A gender role socialization model of explicit and implicit biases in driving self-enhancement

https://doi.org/10.1016/j.trf.2009.08.006Get rights and content

Abstract

We present a gender role socialization model of explicit and implicit biases in driving self-enhancement. The model proposes that men’s higher levels of driving self-enhancement (relative to women) results from socialization experiences, such as those that idealize driving skill and risk taking as core aspects of male identity. This socialization process produces reasoned explicit (declarative or propositional) associations, but also implicit (automatic, non-conscious) associations between masculinity and driving self-enhancement, and these two processes are theorized to operate relatively independently. Structural Equation Modeling of a large sample of young male (n = 200) and female (n = 200) drivers supported the model, and indicated that (a) gender role identification fully mediated the effect of gender on driving self-enhancement and (b) that this effect occurred simultaneously but relatively independently in both explicit or reasoned belief structures and also in implicit non-conscious associations assessed using the Implicit Association Test. The origins of gender differences in traffic safety behaviour and the implications of our model for shifting conceptions of what it means to be a “real man” with regard to driver behaviour are discussed.

Introduction

There is abundant evidence of gender differences in traffic injury rates. Men, and particularly young men, are involved in more crashes than young women. This is true both in New Zealand (Ministry of Transport, 2008), where the current research was conducted, and elsewhere in the world (e.g. Department of Infrastructure, Transport, Regional Development and Local Government, 2008, Hanna et al., 2006, Özkan and Lajunen, 2005). Even the presence of male, compared to female passengers seems to increase crash risk for young drivers (Williams, Ferguson, & McCartt, 2007). Numerous studies have found differences in men’s and women’s attitudes and behaviours that may help explain these patterns (e.g. Forsyth, 1992, Harré et al., 1996, Shinar and Compton, 2004).

One attitude of interest is the belief that one is an above average driver. Various forms of driving self-enhancement bias have been found, and both men and women are prone to reporting themselves as above average on some dimensions (e.g. DeJoy, 1989, Horswill et al., 2004, Sümer et al., 2006). A recent series of studies on young New Zealand drivers demonstrated that there are two forms of this bias; perceived superior driving ability (e.g. feeling more skilled than others) and perceived superior driving caution (e.g. considering oneself a safer driver than others; Harré et al., 2005, Harré and Sibley, 2007, Sibley and Harré, 2009). These studies generally showed men to have higher levels of self-enhancement on the ability dimension than women. A sense that one is more skilled than average is problematic as it may lead to overconfidence (Deery, 1999) and risk taking (Clarke et al., 2005, Harré, 2000). It has also been associated with crash-risk optimism. That is, drivers who think they are more able than others also tend to think they are at less risk of a crash than others (DeJoy, 1989, Harré et al., 2005, Harré and Sibley, 2007). Furthermore, an inflated belief in one’s driving ability and the crash-protection provided by that ability, may encourage drivers to ignore safety advertisements, which are seen as more likely to influence others, a phenomena known as the “third person effect” (Lewis, Watson, & Tay, 2007).

In addition to measuring and showing that these biases exist explicitly, that is, when drivers are asked to complete a questionnaire that requires them to consciously judge and rate themselves relative to others, the two most recent of these studies (Harré and Sibley, 2007, Sibley and Harré, 2009) showed this bias also existed implicitly. That is, these biases exist at the automatic level of processing, and are not merely a product of propositional reasoning or declarative belief structures and attitudes.

To test implicit attitudes, Harré and Sibley, 2007, Sibley and Harré, 2009 used the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998). The IAT is a computer-based reaction time task in which participants are required to match concepts as quickly as possible. The assumption behind the IAT is that the more quickly concepts can be matched, the more closely associated they are in the brain. One of the advantages of using implicit association measures is that unlike explicit measures, it is difficult to suggest they are influenced by socially desirable responding. In these studies, the test involved measuring how quickly participants could match words associated with being an able (e.g. skilled) driver with the self (relative to others), and conversely how quickly they could match words associated with not being an able (e.g. useless) driver with the self (relative to others). Harré and Sibley (2007) found that men had quicker reaction times than women when matching the self with words representing driving ability and skill. (Biases in driving caution were also measured but are not of interest here.) Sibley and Harré (2009) did not find the same gender difference on this implicit measure, however when taken alongside the general evidence for gender differences in driving attitudes discussed earlier, these results suggest that there is an association between perceived superior driving ability and being a man, but that it is not strongly present in all men (or absent in all women).

It is also important to note that the two studies above, consistent with other research using the IAT, found that explicit and implicit measures were only weakly correlated. This suggests that they reflect quite distinct processes. In the study by Harré and Sibley (2007), both explicit and implicit self-enhancement independently predicted crash-risk optimism (i.e. the belief that one is at less risk of a crash than others), and both explicit and implicit driving caution predicted driving violations. This suggests that both how people consciously think about themselves as a driver, and the automatic associations they make about their driving are potentially important, and also quite unique, contributors to risky driving choices. It would therefore seem prudent to study both processes simultaneously in order to develop a more detailed understanding of the psychological factors and processes underlying driver self-perceptions and more general risk taking motivations.

Because we know that there are strong gender differences in crash rates and driving attitudes (and there is evidence that these operate both explicitly and implicitly), the puzzle that remains is in determining the mechanisms through which these occur. In other words, driving behaviour and the attitudes that underlie it (like skill self-enhancement), appear to be “gendered”, but it is unclear exactly how and why this is so. Part of the answer may lie in biological differences between men and women. For example, testosterone, a male steroid hormone, has been associated with sensation seeking (Aluja & Torrubia, 2004), aggression (Book & Quinsey, 2005) and venturesomeness (Coccaro, Beresford, Minar, Kaskow, & Geracioti, 2007). However, these relationships are complex and appear to be influenced by socialization. For example, it has been suggested that aggression may only be related to testosterone insofar as aggression is a socially endorsed strategy for achieving dominance (Rowe, Maughan, Worthman, Costello, & Angold, 2004).

Social norms of masculinity and femininity are therefore likely to play a crucial role and several authors have suggested that the tendency for men to display riskier behaviours and attitudes than women is due to attributes associated with masculinity, such as the importance of being “skilled” and “fearless” and the association of these characteristics with high-risk driving maneuvers (Harré, 2000, Papadakis and Moore, 1991, Raithel, 2001). Unsurprisingly, men have been shown to demonstrate more “macho” driving attitudes than women (Harré et al., 1996) and more “macho” men have reported more driving aggression (Krahé & Fenske, 2002). This research, strongly suggests that there is a socialization mechanism at work, and that men who adopt risky driving attitudes may do so through identification with being a man and the package of characteristics that goes along with that.

One way to test for this is to see if observed gender differences in driving risk attitudes are explained (or mediated) by differences in gender role identification. If observed gender differences are mediated by gender role attitudes (that is, gender does not retain a direct effect on driving attitudes when controlling for gender role identification), then this would suggest that gender differences in this domain occur via gender socialization. There have however, been very few studies (three to our knowledge, as discussed below) that have directly examined gender role identity to see if identification with “being a man” predicts risky driving attitudes or behaviours.

The three studies that have been done have all used the Bem Sex Role Inventory. This treats masculinity and femininity as separate dimensions. Masculinity is “an instrumental orientation, a cognitive focus on ‘getting the job done’; and femininity has been associated with an expressive orientation, an affective concern for the welfare of others” (Bem, 1974, p. 156). An individual who is high on masculinity and low on femininity is assumed to have a masculine identity, whereas the opposite implies a feminine identity. Individuals high on both dimensions are assumed to be androgynous. It is important to note that a man can be predominantly feminine and woman predominantly masculine on this scale, as biological sex is only associated with, not determinant of, one’s sex role orientation. Studies using the inventory have found masculine sex role orientation to be associated with sensation seeking (Öngen, 2007), and aggression (Weisbuch, Beal, & O’Neal, 1999), both of which have been linked to risky driving (Deffenbacher et al., 2003, Lonczak et al., 2007).

The first study that used the Bem Sex Role Inventory to examine gender and driving involved German women (Krahé, 2005). Regression analyses showed decreased aggressive driving was predicted by increased femininity, but masculinity was not a significant predictor. A further study with young Turkish drivers found that masculinity predicted driving violations and offences and femininity predicted the opposite, as well as reduced crashes and driving errors (Özkan & Lajunen, 2005). A second study by these authors found that masculinity predicted perceived perceptual motor skills and femininity predicted perceived safety skills (Özkan & Lajunen, 2006).

The present study extends research on the function that gender role identification plays in risky driving attitudes, and in particular to examine its function in both explicit and implicit driving self-enhancement biases. We propose and evaluate a socialization model in which gender is operationalized as a distal factor that affects both explicit and implicit self-enhancement biases in driving ability via gender role identification. Our IAT measure of gender role identification necessarily used opposing categories of masculinity and femininity and thus provided a one-dimensional measure (rather than providing two separate dimensional measures of both masculinity and femininity, as does the Bem Inventory). In order to maintain consistency, for the explicit measure, we therefore also used a unidimensional measure: the Masculinity–Femininity sub-scale of the Personal Attributes Questionnaire (PAQ; Spence & Helmreich, 1978). This unidimensional sub-scale of the PAQ assesses the extent to which people identify with stereotypically feminine traits versus stereotypically masculine traits by using opposing endpoints theorized to reflect these two trait sets (e.g. a communal versus an agentic orientation).

Our model proposes that men should only display stronger driving ability self-enhancement biases than women to the extent that they more strongly identify with stereotypically masculine versus feminine gender roles. Thus, any observed gender differences in driving self-enhancement should be fully mediated (that is, occur indirectly via) gender role identification. A finding that gender retained a direct and proximal effect on driving self-enhancement biases beyond that explained by gender role identification would instead indicate that gender differences were not entirely attributable to gender role identification and might instead arise through some, as yet, unidentified process, such as perhaps a direct effect of high levels of testosterone independent of masculine identification (we did not expect this to be the case). We tested our model using Structural Equation Modeling of both explicit and implicit gender role identification and self-enhancement biases in driving ability in a large sample of female and male undergraduates.

As noted earlier, previous studies have shown self-enhancement biases in driving ability at both an implicit and explicit level, and indicate that these biases were independent (Harré and Sibley, 2007, Sibley and Harré, 2009). We therefore further predicted that comparable gender role socialization effects would occur for both explicit and implicit processes, and that these two processes would be relatively independent. That is, explicit gender role identification (peoples’ reasoned declarative beliefs about their gender role identification) should explain gender differences in explicit driving self-enhancement biases, whereas implicit gender role identification (automatic non-conscious associations between concepts of the self and gender roles) should explain gender differences in implicit driving self-enhancement.

Section snippets

Participants

Participants were 200 female and 200 male psychology undergraduate students studying at a New Zealand university. All participants held a driver’s licence and 36% of participants owned a personal motor vehicle. Participants ranged from 17 to 52 years of age (M = 20.31, SD = 5.01). Participants reported driving on average for 5.35 h per week (SD = 6.56). Forty eight percent of participants had at some stage during their life been involved in a motor vehicle accident of some sort, whether as a driver or

Gender differences in explicit and implicit driving self-enhancement

One sample t-tests indicated that participants were, on average, significantly above the theoretical scale midpoint of 4.00 in terms of their explicitly estimated driving ability (M = 4.72, SD = 1.01) (t(399) = 14.24, p < .01, Cohen’s d = .71). Participants also displayed a significant implicit self-enhancement bias in driving ability, as the average IAT D effect (M = .37, SD = .39) was significantly greater than a score of 0 (recall that an IAT D score of 0 indicates no bias) (t(399) = 19.24, p < .01, Cohen’s d =

Discussion

Prior research has raised the interesting possibility that gender differences in risky driving and crash rates are due to characteristics associated with masculinity (Harré, 2000, Krahé and Fenske, 2002, Papadakis and Moore, 1991, Raithel, 2001). The current study is one of very few however (the others being Krahé, 2005, Özkan and Lajunen, 2005, Özkan and Lajunen, 2006) that has directly examined the relationship between gender role identification and risky driving attitudes. We focused on

References (43)

  • I.M. Lewis et al.

    Examining the effectiveness of physical threats in road safety advertising: The role of third-person effect, gender, and age

    Transportation Research Part F: Traffic Psychology and Behavior

    (2007)
  • H.S. Lonczak et al.

    Predicting risky and angry driving as a function of gender

    Accident Analysis and Prevention

    (2007)
  • T. Özkan et al.

    What causes the differences in driving between young men and women? The effects of gender roles and sex on young drivers’ driving behaviour and self-assessment of skills

    Transportation Research Part F

    (2006)
  • R. Rowe et al.

    Testosterone, antisocial behavior, and social dominance in boys: Pubertal development and biosocial interaction

    Biological Psychiatry

    (2004)
  • D. Shinar et al.

    Aggressive driving: An observational study of driver, vehicle, and situational variables

    Accident Analysis and Prevention

    (2004)
  • C.G. Sibley et al.

    The impact of different styles of traffic safety advertisement on young drivers’ explicit and implicit self-enhancement biases

    Transportation Research Part F: Traffic Psychology and Behavior

    (2009)
  • N. Sümer et al.

    Asymmetric relationship between driving and safety skills

    Accident Analysis and Prevention

    (2006)
  • A.F. Williams et al.

    Passenger effects on teenage driving and opportunities for reducing the risks of such travel

    Journal of Safety Research

    (2007)
  • A. Aluja et al.

    Hostility–aggressiveness, sensation seeking, and sex hormones in men: Re-exploring their relationship

    Neuropsychobiology

    (2004)
  • S.L. Bem

    The measurement of psychological androgyny

    Journal of Consulting and Clinical Psychology

    (1974)
  • Department of Infrastructure, Transport, Regional Development and Local Government (2008). Road deaths Australia: 2007...
  • Cited by (32)

    • The development and application of the drivers’ attitudes of right-of-way questionnaire (DARQ)

      2023, Transportation Research Part F: Traffic Psychology and Behaviour
    • Implicit and explicit attitudes in transportation research: A literature review

      2021, Transportation Research Part F: Traffic Psychology and Behaviour
      Citation Excerpt :

      For example, Manca et al. (2019) used a persuasive strategy considering characteristics of the message which appear to favor possible changes in implicit attitudes (e.g. valence, source expertise, and argument strength). On the other hand, the study that used conventional interventions without attention to these issues did not report changes (Sibley and Harré, 2009a; 2009b). Furthermore, this study used very brief advertisements which did not take into consideration that implicit associations are the result of slower learning processes.

    • Traffic climate and driver behaviors: Explicit and implicit measures

      2019, Transportation Research Part F: Traffic Psychology and Behaviour
      Citation Excerpt :

      In the literature, male drivers reported higher perceptual-motor skills than female drivers (e.g. Özkan, Lajunen, Chliaoutakis, Parker, & Summala, 2006; Martinussen, Møller, & Prato, 2014). Additionally, when measured both implicitly and explicitly, male drivers showed higher driving self-enhancement than female drivers (Harré & Sibley, 2007; Sibley & Harré, 2009). If drivers perceive being able to show positive driver behaviors as a skill dimension, it might be plausible to expect that male drivers might also have higher self-enhancement for positive behaviors.

    • Implicit evaluations about driving skills predicting driving performance

      2018, Transportation Research Part F: Traffic Psychology and Behaviour
    • Age, gender and deterrability: Are younger male drivers more likely to discount the future?

      2017, Accident Analysis and Prevention
      Citation Excerpt :

      It is noteworthy that males were also statistically more likely to report knowledge of indirect punishment avoidance, and thus, it would also be beneficial if future research endeavors examined if group dynamics influence perceptual levels. More broadly, a sizeable body of literature has revealed the powerful influence of group norms on a range of offending behaviors (Ogilvie et al., 2014) even among young male speeders (Møller and Haustein, 2014; Sibley and Harre, 2009). Research has also demonstrated that younger males have less compliant attitudes towards traffic rules compared to other drivers (Bergdahl, 2005; Laapotti et al., 2003) and perceive lower levels of risk with certain driving situations (Tränkle et al., 1990).

    View all citing articles on Scopus
    1

    These authors contributed equally to this manuscript.

    View full text