Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Deep Learning Series — Module 05

Driver Behaviour
& Psychology

Based on MIT CSAIL Social Value Orientation research, MIT AgeLab driver behaviour studies,
Dual Process Theory, risk homeostasis, young and older driver neuroscience, and behaviour change science

System 1
Fast, automatic thinking drives most everyday driving decisions
SVO
Social Value Orientation — MIT CSAIL research on cooperative vs. competitive drivers
Risk
Home.
Drivers unconsciously offset safety improvements with riskier behaviour
25
The age at which the prefrontal cortex — the risk-assessment brain — is fully developed
Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
How Drivers Actually Think — Kahneman's Framework

Dual Process Theory: System 1 and System 2 Thinking

Nobel laureate Daniel Kahneman's Dual Process Theory is fundamental to understanding driver behaviour. Kahneman identified two modes of thinking: fast, automatic, intuitive System 1, and slow, deliberate, analytical System 2. The critical insight for road safety is that driving — especially for experienced drivers — is almost entirely System 1, which means it is fast and efficient but also prone to systematic biases and errors that System 2 would catch.

System 1 — Fast, Automatic

  • Operates below conscious awareness
  • Pattern-matching: "this looks like a normal junction"
  • No deliberate effort required
  • Draws on learned heuristics and experience
  • Cannot be turned off by willpower
  • Makes 95%+ of driving decisions

Benefits: Fast enough to respond to hazards. Frees conscious mind for route planning and conversation.

Risks: Over-familiar patterns, confirmation bias, optimism bias — all System 1 phenomena.

System 2 — Slow, Deliberate

  • Requires conscious effort and attention
  • Can override System 1 impulses
  • Used for: learning new routes, unfamiliar vehicles, complex situations
  • Engaged when System 1 signals uncertainty
  • Fatigues easily — degrades under cognitive load

When it activates: New situation, something seems wrong, learning mode, high-stakes decision.

Risk when absent: Familiar route with changed conditions (roadworks, new hazard) — System 1 uses the old "safe" template.

MIT Application — Expert Driver

Advanced driving training (IAM, ROSPA, Garda pursuit training) works by deliberately engaging System 2 on roads where System 1 has become too comfortable. IPSGA and commentary driving are System 2 techniques that make unconscious decisions conscious again — recalibrating the automatic responses that have drifted toward complacency.

The Familiar Route Trap

Crashes are disproportionately likely on roads drivers know well — their daily commute, routes driven hundreds of times. System 1 has learned these routes as "safe" and stops scanning for new hazards. The child who steps off a pavement for the first time on a familiar road is invisible to a driver in full System 1 mode.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Why We Misjudge Danger

Risk Perception: Why We Get It Wrong

Humans are systematically poor at estimating risk. We over-fear dramatic, vivid, unfamiliar dangers (plane crashes, shark attacks) and under-fear familiar, everyday ones (car journeys, unhealthy diet). This is not stupidity — it is how the brain's threat-detection system evolved. For road safety, it means drivers routinely underestimate the risks of their own habitual behaviours.

Optimism Bias

The universal human tendency to believe we are less likely than average to experience negative events. Studies consistently show that when asked "What is your risk of a road crash compared to average?", over 70% of drivers rate themselves as below-average risk — a statistical impossibility.

For road safety this means: most drivers believe the crash statistics apply to "other drivers" — not to themselves. This directly reduces motivation to change behaviour.

Familiarity Reduces Perceived Risk

The more often a driver does something without incident, the less risky it feels. Driving the same route at 100 km/h every day for 5 years with no crash → that speed on that road feels "safe." The actual physics of a crash has not changed; the perception has.

Dread vs. Exposure

Risks we "dread" feel bigger than their statistics suggest. Risks we are exposed to daily feel smaller. Flying feels riskier than driving — even though per km, driving is 20× more deadly — because plane crashes are vivid and rare, car crashes are invisible and common.

MIT Research — Risk Calibration

MIT AgeLab research on hazard perception training found that giving drivers accurate risk statistics — "at your speed, if a child stepped out 20 metres ahead, you would hit them" — measurably improved risk calibration and subsequent behaviour. Abstract statistics ("crashes cause X deaths per year") did not change behaviour. Personal, specific, concrete risk information did.

Factors That Increase Perceived Risk (Good)

  • Personal experience of a crash or near-miss
  • Knowing someone killed or injured in a crash
  • Vivid, personalised safety communications
  • Real-time feedback from telematics
  • Hazard perception training with immediate feedback
The Illusion of Control

Drivers consistently overestimate their ability to control outcome in risky situations — "I'm a good driver, I could handle it." Control illusion inflates confidence and reduces precautionary behaviour, particularly in high-skill domains where drivers have seen themselves perform well in normal conditions.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Gerald Wilde's Controversial but Important Theory

Risk Homeostasis: Drivers Seek a Target Level of Risk

Prof. Gerald Wilde (Queen's University) proposed Risk Homeostasis Theory in 1994: drivers do not try to minimise risk — they seek a subjectively acceptable target level of risk. When safety improves (better brakes, safer roads), drivers unconsciously increase their risk-taking to maintain their preferred level. This theory is debated, but the core phenomenon — behavioural adaptation to safety measures — is well-documented and important for safety design.

The Homeostasis Mechanism

Think of a thermostat: the system continually adjusts to maintain a target temperature. Wilde's model treats risk perception as a thermostat. Safety improvements → driver feels "safer" → driver speeds up, follows closer, accepts more gaps — until perceived risk returns to the target level.

The implication: engineering solutions alone may not reduce crash rates proportionally, because driver behaviour compensates. The total risk in the system may remain roughly constant.

ABS — The Classic Example

When ABS (anti-lock braking) was introduced, it was expected to reduce crash rates significantly. Wilde predicted drivers would follow more closely on wet roads because braking felt "safer." Studies confirmed shorter following distances in ABS-equipped vehicles, partially offsetting the safety gain.

Motorway Safety Paradox

Motorways are statistically the safest roads per km. Yet some drivers drive faster on motorways than rural roads, partly because the road environment (wide lanes, barriers, no junctions) makes high speed feel safe. The safety feature creates the speeding behaviour.

Evidence For and Against

Supporting evidence: Cycle helmets → riskier cycling behaviour. Seatbelts → some evidence of faster driving. ABS → shorter following distances. Speed cameras removed → speeds increase immediately.

Counter-evidence: Many safety improvements DO reduce absolute crash rates (AEB, median barriers). Not all drivers show behavioural adaptation. The effect is real but does not fully offset safety gains in most cases.

The Practical Implication

Safety measures work best when they reduce crash severity rather than rely on changing driver behaviour. A median barrier prevents head-on crashes regardless of driver behaviour — no adaptation possible. AEB brakes for you — no adaptation. These are more robust than measures relying on drivers to maintain changed behaviour.

This is why Safe System engineering prioritises physical protection over driver behaviour change alone.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
MIT CSAIL Original Research

Social Value Orientation: Why Some Drivers Are Cooperative and Others Competitive

MIT CSAIL (Computer Science and Artificial Intelligence Lab) applied Social Value Orientation theory to driving to explain why drivers respond so differently to the same traffic situations. SVO is a personality dimension that measures whether a person prioritises their own outcomes only (individualist/competitive) or also values others' outcomes (prosocial/cooperative). In traffic, this predicts whether a driver will let others merge, maintain safe gaps, or compete aggressively for position.

The Three SVO Profiles

Prosocial (≈60% of drivers): Seeks good outcomes for self and others. Yields, maintains gaps, cooperative in merging. Values social harmony in traffic. Traffic flows better when these drivers interact.

Individualist (≈30% of drivers): Focuses on maximising own journey time with little concern for effect on others. Will merge aggressively, exploit gaps, but not necessarily maliciously — just self-focused.

Competitive (≈10% of drivers): Seeks to outperform other drivers — their measure of success is relative position, not absolute journey time. Will tailgate, refuse to yield, cut up. These drivers actively worsen traffic flow for everyone.

MIT CSAIL — CarTel & SVO Research

MIT CSAIL researchers used instrumented vehicles and the CarTel mobile sensing system to measure driving behaviour in Boston traffic. They correlated driving behaviour patterns with SVO scores measured in pre-drive psychological testing. Finding: competitive SVO drivers were 40% more likely to make aggressive lane changes, 60% more likely to drive within 1 second following distance, and had 2.3× the near-miss rate of prosocial drivers — despite reporting equal subjective confidence in their driving ability.

SVO and Traffic Flow

MIT CSAIL modelling showed that if 15% of drivers exhibit competitive SVO in a dense traffic stream, it degrades overall flow for ALL drivers by up to 30%. A small minority of competitive drivers imposes significant time and risk costs on the majority. This is the "tragedy of the commons" in traffic.

This finding supports the case for automated vehicles — an AV with prosocial programming could individually outperform human drivers and collectively improve system efficiency.

SVO Is Not Fixed

SVO in traffic can shift situationally: drivers under time pressure shift toward individualist/competitive behaviour. Drivers who have recently been cut up show competitive retaliation patterns. Running late is a temporary SVO shift that permanently increases crash risk. The safest drivers maintain prosocial behaviour even when provoked or rushed.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
From Frustration to Danger

Aggressive Driving & Road Rage: The Psychology

Aggressive driving sits at the intersection of competitive SVO, situational frustration, stress, and disinhibition. It ranges from persistent tailgating and aggressive lane changes (aggressive driving) to verbal confrontation, vehicle ramming, and physical assault (road rage). Research shows the transition from frustration to aggression follows a predictable psychological pathway — and understanding it helps drivers recognise and break the cycle.

1

Trigger Event

Another driver cuts in, moves too slowly, fails to signal. The event is interpreted as intentional rather than accidental ("that idiot deliberately did that").

2

Attribution Error

We blame character, not situation. The slow driver is "inconsiderate" — we don't consider they might be lost, ill, or learner. This is the Fundamental Attribution Error.

3

Autonomic Arousal

Cortisol and adrenaline rise. Heart rate increases. The amygdala (emotional processing) activates. Prefrontal cortex (rational decision-making) is suppressed. The driver is physiologically primed for fight-or-flight.

4

Vehicle as Armour

Anonymity and physical protection of the car create disinhibition. Behaviours that would be unthinkable face-to-face (screaming abuse, physical intimidation) become thinkable behind glass.

5

Retaliation Behaviour

Tailgating, flashing lights, cutting up, verbal abuse, or physical confrontation. The focus shifts entirely from safe journey completion to "winning" the interaction. Both parties are now at dramatically elevated crash risk.

The Physiological Impairment of Anger

Anger and emotional arousal produce measurable driving impairment comparable to mild alcohol consumption. Studies show:
• Reaction time increases by 20–40%
• Following distance decreases
• Risk-taking decisions increase
• Attention narrows (tunnel vision on the target vehicle)
• Other hazards are missed

An angry driver is an impaired driver. The impairment is invisible to anyone watching from outside.

Breaking the Cycle

  • Re-attribute: "they're probably lost" not "deliberate idiot"
  • Breathe: 5 deep breaths activates parasympathetic system, reduces cortisol
  • Increase distance from the other vehicle — physical separation reduces arousal
  • Never retaliate — escalation always increases risk to yourself and innocent third parties
  • If you feel rage: pull over for 5 minutes before continuing
Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
The Biology of Young Driver Risk

Young Drivers: The Neuroscience of High Risk

Young driver crash rates are not simply the result of inexperience — they have a neurological basis. The adolescent and young adult brain is structurally different from the mature adult brain in ways that directly affect risk assessment, impulse control, and sensitivity to social influence. Understanding this biology transforms the conversation from "irresponsible young people" to "predictable brain development mismatch."

The Prefrontal Cortex (PFC) Timeline

The PFC is the brain region responsible for: impulse control, risk assessment, consequence evaluation, long-term planning, and overriding emotional reactions. It is the brain's rational executive.

Critical fact: the PFC is not fully mature until approximately age 25. It is the last brain region to complete development.

This means a 17-year-old driver has an incomplete neural tool for risk assessment — regardless of how mature or intelligent they are.

The Limbic System Imbalance

While the PFC matures slowly, the limbic system (emotion, reward-seeking) is fully developed in adolescence — and is actually hyperactivated during teenage years. The result: strong reward-seeking and emotional reactivity with inadequate inhibitory control. A sports car, peer dares, or emotional excitement can override safety reasoning entirely.

Sensitivity to Social Rewards

Adolescent and young adult brains show heightened activation of the dopamine reward system in social contexts. The peer approval reward is neurologically more powerful than in adults. This explains why young drivers take risks in front of peers that they do not take alone — the social reward literally overrides risk inhibition at a neurological level.

The peer passenger effect: Each additional teenage peer passenger multiplies a novice driver's fatal crash risk by approximately 1.5×. Three passengers → 3.4× baseline crash risk.

MIT AgeLab Young Driver Programme

MIT AgeLab's longitudinal study followed novice drivers from licence acquisition. Key finding: the first 6 months of independent driving carries 5× the crash rate of experienced drivers. The crash rate halved after 12 months and continued declining to age 25. The curve follows PFC maturation. This provides the scientific basis for Graduated Driver Licensing restrictions during the high-risk developmental window.

crash rate of novice young drivers vs. experienced adults in first 6 months
3.4×
crash risk with 3+ teen passengers vs. driving alone (novice driver)
25
age at which prefrontal cortex fully matures — crash rates stabilise
40%
reduction in young driver fatalities from Graduated Licensing (Australia, NZ)
Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
MIT AgeLab — The World's Leading Older Driver Research

Older Drivers: Strengths, Changes & Adaptations

MIT AgeLab is internationally recognised as a world leader in older driver research. Their LNTP (Life-Span Naturalistic Driving Programme) and Advanced Vehicle Technology studies provide the most comprehensive picture available of how driving capability changes with age — and crucially, how older drivers successfully adapt. The picture is more nuanced than "older drivers are dangerous."

Changes That Occur with Age

  • Vision: Contrast sensitivity declines, especially at night. Useful field of view contracts (selective attention narrowing). Recovery from glare slows. Visual acuity falls.
  • Processing speed: Central processing slows. Reaction time increases. Under time pressure, decisions degrade more rapidly than in younger adults.
  • Head/neck mobility: Reduced rotation limits mirror checking and shoulder checks.
  • Divided attention: Simultaneous management of multiple traffic situations becomes harder.
  • Frailty: Older bodies sustain more severe injuries from the same impact force.

Strengths That Compensate

  • Experience: Deep pattern recognition. "Seen this before" — faster appropriate response in familiar situations.
  • Self-regulation: Older drivers voluntarily restrict their driving to conditions they handle well: daylight, dry roads, familiar routes, low-traffic times.
  • Risk attitude: Lower competitive SVO, less sensation-seeking. Less likely to speed, tailgate, or drive aggressively.
  • Hazard anticipation: Long experience builds excellent predictive models of what hazards will appear at what locations.
MIT AgeLab LNTP Finding

MIT AgeLab's longitudinal naturalistic data shows that older drivers who engage in deliberate self-regulation — voluntarily avoiding their known challenging conditions — can maintain safe driving well into their 80s in many cases. The most dangerous older drivers are not those who drive slowly, but those who lack insight into their own limitations and continue driving in conditions that exceed their current capability.

The Mobility-Safety Balance

Driving cessation for older adults is associated with increased social isolation, depression, and even reduced life expectancy. The goal should not be to remove older drivers from roads — it is to help them drive safely for as long as possible. Assessment, training, vehicle adaptations (wider mirrors, better lighting, adaptive cruise) are more humane and effective than blanket restrictions.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Do Driving Personalities Exist?

Driver Personality: What the Research Says

Popular culture recognises "the aggressive driver," "the careful driver," and "the reckless driver" as personality types. But does the scientific evidence support stable driving personalities? MIT CSAIL and AgeLab research — combined with Big Five personality psychology — shows that personality does predict driving behaviour, but the relationship is complex and context-dependent.

Big Five Personality & Driving

TraitHigh score effect
NeuroticismAnxious driving, over-caution, poor emergency response
ExtraversionHigher speeds, more social driving, more distractions accepted
OpennessLittle direct effect on crash risk
AgreeablenessMore cooperative, lower aggressive driving
ConscientiousnessStrongest predictor of safe driving — rule-following, attention to detail

Sensation-Seeking

Sensation-seeking — the desire for novel, intense experiences — is strongly associated with speeding, running red lights, and drink driving. Highest in 18–24-year-old males. Declines with age. Partially genetic. Correlates with competitive SVO in traffic.

Sensation-seeking drivers do not respond well to fear-based campaigns — the arousal of fear can paradoxically increase the appeal of the forbidden behaviour. Educational approaches emphasising skill and mastery work better.

MIT CSAIL — Driving Style Clustering

MIT CSAIL researchers used machine learning on naturalistic driving data to identify natural clusters of driving behaviour. They found three robust styles:

Calm/Safe: Smooth acceleration, long following distances, low speed variance. Low crash rate.

Normal: Average profile, situationally variable. Average crash rate.

Aggressive: Hard braking, rapid acceleration, short following distance, high speed variance. 3× crash rate of calm drivers.

Critically: style was stable across time and road types. It is a genuine behavioural trait, not just situational.

Personality Does Not Determine Behaviour

Personality is a tendency, not a destiny. High sensation-seekers with good training and strong safety culture show significantly lower crash rates than untrained sensation-seekers. Personality identifies who needs more targeted intervention — it doesn't make intervention futile.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
How Your Mood Changes Your Driving

Emotional State & Driving Performance

Drivers do not step into a vehicle as emotionally neutral operators. They bring their day's stress, their anxieties, their grief, their anger, and their elation. Research shows that emotional state at the time of driving significantly affects risk-taking, attention, and decision quality — and that the vehicle itself can amplify emotional states through isolation, perceived anonymity, and the frustrations of traffic.

Negative Emotions & Risk

Anger: Increases speeding, tailgating, aggressive manoeuvres. Narrows attention. Elevated cortisol impairs judgment. The most dangerous single emotional state for driving.

Anxiety/Worry: Rumination occupies working memory. Driver may be technically competent but cognitively elsewhere. Slow, hesitant responses at junctions. Paradoxically, very anxious drivers are also crash-prone.

Sadness/Grief: Impairs attention, slows processing, reduces motivation to drive carefully ("nothing matters"). Grief is documented in driving crash data — bereaved drivers have elevated crash risk in the 3–6 months after loss.

Positive Emotions — Also Risky

Elation: Euphoria and high positive affect can also impair driving — through overconfidence, risk underestimation, and reduced attention to mundane driving tasks.

Excitement: Similar to anger in physiological terms — elevated heart rate, reduced inhibition. Young drivers receiving good news, or excited by music, show similar impairments to mildly angry drivers.

MIT AgeLab — Emotional Driving Research

MIT AgeLab equipped research vehicles with physiological monitoring (heart rate, skin conductance, facial expression cameras) alongside driving performance metrics. They found that emotional state at journey start predicted driving behaviour throughout the journey. Drivers who were emotionally dysregulated at journey start showed elevated hard braking events, lane departures, and speed variance compared to emotionally neutral baseline drives — even 30+ minutes into a journey.

The "Take 5" Rule

If you have just had a major argument, received shocking news, or are in a state of emotional agitation — wait 5 minutes before driving. Take deep breaths. The journey can wait. Your emotional state at the wheel is a safety-critical variable.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
How Driving Becomes Automatic — and Why That's Risky

Habit, Automaticity & the Danger of Routine

Habit formation is one of the most powerful forces in human behaviour. When a behaviour is repeated in the same context, it transitions from deliberate (System 2) to automatic (System 1). This is how we learn to drive — and it is how driving becomes efficient. But habits are extraordinarily resistant to change, and they continue running even when context has changed in a way that makes the old habit dangerous.

The Habit Loop (Duhigg/MIT)

CueRoutineReward. Every habitual driving behaviour follows this structure. The morning commute cue triggers the routine of "drive the familiar route at familiar speed." The reward is efficient, low-effort journey completion.

Why Habits Are Hard to Change

Habit neurological pathways are in the basal ganglia — below the cortex. Once established, they can be suppressed by cortical intention but not erased. Under stress or cognitive load, cortical suppression fails and the old habit reasserts itself. This is why drivers "revert" to unsafe behaviours when tired or distracted.

When Context Changes — Habit Doesn't

A driver habitually rolls through the stop sign on their quiet street at 7am. One day a cyclist appears for the first time. The habit continues — the cyclist is hit. The stop sign had been rolled through 1,000 times safely. The habit was built on the absence of hazard, not on safety of the manoeuvre itself.

MIT Research — Habit in Commuting

MIT research on commuting behaviour found that driving habits are among the most stable human behaviours. Even when commuters change jobs (and therefore route), they revert to old route patterns within weeks if there is any possibility of using the original route. The habit cue (leaving the house in the morning) is so strong it overrides new intentions. This has significant implications for introducing new road layouts — drivers will use the old routing behaviour for months before adapting.

Using Habit Formation for Safety

Habit formation can be harnessed for safety as well as against it. Consistently performing a safety routine (mirror check before moving off, 3-second commentary at junctions) in the same context eventually automates that behaviour — it becomes a default, low-effort habit rather than requiring conscious effort each time.

This is why the best driving training emphasises consistent routines rather than just knowledge of rules. Rules fade; habits persist.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Why Impaired Drivers Get Behind the Wheel

Drink & Drug Driving: The Psychological Mechanisms

The decision to drive while impaired is not usually made by a person who thinks "I'll drive drunk." It is made by a person whose judgment — including their judgment of their own impairment — has been degraded by the substance they have consumed. Understanding this self-reinforcing impairment helps explain why awareness alone does not prevent drink driving, and what interventions actually work.

Alcohol's Dose-Response Effects

BACDriving Effects
0.02%Divided attention begins to degrade, relaxation reduces vigilance
0.05%Reaction time +15%, hazard detection reduced, risk tolerance increases
0.08%Crash risk 4×, coordination and brake response significantly impaired
0.10%Crash risk 7×, severe judgment impairment
0.15%Crash risk 25×, major motor and visual impairment
The Confidence Paradox

Alcohol simultaneously impairs driving ability AND reduces the driver's perception of their own impairment. A drunk driver feels more confident, not less — lower anxiety, reduced self-monitoring, overestimation of competence. The person least qualified to judge whether they should drive is the intoxicated person deciding whether to drive.

Morning-After Impairment

Alcohol metabolises at approximately 1 unit per hour. After heavy drinking (10+ units), BAC can still be above 0.05% the following morning. Many drink-drive prosecutions occur the morning after a night out — the driver genuinely believes they are "fine."

Sleep does not speed metabolism. The only thing that removes alcohol is time and liver function. Coffee, water, and food have no effect on BAC.

Cannabis — The Underestimated Risk

THC (cannabis) increases crash risk by approximately 2×. Unlike alcohol, cannabis impairment cannot be precisely measured by any roadside test currently in use. Critically, cannabis users significantly underestimate their impairment — studies show users rate their driving as normal when simulator performance shows significant degradation. Cannabis impairs: reaction time, lane keeping, speed regulation, and risk assessment. Combined with alcohol, the effect is multiplicative, not additive.

Why Deterrence Works (and Doesn't)

Legal consequences deter drink driving when perceived detection probability is high. In Ireland, mandatory alcohol testing checkpoints (MATs) are the single most effective deterrence measure because they create visible, credible detection risk. When enforcement is invisible, deterrence fails regardless of penalty severity.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
What Actually Changes Driver Behaviour

Behaviour Change: From Theory to Road Safety

Behaviour change is one of the most studied topics in psychology, and the road safety field has repeatedly failed to apply its findings. Decades of shock advertising, speed awareness courses, and knowledge campaigns have produced limited results. Understanding what actually changes behaviour — versus what feels like it should — is essential for effective road safety work.

What Doesn't Work Well

  • Fear appeals: Graphic crash images produce short-term attitude change but minimal long-term behaviour change. High-fear messages are processed and then psychologically defended against.
  • Knowledge provision alone: People already know speeding, phone use, and drunk driving are dangerous. Knowledge is not the barrier to behaviour change.
  • One-time training: Skills and intentions decay rapidly without reinforcement.
  • Punishment alone: Effective only with high perceived detection probability — rare in road transport.

What Does Work

  • Immediate personalised feedback: Telematics showing your own speeding, braking, and phone use events. Personal data is more motivating than statistics.
  • Social norm correction: Telling drivers that "most drivers in your area do not use phones" — correcting the over-perceived norm — reduces behaviour.
  • Default changes: Making the safe choice the default (Do Not Disturb on by default, seatbelt reminder chimes) exploits status quo bias.
  • Commitment devices: Public or written commitments ("I will not use my phone while driving") increase follow-through.
  • Peer modelling: Seeing people we respect behave safely. Most powerful in young drivers.
MIT AgeLab — The Nudge Approach

MIT AgeLab research on behaviour change in older drivers found that "nudges" — small environmental design changes that make safe choices easier — were more effective than traditional education. Examples: making the seatbelt buckle more visible and accessible (increased use by 8%), changing default navigation announcement to include road condition alerts (increased compliance with speed adjustments). Design for safety, not just education for safety.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
What Good Driving Training Actually Teaches

The GDE Matrix: Goals for Driver Education

The Goals for Driver Education (GDE) matrix is the EU's evidence-based framework for what comprehensive driver education must achieve. Developed by Hatakka et al. and adopted across Europe, it identifies four levels of driving competence — from basic vehicle control to self-awareness and lifestyle. Most driving tests only assess the bottom two levels; crashes are mostly caused by failures at the top two levels.

The Four GDE Levels

Level 1 — Vehicle manoeuvring: Physical control of the vehicle. Steering, braking, gears, parking. This is what driving tests primarily assess.

Level 2 — Mastering traffic situations: Handling interactions with other road users. Junctions, overtaking, merging, following distance. Most of road safety education operates here.

Level 3 — Goals and context of driving: Why am I driving? What goals am I trying to achieve? Am I in the right state? Should I even be driving now? Trip planning and journey decisions.

Level 4 — Self-awareness: What are my personal strengths and weaknesses as a driver? How do my personality, emotions, attitudes, and values affect my driving? This is the level most missed by traditional education.

Why Level 4 Matters Most

Research consistently shows that most crashes do not result from lack of vehicle control skill (Level 1) or even traffic knowledge (Level 2). They result from Level 3 and 4 failures: driving when tired, rushing because running late, driving while emotionally aroused, overestimating one's own skill, normalising risky behaviour.

A driver who knows they are a sensation-seeker and has high aggression (Level 4 self-awareness) and therefore consciously applies more conservative strategies is safer than a technically excellent driver who is blind to their own emotional vulnerabilities.

Ireland's ADI Training Curriculum

Ireland's Approved Driving Instructor curriculum and the Essential Driver Training (EDT) programme both incorporate GDE Level 3 and 4 elements: journey planning, attitude assessment, understanding own risk factors. These are tested in the driving test via discussion and reflective questions — not just vehicle handling.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Traffic as a Social System

Social Influence on Driving Behaviour

Traffic is a profoundly social environment — we are continuously influenced by the behaviour of drivers around us, by cultural norms about "normal" driving, and by the expectations of people inside our vehicle. MIT CSAIL research on the social dynamics of traffic shows that individual driver behaviour cannot be fully understood without understanding the social context in which it occurs.

Speed Norms — Social Learning

Drivers calibrate their speed partly to the speeds of surrounding vehicles. If everyone around you is travelling at 130 km/h on a 120 km/h motorway, travelling at 120 km/h feels actively dangerous (risk of rear-end collision). Social norm has overridden the legal limit.

This creates a self-reinforcing cycle: if most drivers speed, new drivers learn that speeding is the "real" norm. Compliance with limits becomes the deviant behaviour.

The "Everyone Does It" Rationalisation

The most common justification for unsafe behaviour is descriptive norm reference: "everyone runs that junction," "everyone speeds on that road," "everyone uses their phone." Research shows perceived peer behaviour is more predictive of individual behaviour than personal attitudes or legal knowledge.

Interventions that correct over-estimated norms ("actually, 70% of drivers on this road comply with the speed limit") are effective specifically because they undermine this rationalisation.

MIT CSAIL CarTel — Social Norm Feedback

MIT CSAIL's CarTel project (mobile traffic sensing) ran an experiment where drivers received feedback showing how their driving compared to other drivers on the same roads. Drivers who received feedback that they were in the top 20% most aggressive (measured by hard braking, acceleration, speed) showed a 25% reduction in aggressive driving events over 4 weeks. Social comparison — particularly downward comparison ("you are more aggressive than most") — is a powerful motivator for change.

Modelling Safe Behaviour

Parents are the single most influential model for their children's future driving behaviour. Research shows parental driving style is more predictive of child driving style than formal instruction or peer influence. The "do as I say, not as I do" parent creates a dangerous mismatch between the rules they teach and the norms they demonstrate.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
The Decision to Drive Tired

Fatigue Psychology: Why We Drive When Tired

We covered the neuroscience of fatigue in Module 01. This slide focuses on the psychology of the decision to drive fatigued — why people knowingly get behind the wheel when tired, and what that decision process looks like. It is rarely straightforward recklessness — it is usually a series of small rationalisations, each of which seems reasonable, leading to a collectively dangerous outcome.

1

Sleep debt accumulation (often unnoticed)

5–6 hours sleep per night over a week creates massive sleep debt. Driver feels "a bit tired" but not "dangerously tired." The gap between felt tiredness and actual impairment is large.

2

"I'll be fine, it's not far"

Distance rationalisation. Microsleep can occur within 10 minutes of driving onset. Distance does not protect — sudden onset at any point on the journey is equally possible.

3

"I've driven tired before without incident"

Normalisation of deviance. Previous survival creates false safety signal. Prior exposures to the risk with no consequence increase confidence, not safety.

4

"I'll open the window / put music on"

Countermeasure rationalisation. These measures provide minutes of relief — not sufficient for a journey of any length. Driver overestimates their effectiveness.

5

No alternative perceived

"I have to get home." "There's nowhere to stop." "I can't afford a hotel." Economic and logistical pressures override safety judgment — especially at night on motorways.

The Impairment You Can't Feel

After 20 hours awake, driving performance is equivalent to a BAC of 0.08% — the legal drink-drive limit. After 24 hours, performance matches BAC 0.10%. Unlike alcohol, severe fatigue does not produce the subjective experience of being impaired. Fatigued drivers feel they are driving normally up to the point of microsleep. The impairment is invisible from the inside.

Breaking the Decision to Drive Tired

  • Pre-plan: identify motorway services, hotel locations on long journeys
  • Set a personal rule: "If I yawn more than 3 times in 5 minutes, I stop"
  • 20-minute nap + coffee at motorway services is genuinely effective for 1–2 hours
  • Share driving — swap every 2 hours on long journeys
  • Budget time as you would budget fuel — running out has the same consequence
Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Technology Meets Behaviour Change

Telematics & Behavioural Feedback: What the Data Shows

Insurance telematics (black boxes), fleet monitoring systems, and smartphone driving apps represent the most significant recent development in driver behaviour change. By providing drivers with personalised, accurate, immediate feedback on their own driving behaviour, telematics addresses the single biggest barrier to behaviour change: the gap between what drivers believe they do and what they actually do.

How Telematics Feedback Works

  • GPS tracking records speed, route, time of day
  • Accelerometers detect hard braking, harsh cornering, rapid acceleration
  • Phone sensors detect screen activity (some systems)
  • Data compiles into a driver score (e.g. 1–100)
  • Driver receives regular reports: daily summary, weekly trends
  • Comparison to peer group average (social comparison)

Evidence of Effectiveness

  • Young driver telematics policies: 20–40% reduction in crash rate (UK data)
  • Fleet telematics: 30–50% reduction in aggressive driving events (TRL)
  • MIT CarTel social comparison: 25% reduction in harsh driving when shown relative ranking
  • Speed reduction: average 6–10 km/h reduction in excess speed with real-time in-car feedback
Why Telematics Works — Psychology Explanation

Telematics is effective because it addresses the psychological barriers to change:

Optimism bias: "I'm a good driver" is confronted by personal data showing actual speeding events.

Social comparison: Showing your score relative to peers activates competitive prosocial motivation.

Immediacy: In-car real-time alerts provide consequence in the moment — not after a crash.

Gamification: Score improvement creates a reward loop that sustains motivation beyond initial awareness.

Telematics Limitations

Telematics measures inputs (hard braking, speed) not outcomes. A driver can have a perfect telematics score while driving inattentively at legal speeds. Cognitive distraction, fatigue, and inattention are not captured. Telematics improves the measurable behaviours it measures — it doesn't make a driver attentive or alert.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Personal Application

Applying Driver Psychology to Your Own Behaviour

The value of this module is not in describing other drivers' psychology — it is in recognising your own psychological vulnerabilities and designing your behaviour around them. Every driver has biases, habits, and emotional triggers. The difference between safe and unsafe drivers is often not the presence or absence of these vulnerabilities — it is the level of self-awareness about them.

Self-Assessment Questions

  • Optimism bias: Do you believe you are a below-average crash risk? What evidence do you have for this?
  • SVO: Do you feel competitive in traffic? Do you feel angry when someone overtakes you?
  • Risk homeostasis: Did you start driving faster after getting a safer car?
  • Familiar route: Do you drive less attentively on roads you know well?
  • Emotional driving: Do you drive differently after an argument or stressful meeting?
  • Normalised deviance: What rules do you regularly break "because nothing happens"?

Protective Strategies — Based on Your Vulnerabilities

  • If you're competitive: consciously adopt a "mission complete" goal — the goal is safe arrival, not position
  • If you're a sensation-seeker: book track days; get the speed on a circuit, not the road
  • If you drive emotionally: create a 5-minute pre-drive breathing ritual for high-stress days
  • If you're complacent on familiar roads: use commentary driving on known routes to re-engage System 2
  • If you normalise deviance: audit your habits against the rules and identify the gap honestly
The Expert Driver's Psychology

Research on advanced drivers (IAM, ROSPA) shows they share specific psychological traits: high self-awareness about their own limitations, consistent prosocial SVO in traffic, strong commitment to journey completion over speed, and deliberate use of System 2 thinking in familiar environments. These traits are trainable — they are not personality fixed. The expert driver's key skill is knowing when their automatic thinking needs to be overridden.

Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Reference Summary

Key Researchers & Frameworks Reference Card

Researcher / ModelContributionKey Concept
Daniel KahnemanDual Process Theory (2011)System 1 vs. System 2 thinking in driving decisions
Gerald WildeRisk Homeostasis Theory (1994)Drivers seek a target risk level; safety measures can be offset by behaviour change
MIT CSAILSocial Value Orientation in trafficProsocial vs. competitive drivers; social dynamics of traffic flow
Bryan Reimer (MIT AgeLab)Distraction & automation research27-second cognitive residue; PERCLOS; ADAS complacency
Joseph Coughlin (MIT AgeLab)Ageing and mobility researchOlder driver self-regulation; LNTP longitudinal data
Hatakka et al.GDE Matrix (2002)4-level driver education from vehicle control to self-awareness
Paul SlovicRisk perception (1987)Systematic biases in how humans estimate risk
Laurence SteinbergAdolescent risk neurosciencePFC immaturity, social reward sensitivity in young drivers
Michie et al.Behaviour Change Wheel (2011)COM-B model: capability, opportunity, motivation needed for change
Smart Driving Academy
MIT CSAIL · AgeLab — Behaviour Research
Module Summary

Summary: What You Now Know About Driver Psychology

How Drivers Think

  • System 1 (automatic) makes 95% of driving decisions — fast but bias-prone
  • Optimism bias: most drivers think they are below-average crash risk
  • Familiar roads trigger complacency — System 1 uses outdated safe templates
  • Risk homeostasis: safety improvements can be unconsciously offset
  • Emotional state profoundly affects driving — anger = physiological impairment

Young & Older Drivers

  • Young: PFC not mature until 25 — neurological basis for risk-taking
  • Peer passengers multiply young driver crash risk
  • Older: self-regulation is the key protective factor, not speed or avoidance
  • GDL saves young driver lives by protecting the vulnerable developmental window

Social & Personality Factors

  • MIT CSAIL: 3 SVO profiles — prosocial, individualist, competitive
  • Competitive 10% of drivers cause 30% of flow degradation
  • Social norms are more predictive than personal attitudes
  • Conscientiousness = strongest personality predictor of safe driving

Changing Behaviour

  • Fear + knowledge alone: minimal effect
  • Personalised feedback, defaults, commitment: evidence-based
  • Telematics: 20–50% reduction in dangerous driving events
  • GDE Level 4: self-awareness is the highest and most neglected skill
The Core Insight

Crashes happen not just because of what drivers do — but because of who they are, how they think, and what they feel at the moment they drive. Psychology is not a soft topic in road safety — it is the central one.

Instructor Notes

End with the GDE Level 4 self-assessment questions from Slide 18. Ask participants to privately identify their single biggest psychological vulnerability as a driver. Then ask: "What one structural change could you make to your driving environment that would make that vulnerability less dangerous?" This is the practical payoff from the entire module.