Rp

Ia

db

W

a

ARRA

KIrRTPC

1

fAuciRlgtcic

a2

0d

Accident Analysis and Prevention 49 (2012) 96– 104

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

jo ur nal homep a ge: www.elsev ier .com/ locate /aap

isk and threat factors in prior representations of driving situations amongowered two-wheeler riders and car drivers

sabelle Ragot-Courta,∗, Christophe Mundutéguyb, Jean-Yves Fourniera

French Institute of Science and Technology for Transport, Development and Networks (IFSTTAR) – “Accident Mechanisms” Research Unit, Chemin de la croix blanche, 13300 Salone Provence, FranceUniversity Paris Est – French Institute of Science and Technology for Transport, Development and Networks (IFSTTAR) – Production Systems, Logistics, Transport Organisation andork Research Unit, Descartes 2-2, rue de la butte verte, 93166 Noisy-le-Grand Cedex, France

r t i c l e i n f o

rticle history:eceived 21 December 2010eceived in revised form 24 August 2011ccepted 9 September 2011

eywords:nteraction between powered two-wheeleriders and car driversisk perception

a b s t r a c t

Our research objective is to contribute to gaining a better understanding of the difficulties inherentto managing interactions between power two-wheeler riders and car drivers. 132 power two-wheelerriders and 94 car drivers have been asked what they perceive as being riskiest in driving situations for therepresentatives of their generic reference group. From all the answers produced, only those which provideinformation on risk factors related to the other and the occurrence of interferences due to the dynamicsituation have been retained and analysed. The results provide a partial explanation of the difficultiesrelated to the two types of users’ taking each other into account. On the one hand, the frequency withwhich the other is mentioned as a source of risk is linked to the concept of conspicuity. This interpretation

hreatrior representationsonspicuity

is related to the relative frequency of the interactions and the difference of real and perceived vulnerabilitybetween the users. On the other hand, the specific risk factors attributed to the other illustrates a mutualmisunderstanding or ignorance of the driving situation’s determinants. These include various practicesamong users which can be related to the physical and dynamic characteristics of their vehicle and theirlevel of familiarity with other users. The potential road safety consequences of prior representations ofrisk factors among users are presented and suggestions for improving road user safety are proposed.

© 2011 Elsevier Ltd. All rights reserved.

. Introduction

On French roads, power two-wheeler (PTW) users still accountor less than 2% of all road traffic in terms of kilometres driven.nd yet, the road safety results for this category of vulnerablesers1 is highly negative compared with other road users, notablyar drivers. This characteristic justifies improving our understand-ng of the factors that could explain their accident involvement.oad safety challenges and, more generally, public health chal-

enges are increasingly important because PTW traffic has beenrowing constantly over the past 20 years, notably related toraffic congestion in large urban areas (Haworth, in press). The

rash rate for PTWs includes a large number of accidents alsonvolving a car driver (ONISR, 2005). Our research objective is toontribute to understanding the difficulties inherent to the inter-

∗ Corresponding author. Tel.: +33 4 90 56 86 23; fax: +33 4 90 56 86 18.E-mail addresses: isabelle.ragot-court@ifsttar.fr (I. Ragot-Court).

1 The motorcyclists account for more than 10% of the drivers involved in an injuryccident and 18% of the persons killed. The risk of being killed during an accident is3 times more raised for a motorcyclist than for a car driver (ONISR, 2007b).

001-4575/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.oi:10.1016/j.aap.2011.09.011

actions between these two types of users. For this, we soughtto gather their knowledge and beliefs as to what they perceiveas being riskiest in driving situations. Their prior knowledge andbeliefs determine part of their awareness of the situation and thusorient their risk perception in the situation (Smith and Hancock,1995).

PTWs are rare in traffic, but they bring a differential to the trafficsystem through the vehicle’s very design and the type of drivingthey entail. In the most commonly observed accident scenariosinvolving a motorcycle and a car, it can be noted that problemsoften relate to taking others into account (ONISR, 2007a). The weakconspicuity of PTWs has been identified as a major risk factor ininteractions between PTW riders and car drivers (Williams andHoffmann, 1979; Thomson, 1980). Conspicuity usually refers to anobject’s capacity to draw attention and be easily located thanksto its physical properties (Wulf et al., 1989). This concerns theobject’s visual salience, i.e. the degree to which it can be distin-guished from its environment (Hancock et al., 1990). The physical

characteristics of PTWs riding in a complex environment constitutean explanatory element for their weak conspicuity in view of per-ceptual constraints and information processing of other road users(Hole et al., 1996; Horswill et al., 2005; Crundall et al., 2008).

lysis a

cctIoiiaiwenfoisaodtettfir

dvitnaaaothufibewtaTttV

iUVesaatktrsoepa

I. Ragot-Court et al. / Accident Ana

PTW riders are also subject to the effects of another form ofonspicuity: cognitive conspicuity. Beyond their sensory or visualonspicuity, the cognitive conspicuity of objects can be defined byheir capacity to draw the attention of observers in a given situation.n other words, when an individual performs a particular task, anbject that is subjectively considered irrelevant to the task at hand,.e. having weak cognitive conspicuity, will have a lower probabil-ty of being detected and taken into account by this individual thann object with strong conspicuity. An object’s cognitive conspicu-ty is based on several dimensions relative to the observers and

hich are changeable. This refers to their short-term intentions orxisting goals (Hancock et al., 1990). In other words, an object’s cog-itive conspicuity varies from one individual to another, but also,

or a given individual, from one moment to another and/or fromne situation to another. In the literature on PTWs’ traffic accidentnvolvement and its underlying causes, the notion of cognitive con-picuity has not yet been much explored. It is dealt with from twongles of approach which are nonetheless closely linked. On thene hand, the weak cognitive conspicuity of PTWs is evoked asepending on the car drivers’ personal and social interest and/orhe significance of the PTWs in terms of affect, valence, experience,tc. (Hancock et al., 1990). On the second hand, the weak cogni-ive conspicuity of PTWs can be explained by the low expectationshey provoke among other users due to their relative rarity in traf-c and therefore the low probability of encountering them on theoad (Van Elslande et al., 2008).

Besides problems caused by visual and cognitive conspicuity, carrivers may also have difficulty in understanding PTWs’ manoeu-res. This may have some incidences on how they take PTW ridersnto account and how they foresee their behaviour. Foresight ishe result of the combination of circumstantial data and perma-ent knowledge and beliefs (Mundutéguy, 2001). While the formerre dependent upon and concomitant to the interaction, the latterre prior to it. Knowledge and beliefs include categories of affili-tion corresponding to vehicle characteristics (Trucks, PTWs, etc.)r stereotypes referring to prototypical behaviours associated withhese objects (PTW riders are speeders, PTWs couriers are fool-ardy riders, taxis’ drivers drive aggressively, etc.). In situations ofncertainty in which behaviour by one user leads to predictive dif-culties, Mundutéguy and Darses (2007) pointed out that foreseenehaviour is determined by the behaviour that the subject wouldxpect to be confronted with from the group’s prototype. Alongith the conspicuity factors indicated above, it is relevant to add

hat PTW riders often adopt behaviours that are specific to themnd which do not correspond to car drivers’ behavioural standards.hese behaviours do not make it easy for other users to foreseehem. These behaviours could, in fact, surprise car drivers and keephem from reacting effectively in critical situations (Obenski, 1994;an Elslande, 2002).

Different levels of analysis are used to study the difficultiesnherent to interactions between PTW riders and car drivers.sing In-depth Accident Studies and Police Reports, Jaffard andan Elslande (in press) has studied failures affecting PTW rid-rs and car drivers in critical interactions’ situations. They havehown the differences between dysfunctional processes that char-cterise the formers and those of the latters. As a complementarypproach at an earlier stage, the present research’s ambition iso look into what comes before these processes, such as drivers’nowledge and beliefs about risk. In general, this knowledge andhese beliefs prior to driving situations shape the overall socialepresentation of the risk in the road environment. As for allocial representations (Abric, 1993), they structure thoughts that

rient understanding of one’s own behaviour and that of oth-rs. More specifically, knowledge and beliefs orient expectations,erception, information gathering and, a fortiori, decisions forction (Smith and Hancock, 1995).

nd Prevention 49 (2012) 96– 104 97

The literature gives a few clues as to this way of functioningmarked by the drivers’ practices. This notably concerns the impactof dual use of a PTW and a car on drivers’ performances when actingas car drivers. For example, by concatenating data from car driversconcerning their hazard perception, Armsby et al. (1989) put for-ward the idea that car drivers who also ride a PTW (or have riddenone) are able to identify specific road features and specific actionsby other users as risk factors for motorcyclists, whereas car driverswho do not have any experience with PTWs focus on other drivers’behaviours. In other words, experience with the specific risks runwhen driving a particular vehicle orients prior representations ofrisk factors. Very recent studies have shown that dual drivers, com-pared with car drivers who have no experience with PTWs, havebetter skills and performances in their visual search for informationand behaviours toward PTWs that are better suited (Crundall et al.,in press; Shahar et al., in press). Dual drivers, for example, use theirrear-view and right-side mirrors more and adopt an information-gathering strategy that is better suited to PTWs coming from theright (Shahar et al., in press). The effects in driving situations arebacked up by several analyses of accident involvement, which showthat car drivers who have or have had dual driving experience areless represented in accidents involving PTWs than those who donot have such experience (Comelli et al., 2008; Magazzù et al.,2006; Weber and Otte, 1980). This phenomenon also exists whenthey themselves do not ride a PTW, but a member of their fam-ily or a close friend does (Brooks and Guppy, 1990). According toCrundall et al. (2008), experience with and exposure to PTWs sup-plies schemata for managing or dealing with a variety of drivingsituations. A schema represents the know-how accumulated by anindividual concerning a situation and provides directives, guide-lines for action and self-imposed rules on the behaviours to beadopted in these situations.

Exclusive driving of one vehicle or dual use of another typeof vehicle stands out as an influential characteristic in develop-ing knowledge and beliefs prior to driving situations concerningroad risk factors. Other parameters are also useful in understandingthese prior representations. The type and category of PTW riddenconstitute two of these parameters because they largely determinethe use made of this vehicle. In fact, the use made of a small scooteris not the same as that of a high-powered motorcycle. The usemade of a vehicle leads to overexposure to certain particular drivingsituations (familiar situations) and underexposure to others.

Knowledge and beliefs prior to interaction situations are elab-orated or adopted during driving and therefore are largely shapedby practices (Mundutéguy and Ragot-Court, in press). Most studiesusually do not take into account the very wide diversity of PTWs,generically mentioning “PTW riders” or “motorcyclists” in theircomments and put forward conclusions that are often generalisedto cover all PTW riders, as is also the case for crash data in manyjurisdictions (Haworth, in press).

Furthermore, knowledge and beliefs about risks may depend onthe threat constituted by other road users. The representation ofthis threat related to the presence of other road users and the occur-rence of interferences linked to their movement is not insignificantin the context of road interactions between PTW riders and cardrivers. Threatening situations “can be characterised in general bythe presence of a perceived danger in a subject’s (or group’s) envi-ronment. This danger may be real or imaginary, and may have anociceptive or stressful action or influence on the subject (or sub-jects) in question” (Mannoni, 2004). The threat that PTWs or carsrepresent can be linked to their frequency in traffic and to the cor-responding expectations of encountering them. This situation can

be combined with the design of the vehicles themselves. This leadsto different driving behaviours which can increase their drivers’accident risks. It also leads to a variation in real and perceivedvulnerability between the different users.

9 lysis a

ptscOafdsetawTfht

2

2

�Tdptab

-

-

2

dtoe

vrw

olew

8 I. Ragot-Court et al. / Accident Ana

Thus, the assertion that practices orient expectations and theerception of risk factors in situations and their prior representa-ions gives rise to several hypotheses in this study. (1) PTW ridershould be perceived less and mentioned less as a threat by thear drivers interviewed than car drivers are by PTW riders. (2)n the other hand, the improved cognitive conspicuity of PTWsmong dual drivers should attenuate this effect. We should there-ore observe that, compared with car drivers who only drive a car,ual drivers mention PTW riders more often as being a possibleource of risk. (3) Drivers’ prior knowledge and beliefs are alsoxpressed through the specificity of the risk factors they attributeo other road users. More specifically, PTW riders should express

wider variety of risk factors attributed to car drivers due to theider variety of PTWs and related practices compared to cars. (4)

his heterogeneity of risk factors should correspond to drivers’ dif-erent practices, notably in relation to the specific road networkabitually used and the behaviours adopted and/or observed onhese networks.

. Method

.1. Sample

The initial sample comprised 132 PTW riders (average age = 35; = 9.2) and 94 car drivers (average age = 37.6; � = 10.8), all males.he purpose of this sampling was to constitute sub-groups ofrivers with particular characteristics. Approximately 30 driverser sub-group were recruited in order to be able to perform sta-istical comparisons between these sub-groups. The characteristicsdopted for their probable influence in establishing knowledge andeliefs prior to driving situations are:

Exclusive PTW riding (n = 62) or car driving (n = 33) vs. dual driv-ing of a PTW and a car. According to this characteristic, drivers arecalled “exclusive drivers” or “dual drivers”. The dual driver sub-group was then divided in 2 with the random selection method:“dual” drivers who were questioned according to their PTW riding(these will be called “dual” PTW riders, n = 70) and “dual” driverswho were questioned according to their car driving (these will becalled “dual” car drivers,2 n = 61).

The type and engine capacity of the PTW used. On the one hand,a distinction is made between motorcycle riders and scooter rid-ers (n = 67 and n = 65, respectively). For these two types of PTW,the engine capacity was defined as 125 cm3 for the “small” cat-egory (n = 69, n = 32 for motorcycles and n = 37 for scooters). Forthe medium/large categories (n = 63), it was defined as �500 cm3

for motorcycles (n = 35) and �200 cm3 for scooters3 (n = 28).

.2. Material

In a questionnaire, respondents were first asked about type andate license, years of driving experience,4 own driving and difficul-

ies on the road, and then were asked to express themselves freelyn what they consider as “riskiest” for drivers in their generic refer-nce group. The generic reference group corresponds to PTW riders

2 So as not to skew the sample of dual car drivers in relation to their non-targetehicle driving, a homogenous distribution between motorcycle riders and scooteriders and between medium/large and small motorbikes for the two types of PTWas required during recruitment.3 Initially, the medium and large category was defined as ≥500 cm3 for both types

f PTW. The scarcity of scooters with ≥500 cm3, even in the Paris region, however,ed us to extend the “medium or large” scooter target to 200 cm3 and greater. Consid-ring the wider range of engine capacity for motorcycles (up to 1600 cm3 comparedith a maximum of 850 cm3 for scooters), this extension only concerns scooters.4 How long they had been driving was not taken into account in our analysis.

nd Prevention 49 (2012) 96– 104

or car drivers in general. It is the group of PTW riders for exclusiveor dual PTW riders and car drivers for exclusive or dual car drivers.The open-ended question adapted to each driver group was thefollowing:

For exclusive and dual PTW drivers:“In your opinion, what is riskiest for PTW riders?”For exclusive and dual car drivers:“In your opinion, what is riskiest for car drivers?”

The open-ended question method, with no particular orien-tation, is of interest for recording the drivers’ perceived risks. Itprovides access to the most accessible elements in each person’smemory in a way that is the least constraining as possible forthe respondents. This accessibility is determined by the central-ity of the element related to the subject in question, here “whatis riskiest”, or by the proximity in time of an event that may berelated to the question. The question asked activates the partici-pants’ associated “central nucleus”, as the term is used by Abric(1993) in the theory of social representations. Starting with a smallnumber of elements, the whole representative field of risks indriving situations is structured. We can also see that the formula-tion (“riskiest”) is unambiguous in people’s everyday language andthat the question deals with the perceived risk for each driver’sgeneric reference group rather than for themselves. This methodis designed to encourage their verbal account while reducing theusual biases involved in presenting oneself.

2.3. Interview process

The interviews were carried out in the Paris region betweenSeptember 2006 and the end of February 2007. These were individ-ual face-to-face interviews with PTW riders and automobile driversselected according to the aforementioned criteria. These interviewswere carried out by some 10 interviewers, all men and all profes-sionals. They introduced themselves to the participants by sayingthat they were carrying out a totally anonymous survey of differentroad users. The interviews took place at the participants’ homesor in any other location proposed by them for their convenience(office, etc.). The data from the entire questionnaire were gatheredusing the “pencil-on-paper” method, i.e. by real-time gatheringof the participants’ responses, reactions and comments. Specifi-cally concerning the question that interests us in this study, theinterviewers received a strict instruction to make a word-for-wordtranscription of what the respondents said about perceived risk,with no interpretations and no biased interventions.

2.4. Processing

An initial processing phase consisted in analysing the thematiccontent of all the answers given by the respondents. Encodingwas performed by three researchers from three disciplines incomplementary human and social sciences (social psychology,ergonomics and cognitive psychology). First, the three researchersindependently encoded the responses provided by the same 50respondents. Then the initial encoding results were analysed collec-tively and turned out to be homogenous for the three researchers.The answers from the other respondents were then randomlydistributed among the researchers and encoded individually. Thecases for which there were doubts were discussed collectively. Theanswers gave the risk factors perceived by the drivers interviewedin the road environment for their generic reference group. In this

paper, only those risk factors related to the target-other have beenretained. Thus, we have retained those risk factors attributed to cardrivers in the verbal accounts of PTW riders, and the risk factorsattributed to PTW riders in the verbal accounts of car drivers. The

I. Ragot-Court et al. / Accident Analysis and Prevention 49 (2012) 96– 104 99

Table 1Categorisation of the specific risk factors mentioned in the respondents’ verbalaccounts.

Categories Specific risk factors

Internal Conditionattributed

Distraction/inattention (telephone, GPS, etc.),Fatigue, Overconfidence, Apprehension

Type of Driving Dangerous/irresponsible, Not obeying trafficregulations (signalling, right-of-way, etc.),Speeding, Unpredictable, Ill-suited driving dueto a misunderstanding of oneself by the other

Detectability orDetection

Not seen by others, Not seeing others, Notchecking behind (rear view mirror, blind spot)

Situations (more orless continuingconditions)

Overtaking a vehicle or a line of vehicles,Overtaking on the right, Weaving throughtraffic, Running up the line, Behaviour blockingothers’ movement (following somebodyclosely)

Events (more orless suddenconditions)

Not signalling a manoeuvre, Sudden braking,Failure to yield, Changing lanes/leaving a lineof traffic, Opening a door, Behaviour blocking

rto

tooatqTt

3

3

qntsdd

tTsjtod(bo

oDecbm

gw

others’ movement (cutting in front ofsomebody voluntarily)

isk factors were then put into various risk categories. Part 3.1 ofhe results presents the risk factors taken from the verbal accountsf the drivers interviewed and their grouping into risk categories.

In a second processing phase (parts 3.2 and 3.3 of the results),he scope and particularities of the threat represented by the target-ther were studied. The groups of respondents were compared (a)n the frequency with which the target-other was mentioned as

potential source of risk, (b) on the risk factors attributed to thearget-other as such in the verbal accounts and (c) on the subse-uent grouping of the risk factors mentioned into risk categories.he Chi-square test for independence was used for its relevance tohe nature of the comparisons made.

. Results

.1. Risk factors mentioned

In the verbal accounts of the participants who answered theuestion asked, the mention of the target-other is always con-ected to specific risk factors associated with him. Table 1 showshe different categories. Within these (first column of the table), thepecific risk factors mentioned (second column of the table) mayiffer depending on the type and practices of the respondents (carrivers, PTW riders, exclusive drivers, etc.).

Risk factors associated with other users may be brought uphrough the characteristics attributed to them or to their driving.his type of attribution can orient expectations about interactionituations. In the respondents’ verbal accounts, we can see valueudgements on a permanent or transient Internal Condition of thearget-other (distracted, inattentive, tired, etc.). Value judgementsn “Types of Driving”, in other words the factors perceived asetermining the driving behaviour in relation to the risk it causeshigh speed, unpredictable driving, not obeying the rules, atypicalehaviour, ill-suited driving due to the other’s misunderstandingf oneself) also appear.

The representation of the risk can also refer to various phasesf the expected interaction. In the verbal accounts, problems ofetection or Detectability are mentioned (not being seen by oth-rs, not seeing others, not checking in the rear-view mirrors). Thisategory includes risk factors that are related to these problems,ut are mentioned in a decontextualised manner, i.e. outside any

ention of a specific interaction.Lastly, when the interaction is mentioned, it can be distin-

uished by its duration. Risk factors may correspond to Situations,hich are more or less continuing conditions (overtaking a vehicle

Fig. 1. Distribution of the mention of the target-other as a source of risk in the PTWriders and car drivers’ (dual and exclusive) verbal accounts.

or a line of vehicles, overtaking on the right, weaving through traf-fic, running up the line), or Events, which are more or less suddenconditions (not signalling a manoeuvre, sudden braking, failure toyield, changing lanes, leaving a line of traffic, opening a door).

The categorisation presented here encompasses the difficultyof the inclusive or exclusive character of the factors mentioned.For example, while some drivers mentioned the “unpredictability”of the target-other as a risk factor for their own generic referencegroup, others did not explicitly make this value judgement eventhough they mentioned specific interaction situations that consti-tute a threat because they are unpredictable. This difficulty, whichis related to determining cognitive and motivational factors under-lying the mention of elements in terms of precision, will be takeninto account in the interpretation of the results.

3.2. Other-threat perception: car drivers vs. PTW riders

The results show that the exclusive and dual PTW riders inter-viewed mention car drivers as being a source of risk for their genericreference group more than exclusive and dual car drivers mentionPTW riders (�2 = 8.44; p < .01). This differentiating effect is statisti-cally greater between the verbal accounts of exclusive car driversand those of all PTW riders (�2 = 9.10; p < .001) than between theverbal accounts of dual car drivers and those of all PTW riders(�2 = 3.60; p = NS) (see Fig. 1).

The following results only concern those drivers, both exclusiveand dual, who mentioned the target-other as potentially repre-senting a threat on the road for drivers in their generic referencegroup.

The risk factors that car drivers attribute to PTW riders mainlyconcern their Type of Driving (62.8%) and their behaviours, whichare categorised in Situations (60.5%). Proportionally, much fewerof them mentioned factors categorised in other risk categories(PTW riders’ behaviours categorised as Events (18.6%), Detec-tion/Detectability problems (11.6%) and Internal Condition (4.7%).

The risk factors that PTW riders attribute to car drivers aremore diversified. They firstly concern car drivers’ behaviours corre-sponding to Events (66.3%) and then, more generally, their Type ofDriving (46.5%), Detection/Detectability problems (33.7%) and thecar drivers’ Internal Condition (23.3%). Car drivers’ behaviours cor-responding to Situations are only mentioned by 5.8% of the PTWrider respondents (see Table 2).

3.3. Risk factors attributed to other: car drivers vs. PTW riders

This section provides information on the frequency with whichthe drivers interviewed mentioned risk factors, summed up incategories, to characterise the target-other. Then, within the risk

100 I. Ragot-Court et al. / Accident Analysis and Prevention 49 (2012) 96– 104

Table 2Distribution of risk factors attributed to the target-other and related risk categories.

Risk factors attributed tocar drivers by PTW riders(n = 86) and associated riskcategories (%)

Risk factors attributed toPTW riders by car drivers(n = 43) and associated riskcategories (%)

p

Internal Condition attributed 23.3 4.7 <0.01a

Distraction/Inattention 22.1 2.3 <0.01b

Fatigue 1.2 0.0 0.7b

Overconfidence 0.0 2.3 0.7b

Apprehension 1.2 0.0 0.7b

Type of Driving 46.5 62.8 0.08a

Dangerous/Irresponsible 10.5 9.3 0.9b

Not obeying traffic regulations 5.8 9.3 0.7b

Speeding 1.2 23.3 <0.01b

Unpredictable 15.1 34.9 0.02a

Ill-suited driving due to a misunderstanding of oneself by others 25.6 0.0 <0.01b

Detectability or Detection 33.7 11.6 <0.01a

Not seen by others 10.5 0.0 0.07b

Not seeing others 0.0 11.6 <0.01b

Not checking behind 22.1 0.0 <0.01b

Situations 5.8 60.5 <0.01a

Overtaking 3.5 18.6 0.01b

Overtaking on the right 0.0 11.6 <0.01b

Weaving through traffic 0.0 30.2 <0.01b

Running up the line 0.0 9.7 <0.01b

Behaviour blocking others’ movement (following somebody closely) 2.3 2.3 0.5b

Events 66.3 18.6 <0.01a

Not signalling a manoeuvre 33.7 4,7 <0.01b

Sudden braking 5.8 2.3 0.7b

Failure to yield 7.0 2.3 0.5b

Changing lanes/leaving a line of traffic 45.3 9.3 <0.01b

Opening a door 8.1 0.0 0.1b

Behaviour blocking others’ movement (cutting in front of somebody voluntarily) 1.2 0.0 0.7b

ch

3

IrtPao

3

DPsat

ttgr

dct“t

a �2-test.b �2-test with Yates’ correction.

ategories, the results show the specific risk factors attributed toim (see Table 2).

.3.1. “Internal Condition of the target-other” categoryElements in the respondents’ verbal accounts referring to the

nternal Condition of the target-other are mentioned more by PTWiders than by car drivers (�2 = 7.01, p < .01). This mainly concernshe “distracted/inattentive” character of car drivers. 22.1% of theTW rider respondents mentioned this characteristic of car driverss constituting a risk factor for the PTW drivers that they encountern the road.

.3.2. “Types of driving by the target-other” categoryOn the other hand, risk factors corresponding to the Types of

riving by the target-other are mentioned practically as much byTW riders as by car drivers (�2 = 3.04, p = .08). Nonetheless, thepecificity of these factors, or at least the frequency with which theyre mentioned, differs between the PTW riders’ verbal accounts andhe car drivers’ verbal accounts.

PTW riders mainly mention “ill-suited driving by car drivers dueo their misunderstanding of PTWs” (25.6%). To a lesser degree,hey also mention their “unpredictable” driving (15.1%), their “dan-erous/reckless” driving (10.5%) and their “not obeying trafficegulations” (5.8%).

From the car drivers’ point of view, it is mainly “unpredictable”riving by PTW riders (34.9%) and their “high speed” (23.3%) that

onstitute risk factors for the automobile drivers they meet onhe road. Their “not obeying traffic regulations” (9.3%) and theirdangerous/irresponsible” driving (9.3%) are also mentioned andherefore feared, but to a lesser degree.

3.3.3. “Detectability/Detection problem” categoryDetectability/Detection problem are indicated more by PTW

riders than by car drivers (�2 = 7.21, p < .01). More particularly,22.1% of PTW riders mentioned car drivers’ “not checking intheir rear-view mirrors”. To a lesser degree, they point out theirown lack of detectability, rather considering “not being seen”as a problem (10.5%). Car drivers sometimes admit “not seeing”PTWs (11.6%).

3.3.4. “Situation” categoryCar drivers (vs. PTW riders) are those who mention certain Sit-

uations more (�2 = 46.9, p < .0001). PTW riders’ behaviours whichconsist in “weaving through traffic” (30.2%) and, to a lesser degree,“overtaking” (18.5%), including “overtaking on the right” (9.7%) and“running up the line” (9.3%) are perceived as risk factors for cardrivers.

3.3.5. “Events” categoryBehaviours categorised as Events, on the other hand, are

mentioned more by PTW riders than by car drivers (�2 = 28.23,p < .0001). More particularly, PTW riders fear lateral movementsrepresented by “changing lanes/pulling out” (45.3%). In nearly58% of cases, these behaviours are associated with “not signallingmanoeuvres” by car drivers. Furthermore, 33.7% of PTW riders men-tioned this specific risk factor.

3.4. Other threat perception dependent upon car driver or PTWrider characteristics

The results indicate that the threat that PTW riders representdoes not vary depending on the characteristics of the car driversinterviewed. 50.8% of dual car drivers mentioned PTW riders as a

I. Ragot-Court et al. / Accident Analysis a

56.7%

73.8%

63.8%66.7% 66.1% 64.3%

0%

10%

20%

30%

40%

50%

60%

70%

80%

motorcycle scooter small medium/large exclusive dual

Type of PTW Engine capacity Practice

Frp

s1

of

Pdc

mp

3g

c

fit

ta

at

Epa(iasepdmOaw

Dmec

ig. 2. Distribution of the mention of the car drivers as a source of risk in the PTWiders’ verbal accounts, depending on the vehicle’s caractéristics and the riders’ractice.

ource of risk compared with 36.4% of exclusive car drivers (see Fig.). This difference is not statistically significant, however.

Nor did dual and exclusive car drivers differ for the specificityf the risk factors that they attribute to PTW riders, or in how theseactors were grouped into risk categories.

The results also indicate that whether the rider drives a smallTW or a medium/large PTW and whether they are exclusive or dualrivers does not lead to any significant variation in their mentioningar drivers as a source of risk.

On the other hand, we can see that car drivers are mentionedore often by scooter riders than by motorcyclists (�2 = 4.26,

< .05) (Fig. 2).

.5. Risk factors attributed to car driver-other by different PTWroups

The following results only concern PTW riders who mentionedar drivers as potentially representing a threat (see Table 3)

Dual and exclusive PTW riders do not differ in terms of the speci-city of the risk factors they attribute to car drivers, nor in howhese factors were grouped into risk categories.

On the other hand, some differences can be observed betweenhose who ride small PTWs and those who ride medium/large PTWss well as between those who ride scooters and motorcycles.

Medium/large PTW riders mention “ill-suited driving due toutomobile drivers’ misunderstanding of them” significantly morehan do small PTW riders (�2 = 4.43, p < .05).

Scooter riders mention the risk factors associated to thevents category significantly more than do motorcyclists (�2 = 8.07,

< .05). Scooter riders thus tend more to mention behaviours byutomobile drivers that consist in “not signalling a manoeuvre”39.6% vs. 26.3%), “refusing to yield” (10.4% vs. 2.6%), “chang-ng lanes” (47.9% vs. 42.1%) and, even more specifically, “opening

door” (14.6% vs. 0%, �2 = 5.63, p < .05). These results, whichhow a distinction between scooter riders and motorcyclists, areven more specifically marked for exclusive drivers (�2 = 5.60,

< .05) and medium/large PTW riders (�2 = 9.81, p < .005). Theseifferences can specifically be observed for the “not signalling aanoeuvre” factor, which is pointed out more by scooter riders.n the other hand, no difference is observed in the risk factorsttributed to car drivers between scooter riders and motorcyclistshen they ride a small PTW and when they are dual drivers.

Lastly, the category of risk factors related to Detection and

etectability problems is mentioned equally by scooter riders andotorcyclists (�2 = 1.67, p = NS). It is interesting to observe, how-

ver, that, from motorcyclists’ point of view, “not being seen” byar drivers is more what constitutes the risk factor (15.8% vs. 6.3%,

nd Prevention 49 (2012) 96– 104 101

�2 = 1.27, p = NS), whereas scooter riders blame car drivers for their“not checking” for them (31.3% vs. 10.4%, �2 = 4.16, p < .05).

4. Discussion

The results presented in this paper are a partial analysis of theanswers provided by PTW riders and car drivers on what theyperceive as being “riskiest” for their respective generic referencegroups. These results focus on the risk factors attributed to cardrivers in PTW riders’ verbal accounts and the risk factors attributedto PTW riders in car drivers’ verbal accounts. These data deal withpart of the knowledge and beliefs that the different individuals haveformed or adopted while driving. They orient drivers’ perceptionof the situation and underlie their interactions with other users(Mundutéguy and Darses, 2007; Mundutéguy and Ragot-Court, inpress; Ragot et al., 2009).

4.1. The threat represented by others: a determining factor incognitive conspicuity?

First of all, we should point out that, in general, the car driversinterviewed mentioned PTW riders less than PTW riders mentionedcar drivers. In other words, and in keeping with our first hypothesis,PTW riders are perceived less as a threat for car drivers than cardrivers are for PTW riders. This result may seem like a truism, butis deserves a certain number of considerations in order to gain adeeper understanding of the concept of cognitive conspicuity.

In the literature, the cognitive conspicuity of PTWs is mainlyexplained by the low probability of encountering them due to theirrelative rarity in traffic and/or by the interest they attract and theirsignificance for other users (Van Elslande et al., 2008; Hancock et al.,1990). Seen from these two angles, indicators of the weak conspicu-ity of PTWs can be observed in our results, even though the studywas located in the large urban area of Paris region where car-PTWinteractions are more frequent than elsewhere. Indeed, there aremore PTWs in large urban areas than elsewhere, and as a corollarythe frequency of interactions between car drivers and PTW ridersis greater. As is also the case elsewhere, this type of interaction isstill less common here than interactions between car drivers. Andtherefore, contrary to what can be said for PTW riders, the scenariosactivated in car drivers’ prior representations may take less accountof the possibility of interacting with PTWs (Van Elslande et al., 2008)and the related risk factors. Furthermore, our results are in agree-ment with the argument by Hancock et al. (1990). Indeed, we haveobserved that exclusive car drivers mention PTW riders less thandual car drivers do, compared with car drivers being mentioned byPTW riders. This result can be interpreted as an indicator of PTWs’weaker cognitive conspicuity for exclusive car drivers than for dualcar drivers. This result corroborates our second hypothesis and canbe interpreted in line with the results by Brooks and Guppy (1990),meaning that riding a PTW increases the presence of PTW riders incar drivers’ prior representations.

Alongside what has been said, another determinant of weakPTW conspicuity can be suggested given the specific resultsobtained in this study. Whether or not the target-other is men-tioned in the answers provided by the drivers may be based onthe difference in real and perceived vulnerability between the twotypes of users. Indeed, between cars and PTWs, there are interac-tions between two categories of vehicles with different masses anddifferent designs in terms of user exposure. Experimental resultssupport the idea that the difference in real and perceived vulner-

ability between the two types of users can lead to variations inhow they apprehend their interactions. For example, Keskinen et al.(1998) show that the space for merging into the traffic stream cho-sen by a car driver faced with an oncoming PTW is smaller than

102 I. Ragot-Court et al. / Accident Analysis and Prevention 49 (2012) 96– 104

Table 3Distribution of risk factors attributed to car drivers by PTW rider groups, and distribution of the associated risk categories.

Risk factorsattributed to cardrivers by scooterriders (n = 48), andassociated riskcategories (%)

Risk factorsattributed to cardrivers bymotorcyclists(n = 38), andassociated riskcategories (%)

p Risk factorsattributed to cardrivers by smallPTW riders (n = 44),and associated riskcategories (%)

Risk factorsattributed to cardrivers bymedium/high-powered PTWriders (n = 42), andassociated riskcategories (%)

p

Internal Conditionattributed

20.8 26.3 0.6a 18.2 28.6 0.3a

Distraction/Inattention 18.8 26.3 0.4a 15.9 28.6 0.2a

Fatigue 2.1 0.0 0.9b 0.0 2.4 >0.9b

Overconfidence 0.0 0.0 – 0.0 0.0 –Apprehension 2.1 0.0 0.9b 2.3 0.0 >0.9b

Type of Driving 50.0 42.1 0.5a 36.4 57.1 >0.05a

Dangerous/Irresponsible 8.3 13.2 0.7b 6.8 14.3 0.4b

Not obeying trafficregulations

8.3 2.6 0.5b 2.3 9.5 0.3b

Speeding 2.1 0.0 0.9b 2.3 0.0 >0.9b

Unpredictable 14.6 15.8 0.4a 15.9 14.3 0.8a

Ill-suited driving due to amisunderstanding ofoneself by the other

22.9 28.9 0.5a 15.9 35.7 0.03a

Detectability or Detection 39.6 26.3 0.2a 38.6 28.6 0.3a

Not seen by others 6.3 15.8 0.3b 9.1 11.9 >0.9b

Not seeing others 0.0 0.0 – 0.0 0.0 –Not checking behind 31.3 10.5 0.02b 27.3 16.7 02a

Situations 8.3 2.6 0.5b 4.5 7.1 >0.9b

Overtaking 6.3 0.0 0.2b 2.3 4.8 >0.9b

Overtaking on the right 0.0 0.0 – 0.0 0.0 –Weaving through traffic 0.0 0.0 – 0.0 0.0 –Running up the line 0.0 0.0 – 0.0 0.0 –Behaviour blocking others’

movement (followingsomebody closely)

2.1 2.6 0.6b 2.3 2.4 0.5b

Events 79.2 50.0 0.04a 68.2 64.3 0.7a

Not signalling a manoeuvre 39.6 26.3 0.2a 38.6 28.6 0.3a

Sudden braking 6.3 5.3 0.8b 2.3 9.5 0.3b

Failure to yield 10.4 2.6 0.1b 4.5 9.5 0.6b

Changing lanes/leaving aline of traffic

47.9 42.1 0.6a 47.7 42.9 0.7a

Opening a door 14.6 0.0 0.04b 11.4 4.8 0.5b

Behaviour blocking others’movement (cutting infront of somebodyvoluntarily)

2.1 0.0 0.9b 0.0 2.4 >0.9b

a 2

tvtwcwsrtcorowpMncatc

� -test.b �2-test with Yates’ correction

he choice the same car driver would make faced with a largerehicle. We think, like Rogé et al. (2009), that the seriousness ofhe consequences perceived by the car driver in case of a crashith the oncoming vehicle can explain this result. Indeed, the per-

eived consequences of a crash with a PTW are less serious. Thus,e have formulated the idea that the design of the vehicles them-

elves could contribute to the level of threat they represent on theoad for the other users. In other words, prior to real driving situa-ions, the physical threat level that an object represents for oneselfould in part determine its prominence in our representative fieldf perceived risks. In driving situations, the threat that an objectepresents therefore would constitute an attentional goal for thebserver. This would improve its detection and, more generally,ould modify how it is taken into account. This process may explainart of the “looked but didn’t see” accidents (Koustanaï et al., 2008).oreover, Näätänen and Summala (1976) suggest that operators do

ot deal with small risks because, through an unintentional pro-

ess, they do not see them. In the road environment, where risksre everywhere whereas the user’s cognitive resources are limited,he threat represented by an object constitutes a determinant of itsognitive conspicuity.

4.2. Risk factors attributed to others: an illustration of themisunderstanding of determinants in driving situations

On the road, PTWs lead to large differentials (size, speed, specificbehaviours) that contribute to their vulnerability. Consequently, indriving situations, PTW riders have to apply a maximum amount ofattentional resources and concentration to ensure their safety. Thisis especially necessary since car drivers may adopt behaviours thatput them in danger (Van Elslande, 2003). The elements that PTWriders mention as being “riskiest” for their generic reference grouprefer to these failings by car drivers. These failings are mentionedon several levels in the PTW riders’ verbal accounts and the hetero-geneity of risk factors mentioned confirms our third hypothesis.These are judgements concerning car drivers’ Internal Conditions(distracted/inattentive attitude by car drivers due to activities thatsidetrack them from the driving task) or their Type of Driving,with a strong feeling of ill-suited driving due to car drivers’ mis-

understanding of PTWs. This last point emphasises something thatthe other points deflect: car drivers’ ignorance of the constraintsand driving styles that result from the physical and dynamic char-acteristics of PTWs. In terms of interactions, sudden Events are

lysis a

caPwtcnaiaooditpisrmct

4d

(pufimtrotoastd

4rt

cTm

eFinuPicmv

tbatts

I. Ragot-Court et al. / Accident Ana

onsidered “riskiest” for PTW riders. The unpredictability of thections performed by car drivers constitutes what is “riskiest” forTW riders. Among these Events in interaction situations, those inhich the car driver deviates from expected behaviour according

o the traffic regulations (signalling a change of lanes or direction,omplying with right-of-way at intersections, sudden braking) areotably mentioned. These deviations lead to changes in the char-cteristics of the driving situation that are hard to foresee and cannterfere with other drivers’ objectives. Dealing with these devi-tions also requires a great deal of manoeuvrability, which PTWsften lack in critical situations. Car drivers’ ignorance may be at therigin of misunderstandings or the adoption of ill-suited or evenangerous behaviours toward PTW riders. On the other hand, it is

nteresting to see how PTW riders generally talk about the detec-ion problems that affect their peers. In fact, they tend to link thisroblem to car drivers’ failure to check for them rather than their

ntrinsic weak detectability, which has been objectively demon-trated in the literature (Van Elslande, 2003). In other words, PTWiders also show an ignorance or misunderstanding of the deter-inants in driving situations. This ignorance or misunderstanding

an lead to unintentional risks through the behaviours that theyhemselves adopt.

.2.1. A few clues as to the perception of the risks marked byrivers’ practices

Interactions are mentioned in reference to drivers’ practicesMundutéguy and Ragot-Court, in press). Here again, it should beointed out that the study was carried out in the Paris region. Inrban environments and on the “Périphérique” (ring road), traf-c density is often heavy and interactions involving quick lateralovements are all the more common. Frequent driving in this

ype of environment increases exposure to and awareness of theisks related to other users’ not signalling manoeuvres. The resultsbtained on the scooter rider population paint a picture of essen-ially urban driving, with the type of risks associated with this typef environment. Generally, scooter riders mention Events in inter-ction situations significantly more than motorcyclists do. Morepecifically, they also mention opening a door significantly morehan motorcyclists do, which is a purely urban behaviour and pro-uces a risk for them. This result supports our fourth hypothesis.

.2.2. Indicators of the weak behavioural conspicuity of PTWiders identified through the risk factors that car drivers attributeo them

The risk factors that car drivers attribute to PTW riders are con-entrated in two categories. These are value judgements on theirype of Driving and certain Situations in the context of specificallyentioned interactions.PTW riders are specifically characterised by their speed, consid-

red too fast, and by their driving style, considered unpredictable.rom the car drivers’ point of view, these are the “riskiest” factorsn driving situations. Car drivers’ ignorance of part of the determi-ants of the situation is expressed in the mention of the PTW riders’npredictability. We can suppose that a better understanding ofTWs and of their characteristics would reduce the unpredictabil-ty of their behaviour by improving how they take into account theonstraints that these PTW riders are faced with and their usualode of driving. This would make it easier to foresee their manoeu-

res, whereas ignorance and misunderstanding make this harder.This argument is strengthened through the specificities of

he interactions that car drivers fear the most. These are PTWehaviours that consist in weaving between cars, overtaking with

n inferred risk – partially in reference the fact that they overtake onhe right – and running up a line of traffic. All these Situations refero behaviours specific to PTWs and which deviate from behaviouraltandards. That is with this meaning that they are mentioned. Here

nd Prevention 49 (2012) 96– 104 103

again, we can consider that these situations, which are mentionedspecifically, have in common the unpredictability of PTW ridersand the fear of their coming out of nowhere. These situations can-not be foreseen unless one is fully aware of the potential presenceof PTWs and the behaviours associated with them. In any case, thesudden appearance requires the car driver to adopt the operatingprocedure that is best adapted to the situation without having thebenefit of time to prepare for it or to compensate for it, given howclose the PTW comes.

4.3. Conclusions and perspectives

The results for car drivers show the need to encourage theirawareness of the potential presence of PTWs in the road envi-ronment. Gershon et al. (in press) also reached this conclusion.While this action may have a positive effect on car drivers’detection of PTWs, it will not, however, reduce the sources ofmisunderstandings. Indeed, our results also point out car drivers’misunderstanding of PTWs and of their characteristics and theeffect they have on how the riders of these vehicles behave. Thisaction should therefore be accompanied by information on thedifferent dimensions of the interaction. More particularly, thisinformation should concern the different types of PTWs and thebehaviours which are associated with them and are related to theirrespective dynamic characteristics.

Driver awareness of the PTW presence, their behavioural char-acteristics and their vulnerability has to begin from the initialdriver training. In order to achieve it, experience feedbacks of var-ious PTW types’ riders could be planned during the training. Anexperience on road as passenger of one PTW in order to experi-ment different interaction situations with the others users could bealso considered with (or not) the previous proposal. But the safetyroad strategy does not have concerned only the novices. To awarealso more confirmed drivers, mass media campaigns could be per-formed. Each campaign could be split up into various TV spots orinserts in newspapers. Each TV spot or insert should be focused on apotential conflicting situation by considering various types of PTWand environments. So, it could be highlight the more effective visualstrategies of information search to increase the probability to detecta PTW. In addition, the generated risk when the car driver does notindicate or indicate not enough early his changes of way or direc-tion could be also highlight. The frequency of the campaigns couldimprove the PTW cognitive conspicuity by compensating their rar-ity in the traffic. By their contents, these campaigns should increasetheir behavioural conspicuity.

In terms of the research and in terms of prevention, PTW ridersneed to be considered as a heterogeneous group or as several verydifferent groups. Their relationship to risk and to the perception ofrisk are the direct result of their choice of a type of machine, whichitself is related to practices. The results show the need to reinforcePTW riders’ awareness of their intrinsic difficulty in being seen,beyond the failings by car drivers to check for them. On the otherhand, PTW riders would gain in safety if they took greater accountof the difficulties that car drivers have in foreseeing and adapting totheir behaviours. PTW riders tend to think that the behaviours asso-ciated with their type of vehicle are normal operating procedures.But for car drivers, these behaviours are very uncommon in thatPTW riders are rare in their environment and therefore constitutea low level of expectation.

This research does not allow reporting all the behaviours of thePTW riders which can surprise the car drivers. But it maintains theinterest to integrate more practice during the training of all the PTW

riders. This training should be done in so far as various environ-mental configurations in order to confront the riders with a largerpanel of potential risks which could be analysed with the trainer.The objective would be to lead the PTW riders to anticipate better

1 lysis a

to

oc

A

wBcd

R

A

A

B

C

C

C

G

H

H

H

H

J

K

K

wesen, Köln, West Germany.Williams, M.J., Hoffmann, E.R., 1979. Motorcycle conspicuity and traffic accidents.

04 I. Ragot-Court et al. / Accident Ana

hese risks and to adopt a less unpredictable driving style for thether users.

A study like the one we have presented in this paper, but carriedut in a rural environment where interactions between PTWs andars are less frequent, should reinforce these conclusions.

cknowledgements

This research was supported by the FONDATION MAIF. Weould like to thank Mrs. Marie-Hélène Sirop and Mrs. Genevièvearritault and their colleagues of the Society MV2 ETC for theironsciousness contribution in recruiting participants and collectingata.

eferences

bric, J.C., 1993. Central system, peripheral system: their functions and roles in thedynamic of social representations. Papers on Social Representations 2, 75–78.

rmsby, P., Boyle, A.J., Wright, C.C., 1989. Methods for assessing drivers’ perceptionof specific hazards on the road. Accident Analysis and Prevention 21, 45–60.

rooks, P., Guppy, A., 1990. Driver awareness and motorcycle accidents. In: Proceed-ings of the International Motorcycle Safety Conference, vol. 2, pp. 27–56.

omelli, M., Morandi, A., Magazzù, D., Bottazzi, M., Marinoni, A., 2008. Brightlycoloured motorcycles and brightly coloured motorcycle helmets reduce theodds of a specific category of road accidents: a case–control study. BioMedicalStatistics and Clinical Epidemiology 2, 71–78.

rundall, D., Bibby, P., Clarke, D., Ward, P., Bartle, C., 2008. Car drivers’ attitudestowards motorcyclists: a survey. Accident Analysis and Prevention 40, 983–993.

rundall, D., Crundall, E., Clarke, D., Shahar, A. Why do car drivers fail to give way tomotorcycles at t-junctions? Accident Analysis and Prevention, in press.

ershon, P., Ben-Asher, N., Shinar, D. Attention and search conspicuity of motorcy-cles as a function of their visual context. Accident Analysis and Prevention, inpress.

ancock, P.A., Wulf, G., Thom, D., Fassnacht, P., 1990. Driver workload during differ-ing driving maneuvers. Accident Analysis and Prevention 22, 281–290.

aworth, N. Powered two wheelers in a changing world – challenges and opportu-nities. Accident Analysis and Prevention, in press.

ole, G.J., Tyrell, L., Lanham, M., 1996. Some factors affecting motorcyclists’ con-spicuity. Ergonomics 39 (7), 946–965.

orswill, M.S., Helman, S., Ardiles, P., Wann, J., 2005. Motorcycle accident riskcould be inflated by a time to arrival illusion. Optometry and Vision Science82, 740–746.

affard, M., Van Elslande, P. Typical human errors in traffic accidents involvingpowered two-wheelers. In: Proceedings of the 27th International Congress ofApplied Psychology, Melbourne 11–16 July 2010, in press.

eskinen, E., Ota, H., Katila, A., 1998. Older drivers fail in intersections: speed dis-

crepancies between older and younger male drivers. Accident Analysis andPrevention 30, 323–330.

oustanaï, A., Boloix, E., Van Elslande, P., Bastien, C., 2008. Statistical analysis oflooked-but-failed-to-see accidents: highlighting the involvement of two distinctmechanisms. Accident Analysis and Prevention 40, 461–469.

nd Prevention 49 (2012) 96– 104

Magazzù, D., Comelli, M., Marinoni, A., 2006. Are car drivers holding a motorcyclelicence less responsible for motorcycle—car crash occurrence? A non-parametricapproach. Accident Analysis and Prevention 38, 365–370.

Mannoni, P., 2004. Les logiques du terrorismes (In Press eds). Collection psycho-polis, Paris.

Mundutéguy, C., 2001. Reconnaissance d’intention et prédiction d’action pour lagestion des interactions en environnement dynamique. Doctorat en ergonomie,Conservatoire National des Arts et Métiers.

Mundutéguy, C., Darses, F., 2007. Perception et anticipation du comportementd’autrui en situation simulée de conduite automobile. Le Travail Humain 70,1–32.

Mundutéguy, C., Ragot-Court, I. A contribution to situation awareness analysis:understanding how mismatched expectations affect road safety. Human Factors,in press.

Näätänen, R., Summala, H., 1976. Road User Behavior and Traffic Accidents. NorthHolland Publishing Company, Amsterdam.

Obenski, K.S., 1994. Motorcycle Accident Reconstruction: Understanding Motorcy-cles. Lawyers and Judges Publishing Co, USA.

ONISR, 2005. Les motocyclettes et la sécurité routière en France en 2003. La Docu-mentation Franc aise, Paris.

ONISR, 2007a. Les motocyclettes et la sécurité routière en France en 2005. La Docu-mentation Franc aise, Paris.

ONISR, 2007b. Sécurité routière. Les grandes données de l‘accidentologie 2007.http://www2.securiteroutiere.gouv.fr/IMG/Synthese/dep accidentologie.pdf.

Ragot, I., Mundutéguy, C., Fournier, J.-Y., 2009. Evocation du risque routier dansle discours spontané des conducteurs de deux-roues à moteur et des automo-bilistes. Collections de l‘INRETS, Marseille, France.

Rogé, J., Otmani, S., Bonnefond, A., Pébayle, T., Muzet, A., 2009. Effect of a short napon the alertness of young drivers: repercussion on the perception of motorcy-cles according to extent of the useful visual field of the driver. TransportationResearch Part F 12, 143–154.

Shahar, A., Von Loon, E., Clarke, D., Crundall, D. Attending overtaking cars andmotorcycles through the mirrors before changing lanes. Accident Analysis andPrevention, in press.

Smith, K., Hancock, P.A., 1995. Situation awareness is adaptative, externally directedconsciousness. Human Factors 37, 137–148.

Thomson, G.A., 1980. The role frontal motorcycle conspicuity has in road accidents.Accident Analysis and Prevention 12, 165–178.

Van Elslande, P., 2002. Specificity of error-generating scenarios involving motor-ized two-wheel riders. In: Wang, K., Xiao, G., Nie, L., Yang, H. (Eds.), Traffic andTransportation Studies, vol. 2. ASCE, Reston, pp. 1132–1139.

Van Elslande, P., 2003. Scénarios d‘accidents impliquant des deux-roues à moteur:une question d‘interaction. In: Deuxièmes Journées d‘Etude en PsychologieErgonomique-EPIQUE‘2003. INRIA, Rocquencourt, pp. 71–83.

Van Elslande, P., Page, Y., Lermine, P., 2008. La détectabilité du deux-roues motorisé.In: Gisements de sécurité routière pour les deux-roues motorisés. La Documen-tation Franc aise, Paris, pp.73–86.

Weber, H., Otte, D., 1980. Unfallauslösende Faktoren bei motorisierten Fahrrädern(Factors causing accidents with motorized bicycles). Bundesanstalt für StraBen-

Accident Analysis and Prevention 11 (3), 209–224.Wulf, G., Hancock, P.A., Rahimi, M., 1989. Motorcycle conspicuity: an evaluation and

synthesis of influential factors. Journal of Safety Research 20, 153–176.