Outcome Devaluation

Why the Outcome Matters

A rat can learn an action because that action produces a desirable consequence. In the Adams and Dickinson experiment, rats learned to press a lever for sucrose pellets. The central question was not only whether the rats had learned to press the lever. It was whether later pressing would still depend on the current value of the sucrose pellets.

Training the Lever Response

During instrumental conditioning, lever pressing was established as a behavior that produced sucrose pellets. This training stage shows that the rats learned the response. However, training alone does not show why the rats would continue pressing later. Pressing might be controlled by the learned relationship between the action and its outcome, or it might continue simply because pressing had become familiar.

Changing Pellet Value

The researchers devalued the sucrose pellets without requiring the rats to press the lever. Rats received free access to the pellets, so the food was available independently of their actions. After 15 minutes of access, one group received lithium chloride, a nausea-inducing poison. This procedure was repeated for three sessions. By the final session, the injected rats consumed none of the freely available pellets, indicating that the pellets had lost value for them.

Testing During Extinction

After the pellet value was changed for one group, both groups were tested with extinction trials. During extinction, the lever was disconnected from the pellet dispenser. Rats could still press the lever, but pressing no longer delivered sucrose pellets. This allowed researchers to measure lever pressing without giving the outcome again.

The order of operations is essential: first the rats learn that pressing produces pellets, then the pellet value is changed for one group, and finally both groups press while pellets are withheld. Because the test does not provide pellets, a difference in response rate shows how the changed outcome affected performance of the learned action.

Predicted Group Difference

The devalued group pressed the lever at a lower rate than the non-injected group from the start of extinction. This result shows that changing the value of the sucrose outcome affected performance of the learned action, even though the rats had already learned how to press the lever.

Common Interpretation Errors

• Treating training as proof that the outcome will always control the behavior.

  Training establishes the response but does not reveal whether later pressing depends on the current value of the outcome or merely on familiarity with the response.

  Fix: Use devaluation followed by an extinction test to examine the effect of the outcome's current value.

• Saying that devaluation occurred because the rats had to press the lever repeatedly.

  The rats received free access to pellets independently of their actions.

  Fix: Identify the devaluation procedure as free pellet access followed by lithium chloride-induced nausea.

• Assuming that extinction means the rats cannot press the lever.

  The lever remained available for pressing; it was disconnected from the pellet dispenser.

  Fix: Extinction measured pressing while withholding the sucrose-pellet outcome.

• Predicting that the devalued group should press more because it had received more experience with pellets.

  By the final devaluation session, the injected rats consumed none of the freely available pellets, indicating that the pellets had lost value.

  Fix: Predict a lower response rate for the devalued group than for the non-injected group during extinction.

Apply the Experimental Logic

Key Takeaways

1. Instrumental conditioning established lever pressing as a behavior that produced sucrose pellets.
2. Pellet value was reduced through free access to pellets followed by lithium chloride-induced nausea, repeated across three sessions.
3. Extinction disconnected the lever from the pellet dispenser, allowing pressing to be measured without delivering pellets.
4. Rats with devalued pellets showed lower response rates than non-injected rats from the start of extinction.
5. The group difference shows that changing the value of the outcome affected performance of the learned action.