Driving Behaviors That Cause Crashes—Even When Drivers Think They're Being Safe

The Confidence Problem at the Heart of American Driving

Ask most American drivers to rate their own skill behind the wheel, and the majority will place themselves above average — a statistical impossibility that researchers call the illusory superiority bias. A 2018 study published in the journal Transportation Research Part F found that roughly 80% of drivers considered themselves safer than the median. The National Highway Traffic Safety Administration (NHTSA) has identified overconfidence as a behavioral underpinning in a significant share of at-fault crashes, because drivers who believe they are skilled tend to monitor their environment less rigorously and accept higher-risk situations more readily.

This isn't a story about reckless drivers weaving through traffic at triple-digit speeds. It's a story about ordinary people — commuters, parents on school runs, retirees on familiar routes — who are making small, habitual errors they are completely unaware of. The crashes that result feel like accidents. The research suggests many of them are anything but.

Understanding which behaviors cause crashes — and why those behaviors feel safe in the moment — is the first step toward actually driving more safely. What follows draws on NHTSA crash data, IIHS research, and established traffic psychology to identify the mistakes most likely to put you at risk, even when you believe you're doing everything right.

For a broader framework on evidence-based safe driving, see our guide to what defensive driving actually means on American roads.

The Mistakes Behind the Crashes

Each of the following behaviors has a documented relationship to crash causation in U.S. research. None of them requires speeding, intoxication, or obvious recklessness. They are the quiet errors of otherwise conscientious drivers.

Intersection behavior deserves particular attention here. Research into intersection crash patterns shows that over 40% of all U.S. crashes occur at these junctions — and driver behavior, not road design, is the dominant factor in most of them. Drivers who feel they are proceeding legally and safely still collide because they scan too narrowly and assume oncoming traffic will behave predictably.

Why Safe-Feeling Behaviors Can Still Be High-Risk

Traffic psychology offers a concept called risk homeostasis: when drivers perceive themselves to be safer — because they are driving a large SUV, have advanced safety tech, or are on a familiar road — they unconsciously compensate by accepting greater behavioral risks. The net result is that safety improvements in one dimension get offset by behavioral degradation in another.

Modern vehicles equipped with automatic emergency braking, lane-keep assist, and blind-spot monitoring are measurably effective at preventing crashes. But IIHS and NHTSA testing of road safety technology consistently finds that drivers with these systems installed tend to monitor their mirrors less frequently and maintain shorter following distances, partially negating the technology's protective benefit.

The same dynamic applies to familiarity. Studies using naturalistic driving data — instrumented vehicles that record real behavior over months — show that drivers are significantly more likely to run stop signs, exceed speed limits, and skip shoulder checks on routes they drive every day compared to unfamiliar roads. Familiarity breeds inattention, and inattention breeds crashes on roads where the driver is absolutely certain nothing will go wrong.

Fatigue compounds every one of these patterns. Drowsy driving causes over 100,000 crashes per year, according to NHTSA estimates — and crucially, fatigued drivers are among the least likely to self-identify as impaired. The warning signs arrive well before the eyelid-drooping stage that most people associate with dangerous tiredness.

Behaviors Specific to High-Frequency Crash Scenarios

Certain driving environments concentrate behavioral risk in ways that deserve individual attention. Parking lots, for instance, are responsible for approximately 20% of all vehicle accidents — a share that surprises most drivers. The low speeds involved create a false sense of safety that leads to reduced scanning, abrupt reversals without checking mirrors, and pedestrian conflicts. Our guide to backing up safely documents the specific errors that turn parking lots into collision zones.

Construction zones present a different risk profile. Drivers who correctly slow down often still follow too closely in work zones, misjudge the reduced lane widths, and fail to anticipate sudden stops from lead vehicles. Work zone crashes kill more than 800 people annually in the U.S., with a significant share of fatalities involving workers struck by errant vehicles whose drivers had adequate warning distance to stop.

Tailgating is perhaps the behavior most commonly rationalized by its practitioners. Drivers who follow closely often believe they are maintaining traffic flow, pressuring a slow driver to move over, or simply matching the pace of traffic. None of these justifications hold up against crash data. Common tailgating beliefs don't hold up under crash research, and the stopping-distance math is unambiguous: at 65 mph, a vehicle travels approximately 95 feet per second, and human reaction time alone consumes 1.5 seconds before brakes are even applied.

For teen drivers, who lack both experience and a calibrated sense of personal risk, these behavioral patterns are amplified. Drivers aged 16–19 have crash rates nearly three times higher than adults, and the underlying causes map closely to the mistakes outlined above: overconfidence, under-scanning, and inadequate following distance chief among them.

Building Habits That Actually Reduce Your Risk

The research on crash causation is sobering, but it is also actionable. The behaviors that cause crashes are, by definition, behaviors that can be changed. Traffic safety researchers have identified several evidence-based habits that produce consistent reductions in crash involvement when practiced regularly.

• The two-second rule, applied honestly. Most drivers who believe they maintain a two-second following gap are actually following at less than one second when measured by instrumented vehicle research. Count deliberately: pick a fixed point on the road ahead, watch the vehicle in front pass it, and count. Two full seconds is longer than it feels at highway speeds.
• Structured mirror scanning. Professional driver training programs teach a specific mirror-check cadence — center mirror every five to eight seconds, side mirrors incorporated into a regular sweep. This isn't how most drivers were taught, and it isn't how most drive. A deliberate, timed routine transforms mirror use from reactive to proactive.
• Intersection approach protocol. Treat every green light as potentially yellow. Ease off the accelerator slightly before entering, scan left-right-left for cross traffic, and resist the reflex to accelerate immediately when a light turns green. The first second of green is when intersection-clearing violations most commonly occur.
• Managing your own aggression honestly. Brake-checking, gap-blocking, and extended horn use are forms of aggression that most drivers do not categorize as such in the moment. The test: if the other driver described the same behavior, would you recognize it as aggressive? If yes, it is.
• Reintroducing unfamiliarity to familiar routes. Consciously treating a known route as if it were new — attending to signage, scanning driveways and side streets, checking mirrors at habitual intervals — counteracts the inattention that familiarity generates.

If you are ever involved in a crash despite your best precautions, having a clear procedure ready matters. Our accident procedures guide covers what to do in the immediate aftermath, from scene safety to documentation and insurance contact.

Ultimately, safer driving is less about dramatic interventions and more about consistent, low-level attentiveness. The crashes avoided by good driving habits are invisible — they never appear in any data set because they never happen. That invisibility is what makes the effort easy to undervalue, and what makes the research worth taking seriously.