Peer influence is one of the most studied phenomena in adolescent development and one of the most frequently misdescribed.
The popular model is pressure: a group persuades or coerces an individual into behaviour she would not otherwise choose. This does occur. It accounts for considerably less of the observed similarity within friendship groups than most people assume.
Selection versus influence
Friendship groups are similar on many behaviours. Two processes produce that similarity.
Selection. People choose friends who already resemble them. Similarity precedes the friendship.
Influence. People become more similar to their friends over time.
Studies using longitudinal network data, which observe both friendship formation and behaviour change over time, have consistently found that both operate and that selection accounts for a substantial share of the observed similarity.
The proportions vary by behaviour. For some outcomes selection dominates; for others influence is more prominent.
This matters practically. If similarity within a group is largely selection, then removing a young person from the group addresses a symptom rather than a cause, and she will generally select a similar group in a new setting.
How influence actually operates
Where influence occurs, the research suggests the mechanism is less often explicit pressure than perceived norms.
Studies of adolescent behaviour find that perceptions of what peers do are frequently inaccurate, and systematically so — adolescents tend to overestimate the prevalence of risk behaviours among their peers.
Behaviour then adjusts towards the perceived norm rather than the actual one.
This finding has produced one of the better-evidenced intervention approaches in the field: social norms interventions, which provide accurate information about actual peer behaviour. Evaluations have found effects on several behaviours, though results are mixed and the approach has not worked in every setting.
The mechanism is worth noting because it is not persuasion. It is correction of a factual error.
The risk-taking finding
Experimental work has found that the presence of peers increases risk-taking in adolescents, in laboratory tasks where no communication or pressure occurs.
The effect appears in adolescents and is much weaker or absent in adults, and it has been found even when peers are merely observing.
This suggests something more automatic than social pressure — that the presence of peers changes how rewards are evaluated during a developmental period, rather than that peers argue anyone into anything.
The practical implication is that "who was she with" is a more useful question than "what did they say to her."
The gender pattern
Research on adolescent friendship finds differences in structure that are relevant.
Girls' friendships are generally characterised by higher disclosure, greater intimacy and smaller group size. Boys' groups tend to be larger and more activity-based.
The consequences run both ways. Higher-disclosure friendships provide more support and are associated with better outcomes on several measures.
They also carry higher costs when they end or when conflict occurs, and research on co-rumination — repeatedly discussing problems within a close friendship — finds it associated with both closer relationships and higher internalising symptoms.
That finding is a useful corrective to the assumption that more talking is always better. The same conversation that provides support can sustain rumination, and the difference is in whether it moves towards resolution.
Where peer influence is protective
The literature concentrates on negative influence, which distorts the picture.
Peer effects on academic outcomes are substantial and run in both directions. Studies of school and classroom composition find effects of peer achievement on individual attainment.
Friendship group composition predicts extracurricular participation, help-seeking behaviour and educational aspiration.
And peer support is among the stronger protective factors identified in adolescent mental health research.
A young person with a functioning friendship group is in a substantially better position than one without, and interventions that disrupt groups without providing replacements are frequently net negative.
What follows for parents
Attend to selection rather than only to influence. The question of why a particular group is appealing is more informative than what the group is doing, and it is generally answerable — groups meet needs that are not being met elsewhere.
Provide alternative groups rather than removing existing ones. Structured activity that produces a different peer group is more effective than prohibition, which does not supply a replacement.
Correct norm misperceptions where they arise, since the evidence suggests the perceived norm is frequently wrong and is doing real work.
And be sceptical of the pressure model. An adolescent who is told her friends are making her do things generally knows this is not an accurate description of her experience, and the inaccuracy costs the conversation its credibility.