1-s2.0-S0001457512002631-main

Accident Analysis and Prevention 50 (2013) 830 839

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

jo ur n al hom ep a ge: www.elsev ier .co

Develo e p

Marie-AxIFSTTAR-MA, 3

a r t i c l

Article history:Received 18 AReceived in reAccepted 6 Jul

Keywords:PedestrianBehaviorViolationErrorAggressionPositive behav

lop ae Pedte thearrieds werce ofms aoduce

goodstigat

amoS scorof all

1. Introduction

Even though pedestrian exposure to trafc when crossing rep-resents a small part of total walking activity, 11,000 of the 78,000deaths in trcerned pedwere killed2006 (ONISiors when c

Tools forfar between(Latrmouilway of undbehaviors ido not provbehaviors in

Anotherusing declareport on tr2003; Yagilkind of mebehaviors in2004). Thisgating pede

CorresponE-mail add

which to study a number of important issues, such as which typeof behaviors are involved in road accidents and what psychologi-cal mechanisms explained this behaviors. Few tools are available,however, and they often focus on a few transgressive behaviors by

0001-4575/$ http://dx.doi.oafc accidents in the 23 OECD countries in 2006 con-estrians (SafetyNet, 2009). In France, 535 pedestrians

and 5523 seriously injured in trafc accidents inR, 2007). Therefore, understanding pedestrian behav-rossing remains a road safety challenge.

observing pedestrian behaviors, however, are few and. While ethological observation in road environmentle et al., 2004; Zeedyk and Kelly, 2003), is still the besterstanding the effect of environment on pedestriann one given context (Sisiopiku and Akin, 2003), theyide a complete view of everyday pedestrian injury risk

different contexts. method for studying pedestrian behaviors consists inred behavior questionnaires aimed at providing a self-avel and crossing practices (Evans and Norman, 1998,, 2000; Zhou and Horrey, 2010; Zhou et al., 2009). Thisthod permits to classify the wide ranging pedestrianto a system of behavioral factors (Elliott and Baughan,

method is likely to be benecial for research investi-strian safety, because it would provide a framework in

ding author. Tel.: +33 (0)4 90 57 79 79; fax: +33 (0)4 90 56 86 18.ress: [email protected] (M.-A. Grani).

pedestrians when crossing.One of the most complete questionnaires is the one developed in

the United Kingdom on adolescent pedestrian behaviors by Elliottand Baughan (2004), validated in New Zealand (Sullman and Mann,2009), Spain (Sullman et al., 2011) and France (Abou et al., 2008).Based on a list of 43 behaviors judged to be dangerous by the expertsand on a survey of 2433 adolescents between the ages of 11 and16, it identied 21 pedestrian behaviors, differentiated into 3 axes:unsafe street crossings, dangerous games in the street and plannedprotective behaviors. It cannot be used, however, to measure pedes-trian behaviors in relation to the rules and some of the items areonly adapted for an adolescent population, such as playing chickenduring crossing behaviors. Based on this tool, a questionnaire mea-suring behaviors of pedestrians of all ages was developed in Franceand validated with adults (Grani, 2008) and adolescents (Grani,2009). Using and developing the items from Elliott and Baughan(2004), its aim was to measure pedestrian behaviors in terms ofendangerment and transgression. In its current design, however,this tool did not differentiate between errors and lapses in puttingoneself in danger. This distinction made it possible to gain a moredetailed understanding of behaviors with accident risks, dependingon intentionality and the nature of the deviation from safe behavior.

A rst differentiation of risk behaviors should indeed be madebetween violations intentional and mistakes unintentional

see front matter 2012 Elsevier Ltd. All rights reserved.rg/10.1016/j.aap.2012.07.009ping a self-reporting method to measur

elle Grani , Marjorie Pannetier, Ludivine Guho04 Chemin de la Croix Blanche, F-13300 Salon de Provence, France

e i n f o

pril 2012vised form 4 July 2012y 2012

iors

a b s t r a c t

The objective of this study was to deveiors among pedestrians of all ages. Thitems enabling respondents to evaluabehaviors. The validation study was cthe ages of 15 and 78. Factor analysesion, included items concerning offen3 comprised aggressive behavior itesion of the PBS with 20 items was prfour factors. The 20-item version hadvariables on the PBS scores were inveof these different types of behaviorschological factors used to predict PBpopulations of vulnerable road users m/locate /aap

edestrian behaviors at all ages

nd validate a self-reporting scale to measure injury risk behav-estrian Behavior Scale (PBS) was developed that included 47

frequency with which they had different types of pedestrian out on 343 participants (126 men and 217 women) betweene used to differentiate between 4 axes. Factor 1, transgres-

legal rules and errors. Factor 2 included lapses items. Factornd factor 4 included positive behavior items. A revised ver-d by selecting those items that loaded most strongly on the

internal reliability. The effects of demographic and mobilityed. This instrument will be useful in measuring the frequencyng the pedestrians who are most at risk, analyzing the psy-es and thus better adapt preventive actions to the differentages.

2012 Elsevier Ltd. All rights reserved.

M.-A. Grani et al. / Accident Analysis and Prevention 50 (2013) 830 839 831

which have different psychological origins and require differentmode of remediation (Reason et al., 1990). Thus, mistakes involvefailures of cognitive skills and could be corrected by educationalcampaigns and training courses. Transgressions involve motiva-tional factoof specic g

On a secamong mis(slips and inaction, ill-sintent, ill-sudecision-mobjective an

Among (1980), diffknowledge to nd the rules come rule for actright rule inwrong rule.

Individuin a regulatvidual cogncontext in rules and ndeliberate dthat are socation of a p1316). If thed as an econstructedviolation. Vtion from sbut their obtions are asYagil, 1998)1990).

This distined in sevQuestionnaand young committed that elderly(Gabaude eet al., 2006)developed sive behavi(zkan and

The DBQSullman et a conceptuadifferent tyet al. (2007(MRBQ) witmotorcyclisaccident ris

Based oQuestionnaPedestrian with Chileawith driverlations than2002). ThisBianchi, 20differentiat

(Yildirim, 2007). In both cases, the results showed more violationsby men. Furthermore, more errors were observed among youngpedestrians (1725 years) than among older pedestrians (2549years) (Torquato and Bianchi, 2010).

nowevenetaiall agof risrisk rian Brrorston

ldirime an rians

thod

ateri

ng thales oe drisers,ere

997;elf-rGranen 5 ntionms, es thahat ples (

the red as ik (8 just g wa

greeurin

if thd as ba drivors w6 iteping (7 ite. For ten d

veryng w, age, ow

in ag tim

therderermn, w

The0 minhe pbtainrs and should be only countered by altering attitudesroups to specic kind of transgressive behaviors.ond level, two types of behaviors can be differentiatedtakes: lapses and errors (Reason et al., 1990). Lapsesattention) are dened as involuntary deviations in theuited to the original intent. Errors concern failures inited to the situation. This concerns deciencies in the

aking or inferential processes involved in choosing and/or the resources for achieving it.these errors, Reason et al. (1990), after Rasmussenerentiated between two categories. Errors related toare produced when a trial and error procedure is neededright solution to a new problem. Errors related to thefrom an inappropriate application of a pre-establishedion to the situation: either the individual applies the

the wrong way or fails to apply it, or he applies the

als do not perform acts in isolation, however, but rathered social environment. Thus, while errors concern indi-itive processes, violations are produced in a social

which behavior is governed by procedures, practices,orms (Reason et al., 1990). Violations are dened aseviations from practices, whether formalized or not,

ially considered as necessary to maintain the safe oper-otentially hazardous system (Reason et al., 1990, p.

e deviation is not intentional, the action will be classi-rror; if the difference between the action and socially

practices is voluntary, the action will be classied as aiolations produced on the road mainly concern devia-ocial rules that have a certain degree of intentionality,jective is not to cause injury or damages. These viola-sociated with attitudes and motivations (Aberg, 1998;

and are inuenced by the social context (Reason et al.,

inction between violations, errors and lapses was exam-eral studies on drivers, using the Driver Behaviorire (DBQ) (Reason et al., 1990). They showed that mendrivers committed more violations than women, whomore errors (zkan et al., 2006; Reason et al., 1990),

drivers committed fewer violations but more errorst al., 2010) and that lapses increased with age (zkan. New dimensions of accident risk behaviors have beenin the context of the DBQ to measure drivers aggres-ors (Lawton et al., 1997) and drivers positive behaviors

Lajunen, 2005) toward other road users. has been validated in many countries (Gras et al., 2006;al., 2002; Xie and Parker, 2002) and was also used asl framework to construct tools for investigating these

pes of behavior in other types of road users. Thus, Elliott) developed a Motorcycle Rider Behavior Questionnaireh 43 items. Submitted to a sample of more than 8000ts, it has shown that errors were the main predictor ofks for powered two-wheelers.n the conceptual framework of the Driver Behaviorire (DBQ), Moyano Diaz (1997) validated a 16-itemBehavior Questionnaire (PBQ) in Chile. Using the PBQn pedestrians, he found results similar to those obtaineds: men and young pedestrians committed more vio-

women and adult pedestrians (Moyano Diaz, 1997, tool was recently validated in Brazil (Torquato and10) and a version was developed in Turkey, this timeing between violations, aggressive behaviors and errors

Foreties, more dans of types lyzing Pedesttions, eby Lawand Yiprovidpedest

2. Me

2.1. M

Usiand scpositivroad uitems wDiaz, 1other s2004; betwean inte(10 iteone lessions tlegal ruto run denethe tasI have lookinlight islapse dprecisedenethank behaviusers (In keeadded ters))how ofto 6 =

Alogenderlicenseweeklywalkintrian in

In oble in tlocatioworks.took 1kept. Tafter o, this tool has never been validated in Western soci- though it would be a useful resource for gaining aled understanding of risky behaviors among pedestri-es, studying the relationships between these differentky behaviors and pedestrian accidentology and ana-factors. The purpose of this study was to validate aehavior Scale for all ages to differentiate between viola-

and lapses among these behaviors. Based on the studieset al. (1997) and zkan and Lajunen (2005) on drivers

(2007) on pedestrians, the goal of this tool was also tounderstanding of aggressive and positive behaviors by

toward other road users.

als

e conceptual framework of the DBQ (Reason et al., 1990)f aggressive driver behaviors (Lawton et al., 1997) andver behaviors (zkan and Lajunen, 2005) toward other

a Pedestrian Behavior Scale (PBS) was constructed. Its based on existing validated versions of the PBQ (Moyano

Torquato and Bianchi, 2010; Yildirim, 2007), as well aseported Pedestrian Behavior Scales (Elliott and Baughan,i, 2008). The items in the Likert scale differentiatedtypes of pedestrian behaviors. Offences were dened asal deviation from the legal rules of pedestrian behavior

.g. I cross outside the pedestrian crossing even if there isn 50 meters away). Errors were dened as making deci-ut the pedestrian in danger, but without disobeying the11 items, e.g. I start walking across the street, but I havest of the way to avoid oncoming vehicles). Lapses werell-suited behaviors related to a lack of concentration onitems, e.g. I realize that I do not remember the routetaken). According to this denition, crossing withouts considered as a lapse behavior even when pedestriann, particularly at a crossroad. Then, items concerning

g crossing (e.g. I forget to look before crossing) did note crossing is regulated or not. Positive behaviors wereehaviors that appease social interactions (5 items, e.g. Ier who stops to let me cross). Aggressive or antisocialere dened as conicting behaviors with other road

ms, e.g. I get angry with a driver and hit his vehicle).with Torquato and Bianchi (2010), lter items werems, e.g. I walk in covered areas (such as shopping cen-each item, the participant was asked: As a pedestrian,o you have the following behaviors? from 1 = never

often.ith answering the PBS, information was gathered on, the number of years participants have had a driversnership and use of a motor vehicle, kilometers driven

motor vehicle, weekly frequency of walking and dailye, and the accident history as a driver and as a pedes-

previous 5 years. to obtain the widest, most diversied sample possi-s of age, socio-professional categories and geographicale undertook a web-based collection using social net-

questionnaire was lled out online and its completion. Only results from participants living in France werearticipants answered the questionnaire individually,ing their informed consent. The responses were totally

832 M.-A. Grani et al. / Accident Analysis and Prevention 50 (2013) 830 839

Table 1Sample size by gender for each age group.

Gender Age groups Total

Men Women Total

anonymoustime of com

2.2. Particip

The samwomen) beipants wereof participagroup (2(4

Most of(38.8%), stuity of the pvehicle (63average.

Most of tthe individuof the indivday and 23sample dec5 years.

3. Results

Means, each of themean valuethe most frpositive intwho yields and offenceing diagonastopped or behaviors wand concernas behaviorless frequentoward othothers, behof oneself, wreported inoften appeashowed thatype of beh

3.1. Tool va

In orderponent anaon all 40 iteplot indicataccounted fmeasure oftest of sphethe matrix values >1 wloading valu

The rst axis, transgression, explained 21.09% of the variance.It was dened by 15 items relative to offences (7 items) and errors(8 items) while crossing or travelling. Two offence items (item 20and item 30), loading both on this axis and on the second axis,

xclu in coethes theg: .77r time.g. cordson ers an

secoterm

errorg on oaditer ok to b

crosdingal to

to joding:ors r

thirce, w

othsion ifferesult h

get a

fource, cotioneven

it, pederobles co

resu meae befollon (.8or (.5iabilithermted es os ag

show sca

nin

maiing in1525 2635 3645 4655 56 and above

37 43 20 15 11 12661 96 22 17 21 21798 139 42 32 32 343

, as the individuals were only identied by the day andpletion of the questionnaire.

ants

ple comprised 343 participants (126 men and 217tween 15 and 78 years of age (M = 33.86). The partic-

separated into 5 age groups (Table 1). The distributionnts of either gender was equivalent within each age) = 5.58, ns).

the participants are managerial-level employeesdents (26.8%) or salaried employees (18.8%). The major-articipants have a drivers license (81.6%) and a motor.8%). They have had their licenses for 15.27 years on

he participants walk every day (54.6%). The majority ofals walk between 15 and 30 min a day (40.2%). 34.4%

iduals never drive, 23.9% of the participants drive every.3% cover between 50 and 150 km a week. 81% of thelared that they had not had an accident in the previous

standard deviations and distribution of responses to 47 items in the PBS, ranked in descending order by

are presented in Table 2. Other than 4 lter items,equently reported behavior concerned a behavior oferaction with an automobile driver (thanking a driverhis right-of-way), then behaviors categorized as errorss concerning the place and direction of crossing (cross-lly or ending the crossing diagonally, crossing betweenparked vehicles). Thus, the most commonly reportedere not the most desirables from a safety point of viewed problematic behaviors in terms of visibility as well

s violating trafc rules. Many of the behaviors that weretly reported concerned lapses and aggressive behaviorser road users. Forgetting to look or being aggressive toaviors that showed a lack of control of the situation orere not common, or, in any case, were not frequently

the study sample. Although these behaviors did notr, the analysis of the distribution of responses in Table 2t a signicant share of the sample reported having thisavior.

lidation

to explore the PBSs factorial structure, principal com-lysis with orthogonal Varimax rotation was carried outms in the scale, excluding the 7 lter items. The screeed that the data best t a four-factor solution, whichor 39.82% of the total variance. The KaiserMeyerOlkin

were eall hadior, whI crosloadinplace orules (.74). Acby Reato erro

Thewas detive toloadinitems lcharacthe tasbeforetor loaexternI wanttor loabehavi

Thevariantowardexpreswith dand in(e.g. I.57).

Thevarianinteracgo by, behindlet thethe 8 p32 itemAll the

Thepositivin the gressiobehavinal rel

Furcalculavariabloffencealphasoffence

3.2. Re

Thereport sampling adequacy was satisfactory (0.86), Bartlettsricity was signicant (0.0001) and the determinant ofwas close to zero (5.82E007). Four axes with eigen-ere identied. A cut-off point of .40 was used for itemes.

The 32-itemreporting smeasures, icompletionsuring the fded from the nal scale. The items loading on this axismmon the intentional nature of the dangerous behav-r a deliberate offence contrary to the legal rules (e.g.

street even though the pedestrian light is red, factor) or an erroneous or careless decision concerning thee of crossing, without an offence contrary to the legal

I cross the street between parked cars, factor loading:ing to the classication of aberrant behaviors drawn upt al. (1990), behaviors loading on axis 1 referred bothd violations.nd axis, lapses explained 7.73% of the variance andined by 7 items related to inattention. One item rela-s (item 10) and one item relative to inattention but alsoaxis 4 (item 41) were excluded from the nal scale. Allng on this axis were characterized by the unintentionalf the dangerous behavior and by the omission of part ofe performed, either by distraction (e.g. I forget to look

sing because I am thinking about something else fac-: .79), or by focusing on a competing task or a situation

the task (e.g. I forget to look before crossing becausein someone on the sidewalk on the other side fac-

.74). In the classication by Reason et al. (1990), theseeferred to lapses.d axis, aggressive behaviors, explaining 6.17% of theas determined by 5 items of aggressive behaviorser road users. The items on this axis concerned theof negative emotions leading to aggressive interactionnt types of road users (e.g. I get angry with another userim, factor loading: .76) or specically toward driversngry with a driver and hit his vehicle, factor loading:

th axis, positive behaviors, explaining 4.84% of themprised 5 positive behavior items aimed at facilitatings with other road users, whether drivers (e.g. I let a car

if I have the right-of-way, if there is no other vehiclefactor loading = 61) or other pedestrians (e.g. I stop tostrians I meet by, factor loading = 49). After eliminatingmatic items and another PCA with Varimax rotation, themprised 4 axes that explained 43.83% of the variance.lts are presented in Table 3.n scores for items of transgression, lapse, aggressive andhaviors were calculated and used as composite scaleswing analyses. Cronbachs alphas, calculated for trans-9), lapses (.83), aggressive behavior (.70) and positive3) items indicated that the scales had acceptable inter-ty, except for the positive behavior scale.ore, for the transgression axis, two sub-scores were

to observe the effect of demographic and mobilityn the two dimensions of this score: items concerningainst the legal rules and error items. Here again, theed acceptable internal reliability ( = .84 and = .79 for

le and error scale respectively).

g the scale

n objective of the study was to develop a reliable self-strument for measuring pedestrians risky behaviors.

version may not be practical for use in future self-tudies because, when used with other self-reportingt can lead to questionnaires that are too long in terms of

time. A shorter version that is capable of reliably mea-actor structure reported above would be more practical


Table 2PBS behavior items: means, standard deviations and distribution of responses.

Item (how often do you. . .) M SD Distribution of responses (%)

1 2 3 4 5 6

I thank a driI walk outdoI take publicI walk withoI walk for thI start to croI cross betwI cross the stI look at theI stop to let tI cross the stI cross diagoI cross outsiWhen I am a ot

to bother I walk in covI walk on theI let a car goOn a two-wa

second paI walk accomI walk on the

walking slI cross whileI cross even I walk becauI start walkiI walk on theI avoid usingI cross even I cross even I cross withoI lose my waI realize thatI get angry wI walk in a wI hit a pedesI take passagI walk on cyI cross very sI realize thatI get angry wI forget to loI cross withoI deliberatelI get angry wI forget to loI run across I get angry wI walk on bu

to use. For (offences, eiors) with tand these 2went anoththe items reon the rsttor, aggressthird and fotors explainitems of the(1525 yearcomparableyses are nothat in eachtransgressioon the othetor, and aggver who stops to let me cross ors

transportation (buses, metro, tramway, etc.) ut being accompaniede pleasure of itss on a pedestrian crossing and I nish crossing diagonally to save time een vehicles stopped on the roadway in trafc jams reet between parked cars

trafc light and start crossing as soon as it turns red he pedestrians I meet byreet even though the pedestrian light is rednally to save time de the pedestrian crossing even if there is one less than 50 m awayccompanied by other pedestrians, I walk in single le on narrow sidewalks so as nthe pedestrians I meetered areas (such as shopping centers)

right-hand side of the sidewalk so as not to bother the pedestrians I meet by, even if I have the right-of-way, if there is no other vehicle behind it y street, I cross the rst part and wait in the middle of the roadway to cross thertpanied by other people roadway to be next to my friends on the sidewalk or to overtake someone who isower than I am

talking on my cell phone or listing to music on my headphones though the light is still green for vehicles se I have no other choiceng across the street, but I have to run the rest of the way to avoid oncoming vehicles

curb pedestrian bridges or underpasses, even if one is located nearby though obstacles (parked vehicles, buildings, trees, trash bins, etc.) obstruct visibility if vehicles are coming because I think they will stop for meut looking, following other people who are crossing y because I get lost in my thoughts

I do not remember the route I have just taken ith another user (pedestrian, driver, cyclist, etc.) and I yell at him ay that forces other pedestrians to let me through trian or an obstacle because I am not paying attention eways forbidden to pedestrians to save timecling paths when I could walk on the sidewalk lowly to annoy a driver

I have crossed several streets and intersections without paying attention to trafc ith another user (pedestrian, driver, cyclist, etc.) and I make a hand gesture ok before crossing because I am thinking about something else ut looking because I am talking with someone y walk on the roadway when I could walk on the sidewalk or on the shoulderith another user and insult him ok before crossing because I want to join someone on the sidewalk on the other side the street without looking because I am in a hurry ith a driver and hit his vehicle s lanes when I could walk on the sidewalk

this, the 4 items in each of the 5 dimensions involvedrrors, lapses, aggressive behaviors and positive behav-he highest score for each of the 4 factors were selected,0 items (added to 3 lter items, see the appendix) under-er PCA with Varimax rotation. The results showed thatlated to transgressions (errors and offences) all loaded

factor, that lapse items all loaded on the second fac-ive behavior and positive behavior items loaded on theurth factors, respectively. After rotation, the four fac-ed 55.07% of the variance. Factorial analyses on the 20

PBS were also performed for each gender and age groups, 2535 years, and over 35, in order to obtain groups of

sizes). Given the succinctness of the results, the anal-t presented here. It should be pointed out, however,

analysis performed for each gender and age group, then items only loaded on one factor (i.e. had scores < 40r factors), that lapse items only loaded on a second fac-ressive and positive behaviors only loaded on a third

and fourth for the totathe 4-factor

The longwas used fo

3.3. Inter-s

Considescales, partPBS were p(Table 4).

The corrtransgressiindividualssive behavilapses: theother road 5.46 0.89 0.6 0.9 3.2 6.4 24.8 64.15.06 1.12 0 3.8 8.2 12.8 28.9 46.44.39 1.77 8.7 11.7 12.5 10.5 12.2 44.34.31 1.29 1.2 9.3 15.7 25.9 26.2 21.64.03 1.55 6.4 13.4 17.5 17.8 22.7 22.24.03 1.53 7 13.1 14 22.7 22.4 20.73.96 1.43 5.2 13.7 17.5 21.6 27.7 14.33.89 1.36 2.6 16 20.4 25.9 20.7 14.33.78 1.59 8.2 19 16 18.7 20.4 17.83.74 1.33 5.2 12.5 26.2 24.2 22.4 9.33.73 1.44 5.8 18.4 18.4 25.1 19.5 12.83.7 1.57 9.6 17.5 17.8 18.1 22.7 14.33.69 1.49 6.1 19.8 20.4 20.4 19.2 143.63 1.52 9 19 18.1 20.4 21.3 12.2

3.61 1.34 3.8 19 26.8 23.3 17.2 9.93.57 1.63 15.2 14 17.5 17.8 22.7 12.83.43 1.54 12 19.5 20.7 20.7 15.7 11.43.4 1.52 12.8 19.2 19.2 22.4 16.6 9.6

3.34 1.22 4.7 20.4 35.6 20.7 13.4 5.2

3.34 1.41 9.6 23 20.1 24.2 16.9 6.1

3.14 1.72 23.6 19.8 13.4 17.2 13.7 12.23.07 1.35 11.7 28 22.4 21.9 11.7 4.43.06 1.47 14.6 26.5 23.3 16.6 11.4 7.62.86 1.39 17.5 30.6 17.2 21.3 9.6 3.82.72 1.24 17.5 29.2 28.3 16.6 6.1 2.32.61 1.50 27.4 30.9 15.7 11.4 8.7 5.82.56 1.24 21.3 33.8 21.9 14.9 7 1.22.46 1.31 25.4 36.4 16.9 12 6.7 2.62.37 1.29 31.5 28.3 21 12.8 3.8 2.62.31 1.32 31.5 37 12.5 9.6 7 2.32.24 1.32 36.2 32.7 12.2 10.5 6.4 22.15 1.31 41.4 29.7 10.8 10.8 5.5 1.72.09 1.11 35 37.3 16.6 7.3 2.6 1.22.02 1.01 32.1 47.5 10.5 7.3 1.7 0.91.95 1.20 46.6 30.9 10.5 6.4 4.1 1.51.9 1.14 47.5 31.8 9 7.9 2.6 1.21.76 1.14 58.3 22.7 8.5 6.7 2.6 1.21.75 0.98 50.1 33.8 9.3 5 0.9 0.91.71 1.04 58 23.6 9.9 6.1 2 0.31.66 0.90 52.5 35.9 7 2.6 1.5 0.61.64 0.86 54.2 33.8 7 4.1 0.6 0.31.63 0.98 60.9 24.5 7.9 4.7 1.5 0.61.62 1.05 63.3 23.6 6.7 2.6 2.3 1.51.59 0.86 58 30 7.9 2.6 1.5 01.43 0.80 70 21.9 4.1 3.5 0.3 0.31.34 0.83 79.3 13.7 2.6 2 2.3 01.33 0.71 77 16.6 4.4 0.9 1.2 0

factor, respectively. Thus, the factorial structure foundl sample was robust; independent of age and gender,

solution was a suitable model. version of the Pedestrian Behavior Scale, with 32 items,r the rest of the analysis.

core relations

ring the potential effects of age and gender on certainial correlations between the different dimensions of theerformed, while checking the effects of age and gender

elation matrix showed positive correlations betweenon, lapses and aggression: the more transgressions

declared, the more they also declared lapses and aggres-ors. Positive behaviors were negatively correlated with

more individuals declared positive behaviors towardusers, the less they declared lapses as pedestrians.


Table 3Principal axis factor analysis of the 40 PBS behavior items (Varimax rotation).

Item (how often do you. . .) Factor 1 Factor 2 Factor 3 Factor 4

I cross diagonally to save time .81I cross outside the pedestrian crossing even if there is one less than 50 m awayI cross the street even though the pedestrian light is redI cross the street between parked cars I start to cross on a pedestrian crossing and I nish crossing diagonally to save timeI cross even though the trafc light is still green for vehicles I cross between vehicles stopped on the roadway in trafc jams I walk on the roadway to be next to my friends on the sidewalk or to overtake someone who is walkin

than I amI look at the trafc light and start crossing as soon as it turns red I start walking across the street, but I have to run the rest of the way to avoid oncoming vehiclesI deliberately walk on the roadway when I could walk on the sidewalk or on the shoulder I walk on the curb I take passageways forbidden to pedestrians to save time I cross whileOn a two-wa cond pI forget to loI forget to lo eI cross withoI realize thatI cross withoI hit a pedesI realize thatI get angry wI get angry wI get angry wI get angry wI cross very sI let a car goWhen I am a t to b

pedestrianI stop to let tI walk on theI thank a dri% of variance

3.4. Effects

ANOVAs(5) on all thtransgressioas well as ostandard de

ANOVA between mand a sign2 = .105) otransgressioshowed thmore transinteraction (F < 1).

A signi(F(1.333) = 4effect (F(4.3score. Menwomen and

Table 4Partial correlaton the gender

TransgressioLapses Aggressive b

* p < .05.** p < .01.

of 3 .05). talking on my cell phone or listing to music on my headphonesy street, I cross the rst part and wait in the middle of the roadway to cross the seok before crossing because I am thinking about something else ok before crossing because I want to join someone on the sidewalk on the other sidut looking because I am talking with someone

I have crossed several streets and intersections without paying attention to trafcut looking, following other people who are crossing trian or an obstacle because I am not paying attention

I do not remember the route I have just taken ith another user and insult him ith another user (pedestrian, driver, cyclist, etc.) and I yell at himith another user (pedestrian, driver, cyclist, etc.) and I make a hand gesture ith a driver and hit his vehiclelowly to annoy a driver

by, even if I have the right-of-way, if there is no other vehicle behind it ccompanied by other pedestrians, I walk in single le on narrow sidewalks so as nos I meethe pedestrians I meet by

right-hand side of the sidewalk so as not to bother the pedestrians I meet ver who stops to let me cross

explained

of demographic variables the age45 (p


Table 5Means (and standard deviations) of the scores on the PBS for each gender, the ve age groups and the total sample.

Age group Gender Transgression Offence Error Lapses Aggression Positive N

1525 Men 46.27 (11.76) 23.70 (6.83) 22.57 (5.51) 12.54 (5.19) 7.70 (3.49) 18.14 (4.06) 37

2635

3645

4655

56 and abov

Total

SD = 4.05, relicense var

Furthermvehicle, indnicantly mand M = 22.ing license,(M = 11.90, and withouvariable had

The indiprevious 5 y(t(341) = 2.1for individuvious 5 yeavariable did

ANOVAsquency, kilwalking tim

Driving behaviors aThe post ha week decnever drov(F(5.337) = 2tests showemore lapsea week. Theon offencesaggressions

A primalapses (F(5.tests showea week decdrove fewekilometers (F < 1), transbehaviors (

Walking(F(5.337) = 2showed thaA primaryon transgr(F(5.337) = 4p < .0005, walked 3 tim

offeney wsive

ly w(F(5.37) = 15.33

titud

tain ion oe I heasur

corr wal2) = 1

ple (F(5

wale pa

weepant

the em 1o tered uced wor ScWomen 43.49 (11.25) 22.13 (6.95) 21.36 (5.10) Total 44.54 (11.47) 22.72 (6.91) 21.82 (5.26) Men 48.16 (10.14) 25.02 (6.46) 23.14 (4.82) Women 43.11 (11.74) 21.99 (6.59) 21.13 (5.88) Total 44.68 (11.47) 22.93 (6.68) 21.75 (5.64) Men 42.40 (9.39) 21.60 (4.66) 20.80 (5.14) Women 38.68 (13.53) 18.64 (6.95) 20.05 (7.09) Total 40.45 (11.75) 20.05 (6.09) 20.40 (6.18) Men 34.47 (11.30) 17.07 (6.02) 17.40 (5.64) Women 34.53 (13.13) 16.76 (7.21) 17.76 (6.46) Total 34.50 (12.11) 16.91 (6.57) 17.59 (5.99)

e Men 37.09 (11.59) 19.18 (7.14) 17.91 (4.85) Women 34.86 (10.42) 18.00 (5.92) 16.86 (5.78) Total 35.63 (10.70) 18.41 (6.27) 17.22 (5.42) Men 44.10 (11.65) 22.63 (6.79) 21.46 (5.52) Women 41.30 (12.15) 20.89 (6.93) 20.41 (5.98) Total 42.33 (12.03) 21.53 (6.92) 20.79 (5.83)

spectively) (t(342) = 2.40, p < .05). The having a driversiable did not have any effect on the other scores.ore, compared with individuals who owned a motor

ividuals who did not own a motor vehicle declared sig-ore errors (t(341) = 3.14, p < .005) (M = 20.06, SD = 5.9109, SD = 5.48, for individuals with and without a driv-

respectively) and more lapses (t(341) = 2.96, p < .005)SD = 6.44 and M = 13.48, SD = 4.88, for individuals witht a license, respectively). The owning a motor vehicle

no effect on the other scores.viduals who had been involved in a trafc accident in theears declared signicantly fewer lapses than the others9, p < .05) (M = 11.31, SD = 3.58 and M = 12.74, SD = 4.99,als who had had or had not had an accident in the pre-rs, respectively). The accident in the previous 5 years

not have any effect on the other scores. were carried out to measure the effect of driving fre-ometers driven weekly, walking frequency and dailye on the scores on the Pedestrian Behavior Scale.frequency had a primary effect on the number of errors a pedestrian (F(5.337) = 2.40, p < .05, partial 2 = .034).oc tests showed that individuals who drove 4 timeslared signicantly fewer error than individuals whoe (p < .05). Driving frequency had an effect on lapses.70, p < .05, partial 2 = .039). The Bonferroni post hocd that individuals who drove 14 times a week declareds as pedestrians than individuals who drove 5 times

results did not show any effect of driving frequency (F(5.337) = 1.06, ns), transgressions (F(5.337) = 1.60, ns),

(F < 1) or positive behaviors (F < 1).

fewer that thaggres(F < 1).

Daisions (F(5.33iors (F(

3.6. At

Cerindicatbecausthe plativelyforced(F(5.31for thequencyweeklythat thtimes apartici

Foring (itinto twcompathe forBehaviry effect for kilometers driven weekly was observed on337) = 2.64, p < .05, = 0.38). The Bonferroni post hocd that individuals who drove between 250 and 350 kmlared signicantly fewer lapses than individuals whor than 50 km a week. These was no signicant effect ofdriven weekly on errors (F(5.337) = 1.002, ns), offencesgressions (F < 1), aggressive behaviors (F < 1) or positiveF < 1).

frequency had a primary effect on lapses.64, p < .05, = .038). The Bonferroni post hoc testst declared lapses increased with walking frequency.

effect was also observed for walking frequencyessions (F(5.337) = 4.86, p < .0005, = .067), offences.25, p < .001, = .059) and errors (F(5.337) = 4.84,

= .067). The post hoc tests showed that individuals whoes a week declared signicantly fewer transgressions,

The forgression scthe error scviduals whindividualsmitted morSD = 11.47 and more elow and hig

For pleof the resp(t(341) = 2and on thepared with for pleasurefor pleasur13.80 (4.09) 8.20 (3.62) 19.79 (3.95) 6113.33 (4.55) 8.01 (3.56) 19.16 (4.05) 9811.58 (3.98) 7.79 (2.97) 19.56 (3.93) 4313.01 (5.41) 9.18 (3.83) 19.64 (4.56) 9612.57 (5.04) 8.75 (3.64) 19.61 (4.36) 13910.70 (4.97) 9.95 (4.95) 20.05 (3.75) 2013.00 (5.47) 9.32 (3.67) 20.23 (3.60) 2211.91 (5.30) 9.62 (4.28) 20.14 (3.63) 4210.47 (2.61) 8.47 (3.25) 19.87 (4.85) 1511.12 (4.77) 8.06 (3.29) 22.88 (3.16) 1710.81 (3.86) 8.25 (3.22) 21.47 (4.26) 3212.18 (5.64) 7.91 (3.75) 20.55 (3.72) 1111.62 (3.04) 8.90 (3.08) 20.95 (3.43) 2111.81 (4.04) 8.56 (3.30) 20.81 (3.48) 3211.64 (4.55) 8.20 (3.63) 19.34 (4.06) 12612.95 (4.86) 8.80 (3.65) 20.12 (4.17) 21712.47 (4.79) 8.58 (3.65) 19.83 (4.14) 343

ces and fewer errors than individuals who declaredalked daily. There was no walking frequency effect onbehaviors (F(5.337) = 1.34, ns) and positive behaviors

alking time did not have an effect on transgres-37) = 1.78, ns), offences (F(5.337) = 1.75, ns), errors.66, ns), lapses (F(5.337) = 1.08, ns), aggressive behav-

7) = 1.47, ns) or positive behaviors (F(5.337) = 1.40, ns).

es toward walking

lter items in the Pedestrian Behavior Scale provided anf whether walking was a constraint (item 19 I walkave no other choice) or a choice (item 7 I walk fore of it). The scores for these two items were neg-elated (r = .28, n = 343, p < .0005). The score for theking item did not vary in relation to walking frequency.81, ns) nor of weekly walking time (F < 1). The scoreasure walking item varied in relation to walking fre-.312) = 5.14, p < .0005) but did not vary in relation toking time (F < 1). The Bonferroni post hoc tests showedrticipants who declared that they walked fewer than 2k declared signicantly less pleasure walking than the

s who walked more than 2 times a week.rest of the analyses, the scores for the forced walk-9) and pleasure walking (item 7) items were recodedrms: low score (13) and high score (46). Means weresing Students t distribution to measure the effect ofalking and pleasure walking scores on the Pedestrian

ale scores.

ced walking score had a signicant effect on the trans-ore (t(341) = 2.01, p < .05), and more particularly onore (t(341) = 2.24, p < .05). Compared with those indi-o declared that they were rarely forced to walk, the

who declared that they were often forced to walk com-e transgressions (M = 41.36, SD = 12.24 and M = 44.08,for low and high forced walking scores, respectively),rrors (M = 20.27, SD = 5.90 and M = 21.74, SD = 5.60 forh forced walking scores, respectively).asure walking, the results showed a signicant effectonse to item 7 (pleasure walking) on the error score.54, p < .01), on the lapses score (t(341) = 2.83, p < .005)

positive behavior score (t(341) = 2.66, p < .01). Com-those individuals who declared that they rarely walked, the individuals who declared that they often walkede committed more errors (M = 19.77, SD = 5.36 and


M = 21.41, SD = 6.01 for low and high pleasure walking scores,respectively), more lapses (M = 11.53, SD = 3.77 and M = 13.03,SD = 5.23 for low and high pleasure walking scores, respec-tively) and had more positive behaviors (M = 19.07, SD = 3.93 andM = 20.29, Srespectively

4. General

4.1. Constru

The objself-reporteferentiate bbehaviors, aby pedestri

The validin terms of tion betweelapses, posientiation be

A reliabclassicatiothis study. Aitems relatiof them wita structure in each facpositive behtive behaviomen and w75 showed tion with 4These resulof age thosecents (ElliotMann, 2009bility acrossthis tool aliors aggrefor researchmechanism

Not surpduced by thaberrant befor driver bferences bequestionna

Despite between thdifferent grtion by Reas acts of omendangermferentiated or rules of c

The maiand that prtransgressioas offencesneither Torformed factand Yildirimand lapses.Elliott and Blescents, AR

both error and transgression behaviors, some transgression itemsalso being found on axis 2, dangerous playing in the road. Thus,the lack of distinction between errors and transgression could beobserved both in the study by Elliott and Baughan (2004) among

centsems ors.rs are catedrmort thaten

ianceationn leion (r exype se frespr

ve anore

ules tems, 20tandnd, m

e to ood.en th

foun intoychoerstaer daudy,uishtion

the by Gdangrisk-portematisideror foluntauencgh tt theerrorf uncThe lial nith we the. Thency orelat

explthermis, thon ted foncy aD = 4.21 for low and high pleasure walking scores,).

discussion

cting a research tool

ective of this study was to develop and validate ad behavior scale for pedestrians of all ages that can dif-etween offences, errors and lapses among injury risks well as to measure aggressive and positive behaviorsans toward other road users.ation study was carried out on a heterogeneous sampleage and socio-cultural origin. It provided a differentia-n the behaviors proposed on the scale transgressions,tive and aggressive behaviors. However, a lack of differ-tween offences and errors was observed.le scale of self-reported behaviors providing a usefuln of pedestrian behaviors at all ages was developed in

20-item instrument was produced from the initial 40ve to specic examples of pedestrian behaviors, mosth accident risks. The short version of the tool producedwith 4 factors that was easy to interpret, as the itemstor transgressions, lapses, aggressive behaviors andaviors had good internal reliability, except for posi-rs. Furthermore, the analyses carried out separately onomen as well as on the 3 age groups between 15 andthat, independent of age and gender, the factorial solu-

factors in the Pedestrian Behaviors Scale was robust.ts reproduced among pedestrian from 15 to 75 years

of similar of research published previously on adoles-t and Baughan, 2004; Sullman et al., 2011; Sullman and), demonstrating a level of repeatability and applica-

various nations and various age groups. Furthermore,so added two new dimensions of pedestrian behav-ssive and positive behaviors that can be benecial

investigating pedestrian behaviors and psychologicals which can explain those behaviors.risingly, the pedestrian behavior classication pro-e PBS in this study is not completely in line with thehavior classication produced by Reason et al. (1990)ehaviors. These differences probably reect the dif-tween the two road user populations that these twoires are used to study.these differences, there are a few concordant pointsis study and the studies using the DBQ structure onoups of motorized road users. Both in the DBQ classica-on et al. (1990) and in the PBS constructed here, lapsesission putting the individual in a situation of personalent but with no intent to take a risk was clearly dif-from intentional transgressions against the legal rulesaution.n difference between the factorial structure of the DBQoduced by the PBS is that, in the present study, then axis (axis 1) contains items that can be classied

and items referring to errors. As far as we know,quato and Bianchi (2010) nor Moyano Diaz (1997) per-orization on their versions of the PBQ among adultss version (2007) did not differentiate between errors

This result is consistent, however, with the study byaughan (2004) on the Pedestrian Behavior Scale for ado-BQ, where factor 1, unsafe road crossing, comprised

adolesand sebehavi

ErrotherefoferentiFurthethe facto shorcomplexplanbetweeof cautcles, fosame tresponare widnot haalso, msuch rthese i(Graniunderstrians arelativadulth

Givpletelyresultsthe psTo undwhethrent stdistingdistincamonggestedself-entional self-reor norbe conbehaviany voconseqAlthoube thaand in level ocases. the soccies wbecausceivedfrequeof risk line of

Fursion axeffect observfreque and in the current study among pedestrians of all agesto reect differences between pedestrians and drivers

nd offences against the legal rules by pedestrians areommitted by the same individuals, while they are dif-

among drivers (Lawton et al., 1997; Reason et al., 1990).e, the behaviors comprising axis 1 have in commont their underlying objective is to maintain speed andwalking distance, both factors that explain lack of rule

among pedestrians (Sisiopiku and Akin, 2003). One may be that individuals usually do not differentiategal rules (crossing light, pedestrian crossing) and rulesavoiding views obstructed by stopped or parked vehi-ample) and think that their transgressions are of theand have the same level of consequences. Analyses ofequencies tend to show that these injury risk behaviorsead among pedestrians and that not only do legal rulesy more effect on pedestrians than rules of caution, butgenerally, social norms tend to lead them not to takeinto account and to transgress all such norms. For all, it would be interesting to measure normative beliefs08) and the internalization of rules (Grani, 2009) to

if formal and informal rules are differentiated by pedes-ore generally, to observe how representations of rules

pedestrian behavior develop between childhood and

at the error-lapses-violation classication was not com-d in this study, a theoretical approach that takes these

account needs to be identied in order to understandlogical mechanisms underlying pedestrian behaviors.nd the results, a sharp distinction can be made as tongerous behavior is intentional or not. Thus, in the cur-

transgressions against formal and informal rules areed from lapses by their intentional nature. The samebased on the intentionality of the behavior mentionedpedestrian self-reported injury risk behaviors was sug-rani (2008) where a distinction was found betweenerment behaviors (lapses) and behaviors with inten-taking in each of the sub-scales, whether concerningd behaviors, perception of danger or risk for oneselfve beliefs. These intentional dangerous behaviors caned as risk-taking, dened as deliberately undertaking ar which the accident risk is perceived (Saad, 1988) or asry behavior for which the uncertainty surrounding thees for oneself or others is perceived (Trimpop, 1994).his needs to be conrmed by future studies, it could

level of risk-taking in offences against the legal ruless are considered to be weak by our sample because theertainty is considered by pedestrians to be low in bothevel of uncertainty could be perceived as low becauseorm is to transgress as shown by the mean frequen-hich these behaviors are declared by our sample , or

danger of these transgressive behaviors is not well per- high frequency of transgressions and the low relativef pedestrian accidents no doubt reduce the perceptioned to these behaviors. Research is needed to explore thisanation.ore, while offences and errors form a single transgres-

e potentially explanatory variables have a differentiatedhe two sub-scores. Thus, gender differences werer offences against legal rules, but not for errors, drivingnd attitudes toward walking affected error score but


not offence score. This could signify the psychological mechanismsexplaining offences and errors and then the modes of remedia-tion are not the same (Reason et al., 1990) even these two types ofbehaviors are not differentiated by pedestrian. Therefore, it wouldbe useful, agraphic andagainst the

4.2. Results

4.2.1. GendThe resu

legal rules tmen and wamong the observed inDiaz, 2002)Tom and Ggenders for(Harr et aa different pedestrianspedestrian ing less withwith the spThe resultstoward othtion to stere2009): femiing for otheand domina

The resudimensionsopposition errors. Furtiors than infrequency vthe most (dwhy youngolder peopknown andoptimal socdence, and Silbereisen relation betis linked totive controafter childhself-regulatpositive behin aberranta greater tethere is incbehavioral sons for wathan going question ofexplanationies on mobitheir effects

4.2.2. MobiThe resu

Behavior Scmore errorserrors, and

declare more lapses. Likewise, individuals who walk regularlydeclare more offences and errors. The rst explanation for theseresults comes from the greater opportunity to demonstrate aber-rant behaviors when there is a greater walking frequency. The more

ntly omm. Anoenced andn crover er awors tt it isisky

The eors

resu Pedg forualsapsesot reualsore aore

effehowncurito choek tnmenspor

mo, 201nsafelemlaineriansan s

theid vehy feell, codestoad uppeaencerians

clus

resclassg threntilapse-int

isks, hers

in thughabaseding bffects it was made in the current study of the effect of demo- mobility variables, to differentiate between offenceslegal rules and error behaviors in transgressions.

for the effects of demographic and mobility variables

er and agelts show that men commit more offences against thehan women. There is no difference, however, betweenomen for the error score. The highest offence scoremale pedestrians in the sample conrms the results

other studies, both for the behaviors declared (Moyano and for behaviors observed (Latrmouille et al., 2004;rani, 2011). More generally, the difference between

offences is well known in studies on driver behaviorl., 1996; Simon and Corbett, 1996) and demonstratesrelationship to the legal rules among male and female. This difference has already been shown for observedbehaviors (Tom and Grani, 2011), with men comply-

pedestrian crossing lights than women, but complyingatial rules of pedestrian crossing as much as women do.

also show that women have more positive behaviorser road users. This result can be understood in rela-otypes dening womens role in our societies (Grani,nine stereotypes encourage mutual assistance and car-rs, while masculine stereotypes encourage competitiontion.lts show that age has an effect on several of the

measured here. Individuals under the age of 35 are into those over 45, committing more offences and morehermore, 3545 year-olds have more positive behav-dividuals in other age groups. It is known that walkingaries with age younger people and older people walke Solre and Papon, 2010) but this does not explainer people declare having more aberrant behaviors thanle. Risk-taking behaviors during adolescence are well

have been interpreted as a method of developingial and psychological competence, autonomy, indepen-self-regulation (Baumrind, 1991; Parsons et al., 1997;and Noack, 1988). Research now shows that curvilinearween age and risk-taking from childhood to adulthood

changes in the brains socio-emotional and cogni-l systems. The rst system increases reward-seekingood, whereas the second system improves capacity forion after adolescence (Steinberg, 2008). The increase inaviors at 3545 years of age, followed by the decrease

behaviors after the age of 45, may suggest that, afterndency toward risky behavior among younger people,reasing awareness of sharing spaces, leading to greatercontrol in older people. Change with age in the rea-lking has also to be taken into account (shopping ratherto school and later to work) and walking can become a

choice rather than constraint from the age of 35. Theses obviously need to be conrmed through more stud-lity and attitudes toward walking in relation to age, and

on pedestrian behaviors.

lity and attitudes toward walkinglts show an effect of the type of mobility on Pedestrianale scores. Thus, individuals without a vehicle declare

and lapses, individuals who never drive declare more individuals who rarely drive or drive short distances

frequerisks clationsexperiHollaneffect ocross-oa greatbehavisee thamore r

4.2.3. behavi

Theon thewalkinindividmore ling is nIndividof it mhave m

Thebeen sand seals whand seenviroof tranest andPaponmore usafety be exppedestthe urbtive ofstoppeing mato stroand peother rthey aexperipedest

5. Con

Thetional dividina diffesions, easy-todent rresearc

As and BaPBS is Observhuge eone walks and the longer the distances, the more oneitting errors, being inattentive or breaking trafc regu-ther line of explanation may be the effect of driving

on behaviors as a pedestrian. Thus, in keeping with Hill (2010), who found that driving experience has anssing by elderly pedestrians, it is possible that there is a

of skills from driving to pedestrian behaviors, or at leastareness of interactions with drivers and risk, leading to

hat are less risky when crossing. Furthermore, we can walking frequency, and not walking time, that leads tobehaviors when crossing.

ffect of attitudes toward walking on pedestrian

lts also show an effect of attitudes toward walkingestrian Behavior Scale scores. Being forced to walk or

the pleasure of it leads to more errors. Furthermore, who declare that they walk for pleasure also declare

and more positive behaviors. Moreover, pleasure walk-lated to walking time, but rather to walking frequency.

who walk frequently declare walking for the pleasurend commit more errors and lapses, but at the same timepositive interactions with other road users.cts of whether walking is forced or chosen have already

for perception of the environment, in terms of safetyy (Miaux, 2008). It has also been shown that individu-ose to walk are more sensitive to friendly interactionso establish a privileged relationship with the urbant, while people who have no choice in their modetation walking being more common in the weak-st vulnerable socio-demographic groups (de Solre and0) have a perception that the urban environment ise and are more sensitive to security elements than toents (Miaux, 2008). Results of the current study mayd by a different way of occupying urban spaces: forced

commit more errors because they have more anxiety inpace, avoid interactions with others, focus on the objec-r trip and take the shortest route (between parked oricles). On the other hand, pedestrians who enjoy walk-l more comfortable with the urban space, takes timensider that their territory is not limited to sidewalksrian crossings and more easily accept interactions withsers. These results obviously need to be conrmed, but

r to point to an impact of attitudes toward walking d as a constraint or chosen on risk behaviors among.

ion

ults of this study do not correspond to the tradi-ication of aberrant behaviors by Reason et al. (1990)em into errors, lapses and violation, but do provideation of pedestrian behaviors in terms of transgres-s, aggressive and positive behaviors. They provide anerpret categorization of pedestrian behaviors with acci-and the Pedestrian Behavior Scale can be useful to all

investigating pedestrian safety, whatever the age.e DBQ by Reason et al. (1990) and the ARBQ by Elliottn (2004), classication of behaviors provided by the

on declared behaviors and not on observed behaviors.ehavior is important to take into account and study the

of environment on pedestrian behaviors that cannot


be understand through self-reported questionnaire (Havard andWillis, 2012; Sisiopiku and Akin, 2003). Thus, self-report studiesare not sufcient to observe the effect of more recent road designs,such as shared spaces, which interactions between pedestrians andslow trafc would not be viewed as an unsafe situation but ratheras an evide2012).

Self-repoas valid mThis is notwhich are hor when sttrian behavhas shown declared be(West et aquestionnacomplementiple factorinteraction in a complemotivationa

Behavioinjury-risk used to diffepositive behbehaviors, aa more detaof related ppopulation,and analyziof behavior i.e. adolespedestrian to better advulnerable

Appendix A

Short vethe items pViolation

I cross diagoI cross outsi

(40)I cross the stI cross even

ErrorI cross the stI start to cro

time (9)I cross betwI walk on th

someone w

LapseI forget to loI forget to lo

on the othI cross withoI realize that

attention

AggressiveI get angry wI get angry w

(11)I get angry w

gesture (2I get angry w

PositiveI let a car go by, even if I have the right-of-way, if there is no other vehicle

behind it (34)When I am accompanied by other pedestrians, I walk in single le on narrow

sidewalks so as not to bother the pedestrians I meet (46)I stop to let

on thet (5)

for thpublic

becau

nces

., 1998., Grannts eti, M

ractioneil, ppd, D., tance

C., 2viour315e, R., Ptes?es, no, pp. 1.A., B

escenand Be.A., B

orcylis., Norpplicary an., Norms: an aation, C., vioureived.A., 2

ompo264.

M.A., 2rnaliza128E., Su. Spaarch P., Fielern inarch 2C., Wiviour

c Psy, C.A., Hsions lysis a, I., Bes of peafc Puille, . Le s

ons enehavrit 8

R., Paictingal Psy., 200enviro

marcitions

routesports

Diaz,ngir ntentioologiance of a well-designed multi-use space (Kaparias et al.,

rting measurements are, however, widely recognizedeasurements in the social sciences (Corbett, 2001).ably the case when studying injury risk behaviors,ard to measure through located observational studiesudying psychological factors that can explain pedes-iors. Furthermore, for many social behaviors, researchthat relatively strong associations are found betweenhaviors and more objective behavioral measurementsl., 1993). Therefore, observations and self-reportedires on pedestrian behaviors should be conceived astary approaches to explore and understand the mul-s which intervene in a complex phenomenon: thebetween the pedestrian and his/her road environmentx task which brings psychomotor skills, cognitive andl process into play (Saad, 1988).

rs as measured by the PBS can be a good approach tobehaviors measured more objectively. The PBS can berentiate between transgressions, lapses, aggressive andaviors. This ability to measure positive and aggressivelong with aberrant pedestrian behaviors, will provideiled understanding of pedestrian behaviors as well as

sychological and mobility factors. Validated on a French this instrument may prove to be useful in measuringng differences in the frequency of these different types, notably in pedestrian populations that are most at riskcents and elderly, to understand changes with age inbehaviors and associated factors and thus to be ableapt preventive actions to the different populations ofroad users.

.

rsion of the PBS (23 items) (the numbers correspond tolace in the original version of the PBS)

nally to save time (37)de the pedestrian crossing even if there is one less than 50 m away

reet even though the pedestrian light is red (17)though the light is still green for vehicles (25)

reet between parked cars (35)ss on a pedestrian crossing and I nish crossing diagonally to save

een vehicles stopped on the roadway in trafc jams (6)e roadway to be next to my friends on the sidewalk or to overtakeho is walking slower than I am (16)

ok before crossing because I am thinking about something else (36)ok before crossing because I want to join someone on the sidewalker side (39)ut looking because I am talking with someone (4)

I have crossed several streets and intersections without payingto trafc (22)

ith another user and insult him (23)ith another user (pedestrian, driver, cyclist, etc.) and I yell at him

ith another user (pedestrian, driver, cyclist, etc.) and I make a hand8)ith a driver and hit his vehicle (15)

I walkI me

FilterI walkI take I walk

Refere

Aberg, LAbou, A

pareGraninteArcu

Baumrinsubs

Corbett,beha4, 13

de SolrenqusancBron

Elliott, Madology

Elliott, Mmot

Evans, Dan aTheo

Evans, DtionEduc

Gabaudebehaperc

Grani, Mdes c253

Grani, inte1277

Gras, M.2006Rese

Harr, NconcRese

Havard, behaTraf

HollanddeciAna

KapariastionF: Tr

Latrmo2004pitthe bScu

Lawton,predSoci

Miaux, Sde lde lacondtrianTran

Moyanoinfriet inPsicthe pedestrians I meet by (21)e right-hand side of the sidewalk so as not to bother the pedestrians

e pleasure of it (7) transportation (buses, metro, tramway, etc.) (32)se I have no other choice (19)

. Trafc rules and trafc safety. Safety Science 29, 205215.i, M.A., Mallet, P., 2008. Recherche de sensations, attachement aux

prise de risque dans lespace routier chez ladolescent piton. In:.A., Auberlet, J.-M. (Eds.), Le piton et son environnement: quelless? Quelles adaptations? Actes n115. Les collections de lINRETS,. 121133.1991. The inuence of parenting style on adolescent competence and

use. Journal of Early Adolescence 11 (1), 5695.001. Explanations for understating in self-reported speeding. Transportation Research Part F: Trafc Psychology and Behaviour0.apon, F., 2010. La mobilit pied: que nous apprennent les dernires

In: Grani, M.A., Auberlet, J.M. (Eds.), Le piton: nouvelles connais-uvelles pratique et besoins de recherche. Les Collections de lINRETS,524.aughan, C.J., 2004. Developing a self-report method for investigatingt road user behavior. Transportation Research Part F: Trafc Psychol-haviour 7 (6), 373393.aughan, C.J., Sexton, B.F., 2007. Errors and violations in relation tots crash risk. Accident Analysis and Prevention 39, 491499.man, P., 1998. Understanding pedestrians road crossing decisions:tion of the theory of planned behaviour. Health Education Research:d Practice 13 (4), 481489.an, P., 2003. Predicting adolescent pedestrians road-crossing inten-

pplication and extention of the Theory of Planned Behavior. Health Research: Theory and Practice 18 (3), 267277.Marqui, J.-C., Obriot-Claudel, F., 2010. Self-regulatory driving

in the elderly: relationships with aberrant driving behaviours and abilities. Le Travail Humain 73 (1), 3152.008. Inuence de ladhsion aux strotypes de sexe sur la perceptionrtements pitons chez ladulte. Recherche Transports Scurit 101,

009. Sex differences, effects of sex-stereotype conformity, age andtion on risk-taking among pedestrian adolescents. Safety Science 47,3.llman, M.J.M., Cunill, M., Planes, M., Aymerich, M., Font-Mayolas, S.,nish drivers and their aberrant driving behaviours. Transportationart F: Trafc Psychology and Behaviour 9 (2), 129137.d, J., Kirkwood, B., 1996. Gender differences and areas of common

the driving behaviors and attitudes of adolescents. Journal of Safety7 (3), 163173.llis, A., 2012. Effects of installing a marked crosswalk on road crossing

and perceptions of the environment. Transportation Research Part F:chology and Behaviour 15, 249260.ill, R., 2010. Gender differences in factors predicting unsafe crossing

in adult pedestrians across the lifespan: a simulation study. Accidentnd Prevention, http://dx.doi.org/10.1016/j.aap.2009.12.023.ll, M.G.H., Miri, A., Chan, C., Mount, B., 2012. Analysing the percep-destrians and drivers to shared space. Transportation Research Partsychology and Behaviour 15 (3), 297310.M.-E., Thouez, J.-P., Ranou, A., Bergeron, J., Bourbeau, R., Bussire, Y.,exe est-il une variable pertinente pour ltude du comportement des

intersection urbaine? (Is gender a relevant variable when studyingiors of pedestrians at urban intersections?). Recherche Transports 4, 171188.rker, D., Manstead, A.S.R., Stradling, S.G., 1997. The role of affect in

social behaviours: the case of road trafc violations. Journal of Appliedchology 27, 12581276.8. Comment la fac on denvisager la marche conditionne la perceptionnnement urbain et le choix des itinraires pitonniers. Lexpriencehe dans deux quartiers de Montral (How the way of considering walk

the perception of the urban environment and the choice of pedes-s. The experience of walk in two districts of Montral). Recherche

Scurit 101, 327352. E., 1997. Teoria del Comportamiento Planicado e intencion deormas de transito en peatones (Thorie du Comportement Planin de transgresser les rgles routires chez les pitons). Estudos des

2 (2), 335348.


Moyano Diaz, E., 2002. Theory of planned behavior and pedestrians intentions toviolate trafc regulations. Transportation Research Part F: Trafc Psychologyand Behaviour 5 (3), 169175.

ONISR, 2007. La scurit routire en France. Bilan de lanne 2006 La DocumentationFranc aise, Paris.

zkan, T., Lajunen, T., 2005. A new addition to DBQ: Positive Driver BehavioursScale. Transportation Research Part F: Trafc Psychology and Behaviour 8 (45),355368.

zkan, T., Lajunen, T., Chliaoutakis, J., Parker, D., Summala, H., 2006. Cross-culturaldifferences in driving behaviours: a comparison of six countries. TransportationResearch Part F: Trafc Psychology and Behaviour 9 (3), 227 242.

Parsons, J.T., Siegel, A.W., Cousins, J.H., 1997. Late adolescent risk-taking: effects ofperceived benets and perceived risks on behavioral intentions and behavioralchange. Journal of Adolescence 20, 381392.

Rasmussen, J., 1980. What can be learned from human errors report? In: Duncan, K.,Gruneberg, M., Wallis, D. (Eds.), Changes in Working Life. Wiley, London.

Reason, J., Manstead, A., Stradling, S., Baxter, J., Campbell, K., 1990. Errors and viola-tions on the roads: a real distinction? Ergonomics 33 (1011), 13151332.

Saad, F., 1988. Prise de risque ou non-perception du danger. Recherche Transports Scurit 1819 (Septembre), 5562.

SafetyNet, 2009. Pedestrians & Cyclists.Silbereisen, R.K., Noack, P., 1988. On the constructive role of problem behavior in

adolescence. In: Bolger, N. (Ed.), Person and Context: Developmental Processes.Cambridge University Press, Cambridge.

Simon, F., Corbett, C., 1996. Road trafc offending, stress, age, and accident historyamong male and female drivers. Ergonomics 39, 757780.

Sisiopiku, V.P., Akin, D., 2003. Pedestrian behaviors at and perceptions towardsvarious pedestrian facilities: an examination based on observation and sur-vey data. Transportation Research Part F: Trafc Psychology and Behaviour 6,249 274.

Steinberg, L., 2008. A social neuroscience perspective on adolescent risk-taking.Developmental Review 28, 78106.

Sullman, M.J.M., Gras, M.E., Font-Mayolas, S., Masferrer, L., Cunill, M., Planes, M.,2011. The pedestrian behaviour of Spanish adolescents. Journal of Adolescence34 (3), 531539.

Sullman, M.J.M., Mann, H.N., 2009. The road user behaviour of New Zealand ado-lescents. Transportation Research Part F: Trafc Psychology and Behaviour 12,494502.

Sullman, M.J.M., Meadows, M.L., Pajo, K.B., 2002. Aberrant driving behavioursamongst New Zealand truck drivers. Transportation Research Part F: TrafcPsychology and Behaviour 5 (3), 217232.

Tom, A., Grani, M.A., 2011. Gender differences in pedestrian rule compliance andvisual search at signalized and unsignalized crossroads. Accident Analysis andPrevention 43 (5), 17941801.

Torquato, R.J., Bianchi, A.S.A., 2010. Comportamento de Risco do Pedestre ao Atraver-saar a Rua: Um Estudo com Universitarios. Transporte: Teoria e Aplicac ao 2 (1),1941.

Trimpop, R.M., 1994. The Psychology of Risk-taking Behavior. Elsevier, North Hol-land, New York, Amsterdam.

West, R., French, D., Kemp, R., Elander, J., 1993. Direct observations of driving, self-reports of driver behaviour, and accident involvement. Ergonomics 36, 557567.

Xie, C., Parker, D., 2002. A social psychological approach to driving violations in twoChinese cities. Transportation Research Part F: Trafc Psychology and Behaviour5 (4), 293308.

Yagil, D., 1998. Gender and age-related differences in attitudes toward trafc lawsand trafc violations. Transportation Research Part F: Trafc Psychology andBehaviour 1, 123135.

Yagil, D., 2000. Beliefs, motives and situational factors related to pedestrians self-reported behavior at signal-controlled crossings. Transportation Research PartF: Trafc Psychology and Behaviour 3 (1), 113.

Yildirim, Z., 2007. Religiousness, Conservatism and their Relationship with TrafcBehaviours. Middle East Technical University.

Zeedyk, M.S., Kelly, L., 2003. Behavioural observations of adultchild pairs at pedes-trian crossings. Accident Analysis and Prevention 35 (5), 771776.

Zhou, R., Horrey, W.J., 2010. Predicting adolescent pedestrians behavioral intentionsto follow the masses in risky crossing situations. Transportation Research PartF: Trafc Psychology and Behaviour 13 (3), 153163.

Zhou, R., Horrey, W.J., Yu, R., 2009. The effect of conformity tendency on pedestri-ans road-crossing intentions in China: an application of the theory of plannedbehavior. Accident Analysis and Prevention 41, 491497.

Developing a self-reporting method to measure pedestrian behaviors at all ages1 Introduction2 Method2.1 Materials2.2 Participants

3 Results3.1 Tool validation3.2 Refining the scale3.3 Inter-score relations3.4 Effects of demographic variables3.5 Effects of variables concerning mobility3.6 Attitudes toward walking

4 General discussion4.1 Constructing a research tool4.2 Results for the effects of demographic and mobility variables4.2.1 Gender and age4.2.2 Mobility and attitudes toward walking4.2.3 The effect of attitudes toward walking on pedestrian behaviors

5 ConclusionReferences

1-s2.0-S0001457512002631-main

Documents

Transcript of 1-s2.0-S0001457512002631-main