T

OT

a

ARRA

KRPM

1

aasacact

caanfa1tudtm

0d

Accident Analysis and Prevention 45 (2012) 718– 725

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

jo ur n al hom ep a ge: www.elsev ier .com/ locate /aap

he effects of positive emotion priming on self-reported reckless driving

rit Taubman - Ben-Ari ∗

he Louis and Gabi Weisfeld, School of Social Work, Bar-Ilan University, Ramat Gan 52900, Israel

r t i c l e i n f o

rticle history:eceived 5 December 2010eceived in revised form 31 August 2011ccepted 27 September 2011

eywords:eckless drivingositive emotions

a b s t r a c t

Five studies examined the effects of positive emotion priming on the willingness to drive recklessly. Inall five, young drivers were exposed to one of the following primes of positive affect: a positive moodstory; happy memories; an exciting film; a relaxing film; or thoughts on the meaning in life. Followingthe prime, the participants were asked to report on their willingness to drive recklessly. The responseswere compared to those of groups exposed either to neutral affect, another kind of positive affect, ornegative affect priming. In two of the studies, participants were also asked to report on their drivingstyles (risky, anxious, angry, or careful) as a second dependent variable. Positive affect, especially in

eaning in life the form of arousal, was found to be related to higher willingness to drive recklessly. Although mentended to report higher intentions to drive recklessly, men and women did not react differently to theemotional induction. Most interestingly, positive emotions of a relaxing nature, as well as thinking aboutthe meaning in life, lowered the willingness to engage in risky driving. The discussion emphasizes theimportance of looking for new ways to use positive emotions effectively in road safety interventions, andconsiders the practical implications of the studies.

. Introduction

Most of the emotional appeals in road safety campaigns are neg-tive and fear-based (Lewis et al., 2008a), aimed at provoking fear,nxiety, or apprehension in the target audience. While the use ofuch threat appeals has attracted considerable interest in the liter-ture, the results of years of research into their effects are far fromlear or unequivocal (Wundersitz et al., 2010). On the other hand,ppeals to positive emotions are seldom used in the road safetyontext, and are perhaps considered less effective simply becausehey are so rare (Lewis et al., 2007b).

In a survey of the literature on fear appeals, Elliott (2003)oncluded that road safety media campaigns should adopt thispproach with caution, as fear arousal can have both facilitatingnd inhibiting effects and may lead to defective coping mecha-isms. Indeed, a number of studies have found that exposure to

ear appeals can elicit maladaptive responses (e.g. Schoenbachlernd Whittler, 1996; Orit Taubman - Ben-Ari et al., 2000; Witte et al.,998) that is, responses that do not seek to control or removehe threat implied in the message, but rather to cope with thenpleasant feelings it arouses. Such maladaptive responses include

efensively avoiding or ignoring the message, failing to process thehreat it conveys, and denying its personal relevance. Fear appeals

ay also evoke reactance, so that individuals view the message as

∗ Tel.: +972 3 5318066; fax: +972 3 5347228.E-mail address: taubman@mail.biu.ac.il

001-4575/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.oi:10.1016/j.aap.2011.09.039

© 2011 Elsevier Ltd. All rights reserved.

a challenge and consequently increase the undesired behavior. Allthese outcomes are dangerous because they reduce the sense ofthreat without reducing the actual level of risk, thereby renderingthe fear appeal ineffective.

In addition, the results of a meta-analysis show that the individ-ual’s response efficacy is a major predictor of adaptive outcomesfrom exposure to fear-evoking messages (Floyd et al., 2000), out-weighing the effect of negative affect itself. Furthermore, fearappeals can “expose a person against his or her will to harmfulor seriously offensive images” (Hyman and Tansey, 1990, p.110),thereby creating unnecessary anxiety among viewers. Exposingindividuals to levels of fear that are psychologically uncomfortablecould be considered unethical (Hastings et al., 2004).

As a result of such findings, scholars not only recommend amore cautious attitude toward the use of fear appeals, but alsosuggest that road safety campaign developers consider other typesof appeals (e.g., Wundersitz et al., 2010). Nonetheless, alternativeappeals remain rare.

Another issue that has attracted the attention of researchers isthe question of whether there are gender differences in the reac-tion to emotion-based appeals. Early meta-analyses suggested thatgender and age had little influence on the effectiveness of fearappeals (Witte and Allen, 2000). However, more recent researchindicates that males and females react differently to such messages.

Goldenbeld et al. (2008), for example, found that anti-speeding fearappeals had a positive or neutral effect on females, but evokedcounterproductive negative reactions from males. After viewingthreatening ads, males were less likely to perceive speeding as a

alysis

pipntfjfdftay2

gpnfTpmiintgitdat2

btr(shaembatoe(oapmesas

tpdcmf

t

O. Taubman - Ben-Ari / Accident An

roblem, and had a weaker intention to comply with speed lim-ts. While an association between the perception of speeding as aroblem and less self-reported speeding was found for females,o such relationship was found for males. The authors suggesthat male drivers may dissociate their own speeding behaviorrom the social problem, essentially evidencing a defensive self-ustification response to the fear appeal. Similarly, males wereound to report significantly less intention to improve their futureriving behavior in respect to speeding and drunk driving thanemales (Lewis et al., 2007a). These findings are particularly dis-urbing as males represent the more high risk group of road users,nd are thus the intended target of many fear-based appeals, andet it seems they are not persuaded by such messages (Lewis et al.,008a).

Social psychology theories offer possible explanations for theseender-specific findings. Sex role identity theory maintains thatositive emotions are more stereotype-congruent for males thanegative emotions, and consequently males will respond more

avorably to positive emotional appeals (Fisher and Dubé, 2005).he selectivity hypothesis from information processing theoryosits that males are selective processors while females areore elaborate processors (Dubé and Morgan, 1996). As negative

nformation is thought to be more consequential than positivenformation, women are more likely to attend to and processegative information whereas males are more likely to attendo and process positive information. These theories indicate thatender is an important factor that influences the way in whichndividuals process the relevance of a message, and suggest thathe threat of physical harm is not effective for young malerivers. Other approaches therefore need to be considered, as fearppeals apparently fail to reach and influence one of the mainarget groups for which they were developed (Tay and Ozanne,002).

Despite the fact that the need to explore other approaches haseen identified, very few attempts have been made to do so. Inerms of positive emotions, humor has been used occasionally inoad safety messages, but there is limited research into its effectDelhomme et al., 2009). Evidence from the domain of public healthuggests that humorous appeals are more persuasive than non-umorous appeals for males when addressing the issues of AIDSnd sunscreen use (Conway and Dubé, 2002; Struckman-Johnsont al., 1994). In addition, there are indications that negative appealsight have a diminishing effect over time, while positive appeals

ecome more persuasive over time (Lewis et al., 2008a). It haslso been suggested that for speeding behavior, positive emo-ional appeals are more effective for males than females, while thepposite is true for fear-based negative emotional appeals (Lewist al., 2008b). In order to examine these possibilities, Lewis et al.2008b) conducted an experiment using ads with different typesf appeals, and found greater persuasiveness of negative (fear)ppeals immediately after exposure, but greater improvement forositive humorous appeals over time (up to a month later). Further-ore, for males, message acceptance after exposure to a positive

motional appeal was a more important predictor of subsequentpeeding behavior than was previous speeding behavior. For neg-tive emotional appeals, previous driving behavior was the onlyignificant predictor.

In another study, Santa and Cochran (2008) examined the effec-iveness of three different approaches used in anti-drunk drivingublic service announcements in a sample consisting mostly of stu-ents. They found that the empathy approach (i.e., highlighting theonsequences of one’s behavior for others) was perceived to be the

ost effective and elicited the most negative affect, followed by

ear and informational or rational approaches.Thus, although the literature provides considerable insight into

he effectiveness, or ineffectiveness, of negative emotional appeals,

and Prevention 45 (2012) 718– 725 719

much less attention has been paid to positive emotions. Moreover,there is a tendency to regard positive affect as unidimensional (e.g.,humor leads to good mood), which may represent an oversimplifi-cation of the emotional experience and mask a truer understandingof the potential influence of positive emotions on risky driv-ing. Recent developments in the study of emotions indicate thatalthough positive emotions have been studied less than nega-tive, there is much to learn from them. Fredrickson (2003) notesthat compared to negative emotions, positive affects are rela-tively fewer and undifferentiated. In other words, joy, amusement,and serenity, for example, are not easily distinguishable from oneanother, whereas anger, fear, and sadness are distinctly differentexperiences. In her Broaden and Build Theory, Fredrickson main-tains that experiencing a positive emotion leads to broader andmore flexible states of mind and is related to personal growthand development. Her thesis follows studies by Isen and her col-leagues, who showed that when people are made to feel good,their thinking becomes more creative, integrative, flexible, andopen to new information, and they behave in more socially ori-ented modes (for a review, see Isen (2001)). These findings are inline with those of Mittal and Ross (1998), who found that com-pared to negative mood participants, those in a positive mood weremore likely to interpret a strategic decision they had to make asan opportunity, and displayed lower levels of risk taking in theirdecisions.

The findings regarding the importance of positive affect, along-side the paucity of research into the influence of positive emotionson risk taking while driving, indicate the need to examine the con-nection between positive emotions and reckless driving. Moreover,as studies tend to contrast positive and negative emotional states, itis also important to examine whether different positive emotions,such as excitement and arousal vs. relaxation and serenity, have adifferential influence on the willingness to drive recklessly.

The purpose of the present research was therefore two-fold.First, it examined differences in the willingness to drive recklesslybetween individuals who were primed with positive affect andthose in either a neutral or negative affect condition. Secondly, itsought to discover whether differences exist between the influenceof various positive emotions. These are complex questions, as reck-less driving is usually associated with good mood and excitement,as well as with sensation seeking (e.g., Orit Taubman - Ben-Ari,2008). Thus, inducing highly stimulating positive emotions mightlead to a higher intention to drive recklessly than inducing relaxingemotions.

The research consisted of a series of five studies, each of whichemployed a different technique to prime positive emotion and dif-ferent comparison conditions. As gender has repeatedly been foundto be a strong predictor of reckless driving (e.g., Harré et al., 1996;Orit Taubman - Ben-Ari et al., 2004a,b), both men and women wereincluded in all studies, and gender was considered in the analysis.In addition, as youngsters are at an elevated risk for involvementin traffic accidents, with the crash rates for young drivers (16–19years old) throughout the world, including Israel, exceeding thoseof any other age group (e.g., Lotan and Toledo, 2007; Mayhew et al.,2003; Williams, 2003), this population was chosen for the researchprogram.

2. Study 1—the effect of positive vs. natural emotionpriming on intentions to engage in reckless driving

In Study 1, a sample of young Israeli drivers was divided into two

groups and asked to read a “newspaper article” describing either apositive or a neutral story. They then completed a self-report scaletapping their willingness to engage in reckless driving in a series ofhypothetical scenarios.

7 alysis and Prevention 45 (2012) 718– 725

2

2

6uett

2

patdspqt(wtaaWptwoptwpm

swohTu1pdA(wiwo1

dtmsaiop(Sdg

20 O. Taubman - Ben-Ari / Accident An

.1. Method

.1.1. ParticipantsOne hundred and fourteen young Israeli drivers (52 women and

2 men) ranging in age from 17 to 22 (M = 18.95, SD = 1.56) vol-nteered to participate in the study. All of them were students inither high school or pre-academic preparatory programs. The par-icipants were randomly divided into two groups, with 61 exposedo a positive affect prime and 53 to a neutral prime.

.1.2. Materials and procedureThe study was conducted during regular class time. Partici-

ants were told they would be taking part in a study on variousspects of lifestyle and were promised anonymity and full confiden-iality. Following these general instructions, they were randomlyivided into two groups, each of which was presented with a shorttory, supposedly taken from one of the popular daily newspa-ers. They were asked to read it carefully and then answer theuestions on the following page. In the positive priming condi-ion (26 women and 35 men), the story depicted a young personthe same gender as the respondent) saying that they alwaysanted to be a doctor, but their family and friends in the small

own in which they grew up never took them seriously. Eventu-lly, they graduated from high school at the top of their class andpplied to medical school. They waited anxiously for the answer.hen their birthday came, their friends threw them a surprise

arty. They got home that night and found a letter announcingheir acceptance to med school. Their joy was endless, and theyere sure this year’s birthday would be their most unforgettable

ne. In the neutral priming condition (26 women and 27 men),articipants were presented with a story of the same length onhe history of a street, with no emotional trace. The procedureas based on that employed by Mittal and Ross (1998) to primeositive affect, and was aimed at eliciting happiness and goodoodAfter the priming manipulation, all participants completed a

elf-report scale tapping proneness to reckless driving. The scaleas developed by Orit Taubman - Ben-Ari et al. (1999), and consists

f 10 short hypothetical driving scenarios (e.g., “You are drivingome from a pub. You drank two beers and you are a bit tired.he car is full of your friends who are shouting at you to speedp. What do you think are the chances that you will drive at40 km/h?”). For each scenario, responses are indicated on an 11-oint scale, ranging from 0 (0% chance of behaving in the mannerescribed) to 10 (100% chance of behaving in the manner described).s Cronbach’s alpha for the scale was adequate in this sample

.83), each participant’s responses to the 10 items were averaged,ith higher scores reflecting a higher proneness to reckless driv-

ng. It is important to note that the effects, which were foundhen using this self-report measure were corroborated by results

f driving in a driving simulator (Orit Taubman - Ben-Ari et al.,999).

A manipulation check was carried out at the end of the proce-ure in which participants were asked to rate the degree to whichhey were feeling each of 10 emotions (5 positive: happy, good

ood, pleased, calm, satisfied; 5 negative: angry, scared, frustrated,ad, bad mood). As Cronbach’s alpha for the 5 positive emotions wasdequate in the sample (.83), each participant’s responses to the 5tems were averaged, with higher scores reflecting a higher degreef positive affect. The manipulation check confirmed that partici-ants in the positive affect condition reported higher positive affect

M = 3.40, SD = 1.00) than those in the neutral condition (M = 2.92,D = 1.03), t(108) = 2.47, p < .05. Finally, all participants completed aemographic questionnaire to obtain information such as age andender.

Fig. 1. Comparing proneness to reckless driving means and 95% confidence intervalby priming conditions and gender (Study 1).

2.2. Results and discussion

A 2 × 2 ANOVA for experimental condition (positive, neutral) Xgender (men, women) was performed on the willingness to engagein reckless driving. Means and 95% confidence intervals appear inFig. 1. The analysis yielded significant main effects for condition,F(1,113) = 6.40, p < .01, and for gender, F(1,113) = 11.19, p < .001.Thus, respondents in the positive affect condition expressed greaterwillingness to engage in risky driving (M = 2.73, SD = 1.57) thanthose in the neutral condition (M = 1.99, SD = 1.31). In addition, menreported greater willingness to engage in risky driving than women(M = 2.81, SD = 1.57; M = 1.87, SD = 1.23, for men and women,respectively). No significant interaction emerged, F(1,113) = 1.57,n.s.

The results therefore indicated that a happy positive affectresulted in higher proneness to reckless driving as compared to aneutral affect, and that men reported being more willing to engagein reckless driving than women. However, the affective manipula-tion did not have differential effects on men and women, as bothgenders tended to respond in the same way. Before any conclusionscould be derived from the findings, however, it was necessary toconfirm the results using an independent sample and a differenttype of manipulation to elicit positive affect. Furthermore, it waswished to compare the effect of the positive affect condition notonly with a neutral one, but also with a negative emotion manipu-lation. These aims were addressed in Study 2.

3. Study 2—the effect of positive vs. natural and negativeemotion priming on intentions to engage in reckless driving

In Study 2, young Israeli drivers completed the same self-reportscale to assess proneness to reckless driving as described in Study1. Here, however, they were first divided into three groups andexposed to a positive, negative, or neutral induction, and a differentmethod of emotional priming was used.

3.1. Method

3.1.1. ParticipantsAn independent sample of 135 Israeli high school students with

a driver’s license (67 women and 68 men) ranging in age from 17 to18 (M = 17.5, SD = .50) volunteered to participate in the study. They

were divided into three groups, with 49 participants in the positiveaffect condition, 44 in the negative affect condition, and 42 in theneutral condition.

O. Taubman - Ben-Ari / Accident Analysis

Fb

3

pawhsmci1(wFu

3

ntacergtbctdmF

trntetc

ttwdwadt

ig. 2. Comparing proneness to reckless driving means and 95% confidence intervaly priming conditions and gender (Study 2).

.1.2. Materials and procedureThe study was conducted during regular class time. Partici-

ants received similar instructions to those described in Study 1,nd then were randomly divided into three groups and asked torite as vividly as possible about their memories of either “a veryappy occasion” (24 men, 25 women), “a very aggravating occa-ion” (28 men, 16 women), or “a regular day in their life” (16en, 26 women). After the priming manipulation, all participants

ompleted the self-report scale tapping proneness to reckless driv-ng (Orit Taubman - Ben-Ari et al., 1999), as described in Study. As Cronbach’s alpha for the scale was adequate in this sample.81), each participant’s responses to the 10 items were averaged,ith higher scores reflecting a higher proneness to reckless driving.

inally, all participants completed the demographic questionnairesed in Study 1.

.2. Results and discussion

A 2 × 2 ANOVA for experimental condition (positive, negative,eutral) X gender (men, women) was performed on the willingnesso engage in reckless driving. Means and 95% confidence intervalsppear in Fig. 2. The analysis yielded a significant main effect forondition, F(2,134) = 3.44, p < .05, and a marginally significant mainffect for gender, F(2,134) = 3.54, p = .06. A Simple Main Effects testevealed that respondents in the positive affect condition expressedreater willingness to engage in risky driving (M = 2.87, SD = 1.55)han those in the neutral condition (M = 2.11, SD = 1.12; p < .05),ut no difference was found between the positive and negativeonditions or between the negative and neutral conditions. In addi-ion, men tended to report greater willingness to engage in riskyriving than women (M = 2.91, SD = 1.49; M = 2.17, SD = 1.36, foren and women, respectively). No significant interaction emerged,

(2,134) = 1.54, n.s.The results are in line with those of Study 1, showing once again

hat positive affect, evoked here by recalling happy memories, waselated to higher proneness to reckless driving as compared to aeutral affect. Interestingly, anger-evoking memories did not leado a significant difference in the willingness to drive recklessly thanither positive or neutral memories. Again, although gender tendedo be an additional predictor of proneness to reckless driving, itould not explain the effects of the emotional priming.

Although the findings of Studies 1 and 2 were highly consis-ent, strengthening confidence in their validity and generalizability,here were still two issues that needed to be examined. The first washether all positive affects lead to the same reaction, or whetherifferent kinds of positive affect might result in different extents of

illingness to engage in reckless driving. The second was whether

nother measure assessing aspects of driving behavior, namelyriving styles, would confirm the results yielded by the pronenesso reckless driving scale used in the first two studies.

and Prevention 45 (2012) 718– 725 721

4. Study 3—the effect of positive arousal vs. positiverelaxation emotion priming on intentions to engage inreckless driving and reported driving style

Study 3 again employed a different means to prime affect, andtwo types of positive affect were primed: participants were showna film aimed at eliciting either excitement and arousal or calm andrelaxation. They then reported on their willingness to engage inreckless driving. In addition, they completed a driving style ques-tionnaire. Although driving styles are considered relatively stablecharacteristics, it was presumed that they might be affected by theaffect-inducing manipulations, thus providing another measure forrisky vs. careful driving.

4.1. Method

4.1.1. ParticipantsAn independent sample of 103 Israeli high school students with

a driver’s license (58 women and 45 men) ranging in age from 17to 19 (M = 17.8, SD = .56) volunteered to participate in the study.

4.1.2. Materials and procedureThe study was conducted during regular class time. Participants

received similar instructions to those in Studies 1 and 2 and wererandomly divided into two groups, each of which was shown a 2-min film. Participants in the arousal condition (31 women, 19 men)watched a segment from the movie The Fast and the Furious show-ing young people racing cars. The characters in the film reflected ahigh degree of happiness, thrill, and stimulation. Participants in therelaxation condition (27 women, 26 men) watched a nature movieshowing dolphins swimming gracefully in the open sea. The par-ticipants were all told they would later be asked questions aboutwhat they had seen

The priming instruments had previously been validated on aseparate sample of 15 Israeli high school students who were pre-sented with the two film segments and asked to describe thefeelings each evoked. Content analysis of the responses revealedthat the dominant feelings elicited by the arousal-priming filmwere excitement, enjoyment, and thrill; those evoked by therelaxation-priming film were pleasure, delight, calm, and harmony.

After the priming manipulation, participants completed the 10-item proneness to reckless driving scale described in Study 1. Inthis sample, Cronbach’s alpha for the 10 items was .86. This wasfollowed by the Multidimensional Driving Style Inventory (MDSI, OritTaubman - Ben-Ari et al., 2004a), a validated and reliable 44-itemscale assessing four broad domains of driving styles: Reckless andcareless – 11 items relating to the tendency to seek stimulation,sensation, and risk when driving; Anxious – 19 items tapping thetendency to feel distressed when driving, to display signs of anxi-ety and fear in given driving situations, and to express doubt andlack of driving confidence; Angry and hostile – 5 items relating tothe tendency to be hostile toward other drivers, to behave aggres-sively, and to feel intense anger when driving; Patient and careful –9 items relating to the tendency to be polite toward other drivers,to display patience while driving, and to drive carefully. Partici-pants were asked to rate the extent to which they believe each itemreflects their feelings, thoughts, and behavior on the road, indicat-ing their responses on a 6-point scale ranging from 1 (not at all) to6 (very much).

In previous studies (Orit Taubman - Ben-Ari et al., 2004a), thesefactors were found to reflect coherent and meaningful driving stylesand to have adequate internal reliability and validity. They were

also found in another study to correlate with performance mea-sures collected in a simulator (i.e., driving speed, number of drivingmaneuvers, passing gaps; Farah et al. (2009)). In the current study,Cronbach’s alpha coefficients were .96 for the anxious style, .87 for

722 O. Taubman - Ben-Ari / Accident Analysis

Table 1Means and standard deviations of the proneness to reckless driving, by primingconditions and gender (Study 3).

Arousal Relaxation

M SD M SD

Willingness to drive recklesslyMen 4.88 1.36 3.09 1.80Women 3.18 1.22 2.61 1.82

Anxious driving styleMen 3.95 .76 1.89 .60Women 3.34 .91 2.34 .60

Risky driving styleMen 4.16 .64 2.97 .85Women 3.38 .82 2.58 .71

Angry driving styleMen 4.23 .54 2.95 1.10Women 3.35 .95 2.66 .87

Careful driving style

tsar

u

4

affacpa

acFiFrthaiMS

aFaeSaSaewtSr

Men 4.05 .40 4.62 .54Women 3.85 .52 4.66 .62

he reckless style, .77 for the angry style, and .62 for the carefultyle. Four scores were therefore computed for each participant byveraging the scores on the items in each factor, with higher scoreseflecting greater endorsement of the relevant driving style.

All participants also completed the demographic questionnairesed in Study 1.

.2. Results and discussion

A 2 × 2 MANOVA for experimental condition (aroused positiveffect, relaxed positive affect) X gender (men, women) was per-ormed on the willingness to engage in reckless driving and on theour driving styles (anxious, risky, angry, and careful). The resultsppear in Table 1. The analysis yielded a significant main effect forondition, F(5,95) = 25.22, p < .001, and for gender, F(5,95) = 6.08,

< .001. A significant interaction between condition and genderlso emerged, F(5,95) = 5.28, p < 001.

Two-way ANOVAs conducted for each of the dependent vari-bles revealed significant differences between the experimentalonditions in the willingness to engage in reckless driving,(1,99) = 13.77, p < .001, as well as in each of the driving styles: anx-ous, F(1,99) = 114.12, p < .001; risky, F(1,99) = 43.74, p < .001; angry,(1,99) = 32.01, p < .001; and careful, F(1,99) = 43.83, p < .001. Thus,espondents in the arousal condition expressed greater willingnesso engage in risky driving (M = 4.23, SD = 1.54), and also reportedigher anxious (M = 3.72, SD = .86), risky (M = 3.86, SD = .80), andngry (M = 3.90, SD = .12), and lower careful (M = 3.97, SD = .46) driv-ng styles than those in the relaxation condition (M = 2.85, SD = 1.81;

= 2.12, SD = .09; M = 2.77, SD = .11; M = 2.80, SD = .14; M = 4.64,D = .57, respectively).

The analysis also revealed significant differences between mennd women in the willingness to engage in reckless driving,(1,99) = 11.70, p < .001, and in the risky, F(1,99) = 15.03, p < .001,nd angry, F(1,99) = 11.52, p < .001, driving styles. Thus, menxpressed greater willingness to engage in risky driving (M = 4.06,D = 1.79), and also reported on higher risky (M = 3.62, SD = .95) andngry (M = 3.65, SD = 1.05) driving styles than women (M = 2.85,D = 1.61; M = 2.91, SD = .85; M = 2.94, SD = .96, respectively). Inddition, a significant interaction between condition and gendermerged for the anxious driving style, F(1,99) = 13.73, p < .001, as

ell as a marginally significant interaction between condition and

he willingness to drive recklessly, F(1,99) = 3.67, p = .058. Tests forimple Main Effects indicated that in the arousal condition, meneported higher anxious driving style than women, F(1, 99) = 6.55,

and Prevention 45 (2012) 718– 725

p < 01, whereas in the relaxation condition, women reported higheranxious driving style, F(1, 99) = 7.44, p < 01. Furthermore, the relax-ation condition was associated with lower reports of anxiousstyle for both men, F(1, 99) = 125.69, p < 001, and women, F(1,99) = 20.11, p < 001, with the effect of the interaction being con-siderably stronger for men.

The results of Study 3 reveal significant differences in the reac-tion to different sorts of positive affect inductions, and show thatdrivers’ reactions may depend on the specific nature of the affect.The results also indicate that the affective manipulation influ-enced not only the willingness to engage in reckless driving, butalso reports of driving styles, which replicated and validated thefindings yielded by the willingness to drive recklessly measure.However, one potential limitation of the study is that the film usedto prime arousal actually showed risky driving, potentially provid-ing a model for reckless driving in addition to eliciting a positiveemotion. A different kind of arousal priming was therefore chosenfor the next study.

5. Study 4—the effect of unrelated positive arousal vs.positive relaxation emotion priming on intentions toengage in reckless driving

Study 4 employed a similar means to prime the same two typesof positive affect: participants were shown a film aimed at elicitingeither excitement or calm. This time, neither of the priming filmswas related to driving in any way. Participants then reported ontheir willingness to engage in reckless driving.

5.1. Method

5.1.1. ParticipantsAn independent sample of 104 young Israeli non-combat sol-

diers with a driver’s license (62 women and 42 men) ranging in agefrom 18 to 22 (M = 19.21, SD = 1.25) volunteered to participate inthe study.

5.1.2. Materials and procedureThe study was conducted during a class in the soldiers’ units,

when they were asked to participate in a study on social issues.They received similar instructions to those in Study 3, and wererandomly divided into two groups, each of which watched a 5-min film. Participants in the arousal condition (33 women, 19 men)watched a segment from the film Fight Club showing people fightingaccompanied by loud rock music. The fighters displayed a senseof thrill and satisfaction from the experience. Participants in therelaxation condition (29 women, 23 men) watched a nature movieon communication between animals, which was accompanied byserene music. The participants were all told they would later beasked questions about what they had seen

The priming instruments had previously been validated on aseparate sample of 15 Israeli high school students who were pre-sented with the two film segments and asked to describe thefeelings each evoked. Content analysis of the responses revealedthat the dominant feelings elicited by the arousal-priming filmwere happiness, good mood, and fun; those evoked by therelaxation-priming film were lightness, tenderness, and calm.

After the priming manipulation, participants completed the 10-item proneness to reckless driving scale described in Study 1. In thissample, Cronbach’s alpha for the 10 items was .75. All participantsalso completed the demographic questionnaire used in Study 1.

5.2. Results and discussion

A 2 × 2 ANOVA for experimental condition (aroused pos-itive affect, relaxed positive affect) X gender (men, women)

O. Taubman - Ben-Ari / Accident Analysis and Prevention 45 (2012) 718– 725 723

Fb

wMspbnci(

Snrr

innao

6od

sic

6

6

l(wic

6

mttt

Table 2N’s, means, and standard deviations of the proneness to reckless driving, by primingconditions and gender (Study 5).

Meaning in life Neutral

M SD M SD

Willingness to drive recklesslyMen 2.47 1.46 3.66 2.34Women 1.34 1.03 2.68 1.55

Anxious driving styleMen 2.15 .43 2.23 .63Women 2.24 .55 2.34 .59

Risky driving styleMen 2.75 .65 3.14 .88Women 1.96 .41 2.57 .68

Angry driving styleMen 2.60 .99 2.70 1.01Women 1.83 .67 2.33 .98

Careful driving styleMen 4.82 .46 4.57 .90

ig. 3. Comparing proneness to reckless driving means and 95% confidence intervaly priming conditions and gender (Study 4).

as performed on the willingness to engage in reckless driving.eans and 95% confidence intervals appear in Fig. 3. The analy-

is yielded a significant main effect for condition, F(1,103) = 4.98, < .05, but not for gender, F(1,103) = 1.10, n.s. The interactionetween condition and gender was not significant, F(1,103) = 0.10,.s. In other words, both men and women in the arousalondition expressed greater willingness to engage in risky driv-ng (M = 2.49, SD = 1.52) than those in the relaxation conditionM = 1.91, SD = 1.11).

On the whole, therefore, the results of Study 4 replicate those oftudy 3, indicating that arousal induction leads to a higher prone-ess to risky driving than relaxation induction. The finding waseplicated here even though the priming film was not directlyelated to driving as in the previous study.

Taken together, Studies 1–4 reveal that positive affect resultedn higher willingness to engage in reckless driving than neutral oregative affect, and that arousal priming resulted in greater prone-ess to reckless driving than relaxation priming. In the final study,nother possibility was explored: that thinking about the meaningf life might also serve as a calming experience for young drivers.

. Study 5—the effect of meaning in life vs. neutral primingn intentions to engage in reckless driving and reportedriving style

In Study 5, young Israeli drivers completed the same self-reportcales to assess willingness to engage in reckless driving and driv-ng styles as described in Study 3. Here, however, two differentonditions were primed: meaning in life and a neutral condition.

.1. Method

.1.1. ParticipantsAn independent sample of 115 Israeli youngsters with a driver’s

icense (61 women and 54 men) ranging in age from 17 to 22M = 18.9, SD = 1.40) volunteered to participate in the study. Theyere divided into two experimental conditions, with 55 partic-

pants in the meaning in life condition, and 60 in the neutralondition.

.1.2. Materials and procedureThe study was conducted at various military bases and youth

ovement centers. Participants received similar instructions tohose described in Study 1, and then were randomly divided intowo groups. In the meaning in life condition (34 women, 21 men),hey were asked: (a) to write in detail what meaning they see in

Women 5.04 .56 4.62 .63

their life; and (b) to describe the feelings that thinking about themeaning of life evokes in them. Participants in the neutral condi-tion (27 women, 33 men) were asked to describe: (a) their favoritefood; and (b) a restaurant they like to eat in. After the primingmanipulation, all participants completed the self-report scale tap-ping proneness to reckless driving (Orit Taubman - Ben-Ari et al.,1999) described in Study 1. Cronbach’s alpha for the scale in thissample was .87. They also completed the MDSI (Orit Taubman -Ben-Ari et al., 2004a), described in Study 3. In the current study,Cronbach’s alpha coefficients were .81 for the anxious style, .83 forthe reckless style, .77 for the angry style, and .75 for the careful style.Finally, all participants completed the demographic questionnaireused in Study 1.

6.2. Results and discussion

A 2 × 2 MANOVA for experimental condition (meaning, neu-tral) X gender (men, women) was performed on the willingnessto engage in reckless driving and the four driving styles (anxious,risky, angry, and careful). The results appear in Table 2. The analy-sis yielded a significant main effect for condition, F(5,107) = 3.74,p < .01, and for gender, F(5,107) = 7.99, p < .001. The interactionbetween condition and gender was not significant, F(5,107) = .33,n.s.

Two-way ANOVAs conducted for each of the dependentvariables revealed significant differences between the exper-imental conditions in the willingness to engage in recklessdriving, F(1,114) = 15.65, p < .001, as well as in the risky driv-ing style, F(1,114) = 15.00, p < .001, and the careful driving style,F(1,114) = 6.88, p < .01. Thus, respondents in the meaning in lifecondition expressed lower willingness to engage in risky driv-ing (M = 1.77, SD = 1.32), and also reported lower risky (M = 2.26,SD = .64) and higher careful (M = 4.96, SD = .53) driving stylesthan those in the neutral condition (M = 3.22, SD = 2.07; M = 2.88,SD = .84; M = 4.59, SD = .78, respectively).

The two-way ANOVAs also revealed significant differencesbetween men and women in the willingness to engage in recklessdriving, F(1,114) = 10.97, p < .001, and in the risky, F(1,114) = 28.32,p < .001, and angry, F(1,114) = 10.93, p < .001, driving styles. Thus,men expressed a greater proneness to reckless driving (M = 3.20,

SD = 2.11), and also reported higher risky (M = 2.99, SD = .81) andangry (M = 2.66, SD = .99) driving styles than women (M = 1.93,SD = 1.44; M = 2.22, SD = .62, M = 2.05, SD = .85, respectively).

7 alysis

iaiplo

7

pUiroprpcfi

ioddailwfesoIdio

mcTammefttattiaweabu

neafi

24 O. Taubman - Ben-Ari / Accident An

The results of Study 5 indicate that thinking about meaningn life and the emotions associated with such thoughts results in

lower tendency to drive recklessly and to more careful drivingntentions, compared to the control condition. This insight mightave the way for new interventions that may use the meaning of

ife concept as a tool to highlight the things worth living for, insteadf on what might be lost by reckless driving.

. General discussion

The studies reported here examined the effects of positive affectriming on the willingness to drive recklessly among young drivers.sing various priming procedures to elicit positive affect, the stud-

es yielded consistent results, showing that happiness primingesulted in a greater willingness to drive recklessly than neutralr anger-evoking priming. However, a closer look at differentiatedositive affects indicated that positive affect of a relaxing natureesulted in a lower willingness to engage in reckless driving thanositive affect in the form of arousal. Moreover, it was found thatausing young drivers to think about the meaning of life and theeelings associated with such contemplation resulted in a lowerntention to drive recklessly than a neutral condition.

These results are undoubtedly thought-provoking, particularlyn view of the scarcity of literature on the effects of positive affectn reckless driving. The findings indicate that the positive-negativeimension is not sufficient in predicting a change in the desiredirection, and highlight the need to relate to the kind of positiveffect elicited. Although it is hard to compare the studies directly,t does appear that priming which evokes excitement and arousaleads to a greater willingness to drive recklessly than a priming,

hich evokes calm and relaxation. Moreover, the insights gainedrom thinking about the meaning in life may enhance the calmingffect created by positive affect primings of a relaxing nature. Thisuggests that it might be wise to rethink the use of both fear appealsn the one hand, and simple positive affect appeals on the other.n other words, knowing that the effect of fear appeals is limitedoes not lend immediate validity to positive appeals. Rather, var-

ous kinds of positive affect should be considered in order to findut which of them might serve the aim of reducing risky driving.

It is important to note that although four of the studies founden to have higher reckless driving intentions than women, as

onsistent with previous research (e.g., Laapotti et al., 2003; Oritaubman - Ben-Ari et al., 2004b), this variable was unable toccount for the effects of positive affect in our studies. Further-ore, while certain previous findings indicate that fear appealsight have differential effects on men and women (Goldenbeld

t al., 2008; Lewis et al., 2007a), no such effects were found hereor positive affect. The sole exception was the finding related tohe anxious driving style in Study 3, which showed that althoughhe relaxation induction was associated with lower reports of thenxious style for both men and women than the arousal induction,he effect was significantly greater for men. In addition, whereas inhe arousal condition men reported higher levels of anxious driv-ng than women, in the relaxation condition women endorsed thenxious driving style more than men. It might be speculated thathile the relaxing atmosphere helped to reduce anxieties in gen-

ral, the arousal priming elicited not only positive affect, but alsonxiety, especially for men. However, these findings will have toe more thoroughly examined in future studies before we can fullynderstand their implications.

The current series of studies has certain strengths, as well as a

umber of weaknesses. First, different manipulations were used tolicit positive affect: reading a positive mood evoking newspaperrticle, recalling happy memories, watching exciting or relaxinglms, and thinking about meaning in life and its related emotions.

and Prevention 45 (2012) 718– 725

However, as the studies did not employ material from real cam-paigns, the results warrant examination in future field studies.Secondly, a reliable and validated measure was used to learn aboutparticipants’ willingness to drive recklessly, and in two studies asecond validated measure of driving styles was used to replicate theresults and establish their reliability. Nevertheless, both measuresare self-report scales, which might suffer from social desirabilityand other biases. Future studies might attempt to replicate thefindings by means of more objective measures of actual perfor-mance, such as driving simulators, though it should be mentionedthat a recent study comparing self-reports and actual driving foundthem to be strongly correlated (Boufous et al., 2010), and anotherstudy found driving-related self-reports to be unbiased by socialdesirability (Sullman and Taylor, 2010). Thirdly, the research wasconducted on five independent and sound samples of young driversdrawn from various contexts (high schools, academic preparatoryprograms, army units) throughout Israel. These samples, however,were not constituted using representative sampling methods, andall consist of youngsters. Thus, the application of the results to otherpopulations is limited and warrants further examination. Finally,the studies compared the effects of positive affect not only to aneutral condition, but also to a negative affect condition, and alsocompared different kinds of positive affect. However, they focusedonly on the affective component, presenting no salient message todrive carefully and ignoring other elements, which have emergedas pertinent in the literature on fear appeals, such as self-efficacy,involvement, and the relevance of the message. In addition, otherpositive emotions, such as pride, gratitude, and hope, should alsobe explored. In many ways, therefore, the current research can beconsidered an initial, exploratory examination of the usefulness ofpositive affect in road safety interventions. Future studies shouldtake into account these and other variables, as well as investigatingindividual differences in the reaction to positive appeals. Matchingthe experimental groups according to personality characteristicsknown to be related to risky driving, or according to driving stylescan help to demonstrate more clearly that priming indeed led togroup differences.

Thus, despite the merits of the current series of studies, addi-tional research is needed to increase our understanding of thespecific emotions that may best help to moderate reckless driv-ing, and the specific behaviors that may benefit from the inductionof positive affect. This would not only improve the external validityof the present findings, but would also advance our knowledge ofthe psychological processes underlying decision making and riskybehavior. Nevertheless, the current research already seems to sug-gest at least one specific avenue that should be considered whendesigning safe driving campaigns: encouraging people to thinkabout the meaning of life, that is, reminding them of the thingsworth living for, might help in reducing the willingness to engagein risk taking on the road.

Acknowledgement

The studies were partly supported by the Or Yarok Association,Israel.

References

Boufous, S., Ivers, R., Senserrick, T., Stevenson, M., Norton, R., Williamson, A., 2010.Accuracy of self-report of on-road crashes and traffic offences in a cohort ofyoung drivers: the DRIVE study. Injury Prevention 16, 275–277.

Conway, M., Dubé, L., 2002. Humor in persuasion on threatening topics: effectiveness

is a function of audience sex role orientation. Personality and Social PsychologyBulletin 28, 863–873.

Delhomme, P., De Dobbeleer, W., Forward, S., Simoes, A., 2009. Manual for Designing,Implementing and Evaluating Road Safety Communication Campaigns. BelgianRoad Safety Institute, Brussels.

alysis

D

E

F

F

F

F

G

H

H

H

I

L

L

L

L

L

O. Taubman - Ben-Ari / Accident An

ubé, L., Morgan, M.S., 1996. Trend effects and gender differences in retrospec-tive judgments of consumption emotions. Journal of Consumer Research 23,156–162.

lliott, B., 2003. The psychology of fear appeals re-visited. Paper presented at theRoad Safety Research. In: Policing and Education Conference, 24–26 September,Sydney, NSW.

arah, H., Bekhor, S., Polus, A., Toledo, T., 2009. A passing gap acceptance model fortwo-lane rural highways. Transportmetrica 5, 159–172.

isher, R.J., Dubé, L., 2005. Gender differences in response to emotional adver-tising: a social desirability perspective. Journal of Consumer Research 31,850–858.

loyd, D.L., Prentice-Dunn, S., Rogers, R.W., 2000. A meta-analysis of research onprotection motivation theory. Journal of Applied Social Psychology 30, 407–429.

redrickson, B.L., 2003. The value of positive emotions. American Scientist 91,330–335.

oldenbeld, C., Twisk, D., Houwing, S., 2008. Effects of persuasive communicationand group discussions on acceptability of anti-speeding policies for male andfemale drivers. Transportation Research Part F: Traffic Psychology and Behaviour11, 207–220.

arré, N., Field, J., Kirkwood, B., 1996. Gender differences and areas of commonconcern in the driving behaviors and attitudes of adolescents. Journal of SafetyResearch 27, 163–173.

astings, G., Stead, M., Webb, J., 2004. Fear appeals in social marketing: strategicand ethical reasons for concern. Psychology and Marketing 21, 961–986.

yman, M.R., Tansey, R., 1990. The ethics of psychoactive ads. Journal of BusinessEthics 9, 105–114.

sen, A.M., 2001. An influence of positive affect on decision making in complexsituations: theoretical issues with practical implications. Journal of ConsumerPsychology 11, 75–85.

aapotti, S., Keskinen, E., Rajalin, S., 2003. Comparison of young male and femaledrivers’ attitude and self-reported traffic behavior in Finland in 1978 and 2001.Journal of Safety Research 34, 579–587.

ewis, I., Watson, B., Tay, R., 2007a. Examining the effectiveness of physical threatsin road safety advertising: the role of the third-person effect, gender, andage. Transportation Research Part F: Traffic Psychology and Behaviour 10,48–70.

ewis, I., Watson, B., Tay, R., White, K., 2007b. The role of fear appeals in improvingdriver safety: a review of the effectiveness of fear-arousing (threat) appeals inroad safety advertising. International Journal of Behavioral Consultation andTherapy 3, 203–222.

ewis, I., Watson, B., White, K., 2008a. An examination of message-relevant affect inroad safety messages: Should road safety advertisements aim to make us feelgood or bad? Transportation Research Part F: Traffic Psychology and Behaviour

11, 403–417.

ewis, I., Watson, B., White, K., 2008b. Predicting future speeding behaviour: theappeal of positive emotional appeals for high risk road users. Paper presentedat the High Risk Road Users Conference. In: Joint ACRS-Travel safe NationalConference, Brisbane, 18–19 September, 2008.

and Prevention 45 (2012) 718– 725 725

Lotan T., Toledo T., 2007. Driving Patterns of Young Drivers within a Graduated DriverLicensing System. Preprints of the 86th Annual Meeting of the TransportationResearch Board, Washington, D.C.

Mayhew, D.R., Simpson, H.M., Pak, A., 2003. Changes in collision rates among novicedrivers during the first months of driving. Accident Analysis and Prevention 35,683–691.

Mittal, V., Ross, W.T., 1998. The impact of positive and negative affect and issueframing on issue interpretation and risk taking. Organizational, Behavior andHuman Decision Processes 76, 298–324.

Santa, A.F., Cochran, B.N., 2008. Does the impact of anti-drinking and driving publicservice announcements differ based on message type and viewer characteris-tics? Journal of Drug Education 38, 109–129.

Schoenbachler, D.D., Whittler, T.E., 1996. Adolescent processing of social and phys-ical threat communications. Journal of Advertising 25, 37–54.

Struckman-Johnson, C., Struckman-Johnson, D., Gilliland, R.C., Ausman, A., 1994.Effect of persuasive appeals in AIDS PSAs and condom commercials on intentionsto use condoms. Journal of Applied Social Psychology 24, 2223–2244.

Sullman, M.J.M., Taylor, J.E., 2010. Social desirability and self-reported driving behav-iors: should we be worried? Transportation Research Part F 13, 215–221.

Orit Taubman - Ben-Ari, O., 2008. Motivational sources of driving and their associ-ations to reckless driving cognitions and behavior. European Review of AppliedPsychology 58, 51–64.

Orit Taubman - Ben-Ari, O., Florian, V., Mikulincer, M., 1999. The impact of mortalitysalience on reckless driving: a test of terror management mechanisms. Journalof Personality and Social Psychology 76, 35–45.

Orit Taubman - Ben-Ari, O., Florian, V., Mikulincer, M., 2000. Does a threat appealmoderate reckless driving? A terror management theory perspective. AccidentAnalysis and Prevention 32, 1–10.

Orit Taubman - Ben-Ari, O., Mikulincer, M., Gillath, O., 2004a. The MultidimensionalDriving Style Inventory-Scale construct and validation. Accident Analysis andPrevention 36, 323–332.

Orit Taubman - Ben-Ari, O., Mikulincer, M., Iram, A., 2004b. A multifactorial frame-work for understanding reckless driving—appraisal indicators and perceivedenvironmental determinants. Transportation Research Part F: Traffic Psychologyand Behaviour 7, 333–349.

Tay, R., Ozanne, L., 2002. Who are we scaring with high fear road safety campaigns?Asia Pacific Journal of Transport 4, 1–12.

Williams, A.F., 2003. Teenage Drivers: Patterns of Risk. Journal of Safety Research34, 5–15.

Witte, K., Allen, M., 2000. A meta-analysis of fear appeals: implications for effectivepublic health campaigns. Health Education & Behavior 27 (October), 591–615.

Witte, K., Berkowitz, J.M., Cameron, K.A., McKeon, J.K., 1998. Preventing the spreadof genital warts: using fear appeals to promote self-protective behaviors. Health

Education & Behavior 25, 571–585.

Wundersitz, L.N., Hutchinson, T.P., Woolley, J.E., 2010. Best practice in road safetymass media campaigns: a literature review. In: CASR Reports Series, CASR074.Centre for Automotive Safety Research, The University of Adelaide, SouthAustralia.