An exploration of the perceptions of the average driver's speed compared to perceived driver safety and driving skill

https://doi.org/10.1016/S0001-4575(98)00035-9Get rights and content

Abstract

This study investigates the self-enhancement bias in driver attitudes, the finding that drivers rate themselves better than the average driver on safety and skill perceptions (Svenson, 1978Svenson, 1981; McCormick et al., 1986). A sample of 86 New Zealand drivers were asked their perceptions of their own and others' speeds in two conditions, 50 km/h and 100 km/h. The results established the self-enhancement bias for speed and safety, but not skill. Between 85% and 90% of drivers claimed to drive slower than the `average driver.' A new methodological technique derived from Harré and Gillett (1994) was used to investigate the direction of the self-enhancement bias. The results support the Downward Comparison Theory (Wills, 1981) because drivers consider other drivers negatively, rather than exaggerating their self-perceptions.

Introduction

Studies of drivers' attitudes towards their safety and skill have revealed a seemingly universal phenomenon of self-enhancement bias (Delhomme, 1991). When asked to compare themselves to the average driver, people display an over-confidence in their own abilities and attitudes. Drivers claim to be safer and more skilled than others, notwithstanding that people overestimate the number of deaths by motor vehicle accidents (Slovic et al., 1982). Svenson (1981), for example, found that between 70 and 90% of drivers, in her sample, claimed to be safer and more skilful than the average driver. The finding of this bias has been repeated in several studies, and it is found that between 60 and 90% of people claim to be better than the average driver across a range of phenomena (Delhomme, 1991). People claim to be more considerate, reliable, wise and responsible (McCormick et al., 1986), although the bias was eliminated when the reference was to a `very good driver' rather than the `average driver'. McKenna et al. (1991)found the effect with reference to 20 driving behaviours. The effect has been found in Sweden, the U.S.A. (Svenson, 1978, Svenson, 1981), France (Delhomme, 1991) and New Zealand (McCormick et al., 1986).

The self-enhancement bias is explained as a `positive self' bias rather than a `negative other' by McKenna et al. (1991). Drivers are thought to possess a self-enhancing cognition that operates to distort their perceptions of their abilities and over-inflate their confidence. Wills' `Downward Comparison Theory' (Wills, 1981) suggests that people seek comparisons with those who are worse than themselves and therefore distort their own confidence accordingly. However, the idea that drivers believe themselves to be better than average because their perceptions of the average driver are negative is dismissed by McKenna et al. (1991), who claim that the self-enhancement bias is a `positive self' phenomenon. Notwithstanding, the `negative other' hypothesis has recently found some support in Groeger and Grande (1996)who compared the perceptions of experienced drivers to novice, newly qualified drivers, although the result is speculative, with Groeger and Grande offering two alternative explanations from their results. Further to these questions concerning the direction of the relation that people hold to the `average other', the self-enhancement bias has been criticised as arising from the wording of the questionnaires (Groeger and Brown, 1989) and the ambiguity of the reference group (Dunning et al., 1989; Guirin, 1994; Groeger and Grande, 1996). For example, Groeger and Grande (1996)interpret the absence of the self-enhancement effect in McCormick et al. (1986), when the very good driver was used as a comparison, as indicating that the self-enhancement bias disappears when the reference group is less ambiguous.

There have been several calls for careful theorising about the phenomenon of self-enhancement bias in driver attitudes (Guirin, 1994; Assum, 1997). Perhaps the most important reason for such an analysis is that drivers' attitudes do have implications for accident rates. Assum (1997)reports that `wrong attitudes' (attitudes inconsistent with the traffic code) do increase the rate of accidents, although interestingly, the effect is not found within age groups. Most authors mention the implications that driver attitudes have for traffic safety campaigns (Svenson, 1981; Guirin, 1994). The reasons for the poor attitudes towards driving behaviour need to be understood so that effective intervention can be targeted efficiently.

The methodology adopted in the present study was developed to operationalise a social constructionist, or discursive, ontology of psychology developed by Harré and Gillett (1994). They believe that psychological explanation is set against arrays of people who actively construct their realities through discursive exchanges. Concomitant with a constructionist outline of psychology is much criticism of traditional social psychological investigations (Gergen, 1985, Gergen, 1994, Gergen, 1997; Harré, 1989, Harré, 1994; Harré and Gillett, 1994). The notion that an individual is active in constructing their reality with others, including the researcher, has implications for traditional forms of psychological investigation and theorising.

The use of the `average other' as an index to determine the overestimation of the driver attitudes ignores the relations that an individual may hold to the average driver (however vaguely or precisely the term `average' is defined). Individuals can regard themselves as being better or worse than an average driver and they may, individually, judge the average driver's ability accurately or inaccurately. This fact makes obscure the reasons for the self-enhancing bias when the `average' other is used as an index of comparison. In any sample of drivers' attitudes, indexed with reference to others, any variability in the notion of the `average driver' will ensure that respondents fall into at least four distinct groups: those who overestimate their own ability and the ability of the average driver; those who overestimate their own ability and underestimate the ability of the average driver; and the two corresponding groups for those who underestimate their own abilities.

This methodological ambiguity led Groeger and Grande (1996)to criticise, correctly, McKenna et al. (1991)for their assumption that their results are, “more consistent with a self enhancement bias rather a than a downward comparison theory” (p. 50). McKenna et al. (1991)suggested that drivers do not hold negative evaluations of the average driver since they rated the average driver as slightly above 5 on a scale of 0–10. Groeger and Grande (1996)point out:

It is impossible to say whether the same `average driver' was used as the basis of comparison in all cases. Thus a `better' or `worse' average driver might easily have been used when the driver felt variously secure or insecure about his or her ability to perform a particular manoeuvre (p. 64).

Whether or not that argument succeeds, the McKenna et al. (1991)argument is invalid. It does not follow from the fact that drivers regard the average driver as midway between very poor and very good that they do not hold negative evaluations of drivers when they compare themselves to the average driver. The Downward Comparison Theory argument is equivalent to the `positive self' position that McKenna et al. prefer, but the evidence that subjects rate drivers midway between very poor and very good does nothing to prevent the interpretation that the effect is due to a `negative other' evaluation. McKenna et al. use a statistician's sleight of hand. The average ability need not be represented as midway between very poor and very good on any of the abilities that McKenna et al. present, despite the fact that a group of subjects report this. For example, it would be strange indeed if the average driver were not reasonably good at judging the correct speed for bends or corners since presumably even less-than-average drivers manage to do this effectively most of the time. Only 3.3 per 10,000 drivers had an accident as a result of cornering in 1995 (LTSA, 1996, Svenson, 1981). The standard for average driving is not accurately assessed by a weighting from 0 to 10. A score of 5 may be as much of an indictment on the average driver's ability as an IQ score of 100 is to a potential Mensa candidate. Evaluations of a slightly-better-than-5 weighting simply indicate a weighting that subjects give to the average driver that is proportionally less than the weighting that they give to themselves and do not indicate a judgement of the standard of abilities required for performance of the driving measures.

The present investigation attempts to calibrate the relations between self and average other by assessing actual average speed of driving, the self-reported speed of subjects, the perception of the average driver's speed and the relation between actual speed of driving and the average of self-reported speeds. We sought to establish that the method described below has a broad utility as an alternative to the traditional self-other indices used in social psychological research in isolating subgroups of individuals in the way that they relate to the average reference person. Three hypotheses relate to this goal:

  • 1.

    that the `irrationality' described by Svenson (1978), Svenson (1981)is disproportionately represented by one type of relation that the individual holds to the `average reference person';

  • 2.

    that the self-reported driving speed is subject to the self-enhancement bias;

  • 3.

    that the differentiation of the subjects into the relations that they hold to the actual and average other will relate to the groups' self-enhancement bias of perceived safety and skill.

Section snippets

Subjects

The sample consisted of 86 subjects drawn from staff at the Open Polytechnic of New Zealand or organisations that were associated with the polytechnic through a staff member. We were attempting to test a method, so we simply called for volunteers from those associates in our immediate working environment. These people ranged in occupations and education, they were not all professionals but they comprised a variety of occupational categories that reflected the diversity of functions of the

Results

As a total group, the respondents reported their driving speeds accurately. Compared to the mean driving speeds obtained from the LTSA (personal communication, 5 September 1997) (mean driving speeds, 55.4 and 103.9 km/h, SD=11.9), subjects reported their own speeds at 54.7 km/h (N=86) and 103.2 km/h (N=86). There was no significant difference between actual driving speed of the general population and that self-reported by the subjects in the 100 km/h condition (Z=−1.046, NS). Calculation of the Z

Discussion

We were somewhat disappointed with the results in that we would have liked one group to demonstrate a greater level of overconfidence in their safety estimation relative to other groups. If group F drivers had been sufficiently represented and demonstrated a high proportion of overconfidence as to their self-reports of safety, we would have been able to conclude that those persons who claimed to drive faster than others, when this is true, over-represented the proportion of irrationality

Acknowledgements

We would like to thank: Ken Strongman, Rom Harré, Malcolm Crawley; Tui Patterson and Lynley Povey of the LTSA; and the two anonymous reviewers for their comments on a previous draft. This research was assisted through the Open Polytechnic Contestable Research Fund: READ and M97006.

References (27)

  • D. Dunning et al.

    Ambiguity and self-evaluation: the role of idiosyncratic trait definitions in self-serving assessments of ability

    Journal of Personality and Social Psychology

    (1989)
  • K. Gergen

    The social constructionist movement in modern psychology

    American Psychologist

    (1985)
  • K. Gergen

    Exploring the postmodern: perils or potentials?

    American Psychologist

    (1994)
  • Cited by (58)

    • MAD-DAX: Anger and driver behaviours on the road

      2024, Transportation Research Part F: Traffic Psychology and Behaviour
    • Subjective evaluations of self and others’ driving behaviors: A comparative study involving data from drivers in Japan, China, and Vietnam

      2023, Journal of Safety Research
      Citation Excerpt :

      These findings align with the findings from some previous studies (Healey & Picard, 2005; Kaye et al., 2016; Marshall et al., 2007). However, previous studies from different countries stipulated that the drivers consider themselves “better” and more “skillful” than other drivers (New Zealand: Walton & Bathurst, 1998; Sweden: Warner & Åberg, 2014, Åberg et al., 2005, Svenson, 1981; Turkey: Warner & Åberg, 2014; The United Kingdom: King & Parker, 2008; The United States: Preston and Harris, 1965; Svenson, 1981). These subjectively reported “self” and “other” driving behaviors might have differed depending on the social norms, level of development, and strictness of the road safety laws of the country.

    • Implicit evaluations about driving skills predicting driving performance

      2018, Transportation Research Part F: Traffic Psychology and Behaviour
    • Driver Feedback Mobile App

      2016, Transportation Research Procedia
    View all citing articles on Scopus
    View full text