Youth, riskiness and driving

1
Youth, riskiness and driving
2
Younger drivers are at more risk of a road
accident NHTSA (2003) 16-24 year-olds 14 of
U.S. population, but 25 of traffic
deaths. Age-related decline in accident rates,
with increase for drivers aged 75. Males and
females show same age-pattern, but males have
more accidents than females at all ages. Pattern
has remained constant over time (e.g. McKenna et
al 1998, Laapotti and Keskinen 2004). Same
pattern found in all developed countries.
Percentage of licensed drivers in that age-group
South Australia 1994 -1998, Baldock and Maclean
(2005)
3
Interpretation of statistics is complicated by
exposure issues Brorsson et al (1993) 40x more
dangerous to drive on Friday and Saturday
evenings than at other times. DFT (2010) alcohol
involved in 17 of road deaths, 47 of driver
deaths between 10 pm - 4 am. Young drivers
(17-24) over-represented in alcohol-related
accidents.
4
U.K. deaths in 2008. (DFT (2010) "Reported Road
Casualties Great Britain 2010")
5
Age- and sex-differences in types of accident
(e.g. Laapotti and Keskinen 2004) 2,401 fatal
Finnish car accidents. Running off road 47 of
all fatal accidents for young males, 26 for
young females. Head-on collisions 41 of all
fatal accidents for young males, 58 for young
females. Young males 3x more likely than young
females to be involved in a fatal drink-driving
accident, and nearly 4x more likely to have
previous traffic offences. Females more likely
than young males to have a accidents due to
reversing or loss of control (e.g. skids).
6
Risk Homoeostasis Theory Wilde (1982, 1988,
2002) Individuals adjust their behaviour in
order to maintain a preferred level of
risk. Implications - many "safety" measures are
futile. Only way to affect accident statistics is
by changing target levels of risk (i.e.
motivational manipulations). e.g. provide
incentives for safe driving (insurance discounts)
7
Evidence for and against RHT Accident
statistics are equivocal. Sagberg, Fosser and
Saetemo (1997) some confusion between effects on
accident rates of primary and secondary safety
measures (e.g. anti-lock brakes and seat-belt
use).
8
Evidence for and against RHT McKenna (1988),
Hole (2007) better approach is to use research
on psychology of risk-perception. If
risk-assessment is unreliable, then RHT must be
untenable. Wealth of evidence shows people are
poor at judging risk...
9
1. Drivers get little feedback about the risks of
driving - Accidents are rare on an individual
level, if not on a societal level (Rumar
1988) Risky behaviours may not have
consequences Accident rates are not wholly
under an individual's own control (may be an
innocent victim) Poor at remembering accidents
and near-accidents overall, 80 of
near-accidents are forgotten within a fortnight
(Chapman and Underwood 2000). (forgetting varied
from 95 of events rated as having produced
little danger of accident' to 50 of events that
produced a "large danger of an accident").
10
2. Drivers are poor at assessing risks
- Decision-making is irrational and prone to
bias. e.g. "availability heuristic" (Tversky and
Kahneman 1972). Baksteen (1995) in 1985, the
risk of dying in a car accident in the U.S. was 1
in 5,300, and the risk of dying in a terrorist
attack outside the U.S. was 1 in 1,600,000.
Deaths from terrorism were well publicised and
led to many Americans avoiding foreign travel in
1986 probably few avoided using their car for
that year, the more rational course of action to
avoid being killed.
11
2. Drivers are poor at assessing risks (cont .)
- Risk assessment affected by perceived
consequences (immediacy, severity), control,
frequency, familiarity. Self-serving bias -
individuals consider themselves less likely to
experience unpleasant things and more likely to
experience pleasant things. Most drivers
overestimate their own abilities (Svenson 1981,
Horswill, Wayland and Totfield 2004). Due to
self-serving bias, trait/situation
attributions? Duncan, Williams and Brown (1991)
experience does not inevitably lead to expertise,
perhaps because of lack of feedback.
12
3. Risky drivers do not show risk compensation
- Seat-belt non-users tend to be young, male,
drive more riskily, and have higher accident
rates and violations (Hunter et al
1993). Seat-belt non-users use drugs and alcohol
more (Foss et al 1994 Li et al 1999). Users of
hand-held mobile phones are less likely to wear a
seat-belt (Eby and Vivoda 2003). Speeding
drivers may have "risk-mitigating beliefs" that
reduce the perceived risks of speeding - e.g. "I
can drive safely at speed" , "only really high
speeds are dangerous" (Brown and Cotton 2003).
13
Zero-risk theory (Naatanen and Summala 1974,
Summala, 1988, 1997) Drivers seldom think about
risks or experience risk. Vehicle control is
based largely on habit. Accidents arise
because (a) drivers underestimate the amount of
variance in the traffic system (b) drivers are
motivated to speed - have a distorted "speed
utility function". Safety margins are inadequate
as a consequence. Can reduce accidents by
tackling these two factors (e.g. engineering for
(a), enforcement for (b)).
14
Individual differences and risky driving (a)
Vehicle control (skill) differences? Horswill et
al (2004) hazard perception skills are
correlated with safety, not vehicle control
skills - but drivers believe the
opposite. Williams and O'Neill (1974) racing
drivers have more accidents than
controls. Lajunen, Corry, Summala and Hartley
(1998) Two "personalities"? One skill-oriented,
one safety-oriented. Drivers who rate their
vehicle control skills highly tend to drive more
riskily and drive faster. Self-ratings affect
speed, which affects accident risk?
15
Individual differences and risky driving (b)
Lapses, errors and violations Reason et al
(1990), Aberg and Rimmo (1998) Aberrant driver
behaviours consist of "errors" inappropriate
acts, e.g. missing a "Give Way" sign "slips"
and "lapses" unintended omissions of actions,
e.g. forgetting to cancel an indicator "violatio
ns" deliberate floutings of law or safe
practice, e.g. running a red light.
16
Parker et al (1995), Westerman and Haigney
(2000) Errors more common in women, and do not
decrease with age. Lapses increase with
age. Violations more common in men and in
drivers who rate themselves as skilful show
age-related decline. Propensity to violate is
associated with greater risk of accident
involvement (e.g. Parker et al 1995 West,
Elander and French 1993).
17
Individual differences and risky driving (c)
Sensation seeking (Zuckerman 1994) "...defined
by the seeking of varied, novel, complex and
intense sensations and experiences and the
willingness to take physical, social, legal and
financial risks for the sake of such
experiences". SS higher in males than females,
and declines with age in both sexes. Rimmo and
Aberg (1999) SS associated with violations. SS
leads to violations, which lead to accidents? But
- fairly weak correlations between SS and driving
behaviours (.3 - .4 range). Arnett (1996)
reckless behaviours are widespread amongst young
drivers - not confined to high-SS group.
18
(d) Anger and aggressiveness Arnett (1996)
aggressiveness moderately correlated with
frequency of risky driving behaviours in a
student sample. "Driving Anger Scale"
(Deffenbacher et al 1994) High anger drivers 4x
more likely to show aggressive behaviours while
driving, and 2x more likely to engage in risky
behaviours. "Propensity of Angry Driving Scale"
(DePasquale et al 2001, Maxwell et al
2005) Aggressive drivers more likely to commit
violations scores highest for young male
drivers. No correlations between DAS or PADS and
accident rates.
19
Role of personality factors in accidents Rimmo
and Aberg (1999) "...despite the myriad of
studies designed to relate different constructs
of personality to accident involvement, only
limited success has been attained". Test scores
account for only 5-10 of variance in accident
rates. Reasons - Imperfect measurement of traits
by questionnaires. Accidents usually have
multiple causes personality factors may be only
the starting point in a chain of events
culminating in an accident.
20
Interactionist model of risky driving
(Taubman-Ben-Ari et al, 2004) Personal and
environmental influences on riskiness of
driving Personal - perceiving reckless driving
as a challenge, considering oneself able to cope
with it, disregard for possible negative
outcomes. Environmental - perceiving that
friends and family behaved similarly, and
approved of one's risky actions. Model still
only explained 43 of the variance in risky
driving scores - maybe remainder is accounted for
by personality variables, e.g.. sensation seeking.
21
Interactionist model of risky driving (Ulleberg
2003) Cluster analysis of a battery of measures
of personality and attitudes to risk, accident
records. Six clusters of drivers two "low risk",
two "high risk". High-risk A high on
sensation-seeking, normlessness, driving anger,
low on altruism and anxiety. High scores on
risk-taking behaviour, poor attitudes to risk and
low risk perception measures. Rated own driving
skills highly, but had highest accident rates.
81 were male. High-risk B - high on sensation
seeking, aggression, anxiety and anger, low on
altruism. Risky driving habits, poor attitudes to
road safety, high accident rates. Did not rate
their own driving skills as very good, and
perceived risk of accident as high. 59 were
female. High-risk groups less receptive to road
safety campaigns than low-risk groups.
22
Lifestyle factors in risky driving (Moller
2004) In-depth interviews with 29 Copenhagen
young drivers.
23
Conclusions Riskiness of younger drivers arises
from a lethal combination of Adequate vehicle
control skills, but poor hazard perception skills
(Hall and West 1995). Under-20's underestimate
traffic hazards and overestimate their own
abilities (Brown and Groeger 1988). Peer pressure
(Simons-Morton, Lerner and Singer
2005). Personality factors, especially sensation
seeking. Alcohol and drugs. Increased exposure to
risk (drive more often at night). Need to
improve risk perception, and reduce risk
acceptance.