Correlated reward in mirror drawing

A systematic account of correlated reinforcement has been carried out recently by Logan (1960). The concept of correlated reinforcement emerges as a consequence of the fact that various response indices do not manifest consistent changes in a conditioning process. For instance, though the amplitude measure increases consistently through learning to its asymptote, the latency measure of the same response may undergo an irregular change in a particular experiment. A question of importance then is which response measure does and which does not undergo a change in a particular learning situation. Aside from previous learning, the mode of reward presentation relative to change in that response measure is presumably a responsible factor. In a Skinner box, when the number of pellets is positively correlated with the force with which a rat exerts on the bar, the response measure of force is an aspect of behavior changing through the learning process. Most researches included in Logan's book used infrahuman animals as Ss. Obviously, with verbal instructions, the correlated reward study can be done more effectively in a sense that the data will not be contaminated by the learning of correlated reward per se.