HBM 09: Social Cognitive Theory

A mid-sized city installs new protected bike lanes along a major corridor. In the first month, usage is low. But then something shifts. A few confident cyclists begin using the lanes visibly during peak commute hours. Neighbors see them. Coworkers mention trying it. A local cycling group starts weekly “confidence rides” for beginners along the new route. Within six months, daily cycling counts on the corridor have tripled — not because the infrastructure changed, but because people watched others succeed, built confidence through guided experience, and began to see themselves as capable cyclists.

This pattern — learning through observation, building confidence through experience, and changing behavior as both person and environment shift together — is the central mechanism of Social Cognitive Theory (SCT). Developed by Albert Bandura over several decades beginning in the 1960s, SCT is one of the most influential behavioral theories in the social sciences. It argues that human behavior is not simply a response to environmental stimuli or an expression of internal traits. Instead, behavior emerges from a continuous, dynamic interaction among three forces: the person (cognitive factors, beliefs, expectations), the environment (social norms, physical structures, incentives), and behavior itself.

Why Learning and Confidence Matter

Most behavior models in this series focus on decision-making — how people weigh costs and benefits, perceive threats, form intentions, or progress through stages of readiness. These models treat the person as a deliberator.

Social Cognitive Theory adds a different dimension: the person as a learner. People do not only calculate and decide. They also observe, imitate, practice, adjust, and build (or lose) confidence over time. This learning perspective is essential for understanding behaviors that require skill development, confidence-building, or sustained engagement with a social environment.

In transportation and planning, this matters because many target behaviors involve:

• Skills that must be learned: Navigating a transit system for the first time, cycling in traffic, driving defensively, preparing an evacuation kit.
• Confidence that must be built: Believing one can cycle safely, manage a new commute, drive without a phone, or lead one’s family through an evacuation.
• Social environments that shape behavior: What peers do, what neighbors model, what media portrays, and what is normalized in a community.

The Model in Plain Language

SCT rests on a foundational principle that Bandura called reciprocal determinism: behavior, personal factors (cognitive, affective, biological), and environmental factors all influence each other continuously. None of these three elements is primary. Instead, they form a dynamic triangle in which change in any one element ripples through the others.

A person sees a colleague commute by bike (environment → person). They form the expectation that cycling is feasible and desirable (person). They try cycling to work (person → behavior). The act of cycling exposes them to new social contacts and routes (behavior → environment). Their confidence grows from success (behavior → person). They become visible models for other potential cyclists (behavior → environment for others).

This continuous feedback loop distinguishes SCT from more linear models like the Theory of Planned Behavior or the Health Belief Model, which treat environment primarily as background context rather than as an active, evolving force.

Core Constructs

SCT includes several interconnected constructs, organized around learning, motivation, and the person-behavior-environment triad.

Self-Efficacy

Self-efficacy is SCT’s most influential construct and Bandura’s most enduring contribution to behavioral science. It is not general confidence or self-esteem. It is the belief in one’s ability to successfully perform a specific behavior in a specific context.

A person might have high self-efficacy for driving in familiar conditions but low self-efficacy for cycling in traffic. A homeowner might feel confident about locking doors but not about assembling an emergency preparedness kit. Self-efficacy is behavior-specific, context-sensitive, and changeable.

Bandura identified four sources of self-efficacy, ranked from most to least powerful:

Observational Learning (Modeling)

One of Bandura’s breakthrough findings was that people can acquire new behaviors without direct experience, simply by observing others. This is observational learning, or modeling.

Observational learning involves four sub-processes:

Outcome Expectations

Outcome expectations are beliefs about what will happen if a person performs a behavior. SCT distinguishes three types:

Outcome expectations differ from self-efficacy. A person might believe that cycling to work would be healthy and environmentally beneficial (positive outcome expectations) but not believe they could actually do it safely (low self-efficacy). SCT predicts that both are necessary: high outcome expectations without self-efficacy produce wishful thinking, while high self-efficacy without positive outcome expectations produces capability without motivation.

Reinforcement

SCT includes three types of reinforcement that maintain or extinguish behavior:

• Direct reinforcement: The person experiences positive or negative consequences of their own behavior. A cyclist who arrives at work feeling energized receives direct positive reinforcement. A cyclist who gets soaked in rain receives direct negative reinforcement.
• Vicarious reinforcement: The person observes others being rewarded or punished for the behavior. Seeing a colleague praised for cycling, or seeing a speeder pulled over, provides vicarious reinforcement.
• Self-reinforcement: The person evaluates their own behavior against internal standards and experiences pride or dissatisfaction. A driver who resists checking their phone feels self-satisfaction.

Reciprocal Determinism

The overarching framework of SCT is reciprocal determinism: the continuous mutual influence of person, behavior, and environment.

Causal Logic

SCT’s causal logic is fundamentally cyclical rather than linear.

The key prediction: behavior change is most likely when people observe credible models succeed, form positive outcome expectations, attempt the behavior with adequate support, succeed (or manage failure constructively), and receive reinforcement — creating a positive feedback loop that builds self-efficacy over time.

The corollary: behavior change is inhibited when models are absent, outcome expectations are negative, early attempts fail without support, and self-efficacy erodes.

Data Needed

SCT studies require measurement of multiple constructs across the person-behavior-environment triad.

SELF-EFFICACY SCALE
"How confident are you that you could..."
...cycle to work on a regular day? [0-100]
...cycle to work when it is raining? [0-100]
...cycle on roads with moderate traffic? [0-100]
...navigate an unfamiliar route by bike? [0-100]
...handle a minor mechanical issue during a ride? [0-100]
OUTCOME EXPECTATIONS
Physical: “Cycling to work would improve my physical fitness.” [1-5]
Social: “My coworkers would view cycling to work positively.” [1-5]
Self-evaluative: “I would feel proud of myself for cycling to work.” [1-5]
OBSERVATIONAL LEARNING
“How many people in your neighborhood do you see cycling regularly?”
[None] 0 — 1-2 — 3-5 — 6-10 — [More than 10]
“Do you personally know someone similar to you who cycles to work?”
[Yes / No]
“Have you seen positive media coverage of cycling in your city?”
[Never] 1 — 2 — 3 — 4 — 5 [Very Often]
REINFORCEMENT HISTORY
“In the past, when you have tried cycling, the experience was generally:”
[Very Negative] 1 — 2 — 3 — 4 — 5 [Very Positive]
ENVIRONMENT
“There are safe cycling routes between my home and work.” [1-5]
“My workplace provides bicycle parking and shower facilities.” [1-5]
“My community is supportive of cycling as transportation.” [1-5]

Additional data sources include:

• Longitudinal surveys tracking self-efficacy changes over time, especially before and after interventions (community rides, training programs, infrastructure improvements)
• Observation data on cycling volumes, modeling visibility, and social interaction patterns
• GPS and sensor data linking self-reported efficacy to actual riding behavior, route choices, and trip frequency
• Qualitative interviews exploring how people learned about cycling, who influenced them, and how their confidence developed

Methods

Cross-lagged panel designs are particularly appropriate for testing reciprocal determinism. By measuring person, behavior, and environment at two or more time points, researchers can estimate the cross-lagged effects (e.g., does Time 1 self-efficacy predict Time 2 behavior, and does Time 1 behavior predict Time 2 self-efficacy?).

Hierarchical linear modeling (HLM) is useful when data has a nested structure (individuals within neighborhoods, cyclists within community groups), allowing analysis of how environment-level variables moderate person-level effects.

Transportation Example: Cycling Confidence Through Community Rides

A city transportation department partners with a cycling advocacy organization to launch a “Ride With Confidence” program aimed at increasing cycling commuting in neighborhoods with new bike infrastructure that remains underused.

Program Design (SCT-Informed)

The program is designed around SCT principles:

Hypothetical Evaluation Results

A pre-post study with a matched comparison group surveys 150 program participants and 150 non-participants before and 3 months after the program.

Mediation analysis shows that the program’s effect on cycling frequency is largely mediated by self-efficacy change:

Path a: Program → Self-Efficacy      B = 28.7,  p < .001
Path b: Self-Efficacy → Cycling Days B = 0.09,  p < .001
Direct: Program → Cycling Days       B = 0.42,  p = .183  (n.s.)
Indirect (mediated by self-efficacy): B = 2.58,  95% CI [1.73, 3.61]
Conclusion: The program increased cycling primarily by building
self-efficacy, not through other mechanisms.

Safe Driving Through Peer Modeling

SCT also applies to safety behavior. Consider a fleet safety program for a transportation company where some drivers consistently exceed speed limits.

SCT in Context: Comparison with Other Models

Strengths

Integrates cognition, behavior, and environment. SCT’s triadic reciprocal determinism is one of the most comprehensive frameworks for understanding behavior. It avoids the trap of treating behavior as purely individual or purely environmental.

Strong on self-efficacy. Self-efficacy is one of the most robust predictors in behavioral science. SCT provides a clear theory of where self-efficacy comes from and how to build it, which is directly actionable for program design.

Powerful modeling framework. Observational learning explains how behaviors spread through social networks, how peer programs work, and why visible role models are important in communities transitioning to new transportation modes.

Widely applicable. SCT has been applied to health behavior, education, organizational behavior, media effects, athletic performance, and — increasingly — transportation and sustainability. Its constructs are flexible enough to accommodate diverse behavioral domains.

Dynamic and cyclical. Unlike static models, SCT captures the feedback loops between behavior, cognition, and environment. This makes it suitable for understanding how behavior develops and stabilizes over time.

Limitations

Complex to operationalize fully. Testing the complete SCT model — with all constructs, all sources of efficacy, all types of reinforcement, and reciprocal causation — is methodologically demanding. Most studies test subsets of the theory. Full reciprocal determinism requires longitudinal data and complex statistical models.

Less clear on structural constraints. While SCT includes “environment,” it emphasizes the social and informational environment more than structural factors like income, infrastructure quality, or policy. A person may have high self-efficacy for cycling but no safe route — and SCT does not clearly distinguish between environmental barriers that are perceptual and those that are physical.

Self-efficacy can be hard to change. While SCT provides a clear theory of self-efficacy sources, actually building self-efficacy — especially for populations with histories of failure, discrimination, or limited resources — is challenging in practice. Mastery experiences require opportunity, support, and environments that permit trial and error.

Risk of overemphasizing individual agency. SCT’s focus on personal agency through self-efficacy can inadvertently understate the role of systemic inequities, power structures, and institutional barriers. Not all populations have equal access to models, mastery opportunities, or supportive environments.

Measurement challenges. Self-efficacy scales must be behavior-specific and context-specific, which means new scales must often be developed for each application. Observational learning is difficult to measure retrospectively and hard to isolate from other social influences.

Best Use Case

It is particularly well-suited for community-based programs (cycling confidence, safe driving, disaster preparedness), peer-led interventions, and settings where behavior change requires skill development rather than just attitude change.

Key Takeaway

Key References