Albert Bandura
There are many different ways to approach the thinking/research of Albert Bandura, but we shall focus on his work in relation to the classical neo-behaviorist learning theory tradition that he comes from (student of Spence). So, we shall look at his work and its implications for the necessity of reinforcement for learning among other learning theory claims. If Bandura is known for anything, I guess the concept of observational learning would be foremost amongst them. Bandura did not come up with the concept. Others have hypothesized and tested whether behavior can be acquired by observing another individual perform that behavior. For example, Miller and Dollard’s “Social Learning and Imitation” in 1941, suggested a theory of observational learning, but required you to imitate the observed behavior and for that imitation of the observed behavior to be reinforced. Thus, they held that observational learning would be just one type of operant conditioning--- behavior acquired via reinforcement of that behavior.
For Bandura, observational learning is a different type of learning from classical conditioning (obviously) and operant conditioning. The point is made in Bandura’s classic 1965 article. Although the first phase of this experiment is especially well known, the second phase (called the Acquisition Index in the article) is important for demonstrating the differences between Bandura’s view of learning and traditional notions. The first phase is the classic “Bobo Doll” finding that children who observe a model interacting aggressively with a Bobo Doll would behave aggressively toward the doll themselves when given the opportunity to freely interact with the doll. This was the finding for both the group of children who saw the model rewarded for the aggression, and for the group of children who saw no consequences for the model’s aggression toward the doll. But, even in this phase of the experiment, the story is not so simple. If children saw the model punished for aggression toward the doll, they were less likely to behave aggressively toward the doll themselves when given the opportunity. Suggesting that there are complicating factors toward the idea that we imitate the behaviors we see models perform. So, the model-rewarded and model-no consequences group of children exhibited more aggressive behaviors when given the chance to interact with the doll than the model-punished group of children.
It is this last group of children in the Bandura (1965) experiment that is particularly important in the second phase of the study. In this phase, all the children were encouraged to display the behaviors they had observed from the model in the first phase of the study. For every behavior they exhibited they would receive stickers and juice treats (reinforcements). With proper incentives, all three groups were able to exhibit equal amounts of aggressive behaviors. This finding is contrary to what was found in the first phase of the study where there was a difference in aggressive behaviors produced by the three groups. The obvious interpretation of these findings is that all three groups acquired all the aggressive behaviors they observed, but that the model-punished group was inhibited from demonstrating what they had learned. Given proper incentive this group showed that it had observed and acquired as many aggressive behaviors as either of the other two groups. Once again, we see the distinction between the learning of a behavior and the performance of a behavior.
Bandura’s view of learning differs from traditional learning theory (and the Miller-Dollard theory of imitation) by noting that reinforcement is not necessary for learning. It works to provide incentives to exhibit the behavior, but does not influence the acquisition of a behavior. Also, observational learning is more than imitation. Depending upon what you observe, you may be learning not to imitate the behavior you see (such as when a model is punished). Another difference between Bandura and Miller-Dollard and other traditional learning theorists regards the differences in the effects of reward and punishment. In phase 1 of Bandura (1965), we saw that there was an effect of reinforcement/punishment on the children’s behavior due to the model being reinforced/punished, not the child himself or herself being reinforced or punished. Miller and Dollard thought that the imitator had to have his or her imitative behavior reinforced in order to learn the behavior. The effect of reinforcement/punishment for Bandura’s study was the effect vicarious (or observed) reinforcement or vicarious punishment.