Health Risk Behaviors

Abstract

Health risk behaviors are a type of health behavior: patterns of action that raise a person's probability of disease, injury, or premature death—tobacco and alcohol use, physical inactivity, poor diet, unsafe driving, and unprotected sex among them. What makes them a distinct problem for psychology is that they are not isolated lapses but tend to co-occur, to peak in adolescence, and to persist against full knowledge of their cost. This article develops the construct across levels: how it is defined and why its component behaviors cluster, the decision-theoretic and developmental theories that explain it, the special status of adolescence under the dual-systems account, how risk is perceived and measured in populations, and what current work says about the stubborn gap between knowing a risk and avoiding it.

Keywords: health risk behavior, problem behavior theory, risk perception, sensation seeking, adolescent risk taking

A teenager who smokes is more likely than chance to also drink, to drive fast, and to take sexual risks—the behaviors travel together, as though expressions of one underlying tendency rather than independent habits. Across the lifespan, the propensity to run such risks rises sharply in the early teens, peaks in the middle of adolescence, and declines through the twenties, tracing an arc that no account built on information or deliberation alone can explain. Health risk behaviors are the term the medical and psychological literature uses for this class of action, and their covariation, their developmental timing, and their resistance to warning are the organizing puzzles of the field (Jessor, 1991).

The stakes are concentrated and largely preventable. A handful of behaviors—smoking, heavy drinking, inactivity, poor diet—account for a disproportionate share of chronic disease and early death, and most are established in adolescence, the developmental window in which risk taking is highest and lifelong patterns are set (Kann et al., 2018). This article treats health risk behavior as a problem in the psychology of judgment and development: how it is defined, why its parts cohere, what theories predict it, why adolescence is its peak, and what the evidence says about perceiving and reducing it.

What Health Risk Behaviors Are

Health risk behaviors are activities that increase the likelihood of adverse health outcomes—disease, injury, disability, or premature death. In the Medical Subject Headings vocabulary, Health Risk Behaviors is a descriptor filed at tree position F01.145.488.250, directly beneath health behavior, naming the risk-bearing half of the broader construct: where health behavior as a whole includes the protective acts of a well person, health risk behaviors are the patterns that place health in jeopardy. The canonical list is short and consequential—tobacco use, alcohol and other drug use, physical inactivity, poor diet, risky driving, and sexual behaviors that carry a risk of infection or unintended pregnancy (Kann et al., 2018).

The defining psychological fact about these behaviors is that they cluster. Richard Jessor and Shirley Jessor, studying adolescent development longitudinally, found that problem behaviors—drinking, drug use, early sexual activity, delinquency—were correlated with one another and with a common pattern of personality, perceived environment, and behavior, which they called a problem behavior syndrome (Jessor & Jessor, 1977). A young person high on one risk behavior is disproportionately likely to be high on the others, not because each causes the next but because they share psychosocial roots: low perceived control, weak attachment to conventional institutions, and a value placed on independence over achievement. This covariation is why the field treats health risk behavior as a coherent construct rather than a list of unrelated vices, and why interventions that address the shared profile can affect several behaviors at once (Jessor, 1991).

Figure 1

The Covariation of Adolescent Risk Behaviors

Theories of Health Risk Behavior

Why would a person knowingly choose an action that harms them? The field answers at two levels—the psychology of the decision itself, and the psychosocial and developmental conditions that bias it. No single theory owns the construct; each illuminates one facet of it, and Table 1 sets the major accounts side by side before the sections below take them in turn.

The decision level begins with the discovery that choice under risk departs systematically from the expected-utility ideal. Kahneman and Tversky's prospect theory showed that people evaluate outcomes as gains and losses relative to a reference point, with a value function that is concave for gains, convex for losses, and steeper for losses than for gains—so that the pain of a loss exceeds the pleasure of an equal gain (Kahneman & Tversky, 1979). The practical corollary is the framing effect: the same health choice, described in terms of lives saved rather than lives lost, flips a person from risk-averse to risk-seeking even though the underlying odds are unchanged (Tversky & Kahneman, 1981).

The decision is further distorted by how the risk is perceived. Paul Slovic's psychometric research established that lay judgments of risk are driven less by statistical probability than by qualitative features—above all dread (how catastrophic and uncontrollable a hazard feels) and how unknown or unfamiliar it is—so that familiar, voluntary, everyday hazards such as smoking are systematically underweighted relative to dramatic but rarer ones (Slovic, 1987). This is why the gravest behavioral risks are precisely the ones that fail to frighten: their harms are delayed, probabilistic, and self-imposed, the combination that perception discounts most.

Beneath the decision sits a disposition. Marvin Zuckerman's work on sensation seeking—the trait of seeking varied, novel, and intense experience and accepting risk for its sake—identified a temperamental substrate for risk behavior with a common biosocial basis, linking personality, neurochemistry, and the propensity to run risks (Zuckerman & Kuhlman, 2000). The deliberative route is captured by two belief-based frameworks. The earliest, the Health Belief Model, held that a person's readiness to avoid a health threat turns on their perceived susceptibility to it and its perceived severity, weighed against the perceived benefits and barriers of acting—so that a risk behavior persists when its harm feels neither likely nor grave enough to outweigh its immediate reward (Rosenstock, 1974). Ajzen's later theory of planned behavior generalized this logic, predicting a behavior from the intention to perform it and that intention from attitudes, subjective norms, and perceived control—the framework that links a young person's beliefs and perceived social pressures to whether they take up a risk behavior in the first place (Ajzen, 1991).

Adolescence and the Dual-Systems Model

No fact about health risk behavior is more robust than its developmental timing: the propensity to take risks rises in the early teens, peaks around the middle of adolescence, and falls through the twenties. The puzzle is that this peak coincides with a period when adolescents are, by most cognitive measures, as capable of reasoning about risk as adults—they are not simply ignorant of the dangers. The resolution is the dual-systems model, developed by Laurence Steinberg from developmental social neuroscience, which holds that two neural systems mature on different timetables (Steinberg, 2008). A socioemotional system governing reward sensitivity and sensation seeking is remodeled early, around puberty, and becomes highly reactive; a cognitive-control system supporting self-regulation and the weighing of future consequences matures slowly and is not complete until the mid-twenties. Risk taking peaks in the interval when the first system is fully online and the second is not.

Valerie Reyna's fuzzy-trace theory supplies a complementary account of why the reasoning that is available to adolescents often fails to protect them. The theory distinguishes verbatim representations—precise, quantitative detail—from gist, the bottom-line meaning of a situation, and argues that mature, protective decision making relies on gist rather than on deliberate calculation of odds (Reyna & Farley, 2006). Adolescents, paradoxically, are more likely than adults to reason verbatim—to weigh the small probability of harm against the certain reward—a mode of thought that can actually license a risk a gist-based reasoner would reject out of hand. The two accounts converge on the same practical point: interventions that assume adolescents simply need the facts misread the problem, because the shortfall is in the maturation of control and in the mode of reasoning, not in knowledge of the odds (Zuckerman & Kuhlman, 2000).

Risk Perception and Measurement

Turning from the individual to the population, two measurement problems organize the applied field. The first is the measurement of perceived risk, which the psychometric paradigm maps as a space whose principal axes are dread and the unknown; a hazard's location in that space predicts how strongly the public wants it regulated far better than its actual death toll does (Slovic, 1987). The gap between perceived and statistical risk is itself a target for intervention: behaviors are underregulated and underfeared exactly when they sit low on dread and high on familiarity.

The second is the measurement of behavior at scale. Large surveillance systems such as the Youth Risk Behavior Surveillance System track the prevalence of the priority risk behaviors—tobacco, alcohol and drug use, sexual behaviors, dietary behaviors, and physical inactivity—across successive cohorts of adolescents, providing the epidemiological backbone that tells the field which behaviors are rising, which are falling, and where they concentrate (Kann et al., 2018). Such monitoring matters because the behaviors are both common and clustered: a survey that measures them together can detect the covariation Jessor described and can show, cohort by cohort, whether the developmental peak is shifting. Taken with the problem-behavior tradition, population surveillance reframes health risk behavior as a public-health problem whose levers are as much social and environmental as individual (Jessor, 1991).

Worked Example

The framing effect can be made quantitative with prospect theory's value function. Consider a public-health program for a disease expected to kill 600 people, posed in the classic two-frame form (Tversky & Kahneman, 1981). In the gain frame the options are described by lives saved: program A saves 200 for certain; program B has a one-third chance of saving all 600 and a two-thirds chance of saving no one. In the loss frame the same outcomes are described by deaths: program C lets 400 die for certain; program D has a one-third chance that no one dies and a two-thirds chance that all 600 die. The two frames are arithmetically identical—every option has an expected value of 200 lives saved (equivalently, 400 lost).

Prospect theory assigns each outcome a subjective value through v(x) = x^α for gains and v(x) = −λ(−x)^α for losses, with representative parameters α = 0.88 and a loss-aversion coefficient λ = 2.25 (Kahneman & Tversky, 1979). Evaluating the sure option against the gamble in each frame (using the stated probabilities) gives:

Gain frame: V(A) = 200^0.88 = 105.90, versus V(B) = (1/3) × 600^0.88 = (1/3) × 278.47 = 92.82. Because 105.90 > 92.82, the model prefers the sure option—risk aversion.

Loss frame: V(C) = −2.25 × 400^0.88 = −438.53, versus V(D) = (2/3) × (−2.25 × 600^0.88) = −417.70. Because −417.70 > −438.53, the model prefers the gamble—risk seeking.

The same person, offered the same odds, is pushed toward the safe choice when outcomes are framed as gains and toward the risky choice when they are framed as losses. The reversal is driven entirely by the curvature of the value function—concave for gains, convex for losses—and it is the clearest demonstration that health risk decisions are governed by the description of the choice, not only by its objective structure. For a health campaign this is consequential: a message that frames a behavior in terms of what is lost by not acting can invert the risk preference a gain-framed message would produce.

Discussion

The study of health risk behavior has moved from treating risky acts as isolated moral failings to understanding them as a patterned, developmentally timed, and systematically non-rational class of action. Three findings carry most of the weight. The covariation discovered by the problem-behavior tradition showed that the behaviors share psychosocial roots, so that they are predictable from a common profile and modifiable through it rather than one habit at a time (Jessor & Jessor, 1977; Jessor, 1991). The decision research showed that the choices themselves are lawfully biased—by framing, by loss aversion, and by a perception of risk that answers to dread and unfamiliarity rather than to probability (Kahneman & Tversky, 1979; Slovic, 1987). And the developmental neuroscience explained the field's most reliable regularity, the adolescent peak, as the signature of two systems maturing out of step (Steinberg, 2008).

The convergence of these lines has a sobering implication for intervention. If risk behaviors cluster around a disposition, are chosen through biased evaluation, and peak when self-regulation is still immature, then the most intuitive remedy—supplying accurate information about the odds—targets the one deficit that is largely absent. Adolescents can state the risks; what is immature is the control system that would act on them and, on the fuzzy-trace account, the gist-based reasoning that would reject a bad bet without computing it (Reyna & Farley, 2006). The field's center of gravity has accordingly shifted toward interventions that change the environment and the framing of choices, build self-regulatory capacity, and exploit rather than fight the developmental trajectory.

Current Directions

Contemporary work has complicated the simple dual-systems story in three ways. First, the stereotype of the uniformly reckless adolescent has been challenged: a careful reading of the developmental data argues that much adolescent risk taking reflects adaptive exploration and learning rather than a deficit of control, and that the brain changes once read as pure immaturity may support the flexible, experience-driven learning the period requires (Romer et al., 2017). Second, the generality of the pattern has been tested across cultures: a study of adolescents in eleven countries found that heightened sensation seeking and immature self-regulation appear worldwide, supporting a biological core to the dual-systems account, even as the absolute levels and the specific behaviors vary with context (Steinberg et al., 2018; Duell et al., 2018). Third, the very claim that adolescence is a time of uniformly heightened risk taking has been scrutinized behavior by behavior: an overview across age groups finds that whether adolescents out-risk adults depends heavily on the type of risk and how it is measured, with laboratory risk tasks, self-reported real-world behaviors, and health statistics telling partly different stories (Willoughby et al., 2021). The common thread is a move away from a single deficit narrative toward a differentiated account in which developmental timing, culture, and the specific behavior all condition how much risk is run.

Common Misconceptions

Adolescents take risks because they underestimate the dangers.

Adolescents generally estimate risks about as accurately as adults, and on some tasks even overestimate them. The developmental peak in risk taking reflects heightened reward sensitivity and still-maturing self-regulation, not a knowledge deficit, which is why information-only campaigns have little effect (Steinberg, 2008; Reyna & Farley, 2006).

The riskiest behaviors are the ones people fear most.

Perceived risk tracks dread and unfamiliarity, not death toll. Familiar, voluntary, delayed-harm behaviors such as smoking and poor diet are among the deadliest yet provoke little fear, precisely the combination the psychometric paradigm shows perception discounts (Slovic, 1987).

Each risk behavior has its own separate cause.

The behaviors covary: a young person high on one is disproportionately high on others, because they share a psychosocial profile rather than arising independently. This is the core finding of problem behavior theory and the reason an intervention aimed at the shared roots can move several behaviors at once (Jessor & Jessor, 1977).

Adolescence is uniformly a time of heightened risk taking.

Whether adolescents take more risks than adults depends on the type of risk and how it is measured; the generalization holds for some behaviors and tasks and not others. The uniform-recklessness picture oversimplifies a more differentiated and partly adaptive pattern (Willoughby et al., 2021; Romer et al., 2017).

Glossary

Covariation.

The tendency of distinct health risk behaviors to occur together in the same individuals more than independence would predict; the empirical basis for treating risk behavior as a single construct.

Dread risk.

In the psychometric paradigm, the dimension of perceived risk capturing how catastrophic, uncontrollable, and involuntary a hazard feels; a leading driver of public concern.

Dual-systems model.

Steinberg's account that adolescent risk taking peaks because a reward-driven socioemotional system matures early while the cognitive-control system matures slowly.

Framing effect.

The reversal of risk preference produced by describing identical outcomes as gains rather than losses; a central prediction of prospect theory.

Fuzzy-trace theory.

Reyna's theory that decisions rest on gist (bottom-line meaning) rather than verbatim detail, and that mature, protective choice is more gist-based.

Gist.

In fuzzy-trace theory, the bottom-line meaning a person extracts from a situation, as opposed to its verbatim detail; the basis of mature, protective risk judgments.

Health behavior.

The broader MeSH parent category of action affecting health, under which health risk behaviors are classified as the risk-bearing subtype.

Health Belief Model.

Rosenstock's early framework in which readiness to avoid a health threat depends on perceived susceptibility and severity weighed against the perceived benefits and barriers of acting.

Health risk behavior.

A pattern of action that increases the probability of disease, injury, or premature death, such as tobacco use, heavy drinking, inactivity, or unsafe sex.

Loss aversion.

The tendency for losses to loom larger than equivalent gains, represented in prospect theory by a value function steeper for losses than for gains.

Problem behavior syndrome.

Jessor and Jessor's finding that adolescent risk behaviors covary and are predicted by a shared profile of personality, perceived environment, and behavior.

Prospect theory.

Kahneman and Tversky's descriptive theory of choice under risk, in which outcomes are valued as gains and losses relative to a reference point.

Reference point.

In prospect theory, the baseline against which outcomes are coded as gains or losses; shifting it is what produces the framing effect.

Risk perception.

The subjective judgment of a hazard's threat, shown by the psychometric paradigm to depend on dread and unfamiliarity more than on statistical probability.

Sensation seeking.

Zuckerman's trait of seeking novel, varied, and intense experience and accepting risk for its sake; a temperamental substrate of risk behavior.

Theory of planned behavior.

Ajzen's model predicting behavior from intention, which is shaped by attitudes, subjective norms, and perceived behavioral control.

Verbatim representation.

In fuzzy-trace theory, a precise, quantitative memory of a situation's details; adolescents' greater reliance on it can license risks that gist-based reasoning rejects.

Key Researchers

Icek Ajzen

(University of Massachusetts Amherst, Professor Emeritus of Psychological and Brain Sciences). Developed the theory of planned behavior, the most widely applied account of how attitudes, norms, and perceived control shape the intentions behind health-relevant action. ORCID

Richard Jessor

(University of Colorado Boulder, Distinguished Professor Emeritus; Institute of Behavioral Science). Founder of problem behavior theory, who with Shirley Jessor established that adolescent risk behaviors covary as a syndrome rooted in a common psychosocial profile. Faculty profile

Valerie F. Reyna

(Cornell University, Lois and Melvin Tukman Professor of Human Development). Originator of fuzzy-trace theory, which explains risky decision making through the contrast between gist and verbatim representations and reframed adolescent risk as a matter of reasoning mode. ORCID

Paul Slovic

(University of Oregon and Decision Research, Founder and President). Established the psychometric study of risk perception, showing that dread and unfamiliarity, not statistical probability, govern how threatening a hazard feels. ORCID

Laurence Steinberg

(Temple University, Distinguished University Professor of Psychology). Developed the dual-systems model of adolescent risk taking, locating the developmental peak in the mismatch between early-maturing sensation seeking and slow-maturing self-regulation. ORCID

Marvin Zuckerman

(University of Delaware, Professor Emeritus of Psychology; 1928–2018). Defined and measured the trait of sensation seeking and traced its common biosocial basis with risk taking, linking personality to the propensity to run risks. Wikipedia

Frequently Asked Questions

What are health risk behaviors?

They are patterns of action that raise a person's probability of disease, injury, or premature death—tobacco use, alcohol and other drug use, physical inactivity, poor diet, risky driving, and unsafe sexual behavior. In the MeSH classification they are a direct subtype of health behavior, its risk-bearing counterpart (Kann et al., 2018).

Why do risk behaviors tend to occur together?

Because they share psychosocial roots. Problem behavior theory showed that adolescent risk behaviors covary as a loosely organized syndrome predicted by a common profile of personality, perceived environment, and behavior, rather than each arising from a separate cause (Jessor & Jessor, 1977).

Why is risk taking highest in adolescence?

The dual-systems model attributes the peak to two neural systems maturing on different schedules: a reward-driven socioemotional system becomes highly reactive around puberty while the cognitive-control system supporting self-regulation matures only into the mid-twenties, opening a window of heightened risk (Steinberg, 2008).

Does giving people the facts reduce risk behavior?

Usually only modestly. Adolescents estimate risks about as accurately as adults, so the shortfall is not knowledge but the maturation of self-regulation and the mode of reasoning used; information-only interventions therefore tend to have small effects (Reyna & Farley, 2006).

What is the framing effect and why does it matter for health?

It is the reversal of risk preference caused by describing identical outcomes as gains rather than losses. Because prospect theory's value function is concave for gains and convex for losses, a health message framed around what is lost can invert the choice a gain-framed message would produce (Tversky & Kahneman, 1981).

Why do the deadliest behaviors often cause the least fear?

Because perceived risk tracks dread and unfamiliarity, not death toll. Smoking and poor diet are voluntary, familiar, and slow to harm—the combination the psychometric paradigm shows people systematically underweight (Slovic, 1987).

Is sensation seeking the same as recklessness?

Not quite. Sensation seeking is the trait of pursuing novel, varied, and intense experience and accepting risk for its sake; it is a temperamental disposition with a biosocial basis that raises the probability of risk behavior without being identical to any single reckless act (Zuckerman & Kuhlman, 2000).

Are adolescents everywhere equally prone to risk taking?

The core pattern appears across cultures: a study of eleven countries found heightened sensation seeking and immature self-regulation worldwide, though the absolute levels and the specific behaviors vary with context, and whether adolescents out-risk adults depends on the behavior measured (Steinberg et al., 2018; Willoughby et al., 2021).

References