Frustration-Aggression Hypothesis

Social Learning Theory and Aggression

Bandura's social learning theory (SLT) proposes that aggression, like other complex social behaviours, is learned through observation, imitation, reinforcement, and vicarious learning — not through biological programming. This approach emphasises the role of the social environment, models, and cognitive processes in shaping whether and how aggression is expressed.

Bandura's Bobo Doll Experiments

Bandura, Ross and Ross (1961, 1963) conducted the classic Bobo doll studies demonstrating observational learning of aggression. In the 1961 study, children (aged approximately 3–6) watched an adult model interact with an inflatable Bobo doll. Participants were divided into three conditions:

• Aggressive model condition: children watched the adult model physically and verbally aggress towards the Bobo doll (hitting it with a mallet, punching it, sitting on it, shouting 'Sock him on the nose!')
• Non-aggressive model condition: children watched the adult play quietly with other toys, ignoring the Bobo doll
• Control condition: no exposure to a model

After mild frustration (being shown attractive toys they were not allowed to play with), children were placed with the Bobo doll. Children who had observed the aggressive model showed significantly more aggressive behaviour — including precise imitation of the novel aggressive acts they had observed — than the non-aggressive model or control conditions. Same-sex models were imitated more than opposite-sex models, consistent with SLT prediction that model similarity increases imitation.

In 1963, Bandura et al. extended the design to include a vicarious reinforcement condition: children watched either (a) the aggressive model rewarded (given praise and sweets), (b) the model punished (scolded and hit), or (c) no consequences. Children who had watched the model punished showed less imitation of aggression. However, when all children were offered incentives to reproduce as many of the model's behaviours as possible, the performance differences between groups disappeared — demonstrating that all children had learned the aggressive behaviours equally, but only performed them depending on anticipated consequences. This distinction between learning and performance is theoretically important.

Key SLT Mechanisms in Aggression

Modelling: children (and adults) acquire aggressive repertoires by observing models — parents, peers, media characters — and encoding the specific acts, targets, and contexts in which aggression is expressed.

Vicarious reinforcement: observing models being rewarded for aggression (social approval, obtaining resources, avoiding threats) increases the probability of imitation. Observing punishment reduces it.

Direct reinforcement: when a child's own aggressive acts are directly reinforced (winning a fight, getting attention, obtaining a desired object), aggressive behaviour is strengthened.

Self-efficacy and cognitive mediation: Bandura's later social cognitive theory emphasised that learning is mediated by cognitive processes — including self-efficacy beliefs ('I can successfully act aggressively') and outcome expectations ('Aggression will get me what I want'). These cognitive mediators determine which observed behaviours are translated into performance.

Disinhibition: exposure to aggression in the media or real life may disinhibit restraints against expressing pre-existing aggressive tendencies — the observer learns that aggression is normal, acceptable, or has no negative consequences, lowering their threshold for expressing it.