Social Influence Theory: Compliance, Identification, and Internalization on Social Media

Have you ever shared a post just because it had thousands of likes, even though you hadn’t verified it? Or started using the same slang and posting style as a community you joined online? Or found yourself genuinely believing something because every account you follow seems to say the same thing? These are not random habits – they are predictable, psychological responses to social influence. The mechanics behind them were mapped out decades ago by Harvard social psychologist Herbert C. Kelman, whose landmark 1958 framework explains exactly why and how people change their attitudes and behaviors in response to others. On social media, his theory is more relevant than ever.

Kelman’s three processes of social influence 🔗

Social Influence Theory, as formulated by Kelman, proposes that attitude and behaviour change does not happen uniformly. Instead, it occurs through three qualitatively distinct processes, each driven by a different motivation and producing a different depth of change. These are compliance, identification, and internalization. Understanding the difference between them is key to understanding why some online behaviours are fleeting and why others become deeply held convictions.

Each process is defined by three factors: the importance of the anticipated outcome, the power of the influencing agent, and the nature of the response it triggers. The resulting changes in behaviour and attitude can be either visible (what we post and share) or invisible (what we privately come to believe).

Compliance: performing for reward or approval 🔗

Compliance is the most surface-level form of social influence. According to Kelman’s framework, it occurs when an individual accepts influence and adopts a behavior in order to gain approval or avoid disapproval – not because they genuinely agree with the position. The change is essentially transactional. When the social pressure or the reward disappears, so does the behavior.

On social media, compliance is woven into the platform’s basic architecture. A user posts content that they know will be well-received in their network – a popular opinion, a trending format, a safe take on a news story – to gain likes and positive feedback. The like count is the reward. Conversely, a user may self-censor a genuine opinion to avoid public backlash or the discomfort of being ratio’d. Neither action reflects their true beliefs; both are driven purely by the social effect of the outcome. As Hwang (2016) defines it, compliance happens when someone accepts influence because they hope to gain a positive reaction from a person or group – a dynamic that social media platforms, optimised for engagement, are extraordinarily good at triggering.

Identification: belonging to the tribe 🔗

Identification goes deeper than compliance. Here, a person changes their attitudes or behaviours not for a specific reward, but because they want to establish or maintain a meaningful relationship with a person or group they admire. As Wikipedia’s entry on social influence notes, drawing on Kelman’s original work, identification is about being influenced by someone who is liked and respected – and adopting their behaviors because doing so reinforces a self-defining relationship.

Critically, the person genuinely believes in the behavior while the relationship holds. But if the relationship dissolves, the behavior is likely to fade. On social media, this plays out constantly. A follower adopts the opinions, aesthetic choices, or even the vocabulary of an influencer they admire – not for a reward, but to feel a sense of connection and belonging. Zhou (2011) captures this precisely, explaining that identification reflects a user’s sense of belongingness and attachment to a community. This is what drives the tribal loyalty of fandoms, political echo chambers, and niche communities like #BookTok or fitness subcultures on Instagram. Users adopt the group’s norms, vocabulary, and shared opinions because it affirms their identity as a member of that group.

Internalization: when influence becomes belief 🔗

Internalization is the most permanent and profound level of influence. It occurs when a person adopts a belief or behaviour because it is intrinsically consistent with their own personal value system. There is no external reward at stake, and no relationship to maintain. The person is genuinely convinced. As LibreTexts explains, internalized influence is intrinsically rewarding – it persists even when the original source of influence is long gone, because the belief is no longer external; it is part of the person’s identity.

This is the distinction that sets internalization apart. Research confirms that while all three processes shape online behaviour, internalization produces the most lasting change. A useful illustration is an Online Health Community (OHC). A person newly diagnosed with a chronic condition might first join a forum and comply with its posting norms to get answers. Over time, they identify with the community and its members. Eventually, they begin sharing health information not to fit in, but because they genuinely believe in its value. Research by Zhou (2011) examining online community participation through the lens of social influence theory found that compliance, identification, and internalization all significantly shape user behaviour – but that identification and internalization are stronger and more enduring predictors of continued participation than compliance alone.

How compliance and identification drive social media behavior 🔗

The design of modern social media platforms is, whether intentionally or not, a machine for triggering compliance and identification at scale. Metrics like likes, shares, retweet counts, and follower numbers serve as highly visible social rewards. Users receive near-instant feedback on whether their behaviour is approved or disapproved of by their network. This feedback loop is the engine of compliance – and it shapes what content gets created, what gets shared, and what gets suppressed.

Identification operates on a slightly longer cycle, but is just as powerful. Platform algorithms are designed to cluster users around shared interests and viewpoints, which naturally fosters group identity. Hwang’s (2016) research on social influence in digital settings demonstrates that identification – the drive to realise a self-defining relationship with a group – is a key mechanism through which users engage with online communities, including in e-learning contexts and social networks. Once a user identifies strongly with a community, conforming to its norms feels natural, even desirable.

The dark side: social influence and misinformation 🔗

Kelman’s processes are psychologically neutral tools. They can drive positive outcomes – health information sharing, community support, civic engagement – but they can equally be exploited to spread harmful content. Misinformation on social media weaponises all three levels of influence in sequence.

Research published in the Journal of Marketing Management by Mulcahy et al. (2024) directly examines the spread of health and well-being misinformation through social media influencers. The study highlights that consumers are highly susceptible to sharing misinformation from influencers precisely because social influence mechanisms – particularly identification – bypass critical evaluation. When a follower strongly identifies with an influencer, they extend trust not just to the relationship, but to the information itself. A false health claim lands differently when it comes from someone whose lifestyle a person aspires to.

The process can unfold in a three-stage sequence. First, a user sees a piece of misinformation going viral, accumulating thousands of shares. The sheer volume of engagement signals social approval – compliance kicks in, and the user shares it too, not wanting to be left out. Second, the user continues to see the claim repeated within communities they belong to and identify with; the group consensus makes it feel credible. Third, after repeated exposure within a trusted echo chamber, the false claim may eventually be internalized – it becomes a personal conviction, detached from its original source. At that stage, fact-checking alone rarely works, because the belief is no longer experienced as an external claim but as an internal truth.

A UNESCO study of 500 digital content creators across 45 countries, published in late 2024, found that 62% of surveyed influencers do not verify content before sharing it with their audiences. Roughly one-third said they would share without checking if the content came from a source they already trusted – demonstrating precisely how identification (trust in a source) enables the spread of unverified information at scale.

The solution, as Mulcahy et al. argue, lies not in restricting platforms alone, but in building critical awareness among users. They call for educational campaigns that actively encourage individuals to question and verify content before sharing – fostering the kind of analytical thinking that disrupts blind compliance. Research published in Health Promotion International (2025) similarly concludes that countering health misinformation requires a combination of approaches targeting users, content creators, platforms, and governments simultaneously.

The scalability problem: why this theory matters more than ever 🔗

Kelman developed his framework in the 1950s, observing influence in classrooms, small groups, and communities. The psychological processes he identified were real and consequential at that scale. But the fundamental challenge of the social media era, as Delfanti and Arvidsson note in Introduction to Digital Media, is scalability. Social media platforms do not operate like village squares where influence is bounded by geography or personal acquaintance. They are global infrastructure, capable of amplifying a single message to millions within hours.

This scalability transforms the stakes of each of Kelman’s processes. An influencer activating identification among their follower base with a single video is not influencing a classroom of thirty – they are potentially shaping the attitudes of hundreds of thousands of people simultaneously. A piece of misinformation that triggers compliance-based sharing does not spread through a neighbourhood; it crosses continents before a fact-checker can respond. The algorithms that govern content distribution are optimised for engagement, and engagement is most reliably driven by the emotional triggers that fuel compliance (the dopamine hit of a like) and identification (the tribalism of an in-group). Emotionally resonant content – regardless of its accuracy – is structurally advantaged.

Understanding Kelman’s three processes is therefore not simply an academic exercise. It is a practical tool for navigating an information environment where the architecture of influence has been amplified to a previously unimaginable degree. Recognizing whether your own online behavior is driven by compliance, identification, or genuine internalization is the first step toward more intentional and critical engagement with social media.

What do you think? When you last shared a piece of content online, were you driven by the number of likes it already had, by the identity of the person who posted it, or because it genuinely aligned with your own values? And if social media platforms are structurally designed to exploit compliance and identification, whose responsibility is it to address that – the platforms, the users, or regulators?

References

1. https://scholar.harvard.edu/hckelman/publications/compliance-identification-and-internalization-three-processes-attitude-change
2. https://open.ncl.ac.uk/theories/15/social-influence-theory/
3. https://is.theorizeit.org/wiki/Social_Influence_Theory
4. https://helpfulprofessor.com/compliance-psychology/
5. https://en.wikipedia.org/wiki/Social_influence
6. https://helpfulprofessor.com/social-influence-theory/
7. https://socialsci.libretexts.org/Courses/Pueblo_Community_College/Interpersonal_Communication_-_A_Mindful_Approach_to_Relationships_(Wrench_et_al.)/09:_Conflict_in_Relationships/9.03:_Power_and_Influence
8. https://www.emerald.com/insight/content/doi/10.1108/10662241111104884/full/html
9. https://www.semanticscholar.org/paper/Understanding-social-influence-theory-and-personal-Hwang/cb4801c1fcd58d2dee9085b10063b7e0927fab87
10. https://journals.sagepub.com/doi/10.1177/14413582241273987
11. https://www.cnn.com/2024/11/26/media/social-media-influencers-verify-information-study
12. https://academic.oup.com/heapro/article/40/2/daaf023/8100645
13. https://www.wiley.com/en-us/Introduction+to+Digital+Media-p-9781119276210