Habitual Behavior
The experiment changes reward value after learning and measures whether behavior adjusts.
The Value Test
An animal can continue an action for two broad reasons. It may be responding to what the action is expected to produce and to how valuable that outcome currently is. Alternatively, the action may have become a learned pattern that is less responsive to changes in the outcome's value. An outcome-devaluation experiment separates these possibilities by changing the value of a reward after learning and then measuring whether the behavior changes.
The central question is not simply whether the subject learned the action. It is whether later responding remains connected to what the expected outcome is currently worth.
Following the Experiment
Adams and Dickinson's 1981 study separated learning, value change, and testing into distinct stages. During lever training, the rats learned a relationship between pressing a lever and obtaining pellets. During pellet devaluation, the value of the pellets was changed by linking them with nausea. During extinction testing, the lever was presented again, but it was disconnected from the pellet dispenser. This final arrangement allowed the researchers to measure the learned response without delivering more pellets during the test.
Reading the Adams and Dickinson Design
Determine what each stage contributes to the interpretation of the experiment.
- 1Lever training
The rats learn that pressing the lever normally produces pellets. This establishes the behavior-outcome relationship.
- 2Pellet devaluation
The pellets become associated with nausea, changing their value after the action has already been learned.
- 3Extinction testing
The lever is available again, but pressing it no longer activates the pellet dispenser. Responding can therefore be compared after the value change without newly delivering pellets.
- 4Interpretation
If pressing falls after devaluation, the rats' responding was sensitive to the changed value of the expected pellets.
The experiment tests whether learned lever pressing remains guided by the current value of its expected outcome.
The Cognitive Map
Adams and Dickinson interpreted the result as evidence for a connected representation of the situation. In this representation, lever pressing is related to obtaining pellets, and the pellets are related to their changed value after devaluation. The subject can combine these relationships: pressing the lever would normally produce pellets, but those pellets have become something to avoid. The changed value of the expected outcome can therefore influence whether the subject presses.
The cognitive-map interpretation explains why changing the value of the pellets can change lever pressing even though the lever-pellet relationship was learned earlier. The behavior is not represented in isolation. It is connected to an expected outcome, and that outcome is connected to its current value.
Reading the Behavioral Change
The decisive observation occurred during extinction. Rats whose pellets had been devalued pressed the lever at significantly lower rates than the non-injected rats from the beginning of the extinction trials. The lower response rate supports the conclusion that the rats were sensitive to the current value of the expected pellets.
Reduced responding is informative because the action was learned before the value change. When the behavior decreases after the expected outcome is devalued, the pattern indicates that the subject is using the outcome's current value rather than responding only to a fixed learned action pattern.
Goal Direction and Habit
Under goal-directed control, the expected outcome and its value affect behavior. If the outcome becomes less valuable, responding decreases. A behavior treated as a habit would be expected to be less responsive to the changed value of the outcome and could continue at a similar rate. The experiment therefore distinguishes the two possibilities by observing what happens after devaluation.
| Pattern after devaluation | Interpretation |
|---|---|
| Responding decreases | Evidence that behavior is sensitive to the current value of the expected outcome and is under goal-directed control |
| Responding remains at a similar rate | Behavior is less responsive to the changed outcome value, as expected for a habit-like pattern |
Outcome-devaluation experiments use post-devaluation behavior to distinguish these broad forms of control.
Common Interpretation Errors
Common mistake Treating devaluation as part of the initial learning stage
What it looks likeDescribing the rats as learning the lever response while the pellets were already devalued
Why it failsThe design separates lever training from the later pellet-devaluation stage.
Do this insteadState the order explicitly: lever training, pellet devaluation, and extinction testing.
Common mistake Ignoring the extinction arrangement
What it looks likeExplaining the result as if pressing the lever continued to deliver pellets during the test
Why it failsDuring extinction, the lever was present but disconnected from the pellet dispenser.
Do this insteadUse the test response as the measure after the value change, without describing new pellet delivery during testing.
Common mistake Calling lower responding proof that the action was never learned
What it looks likeConcluding that the devalued rats failed to acquire lever pressing
Why it failsDevaluation occurred after the lever-pellet relationship had been learned.
Do this insteadInterpret the decrease as evidence that the learned behavior was sensitive to the current value of its expected outcome.
Common mistake Equating a habit with any behavior that has been learned
What it looks likeSaying that all learned lever pressing is habitual
Why it failsThe experiment distinguishes behavior that adjusts to outcome value from behavior that is less responsive to that value.
Do this insteadUse the post-devaluation response pattern to discuss the distinction.
Check Your Reasoning
Suppose two groups learned the same lever-pellet relationship. After one group's pellets were devalued, both groups were tested in extinction. The devalued group pressed the lever at a lower rate from the beginning of testing. What does this pattern support, and which part of the experiment makes that conclusion possible?
Hints
- Identify what happened to the pellet value after learning.
- Recall what happened to the lever-pellet connection during extinction testing.
- Connect the lower response rate to sensitivity to the expected outcome's current value.
What do you think happens?
If the expected pellets become something to avoid and the subject uses the current value of the expected outcome, what should happen to lever pressing during extinction?
Reveal answer
It should decrease.
The devalued rats pressed the lever at significantly lower rates than the non-injected rats from the beginning of extinction. This supports sensitivity to the current value of the expected pellets and goal-directed control.
Key Takeaways
- Outcome-devaluation experiments change the value of a reward after an action-outcome relationship has been learned and then measure behavior.
- The Adams and Dickinson (1981) design separated lever training, pellet devaluation, and extinction testing.
- Lower lever pressing after pellet devaluation indicates that responding was sensitive to the current value of the expected pellets.
- A cognitive map can connect lever pressing, pellets, and the changed value of pellets.
- Reduced responding supports goal-directed control, whereas similar responding after devaluation indicates behavior that is less sensitive to outcome value.