Assessing individual differences in driving inattention: Adaptation and validation of the Attention-Related Driving Errors Scale to Spain

Highlights

• ARDES is a self-reported questionnaire to assess the proneness to driving attentional errors.

• ARDES has been adapted to the culture, language, traffic regulations and driving habits in Spain.

• New validity evidence of the cross-cultural equivalence of ARDES is provided.

• Differences in ARDES are found between drivers who reported or not traffic collisions.

Abstract

The Attention-Related Driving Errors Scale (ARDES) is a self-reported questionnaire to assess individual differences in the proneness to make attentional errors while driving. The aims of the current work are to adapt the original Argentinean version of the ARDES to the culture, language, traffic regulations and driving habits of Spain and provide new validity evidence of the cross-cultural equivalence of the scale. In the first step of the validation process, five external independent experts reviewed the original ARDES-Argentina and proposed modifications, adapted to the culture, language, traffic regulations and driving habits in Spain. Secondly, a sample of 320 drivers completed the adapted questionnaire. Descriptive statistics and psychometric properties (corrected item-total correlation, Cronbach’s alpha and factor structure) were performed on the data. Besides, in order to obtain further validity evidence, the relationships between the questionnaire scores and socio-demographic variables (age, sex, educational level, driving experience, crash involvement and traffic fines received) were analyzed. Factor analysis suggested a single factor that exceeded the parallel analysis criterion and accounted for 32.70% of the total variance. All items showed positive loadings on this factor, ranging from .41 to .72. The corrected item-total correlation values extend from .41 to .60, indicating that the items had good discrimination power. Cronbach’s alpha coefficient value was .88. The analysis of the relationships between ARDES-Spain scores and socio-demographic variables provided further validity evidence of the appropriateness of the adapted questionnaire. In particular, differences in ARDES-Spain scores were found between drivers who reported traffic collisions with material damage and participants who did not. In conclusion, results in the current study suggest that the adapted version of the ARDES is a useful tool for evaluating the proneness to attentional errors during driving in the Spanish population. Future studies adapting the questionnaire to other countries with different languages, cultures, traffic regulations and driving habits are encouraged in order to expand the discussion on the cross-cultural equivalence of the ARDES.

Introduction

Current research in road traffic domain shows that driver inattention is one of the main factors explaining accidents and its negative influence on safety is expected to further increase, as a consequence of the proliferation of some potentially distractive in-vehicle technologies (e.g., Klauer et al., 2006, Ranney, 2008, Stutts et al., 2001). Multiple research strategies, including controlled experiments in a laboratory or in driving simulators, test track and field studies and also correlation analyses of questionnaire data, are complementary being applied to better understand driver inattention and thus prevent attention-related accidents (for a review see, for example, Kircher, 2007).

According to a recent review (Regan, Hallett, & Gordon, 2011), driver inattention can be defined as “insufficient or no attention to activities critical for safe driving” (pp. 1775). In addition, Regan et al. (2011) proposed a taxonomy in which the different categories of inattention are distinguished by the different mechanisms that produce driver inattention. For example, among these categories, driver distraction is considered as “the diversion of attention away from activities critical for safe driving toward a competing activity” (Regan et al., 2011, pp. 1776), whereas other forms of driving inattention do not require a competing activity (e.g., when the driver is influenced by biological factors that physically prevent the detection of critical information, such as micro-sleeps).

Ledesma, Montes, Poó, and López-Ramón (2010) suggested that individual differences in driver inattention might reflect a relatively stable pattern of behaviour associated with attentional errors in different dimensions of everyday life and with particular psychological variables. Accordingly, those individuals who are more prone to make attentional errors while driving will also manifest inattention in their daily lives and will present specific psychological traits (Ledesma et al., 2010). Within this theoretical framework, the Attention-Related Driving Errors Scale (ARDES; Ledesma et al., 2010) was developed to assess individual differences in the proneness to driving inattention. Whereas Regan et al. (2011)’s taxonomy was focused on the mechanisms that produce driver inattention (i.e. the different sources of inattention, such as a micro-sleep or a distraction), the work by Ledesma et al. (2010) was aimed at measuring the resulting errors (i.e., the consequences of inattention, such as failing to spot a pedestrian crossing or to notice a leading vehicle slowing down, which can be brought by several of the inattention sources defined in Regan et al.’s taxonomy; see also Stutts et al., 2001). In this regard, a number of attention-related driving errors were used as indicators to measure the construct of driver inattention (particularly, the proneness to driver inattention) by means of the ARDES.

The items in the ARDES, originally developed for Argentinean culture, language, traffic regulations and driving habits, specifically refer to non-deliberate errors in driving performance resulting from an attentional failure, such as failing to notice a pedestrian crossing the street (Ledesma et al., 2010). Items and instructions were based on previous questionnaires, such as the Attentional Lapses subscale of the Driver Behaviour Questionnaire (DBQ; Reason, Manstead, Stradling, Baxter, & Campbell, 1990), and the Multidimensional Driving Style Inventory (MDSI; Taubman-Ben-Ari, Mikulincer, & Gillath, 2004). However, according to Ledesma et al. (2010), in comparison with these questionnaires, the ARDES was specifically developed to measure driving attentional errors and avoid overlapping with other psychological constructs such as planning errors, which are not attentional in nature, or daydreaming, what is not essentially an error (actually, in Regan et al.’s, 2011, taxonomy, daydreaming is considered as a source of inattention that might lead to driving errors, and thus it is not an error per se). Besides, the internal consistency of the original Argentinean version of the ARDES has been reported to be higher than the Attentional Lapses subscale of the DBQ, and it covers a wider range of inattentive driving behaviour than the latter (see Ledesma et al., 2010). In consequence, the ARDES may constitute a promising tool available to researchers interested in analysing driver inattention.

Reliability and validity of the scores obtained from the original Argentinean version of the ARDES were evaluated on a sample of drivers in Argentine (Ledesma et al., 2010). An exploratory factor analysis suggested that all 19 items in the scale are multiple indicators of a one-dimensional construct related to the proneness to attention-related errors while driving (high loadings on the first factor and good discrimination indexes were observed; internal consistency was also high, Cronbach’s alpha = .86). The one-dimensional factor solution was consistent with previous studies with related questionnaires, such as the Driver Behaviour Questionnaire (DBQ; Reason et al., 1990) and the Multidimensional Driving Style Inventory (MDSI; Taubman-Ben-Ari et al., 2004), in which a single inattentive driving factor was differentiated from other dimensions of driver behaviour. First, Reason et al. (1990) developed the DBQ to distinguish between non-intentional driving errors and deliberate traffic violations, supporting the idea that different psychological processes influence these factors. In their original study, a factor analysis provided support for the difference between errors and violations and, additionally, evidence was found of a third factor that mainly included minor attentional failures (“slips and lapses” or simply “lapses”). On the other hand, Taubman-Ben-Ari et al. (2004) developed the MDSI as a multidimensional instrument to evaluate eight driving styles (dissociative, anxious, risky, angry, high-velocity, distress reduction, patient and careful). Among them, the dissociative style was characterized as a tendency to get distracted easily, commit errors due to distractions and exhibit cognitive failures and experiences of dissociation during driving that could modulate inattention errors.

The relationship between the scores obtained from the original Argentinean version of the ARDES and a variety of cognitive and psychological variables was analysed to provide further validity evidence. First, Ledesma et al. (2010) found a significant and theoretically consistent pattern of correlations between the ARDES and different measures of cognitive error proneness (the Attention-Related Cognitive Errors Scale or ARCES, and the Memory Failures Scale or MFS; Cheyne, Carriere, & Smilek, 2006), lack of general awareness and attention (the Mindful-Attention Awareness Scale or MAAS; Brown & Ryan, 2003), proneness to boredom (the Boredom Proneness Scale or BPS; Farmer & Sundberg, 1986) and dissociative personality traits (the Dissociative Experiences Scale or DES; Bernstein & Putnam, 1986). Second, López-Ramón, Castro, Roca, Ledesma, and Lupiáñez (2011) compared ARDES scores and drivers’ attentional performance in a neurocognitive attentional test (the Attention Networks Test or ANT; Fan, McCandliss, Sommer, Raz, & Posner, 2002) and observed that drivers reporting the greatest propensity to experience attention-related errors (i.e., higher ARDES scores) showed an overall slowdown in attentional performance, less endogenous preparation for high-priority alerting signals and a better response to cognitive conflict in the presence of valid orienting cues. Third, a recent study (López-Ramón et al., in preparation) evaluated a sample of taxi drivers’ attentional performance using the Attention Networks Test for Interactions and Vigilance (ANTI-V; Roca, Castro, López-Ramón, & Lupiáñez, 2011) and found that only the drivers with low propensity to make attentional errors (i.e., low ARDES scores) were able to maintain good attentional functioning after a 12-h working shift (i.e., an alerting tone was more effective to reduce cognitive interference). In addition, further evidence of the ARDES validity was provided by exploring the relationship between the drivers’ scores and self-reported crashes and traffic fines (Ledesma et al., 2010). Results showed that ARDES scores were able to discriminate drivers who reported having been involved in at least one traffic collision with material damage or have had traffic fines.

Overall, evidence on the validity and reliability of the original Argentinean version of the ARDES suggests that the scale has good psychometric properties and thus can be considered as a simple and useful measure of individual differences to attention-related driving errors. However, the cross-cultural stability and applicability of the scale is still uncertain. Only the original version of the ARDES has so far been developed and validated, using Argentinean samples, and previous experience with similar questionnaires highlights the importance of developing different versions specifically adapted to the varying cultures, languages, traffic regulations and driving habits of target populations (for example, see Lajunen, Parker, & Summala, 2004, for a cross-cultural study on the DBQ). Considerations of driving context, language and culture are all highly relevant when adapting a questionnaire. According to the guidelines of the International Test Commission (International Test Commission (ITC), 2010), questionnaire developers should fully take into account the linguistic and cultural differences among the populations for whom the adapted versions are intended, and should demonstrate that the language used is appropriate for the populations being tested, as language use can reflect major cultural differences (Hambleton, 2001, van de Vijver and Tanzer, 2004).

The International Test Commission on Adapting Tests (ITC, 2010), reflect the widest consensus among professionals and researchers about best practices to adapt tests and questionnaires. The ITC guidelines D8 recommends researchers and publishers to provide information on the evaluation of validity in all target populations for whom the adapted versions are intended. Language and cultural differences can be frequently found across countries or contexts in which majority language as English or Spanish are spoken. Spanish is the official language in 21 countries and is used by more than 450 million people. This diversity does not prevent the understanding but causes differences in pragmatic meanings reflecting different traditions, social norms and values. Also, it is possible that the psychological construct measured by ARDES was better assessed by different driving behaviour or alternative traffic situations in either cultural context.

Several indicators in traffic and road safety domains reveal clear differences between Argentina and Spain, and thus not only the culture and language but also the traffic regulations and driving habits where the ARDES will be applied would not be directly comparable. For example, according to a recent international study (International Traffic Safety Data, 2012), the absolute number of deaths in road traffic accidents is higher in Argentina (5094) than in Spain (2478). Similarly, when differences in population sizes are considered, the death rate per 100,000 inhabitants is also superior in Argentina (12.6), as compared to Spain (5.4), and analogous results concerning the death rate per 10,000 vehicles are reported (2.9 vs. 0.8, respectively). Regarding attention-related traffic accidents, it has been estimated that driver distraction is a concurrent factor in at least 39% of crashes in Spain (Instituto OPINA, 2011). There is no official figure for this traffic aspect in Argentina, and thus a comparison is difficult. However, some data suggest, for instance, that cellular phone use might be higher in the latter country. According to an observational study carried out in Argentina (Agencia Nacional de Seguridad Vial, 2011), 7.4% of drivers were using their cellular while driving. A similar study in Spain (CONSULTRANS, 2008) found that this figure was 2.6% of drivers. It should be noted that differences between both studies (e.g., different observational protocols) discourage direct comparison between the figures on cellular phone use. For this reason, it is important to create, validate and adapt assessment tools based on participants’ behaviour and responses to the varying contexts, such as it has been currently done with the ARDES.

The differences existing between Argentina and Spain challenge that the same questionnaire (e.g., using the same wording or asking about the same behaviours and traffic situations), would obtain similar psychometric properties or a similar factor structure. Thus, differences between the Argentinean and the Spanish culture, language, traffic regulations and driving habits, along with the well-known lack of sample invariance of the psychometrics based on the Classical Test Theory (Crocker & Algina, 1986), suggest further psychometric analyses to confirm that ARDES is also appropriate in Spain. The current study is aimed at adapting the original Argentinean version of the ARDES (ARDES-Argentina) to the culture, language, traffic regulations and driving habits in Spain, thus providing new evidence of the cross-cultural stability and applicability of the scale. Besides, analysing relationships between the Spanish version of the ARDES (ARDES-Spain) and self-reported accidents and other socio-demographic variables will be of further interest in the current study.

The development of valid, reliable and usable self-reported questionnaires to measure the proneness to attention-related errors while driving can be found of interest in different applied research areas. For example, population reference values could be obtained and then drivers’ scores may be used to assess individual differences in the proneness to inattention. Next, provided that this construct was effectively related to proper driver performance and accident risk indicators, questionnaire scores might be potentially applicable as a part of driver selection processes or as a screening tool to apply customized intervention programmes (see, for example, Knipling, Burks, Starner, Thorne, & Barnes, 2011).

Similarly, the proneness to inattention between particular groups of drivers (e.g., novice or older drivers, long haul drivers, drivers under the influence of specific drugs or diseases, etc.) could be compared. Therefore the questionnaire might be useful to better characterize and consequently intervene on vulnerable groups, for instance, making these drivers aware of their increased proneness to distraction, identifying the potentially distracting activities that they initiate (which activities and when they are performed) and applying interventions that focus on strategic decisions and planning (see, for example, Horrey & Lesch, 2009).

In addition, having versions of the questionnaire adapted to different countries or cultural contexts might allow carrying out cross-cultural comparison studies, discussing potential explanations for the observed differences and providing useful information to be considered by road-safety practitioners (see, for example, Warner, Özkan, Lajunen, & Tzamalouka, 2011). In this regard, self-reported questionnaires can be useful tools for researchers and practitioners to draw up or evaluate road safety interventions.