Journal Note, AAS. 20 sept 94

Implications of Skinner for demand analysis

One approach to demand analysis is to assume that preferences are given. One implication of this is that if prices change the person moves along their demand curve which remains unchanged. Then if price returns to its previous level, the quantity demanded should return as before. This should be empirically testable.

Alternatively, a Skinnerian analysis might be conceived which asks if preferences are a changing product of consumption. Experience with a new quantity of the good and substitutes can reinforce the behavior of buying and using at the new levels.

Relevant empirical observation shows that after a drought or some other supply and price altering event, consumption may not return to its previous level even when price returns to previous levels.


Perspectives on behavior modification

Skinner frames one problem of behavior change as how to bring the individual under the control of the future consequences when it is being currently shaped by more immediate reinforcers. The thrust of Skinner's rat and bird experiments exploit the random occurrence of a desired behavior which the experimenter then reinforces. But what happens if the subject never exhibits the desired behavior in the desired context or if exhibited occasionally, it is not possible to deliver something rivaling that which reinforces the undesired behavior? For example, a high school drug dealer is currently reinforced by receiving gold chains, fashionable apparel, etc. How is the experimenter going to supply these things for good scholarship and cooperative behavior?

Skinner's concept of operant reinforcers may help. The primary reinforcers are food, sex, warmth, etc. This creates the possibility of tying a secondary reinforcer to the primary. An example is money. Money is a means to food and from that experience certain behaviors are reinforced by the delivery of money even when it is not a means to food or in fact to anything--money becomes an end in itself for some. So the experimenter has an economizing problem--how to find an operant reinforcer that is cheaper than gold chains.

Margolis provides a related but different perspective. He notes that the mind is capable of organizing complex stimuli into patterns and leaping to a behavior without much if any step wise calculation. He refers to this as "Seeing that such and such fits a pattern (e.g. Abe Lincoln), rather than "reasoning why" a certain behavior is appropriate. So a person might for example perceive of a situation as one involving family, and then certain behaviors such as non-opportunistic actions follow from it. The object of behavior modification is to get the subject to apply the "family" category to new situations where family type behavior is desired. For example, I might want to increase non-opportunistic behavior in the context of fishery use (cooperate to achieve sustainable yield). If the person regarded other fisher people as family, the non-opportunistic behavior might follow. It might be possible to wait for some cooperative behavior and then reinforce it with money, food or prestige. Is there a cheaper way?

Could we create experiences that change what is included in a category such as family? For example we know experimentally that people quickly form bonds with complete strangers when they are part of an arbitrarily formed team competing with another team. (Military training exploits this phenomenon.) Is is possible that this could spill over (mind make a leap) to seeing other fisher people as family? Many observors advocate participation as a means of solving resource use problems. Perhaps what is going on in participation is a change in what is included in certain conceptual patterns. Is it different to reinforce a behavior than to alter the boundary of some already conceived pattern and its associated behaviors and rely on the generalized reinforcers that support that categorized behavior? Is it cheaper?

There are situations where it is impossible to selectively deliver a reinforcer to just those who are doing what we want them to (and not others with inappropriate behavior). Skinner's first rule is that the reward-reinforcer should never follow undesired behavior. In this case if we can do something to change the perceived pattern (category) of some it won't matter if the same message is received by others--their desired associations would just be reinforced.