bandura (aggression)

The concept of identification in terms of incidental learning was previously studied, revealing that children tend to imitate behaviors exhibited by adult models. Key study: Bandura & Huston (1961).
Other studies (e.g., Blake, 1958; Grosser et al., 1951; Rosenblith, 1959; Schachter & Hall, 1952) show that merely observing a model can significantly facilitate the subject's behavioral responses in a social context.
To further investigate imitative learning, especially in new contexts absent of the original model, an experiment was designed where children observed aggressive and nonaggressive adult models.

Primary Hypothesis: Children exposed to aggressive models will demonstrate more aggressive behaviors compared to children exposed to nonaggressive models and those with no model exposure.
- This hypothesis is grounded in the theory that prior reinforcement from models leads to learned imitative behaviors.
Generalization Effect: It was predicted that exposure to nonaggressive models will inhibit aggressive responses in children compared to control groups, with the expectation of significant behavioral differences.
Additional hypotheses focused on the sex of models and subjects, anticipating that children would imitate behaviors more readily from same-sex models. Specifically:
- Male children would be more inclined to imitate aggression, especially from male models, due to the masculine association with aggression.

Subjects: 72 children (36 boys and 36 girls) from the Stanford University Nursery School, ages ranging from 37 to 69 months (mean age: 52 months).
Models Used: Two adult models (one male and one female).
Experiment Structure: Children divided into eight experimental groups plus a control group of 24 (not exposed to models).
Each experimental group contained same-sex and opposite-sex model exposure, with half observing aggressive models and half observing nonaggressive models.

Subjects were matched based on ratings of aggressive behavior during social interactions at the nursery school. Ratings were conducted on four five-point scales measuring:
- Physical aggression
- Verbal aggression
- Aggression towards inanimate objects
- Aggressive inhibition (assessing tendency to inhibit aggression under provocation)
Interrater Reliability: The composite aggression score (from the four scales) showed a reliability of 0.89, with assessments made by both the experimenter and a nursery school teacher.

Initial Setup:
- The subjects were brought to an experimental room where they engaged in a play activity with a model present but separated.
- Nonaggressive models engaged quietly, whereas aggressive models displayed aggressive behavior towards a Bobo doll, including:
  - Punching
  - Striking with a mallet
  - Tossing the doll
  - Verbally aggressive remarks like "Sock him in the nose…"
Imitative Learning Demonstration:
- For significant learning to occur in this context, the models performed unique patterns of aggressive behavior that would likely not manifest without observation.
- Children’s attention was diverted to ensure they observed model behavior without direct prompts to imitate.

Prior to imitation tests, subjects underwent a mild aggression arousal to stimulate aggression behaviors uniformly across groups. This aimed to highlight differences in aggression post-exposure:
- Attraction to toys was used to engage subjects, followed by a suggestion that those toys were exclusively for other children, encouraging provocation for aggression.

After 10 minutes, subjects were moved to a new room filled with various toys (aggressive and nonaggressive). They were observed for a full 20 minutes, categorized into:
- Physical aggressive imitation
- Verbal aggressive imitation
- Nonimitative aggression
- Nonaggressive behavior responses

Children in the aggressive condition exhibited significant physical and verbal imitative aggression compared to nonaggressive and control groups, marked by clear differences in mean scores (refer to Table 1).
Statistical analysis confirmed significant differences in aggression due to model exposure using Friedman two-way analysis of variance.
- Pairwise comparison revealed that aggressive model exposure led to greater aggression levels compared to subjects observing nonaggressive models.

Children demonstrated partial imitation of observed behaviors, especially more mallet aggression and nonimitative responses when exposed to aggressive models.
Specifically, girls with nonaggressive exposure displayed fewer aggressive responses compared to those with respective control conditions.

Boys exhibited higher levels of imitative physical aggression than girls in response to male models, confirming partial support of the hypothesis regarding sex influence on aggression imitation.
Though no significant differences were seen in the verbal aggression response as a function of sex, stronger overall effects stemmed from male model exposure.

Behavioral differences persisted across conditions, particularly in nonimitative aggression and engagement in nonaggressive behaviors, with significant play behavior alterations based on treatment conditions:
- Nonaggressive models resulted in subjects being less aggressive and exhibiting more nonaggressive play.
- Notable gender-specific play patterns emerged, such as girls spending more time on certain nonaggressive activities and boys on exploratory, aggressive play.

Current social learning research emphasizes behavior shaping through rewards & punishments while noting that mere observation influences initiating certain behaviors.
Findings corroborate that observing aggressive adult behaviors serves as a potent strategy for eliciting such behaviors while weakening inhibitory responses in children.
Imitation occurred distinctly in observable aggressive behaviors showcasing a compelling evidence of learning through observation without reinforcement.

The study challenges past frameworks (notably Miller & Dollard's approach) by suggesting learning via observation can establish complex imitative behaviors independently of reinforcement or direct modeling.
Suggests re-evaluating the mechanisms of observational learning within a social context, indicating that behavior may transfer across contexts based solely on observation of aggression, regardless of model quality.

The outcomes indicated a strong tendency for children, notably boys, to imitate aggression observed in male adult models, while also demonstrating significant effects of nonaggressive exposure in inhibiting potential aggressive responses.
The research sheds light on implications for psychoanalytic interpretations involving identification with aggressors, while paving the way for further exploration into effective modeling for child behavior adaptations.