The Motorcycle Rider Behavior Questionnaire: Psychometric properties and application amongst novice riders in Australia

https://doi.org/10.1016/j.trf.2013.10.005Get rights and content

Highlights

  • The MRBQ psychometric properties differ between novice and experienced riders.

  • Only the speed violations scale had sufficient reliability.

  • Only the protective gear scale demonstrated some content validity.

  • The stunts scale was a significant predictor of police-recorded crashes.

  • The protective gear scale was a significant predictor of police-recorded offences.

Abstract

The Motorcycle Rider Behavior Questionnaire (MRBQ) was developed to measure behavioral factors influencing motorcyclists’ crash risk including errors and violations as well as the use of motorcycle safety equipment via self-report. The aims of the present study were to (1) examine the previously examined psychometric properties of the MRBQ including the factor structure, internal consistency, and predictive validity in terms of self-reported crashes amongst experienced riders in the UK and Turkey; (2) examine the psychometric properties of the MRBQ not yet examined, including its stability, content validity, and predictive validity in terms of police-recorded crashes and offences as well as self-reported near crashes and crashes; and (3) assess the applicability of the MRBQ to a population of novice riders in Australia, to whom the MRBQ has not been applied to date. Novice riders (N = 1305) in the state of Victoria, Australia participated in the present study. Confirmatory factor analyses showed that the present data did not fit with the previously found factor models in experienced riders in the UK and Turkey. Principal axis factoring was performed to respecify the MRBQ factor model amongst novice riders and revealed four scales: errors; speed violations; stunts; and protective gear. The insufficient internal consistency, stability, content and predictive validity demonstrated by the MRBQ in the present study and some inconsistencies amongst the three MRBQ studies suggest that the development and refinement of the MRBQ items are required before wider use of the MRBQ instrument, especially amongst novice riders. Possible causes of the limited reliability and validity of the current MRBQ are discussed to inform further development and refinement of the items, thereby making the MRBQ more useful in future research to understand and evaluate riders’ behaviors.

Introduction

Understanding motorcyclists’ on-road behavior is critical for not only the development but also the evaluation of interventions specifically targeted for motorcyclists. Risky on-road behaviors have been historically dichotomized broadly into errors and violations (Reason, Manstead, Stradling, Baxter, & Campbell, 1990). The Motorcycle Rider Behavior Questionnaire (MRBQ) was developed to measure motorcyclists’ errors and violations, as well as the use of motorcycle safety equipment, via self-report based on the widely used Driver Behavior Questionnaire (DBQ) and Reason et al’s (1990) model of aberrant driver behavior (Elliott, Baughan, & Sexton, 2007). Official records of traffic offences are useful to understand risky on-road behaviors. However, they do not necessarily capture all errors and violations and non-use of safety equipment by the motorcyclists, due to the variability in enforcement (people do not always get caught for breaking road laws) and because in most jurisdictions and countries the use of safety equipment is not mandated except for helmets. Self-report can be a useful additional tool when alternative data collection methods such as direct observations and official records are not feasible due to limited research resources. For these reasons development and evaluation of a self-report measure of on-road riding behaviors such as the MRBQ is beneficial for motorcycle safety research and practice. Ensuring the validity and reliability of any research instruments including the MRBQ is essential before their wider use.

Two studies have so far to some extent examined the validity and reliability of the MRBQ in motorcycle populations in the United Kingdom (UK) (Elliott et al., 2007) and Turkey (Özkan, Lajunen, Dogruyol, Yildirim, & Çoymak, 2012). In both studies principal component analyses identified five components which were named traffic errors, control errors, speed violations, stunts, and safety equipment. These five sub-scales were shown to have good internal consistency; however, the item constituents of each scale were not exactly the same in the two studies. Although both studies reported the predictive validity of some of the MRBQ scales in terms of self-reported crashes, the findings were not consistent between the two studies. The stunts factor was found to be predictive of self-reported crashes in the Turkish rider sample, whereas traffic errors were in the UK rider sample, and additionally control errors and speed violations when restricted to crashes for which the respondents took blame. In the Turkish study, both speed violations and stunts were significant predictors of self-reported traffic offences.

Further, three key psychometric features remain untested for the MRBQ. The ability of a self-report questionnaire to predict police-recorded as well as self-reported crashes and offences would increase its utility in road safety research. These were not examined in the two European studies (Elliott et al., 2007, Özkan et al., 2012). Aside from the possible consistency motif bias (people’s tendency to want to respond consistently across related measures; (Af Wåhlberg, 2011, Podsakoff et al., 2003), MRBQ factor and crash/offence relationships may vary depending on whether the data was obtained via self-report (which includes minor crashes) or the police (typically limited to serious crashes) due to the different nature of crashes represented (Boufous et al., 2010). The two studies also did not examine the content validity or the stability of the MRBQ scales. For the results based on MRBQ to be meaningful it must also be able to measure what is intended to be measured – errors, violations, and use of safety equipment (i.e. content validity). If rider behaviors are expected to be stable over time, the stability of the MRBQ is critical to ensure its ability to measure behaviors consistently over time. The stability of the tool can also reflect the modifiability of behaviors over time, which has implications for the development and evaluation of rider interventions.

Furthermore, no published study has examined the applicability of the MRBQ to novice riders in Australia. Self-report behavioral measures for novice riders are lacking, and it is useful to examine the applicability of the already developed MRBQ amongst novice riders before considering the development of a new measure. The motorcyclists in the two European studies were experienced riders with an average of 11 years of riding (Elliott et al., 2007) or over 60,000 km of riding (Özkan et al., 2012). The MRBQ psychometric properties for novice riders who are in the formative years of riding may be distinct from more experienced riders who have committed themselves to riding and possibly established their behavior on-road. Similarly, climatic and cultural differences amongst UK, Turkey, and Australia may have implications for the MRBQ psychometric properties. Demographic and cross-cultural variations have been identified as contributors to the different DBQ factor structures and DBQ–crash relationships (Bener et al., 2008, Blockey and Hartley, 1995, De Winter and Dodou, 2010, Hennessy and Wiesenthal, 2005, Kontogiannis et al., 2002, Xie and Parker, 2002).

Therefore the present study aimed to (1) examine the previously examined psychometric properties of the MRBQ including the factor structure, internal consistency, and predictive validity in terms of self-reported crashes (Elliott et al., 2007, Özkan et al., 2012); (2) examine the psychometric properties of the MRBQ not yet examined, including its stability, content validity, and predictive validity in terms of police-recorded crashes and offences as well as self-reported near crashes and crashes; and (3) assess the applicability of the MRBQ amongst novice riders in Australia, a population to whom the MRBQ has not been applied to date.

Section snippets

Participants and procedure

Novice riders in the state of Victoria, Australia participated in the present study. They were drawn from a large scale randomized control trial to evaluate a motorcycle rider training program (VicRide) designed for novice riders. VicRoads, the state authority for roads in Victoria, sent letters of invitation to motorcycle riders who had recently advanced from a learner motorcycle permit to a probationary/restricted motorcycle licence (N = 23,696) through the period May 2010–June 2011. The study

Exploratory factor analysis of the MRBQ

Given both the previously found models were rejected, a new model was respecified using exploratory factor analysis. The PAF of the baseline 43-item MRBQ data (N = 1305) revealed four factors. The loadings that make up the final four scales are presented in italics (shaded cells) in Table 2. Based on the underlying themes of the constituent items, these four scales were named errors, speed violations, stunts, and protective gear. Ten items had loadings below 0.3 for all the factors and were not

Discussion

The present study examined the psychometric properties of the MRBQ amongst novice riders in Australia. The present findings were not consistent with the previously found factor structure, internal consistency and predictive validity of the MRBQ in terms of self-reported crashes amongst experienced riders in the UK and Turkey (Elliott et al., 2007, Özkan et al., 2012). This study additionally examined the stability, content validity, and predictive validity of the MRBQ with respect to

Conclusions

The present study examined the psychometric properties of the MRBQ amongst novice riders in Australia. A four-factor model of errors, speed violations, stunts, and protective gear was found to be most appropriate in the present sample. The present findings were not consistent with the previously found five-factor structure, internal consistency and predictive validity of the MRBQ in terms of self-reported crashes amongst experienced riders in the UK and Turkey (Elliott et al., 2007, Özkan et

Role of the funding source

The data for the present research originated from a project that was supported by the former Victorian Motorcycle Advisory Council and is being funded by the Motorcycle Safety Levy in Victoria, Australia. VicRoads, a Victorian statutory authority, provided approval in the writing and in the submission of the paper for publication. TS and RI receive salary funding from the National Health and Medical Research Council of Australia.

Acknowledgements

We would like to thank all study participants and the interviewers at the Survey Research Centre, Edith Cowan University for their support in data collection.

References (37)

  • C.C. Liu et al.

    Hazard perception abilities of experienced and novice motorcyclists: An interactive simulator experiment

    Transportation Research Part F: Traffic Psychology and Behaviour

    (2009)
  • M. Mattsson

    Investigating the factorial invariance of the 28-item DBQ across genders and age groups: An exploratory structural equation modeling study

    Accident Analysis and Prevention

    (2012)
  • A.T. Mccartt et al.

    Overview of motorcycling in the United States: A national telephone survey

    Journal of Safety Research

    (2011)
  • T. Özkan et al.

    Motorcycle accidents, rider behaviour, and psychological models

    Accident Analysis and Prevention

    (2012)
  • T. Özkan et al.

    Driver behaviour questionnaire: A follow-up study

    Accident Analysis and Prevention

    (2006)
  • A.I. Reeder et al.

    The risky and protective motorcycling opinions and behaviours of young on-road motorcyclists in New Zealand

    Social Science and Medicine

    (1996)
  • M.J.M. Sullman et al.

    Social desirability and self-reported driving behaviours: Should we be worried?

    Transportation Research Part F: Traffic Psychology and Behaviour

    (2010)
  • C. Xie et al.

    A social psychological approach to driving violations in two Chinese cities

    Transportation Research Part F: Traffic Psychology and Behaviour

    (2002)
  • Cited by (57)

    View all citing articles on Scopus
    View full text