A new employee can explain the procedure perfectly in a classroom. In the first live encounter, a customer asks an unfamiliar question and the employee hesitates. The issue may not be a lack of knowledge. They may be unsure that they can use what they know under these conditions. That judgment is close to what psychologists call self-efficacy.
Self-efficacy is a belief about one’s capability to organize and carry out actions needed for a particular task or class of tasks. It is usually more informative when the task and conditions are named: “Can I handle this conversation with support available?” is a different question from “Am I a confident person?” A belief can influence effort and persistence, but it does not create skill, time or institutional permission by itself.
A belief about doing, not a verdict on worth
Albert Bandura’s 1977 account distinguished efficacy expectations from expectations about what an action would produce. Someone might believe they can complete a training exercise but doubt that the workplace will reward the behavior. Another might strongly value the outcome while doubting they can perform the steps. Those are different barriers and invite different responses.
The belief is prospective: it concerns an upcoming action. It is not a retrospective grade or a global measure of self-esteem. A person can feel capable in one conversation and uncertain in another because stakes, available help and prior experience differ. Treating efficacy as a permanent personality label loses that specificity.
Bandura’s early research on treatment for feared tasks examined changes in perceived efficacy alongside approach behavior. These studies helped motivate the theory, but a result in one behavioral domain cannot simply be assumed in every school, clinic or workplace. The concept needs a clear task, a reasonable measure and an outcome that matters in the setting.
Where a sense of capability comes from
Successful attempts can provide information that a task is possible. A learner who completes a difficult but manageable step with feedback can revise their estimate of what they can do. Bandura gave particular weight to mastery experiences. Yet a success that depends entirely on hidden help may teach the wrong lesson if the support will disappear later.
Watching a relevant peer can offer another form of evidence. The observer sees a route through the task, especially when the model’s starting point is recognizable. This is not the same as being told to imitate a polished expert whose conditions differ. The model’s struggles, adjustments and access to resources are part of what makes the example informative.
Encouragement can matter, but credibility matters too. “You can do anything” may be less useful than feedback naming the part that worked and the next step that is within reach. Bodily arousal also enters the judgment: a racing heart before a presentation may be interpreted as inability, preparation or ordinary activation, depending on context. A single sensation does not determine actual capability.
Capability and opportunity are separate
A skilled worker may accurately doubt they can complete a task with inadequate staffing. Calling that doubt “low self-efficacy” without examining the available resources can turn a systems problem into an individual deficit. The judgment may contain useful information about the environment. Research and practice should ask whether the task is feasible under the stated conditions.
Likewise, an employee may feel able to raise a concern but lack a safe channel for doing so. Their belief about speaking differs from their belief about what will happen afterward. Bandura’s distinction between efficacy and outcome expectations is helpful here. An institution that wants participation must address consequences and authority as well as skill.
The Institute’s synthesis is that capacity development should join practice opportunities to real permission and support. Increasing confidence while removing the means to act creates a mismatch. This is an institutional design implication, not a trial result showing that one training package solves it.
How to measure the construct
A useful item asks about a defined action under specified difficulty. “How certain are you that you can explain the process when the customer challenges a step?” gives more interpretable information than “How confident are you?” alone. A set of items can cover progressively harder conditions, but the scale must match the actual domain of action.
Bandura’s guide to constructing self-efficacy scales emphasizes the domain and the level of challenge. A measurement instrument should not be borrowed merely because it uses the same word. Efficacy for studying, managing pain, responding to conflict and changing exercise habits can involve different tasks and constraints. Translation across languages also needs more than word substitution.
Reports should give the item wording, response range, timing and population. A high score immediately after a reassuring workshop may differ from a judgment made after a difficult shift. Comparing scores without context can mislead. Self-report is appropriate for a belief, but the belief should be considered alongside performance, access to resources and later behavior.
Why prediction is not proof of a mechanism
If people who report greater efficacy later perform better, it is tempting to say that the belief caused the performance. Yet previous success can raise both the belief and later performance. Skill, feedback and support may affect all three. A longitudinal association helps establish order, but it does not automatically isolate a causal effect.
Sitzmann and Yeo’s meta-analysis asked whether within-person efficacy is a product of past performance or a driver of future performance. Their question itself is a useful corrective to one-way stories. A person does not simply receive confidence and then act; action and results feed back into the next judgment. The strength of each relationship depends on the design and task.
An intervention study can test a more specific claim by changing an experience expected to affect efficacy, measuring the belief, and observing later action. Even then, the intervention may also build skill or provide support. A claim that efficacy alone mediated the change requires evidence for that particular pathway, not merely simultaneous improvement in ratings and behavior.
Calibrated confidence
High self-efficacy can support persistence when a task is difficult but achievable. It is not automatically beneficial when a person overlooks a hazard, declines help or underestimates what the task demands. A sound aim is calibration: a belief that tracks both capability and available conditions closely enough to guide action.
Vancouver and colleagues studied cases in which higher self-efficacy did not straightforwardly improve subsequent performance on a learning task. Their findings do not erase the larger theory; they show that the relationship can be conditional and that excessive assurance may affect effort allocation. A specific result in one experimental context should not be made into the opposite universal slogan.
Good feedback can support calibration. It identifies which part of a task was performed independently, which part required assistance and what changed between attempts. It can also make difficulty legible rather than treating one mistake as proof of inability. The learner gains information, not merely a positive feeling.
Designing mastery without manufacturing success
In training, early tasks can be attainable while still requiring genuine action. A realistic sequence adds complexity and varied cues as capability grows. Feedback arrives near enough to the action to be useful. A learner can retry with a changed strategy and then test the skill in a context that does not look exactly like the practice room.
If every practice attempt is scripted to succeed, the resulting belief may be fragile. The person may have learned the script rather than the judgment needed when circumstances change. Transfer-of-learning research asks whether competence travels to new settings. Self-efficacy can be one part of that transition, but it cannot replace an actual test of performance in the destination context.
Support should be visible. A trainee who succeeds with a colleague’s guidance can reasonably learn that they can act effectively with that support. It would be misleading to score the attempt as proof of independent performance. Institutions often work through teams, and capability in a team is meaningful. The description must match the conditions.
When encouragement becomes pressure
“Believe in yourself” can be offered as kindness. It can also shift responsibility onto a person who lacks time, safety or authority. In education, a learner may need access to instruction or a more appropriate challenge. In health care, a patient may face cost or transport barriers. In work, a staff member may be expected to meet a target that cannot be met safely. A motivational message does not remove those constraints.
A better question is: what evidence would allow this person to judge their capability accurately, and what conditions would make the action possible? Sometimes the answer is a supported practice opportunity. Sometimes it is a changed rule, a reliable colleague or less demand. The point is not to suppress belief but to give it a world in which it can be exercised.
Ethical use also requires consent around measurement. An efficacy survey can reveal uncertainty that a person would not want used in performance ranking. If the purpose is learning, the data should not quietly become a judgment of character. The institution should state what the information will be used for and limit access accordingly.
Relationship to autonomy and agency
Self-efficacy asks whether a person believes they can carry out an action. Autonomy concerns whether the action feels chosen and endorsed rather than controlled. Agency is broader, involving the possibility of acting in a world with other people and structures. Someone can be highly efficacious at a task they resent, or strongly value a choice they do not yet know how to carry out.
These distinctions matter for intervention design. A workshop that teaches a procedure may strengthen perceived capability without giving participants discretion over when to use it. A policy that gives choice may increase autonomy but leave people without training. Combining the concepts in one “empowerment” score can conceal which part of the situation needs to change.
The Institute’s synthesis is to ask three separate questions: Can the person do the action, do they have a meaningful reason to do it, and can the setting support it? The framework is a practical interpretation of multiple research traditions, not a new validated scale. Each question can guide a different kind of evidence and response.
A stronger study or program
Imagine a training program for handling unexpected customer questions. It defines the target behavior, observes performance before training, and asks participants about their efficacy for progressively harder situations. It gives practice with varied cases and records which supports are available. It then measures performance in a later, less scripted setting rather than relying only on a workshop survey.
A comparison condition helps show whether the program adds value beyond time, attention or ordinary learning. Repeated measures can show whether beliefs change before, after and between performances, while still requiring caution about causation. Interviews could reveal whether low ratings reflected missing skill, unclear authority or a genuinely unsafe situation.
The program should report who participated and who left. A rising average can hide the fact that people with the least support dropped out. It should specify whether the purpose was to improve task performance, reduce avoidance or make workers feel more capable. These outcomes can diverge, and none should silently stand in for the others.
Measurement can become circular if a questionnaire asks people whether they can perform a task immediately after a persuasive training session and treats their answers as proof that they can do it. A better evaluation separates belief from demonstrated performance. It also defines what performance means: a first attempt, sustained accuracy, adaptation to a new case, or appropriate help-seeking may tell different stories. Confidence measured after feedback is informative, but it is one part of the learning process, not a substitute for observing it.
Self-efficacy may also be high for the wrong reason. If a task seems easier than it is, a person can rate their capability highly before encountering its real constraints. Conversely, an expert may rate themselves cautiously because they understand the difficult cases. This does not make the construct useless; it means ratings need a specified task, difficulty and context. The Institute’s practical question is whether belief, skill and opportunity are becoming better aligned, rather than whether every score rises.
What the evidence does not show
Self-efficacy research does not show that confidence alone produces competence, that a high score is always desirable, or that people can overcome structural barriers through belief. It does not make one general scale suitable for every task. A correlation between belief and performance does not identify which caused the other.
Nor does an encouraging message count as a mastery experience simply because it feels good. The strongest practical claims require specific tasks, observable opportunities, meaningful feedback and measurement at the point where the skill must be used. A belief can be a resource, but its value depends on what it is a belief about.
Returning to the live encounter
The employee pauses, asks a colleague one question and returns to the customer with a clearer answer. The encounter gives information: part of the procedure was available, part required support, and the unusual question was manageable with that support. That is more useful than either a verdict of failure or a command to be confident.
Self-efficacy becomes useful when it remains specific and revisable. People learn what they can do through action, feedback and the conditions around them. A responsible institution gives them room to test that belief honestly, and changes the conditions when the task exceeds the resources it has provided.
Research references
- Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. View source
- Bandura, A., Adams, N. E., & Beyer, J. (1977). Cognitive processes mediating behavioral change. Journal of Personality and Social Psychology, 35(3), 125–139. View source
- Bandura, A. (2006). Guide for constructing self-efficacy scales. In F. Pajares & T. Urdan (Eds.), Self-Efficacy Beliefs of Adolescents (pp. 307–337). Information Age Publishing. View source
- Sitzmann, T., & Yeo, G. (2013). A meta-analytic investigation of the within-person self-efficacy domain: Is self-efficacy a product of past performance or a driver of future performance? Personnel Psychology, 66(3), 531–568. View source
- Vancouver, J. B., Thompson, C. M., & Williams, A. A. (2001). The changing signs in the relationships among self-efficacy, personal goals, and performance. Journal of Applied Psychology, 86(4), 605–620. View source
- Schunk, D. H. (1991). Self-efficacy and academic motivation. Educational Psychologist, 26(3–4), 207–231. View source