September 8, 2026 / Studio Practice
Kohlberg’s theory of moral development: stages, logic, and limits
Kohlberg’s theory of moral development
A school counselor asks a 10-year-old whether it is worse to steal a loaf of bread to feed a hungry sibling or to let the sibling go hungry. A few years later, that same student, now 15, argues instead about whether a law that criminalizes theft is legitimate in the first place. The shift is not simply a matter of growing vocabulary. It reflects a change in how the child reasons about right and wrong, and that is exactly the territory mapped by Kohlberg’s theory of moral development, first published by the American psychologist Lawrence Kohlberg in 1958 and refined across the following three decades.
Kohlberg’s theory of moral development is one of the most widely taught frameworks in developmental psychology, ethics education, and teacher training. It describes moral thinking as a sequence of six identifiable stages, grouped into three broader levels, that children, adolescents, and adults move through as their reasoning matures. The model is not a description of whether someone behaves well or badly. It is a description of the reasons a person gives for a moral choice, and how those reasons grow more complex with age and experience.
This article walks through the structure of the theory, the research evidence behind it, how educators use it, and the well-known critiques from Carol Gilligan, neuroscience, and cross-cultural research. It is intended for teachers, curriculum designers, parents, and design or product professionals who want a clear, working understanding rather than a textbook summary.
What Kohlberg’s theory actually claims
Kohlberg’s central claim is that moral reasoning develops in a fixed sequence of stages, and that each stage represents a more adequate way of resolving moral dilemmas than the one before it. A stage is not a behavior. Two people who both refuse to cheat on a test may be doing so for very different reasons, and only one of them is reasoning at a higher stage. The unit of analysis is the reasoning, not the act.
Three structural assumptions sit underneath the model:
- Stages are invariant in order. Everyone who moves to a higher stage passes through the same sequence. A person cannot reach stage 4 without first passing through stage 3.
- Stages are hierarchical. Each later stage is more cognitively complex and more adequate at balancing competing moral claims than the one before it.
- Stages form a structured whole. Thinking at a higher stage reorganizes how a person sees earlier problems, rather than simply adding new ideas on top of old ones.
Kohlberg drew this structure from the cognitive-developmental tradition of Jean Piaget, who had earlier proposed that children’s logical thinking moves through predictable qualitative shifts. Kohlberg extended the same logic to moral judgment. A child is not simply told what is right; they actively construct more adequate reasons as they encounter social dilemmas they cannot resolve with their current reasoning.
The three levels and six stages of moral reasoning
The model is usually presented as three levels, each containing two stages. The labels below use Kohlberg’s own terminology, simplified for clarity.
Level 1: preconventional morality
At the preconventional level, moral reasoning is anchored in the self. Right and wrong are defined by the physical or hedonic consequences of an action, not by shared social agreements.
- Stage 1: obedience and punishment. The good action is the one that avoids punishment. A child stops running in the hallway because a teacher will scold them, not because the rule has any deeper justification.
- Stage 2: individualism and exchange. Right is defined by what serves one’s own interests while recognizing that others have interests too. Fairness is treated as an equal exchange rather than a shared standard.
Level 2: conventional morality
At the conventional level, moral reasoning is anchored in social relationships and shared norms. The self is understood as a member of groups whose expectations matter in their own right.
- Stage 3: good interpersonal relationships. Good behavior is what helps others and is approved by them. The person wants to be seen as a good son, friend, or colleague. Intent matters as much as consequence.
- Stage 4: maintaining social order. Right is defined by duty to authority and the maintenance of a functioning social system. Laws are upheld because they keep society predictable, not because they are questioned.
Level 3: postconventional morality
At the postconventional level, moral reasoning becomes self-authored. Social rules are evaluated against deeper principles that can justify or override them.
- Stage 5: social contract and individual rights. Laws are seen as social agreements that exist to protect rights and welfare. They can be changed when they fail to do so. A person at this stage can respect law while also recognizing that an unjust law should be revised.
- Stage 6: universal ethical principles. Right is defined by self-chosen ethical principles that any rational person could endorse, such as justice, equality, and respect for persons as ends in themselves. These principles can override laws that conflict with them.
Empirical work suggests that stage 5 is the ceiling for most adults in Western samples, and that stage 6 is rare even in educated populations. That observation has shaped both the practical use and the critique of the model.
How researchers actually score a stage
Kohlberg’s original method was the Moral Judgment Interview, a roughly 45-minute semi-structured conversation in which the participant is presented with moral dilemmas and asked what the main actor should do and why. The interview is scored by trained coders who match the participant’s reasoning to a stage using a detailed scoring manual.
The most famous stimulus is the Heinz dilemma: a man named Heinz cannot afford the drug that will save his dying wife, the pharmacist refuses to lower the price, and Heinz breaks into the pharmacy to steal the drug. Should he have done so, and why? The point of the dilemma is not the verdict but the reasoning. A stage 2 response treats the dilemma as a calculation of self-interest. A stage 4 response treats it as a question of law. A stage 5 response weighs law against human welfare. A stage 6 response appeals to the principle that a human life outweighs property rights.
Because the interview is labor-intensive, researchers often use the Defining Issues Test, a multiple-choice measure that asks respondents to rank the importance of statements reflecting different stages. It is quicker to administer, though some researchers argue it underestimates stage 5 and 6 reasoning compared with the interview.
What the research evidence shows
Three findings have held up reasonably well across decades of replication and are worth understanding before applying the model:
1. People do reason in qualitatively different ways
Studies across cultures consistently find that not all moral arguments are the same. The same participant can give stage 2 reasons about a family dispute and stage 4 reasons about a civic issue, suggesting that reasoning is sensitive to context, not a single fixed trait.
2. Average stage scores rise with age and education
Longitudinal work, including Colby and Kohlberg’s own 30-year follow-up of the original Chicago sample, shows a gradual upward drift in moral judgment stage across adolescence and adulthood. Most of the change happens in adolescence and early adulthood, then plateaus.
3. Stage use, not stage possession, predicts behavior
Later research by James Rest and colleagues found that people are capable of reasoning at several stages at once and that which stage they rely on in a given situation better predicts their behavior than their highest stage. This is the basis of the Four Component Model of moral behavior, which decomposes moral action into moral sensitivity, moral judgment, moral motivation, and moral character.
Comparing the three levels at a glance
The table below summarizes the orientation, central question, and typical justification style of each level.
| Level | Orientation | Central question | Typical justification |
|---|---|---|---|
| Preconventional | Self | What happens to me? | Punishment, reward, exchange |
| Conventional | Social role and group | What do others expect? | Approval, relationships, duty to law |
| Postconventional | Principled | What is right in principle? | Social contract, universal principles |
A practical way to read this table: when a student justifies a rule by saying “we’ll get in trouble if we don’t,” the response is anchored in stage 1. When the same student says “it’s the rule, and without rules nothing works,” the response is anchored in stage 4. The shift in justification, not the change in behavior, is the development.
How classroom teachers and curriculum designers use the model
Many teacher-education programs still use Kohlberg’s theory of moral development as a scaffold for thinking about moral education. Three concrete applications are common.
1. Selecting appropriate moral dilemmas for discussion
Dilemmas are chosen to be just above the dominant stage in a class, so that students encounter reasons that stretch them. Discussing a dilemma they can already resolve does not produce growth. A useful rule of thumb: if every student gives the same stage 2 answer, the prompt is too easy.
2. Sequencing the “just community” approach
In the Just Community schools Kohlberg helped develop in the 1970s, adolescents ran their own governance and discipline systems, which exposed them to stage 5 reasoning in practice rather than only in interview. The model here is not lecture but participation in democratic decision-making.
3. Curriculum design for ethics in professional education
Medical, legal, and business schools use stage-based reasoning tasks to help students articulate the principles behind professional codes, not just memorize them. A learner at stage 3 may follow the rule; a learner at stage 5 understands why the rule exists and can articulate when it should change.
The strongest critiques of the model
Kohlberg’s theory of moral development is also one of the most heavily criticized frameworks in psychology. The critiques are useful to understand because they often improve how the model is used in practice.
The gender critique from Carol Gilligan
Carol Gilligan argued in 1982 that the model encoded a male perspective on morality and undervalued an ethics of care that focuses on relationships and responsibility to specific others rather than abstract rights. She did not claim women cannot reason at higher stages, but that the scoring manual was blind to a different moral vocabulary. Modern revisions often present the care and justice perspectives as parallel rather than competing.
The cultural critique
Cross-cultural research by scholars including Richard Shweder has shown that the stage sequence may not be universal. Participants in some non-Western samples are more likely to ground moral reasoning in community, divinity, or duty to social hierarchy in ways that the standard scoring manual registers as lower stage. Whether this is a real developmental lag or a category error in the instrument is still debated.
The moral behavior critique
Knowing the right thing to do is not the same as doing it. Critics point out that stage 6 reasoning is rare even among highly ethical professionals, and that character, habit, emotion, and situation often matter more than reasoning in predicting behavior. The Four Component Model mentioned earlier is a partial response to this critique.
The cognitive load critique
Some educational neuroscientists argue that the model underestimates the role of automatic emotional processing in moral judgment. Recent dual-process accounts of morality treat System 1 intuitions and System 2 reasoning as jointly responsible, which sits awkwardly with a strictly stage-based account.
What the model is genuinely good for, and what it is not
A working editor or designer usually does not need the full theoretical apparatus. A narrower summary is more useful in practice.
| Use case | Is the model helpful? | Reason |
|---|---|---|
| Designing ethics discussions for adolescents | Yes | Provides a shared vocabulary for levels of justification |
| Diagnosing why a learner resists feedback | Sometimes | Useful if the resistance is principled, not just defensive |
| Predicting whether a person will act ethically under pressure | Weakly | Reasoning is only one component of moral behavior |
| Comparing moral traditions across cultures | Cautiously | The instrument is biased toward Western liberal assumptions |
| Building a product ethics checklist | Indirectly | Use it to ask “what stage of reasoning does this feature assume of the user?” |
The model is a thinking tool, not a measurement instrument for individuals. Used loosely, it sharpens a conversation. Used rigidly, it produces false precision.
How the model connects to design and product work
For a studio audience, the relevant question is not “what stage is my user at?” but “what assumptions about moral reasoning are baked into this product or service?” A few examples make this concrete.
- Terms-of-service design. A flow that asks the user to “agree” before showing a one-sentence summary assumes stage 1 or 2 reasoning. A flow that summarizes obligations, then offers a right to withdraw consent, treats the user as capable of stage 4 or 5 reasoning.
- Moderation policy framing. A rule framed as “do not post X or your account will be suspended” anchors reasoning in stage 1. A rule framed as “this community excludes X because it harms Y” invites stage 3 or 4 reasoning. The first is faster; the second is more durable.
- Accessibility and consent flows. When a design assumes a user will comply because “everyone does,” it is operating at a conventional level. When it explains why a default protects the user and offers a meaningful alternative, it operates closer to a principled level. The connection to inclusive design is that both rely on moving from rule-following to reasoning about the people behind the rule.
Used in this way, the framework becomes a way to ask whether a product is asking more from users than the design itself rewards. If a privacy policy requires stage 5 reasoning to interpret, but the UI only supports stage 1 compliance, the design is mismatched.
A short worked example: the Heinz dilemma in three stages
To make the levels concrete, here is how a single participant might respond to the Heinz dilemma at three different stages.
| Stage | What Heinz should do | Reasoning given |
|---|---|---|
| 2 (preconventional) | He should not steal | He’ll go to jail, and jail is bad. It is not worth the risk to him personally. |
| 4 (conventional) | He should not steal | Stealing breaks the law. If everyone broke laws that were inconvenient, the system collapses. He should respect the rule even if the outcome is sad. |
| 5 (postconventional) | He should steal, and the law should be reconsidered | Laws exist to protect welfare. When a law fails to protect a fundamental interest like life, the law itself is the problem, not Heinz. Society should change the rule. |
Notice that stages 4 and 5 reach different verdicts. That is the point. The theory is not asking which answer is correct, it is asking which reasoning a person can sustain when challenged.
Practical limits when applying the model in education
Educators who draw on Kohlberg’s theory of moral development should be aware of three operational limits, not just theoretical ones.
- Time and training. The interview method requires roughly an hour of coding per participant. Few schools have that capacity, which is why shorter instruments dominate despite known compromises.
- Resistance as data. A student who refuses to engage with a dilemma is not necessarily at a low stage. They may be at a high stage and rejecting the framing. Treat silence as ambiguous.
- Transfer is not automatic. A student who reasons well in a classroom dilemma may not transfer that reasoning to a real situation with social pressure, time pressure, or emotional stakes. Practice, not just exposure, is required.
How the theory fits alongside other frameworks
Kohlberg’s theory is one of several influential accounts of moral development. The table below positions it against three neighbors, in plain terms.
| Framework | Central focus | Main contribution | Where it differs from Kohlberg |
|---|---|---|---|
| Kohlberg | Reasoning about justice | A stage sequence for moral judgment | Reference point for the others |
| Gilligan | Ethics of care | Highlights relational responsibility | Adds a moral vocabulary Kohlberg underweighted |
| Rest (Four Component Model) | Moral behavior | Decomposes moral action into four skills | Replaces stages with component skills |
| Haidt’s moral foundations | Moral intuitions | Multiple innate foundations, not one ladder | Rejects stage hierarchy, emphasizes intuition |
For practical purposes, most modern ethics curricula combine insights from several frameworks rather than committing to one. A unit might use Kohlberg’s stages to structure discussion, Gilligan’s care perspective to widen the language, and Rest’s components to design practice that goes beyond reasoning.
How to use this article in a working session
If you are a teacher, designer, or studio lead reading this for application, three short reminders are worth keeping.
- Reasoning, not behavior, is the unit of analysis. Two people can reach the same verdict for very different reasons, and the difference is the part that matters.
- Stages are a vocabulary, not a verdict. Calling a person “stage 2” is shorthand for the kind of reasons they currently use. It is not a permanent label and it is not a moral judgment about them.
- Higher stages are not automatically better rules. A stage 4 rule that fits a community may be more humane in practice than a stage 5 principle that ignores context.
If you are applying this to a product or policy decision, the simplest test is to ask what kind of reasoning the design assumes. If the answer is “the user will just comply,” the design is operating at stage 1. If the answer is “the user can evaluate and contest the rule,” the design is operating at stage 5. Most production-ready systems aim for stage 4 with a stage 5 escape hatch for the cases the rules cannot anticipate. For more on framing decisions at the level of a product or a service, the studio’s notes on planning a site that actually performs cover the same logic from a delivery angle.
Frequently asked questions
What are the three levels in Kohlberg’s theory of moral development?
Preconventional morality, in which right and wrong are defined by personal consequences; conventional morality, in which right and wrong are defined by social roles, relationships, and laws; and postconventional morality, in which right and wrong are defined by self-chosen ethical principles that can override unjust laws.
What is the difference between Kohlberg’s stage 4 and stage 5?
Stage 4 treats laws and authority as binding because they keep society functional. Stage 5 treats laws as social agreements that exist to protect rights and welfare, and that should be changed when they fail to do so. The difference is the willingness to challenge the law in the name of the principle behind it.
How is Kohlberg’s theory used in schools today?
It is most often used to structure moral discussion, to design dilemma-based lessons, and to frame whole-school democratic programs like the Just Community approach. It is less often used as a formal assessment, because the interview is too time-consuming for routine classroom use.
Is Kohlberg’s theory of moral development still accepted?
It is still widely taught as a foundational framework, but it is treated as one tool among several. Researchers now combine it with care-based, intuitionist, and component-based models to address its known limits, including gender bias, cross-cultural variation, and weak prediction of behavior.
At what age do children reach each level?
Most children show preconventional reasoning through middle childhood, begin to use conventional reasoning in early adolescence, and only a minority of adults reach the postconventional level. Stage 6 reasoning is rare in any age group and is often treated as an ideal rather than a typical endpoint.
Why was Carol Gilligan critical of Kohlberg’s model?
Gilligan argued that the original scoring manual underweighted an ethics of care grounded in relationships and responsibility to specific others, and that this bias made women’s moral reasoning appear less mature than men’s. Her critique is now often incorporated by treating care and justice as parallel moral languages rather than competing stages.
What is the Heinz dilemma, and why is it used?
The Heinz dilemma describes a man who must choose between letting his wife die or stealing an unaffordable drug. It is used because it has no obviously correct answer and forces participants to articulate the principles behind their choice, which is what a stage-based scoring manual needs to identify their level of moral reasoning.
Can moral reasoning be measured by a simple questionnaire?
Short measures like the Defining Issues Test approximate stage use in research settings, but they are less sensitive to postconventional reasoning than the original interview. For high-stakes claims about an individual, the interview remains the standard method, with the understanding that even it is a snapshot of reasoning, not behavior.
What is the relationship between moral reasoning and moral behavior?
Reasoning is one component of moral behavior, not the whole of it. The Four Component Model breaks behavior into moral sensitivity, moral judgment, moral motivation, and moral character. A person can reason well at stage 5 and still act badly when any of the other three components fail.
How does this theory apply outside psychology?
In ethics education, the model is a scaffold for classroom discussion. In professional training, it is a way to make the principles behind a code of conduct explicit. In product and service design, it is a way to ask what kind of moral reasoning a flow assumes from the user, and to test whether the design and the policy are operating at the same level.