 (3): 624– 632.
Stanton, N. A. 2014. “Representing Distributed Cognition in Complex Systems: How a Submarine Returns to Periscope Depth.”
Ergonomics 57 (3): 403– 418.
Stanton, N. A., P. M. Salmon, L. Raffery, G. Walker, C. Baber, and D. P. Jenkins. 2013. Human Factors Methods: A Practical Guide for
Engineering and Design. 2nd ed. Aldershot: Ashgate.
Stewart, R. J., N. A. Stanton, D. Harris, C. Baber, P. M. Salmon, M. Mock, K. Tatlock, L. Wells, and A. Kay. 2008. “Distributed
Situational Awareness in an Airborne Warning and Control Aircraft: Application of a Novel Ergonomics Methodology.” Cognition,
Technology and Work 10 (3): 221– 229.
VicRoads. 2013. “CrashStats.” Accessed December 16, 2013. [Link]
StatisticsAndResearch/[Link]
Wagenaar, W. A., and J. T. Reason. 1990. “Types and Tokens in Road Accident Causation.” Ergonomics 33 (10 –11): 1365– 1375.
Walker, G. H., H. Gibson, N. A. Stanton, C. Baber, P. M. Salmon, and D. Green. 2006. “Event Analysis of Systemic Teamwork (EAST):
A Novel Integration of Ergonomics Methods to Analyse C4i Activity.” Ergonomics 49 (12– 13): 1345– 1369.
Walker, G. H., N. A. Stanton, C. Baber, L. Wells, D. P. Jenkins, and P. M. Salmon. 2010. “From Ethnography to the EAST Method: A
Tractable Approach for Representing Distributed Cognition in Air Traffic Control.” Ergonomics 53 (2): 184– 197.
Walker, G. H., N. A. Stanton, and P. M. Salmon. 2011. “Cognitive Compatibility of Motorcyclists and Car Drivers.” Accident Analysis &
Prevention 43 (3): 878– 888.