Social Influence Theory and Its Impact on Behavior: Insights by Kelman

Why do people change what they do, say, or believe when others are around? A teenager picks up the slang their friends use. An employee follows an office rule they privately disagree with. A young professional adopts the work ethic of a mentor they deeply admire. These are all forms of social influence, yet they are not the same thing. In 1958, Harvard social psychologist Herbert C. Kelman gave us a clear framework to tell them apart, and his ideas remain one of the most useful tools for understanding how attitudes and behaviour actually shift in social settings.

What social influence theory actually explains
The three processes of influence
Compliance
Identification
Internalization
What determines which process takes hold
The importance of the anticipated effect
The power of the influencing agent
The strength of the induced response
Why looking only at compliance falls short
Where the theory is useful and where it has limits

Social influence theory was first formulated by Herbert Kelman in his 1958 paper published in the Journal of Conflict Resolution. Kelman defined social influence as a change in behaviour brought about within a social setting by another person or a group of people. The question he set out to answer was deceptively simple: when someone changes their attitude or behaviour under social pressure, what is really happening underneath?

The theory was shaped by the social and political climate of its time. Kelman was working during an era marked by civil rights movements and anti-war protests, when it became obvious that people could conform to social norms whether or not they believed in their legitimacy. Earlier research could predict when people would conform, but it could not explain why some changes were shallow and temporary while others ran deep and lasted a lifetime.

Kelman’s central insight was that attitude and behaviour change does not happen in one uniform way. Instead, it occurs through three qualitatively different processes, each driven by a different motivation and each producing a different depth of change. He called these processes compliance, identification, and internalization. The purpose of distinguishing them was practical: to separate public conformity from private acceptance, that is, to tell apart what people do in front of others from what they genuinely come to believe.

The three processes of influence

The heart of Kelman’s theory is the distinction between these three modes of accepting influence. They are not ranked as better or worse, but they differ in how deep the change goes and how long it lasts.

Compliance

Compliance is the most surface-level form of influence. According to Kelman, it occurs when a person accepts influence and adopts a behaviour in order to gain rewards or approval and to avoid punishment or disapproval. The key point is that the person does not necessarily believe in what they are doing. They go along with it because of what they expect to get or avoid.

This is why compliance produces a change in behaviour but not always a change in attitude. A worker who follows a dress code only while the supervisor is watching is complying. A student who finishes an assignment only to avoid a penalty is complying. The change here is real but fragile. As this kind of influence lasts only as long as the person sees compliance as beneficial, the behaviour tends to disappear the moment the reward or threat is removed. It is influence held in place by external pressure, not inner conviction.

Identification

Identification goes a step deeper and lives in the realm of relationships. Here a person changes their attitude or behaviour not for a specific reward, but because they want to establish or maintain a satisfying relationship with a person or group they admire. The behaviour matters because of the bond it represents.

This is the mechanism that makes celebrity endorsements work. When advertisements rely on famous personalities to sell products, they are betting on identification: consumers adopt a preference because they want to feel connected to someone they like or respect. The same process explains why a junior colleague might copy the habits of a senior they look up to, or why someone takes on the opinions and even the vocabulary of a group they want to belong to. The crucial feature is that the person genuinely believes in the behaviour, but only as long as the relationship holds meaning. If the relationship fades, the borrowed behaviour usually fades with it.

Internalization

Internalization is the deepest and most durable form of influence. It occurs when a person accepts influence because the new attitude or behaviour is intrinsically rewarding and fits with their own value system. The change is no longer about pleasing someone or maintaining a relationship. The person adopts the behaviour because they have decided, for themselves, that it is right, useful, or true.

At this level, influence becomes woven into a person’s sense of who they are. Because the change coincides with the individual’s values or fulfils a need they have, this type of influence tends to be long-lasting. A person who recycles because they were once fined for littering is complying. A person who recycles because their admired team leader does so is identifying. But a person who recycles because they have come to believe protecting the environment is genuinely important has internalized the behaviour, and they will keep doing it whether anyone is watching or not.

What determines which process takes hold

Kelman did not stop at naming the three processes. He argued that each one is produced by a distinct set of conditions. According to his framework, every process can be understood as a function of three determinants: the relative importance of the anticipated effect, the relative power of the influencing agent, and the prepotency of the induced response. Although all three processes share these same three ingredients, the ingredients work differently in each case.

The importance of the anticipated effect

This determinant concerns what the person expects to gain or lose from accepting the influence. In compliance, the anticipated effect is external, things like rewards, approval, or avoiding punishment. The behaviour is essentially instrumental. As one analysis of Kelman’s work puts it, in instrumental compliance the level of effort is likely to be only the minimum needed to gain the reward or avoid the punishment. In identification, the anticipated effect is a satisfying relationship. In internalization, the anticipated effect is the intrinsic value of the behaviour itself.

The power of the influencing agent

The source of influence holds different kinds of power, and the kind of power shapes which process unfolds. Kelman drew on the idea that influence depends on the basis of the agent’s power. Compliance is rooted in the agent’s control over rewards and punishments, what later researchers described as reward power and coercive power. Identification depends on the agent’s attractiveness or likeability, while internalization depends on the agent’s credibility. Later work in this tradition confirms that attractiveness tends to foster identification, while credibility drives internalization. A boss who can withhold a bonus influences through compliance; a charismatic figure influences through identification; a trusted expert influences through internalization.

The strength of the induced response

The third determinant is how strongly the response is triggered and how much it depends on the situation. In compliance, the behaviour appears mainly when the agent is present and watching, which is why it is so easily switched off. In identification, the response surfaces when the relationship is relevant or salient. In internalization, the response is prepotent in the sense that it can be activated whenever the underlying value is relevant, regardless of who is around. This is precisely why internalized change is so stable: it does not need an audience to keep going.

Why looking only at compliance falls short

Kelman’s framework is powerful, but it is most valuable when all three processes are kept in view. A common mistake, both in research and in everyday thinking, is to treat social influence as nothing more than compliance, as if getting people to do something is the same as getting them to believe in it. Kelman himself argued the opposite. The whole point of his theory was to show that surface-level conformity and genuine acceptance are different phenomena with different consequences.

Consider what happens when we focus only on compliance. Studies of how people adopt new technology, for instance, were criticised for assuming influence works mainly through social normative compliance, thereby overlooking the identification and internalization processes. This matters because compliance-based change is brittle. The effects of compliance tend to weaken over time, whereas the effects of identification and internalization persist far longer. A policy that secures only compliance buys short-term behaviour but not lasting commitment. The moment enforcement relaxes, the behaviour collapses.

This has real implications for anyone trying to encourage durable change, whether a teacher building good study habits, an organisation shaping workplace culture, or a public health campaign trying to change daily routines. If the goal is behaviour that survives without constant monitoring, compliance alone is not enough. Influence needs to move toward identification and, ideally, internalization, where people carry the behaviour forward because it has become part of who they are.

Where the theory is useful and where it has limits

Kelman’s model has been applied widely, from understanding prejudice and conflict resolution to studying consumer behaviour, organisational change, and how people behave on social media. On digital platforms, the three processes play out constantly: a post liked for the approval it brings is compliance, a trend copied from an admired influencer is identification, and a cause shared because it matches one’s deepest values is internalization. Recognising which process is driving your own online behaviour is a useful step toward more intentional engagement.

The theory is not without criticism, though. Some scholars argue that Kelman’s model can oversimplify the complex processes of social influence and may not fully account for individual differences or context. Others point out that situational factors, such as the timing and crowdedness of the setting in which influence happens, can shape how persuasive an influencing agent appears, and that Kelman’s framework may not fully address the role of such situational factors. The model also focuses on three core processes and does not directly cover every form of influence, such as the influence a committed minority can have on a majority.

These limitations do not undo the theory’s value. What they suggest is that social influence is rarely a single, clean process. In real life, compliance, identification, and internalization often overlap and shade into one another. A behaviour might begin as compliance, deepen into identification as a relationship grows, and finally become internalized as a person makes it their own. Kelman’s lasting contribution is the reminder that not all change is equal, and that understanding the difference is the first step toward influence that lasts.

What do you think? Think of a habit or belief you hold strongly today. Did it start as compliance, identification, or internalization, and has it shifted from one to another over time? And when you try to encourage change in others, are you aiming for behaviour that lasts only while you are watching, or change that they will carry on their own?

References

PDF 📄

Table of Contents

What social influence theory actually explains 🔗

The three processes of influence 🔗

Compliance 🔗

Identification 🔗

Internalization 🔗

What determines which process takes hold 🔗

The importance of the anticipated effect 🔗

The power of the influencing agent 🔗

The strength of the induced response 🔗

Why looking only at compliance falls short 🔗

Where the theory is useful and where it has limits 🔗