Glossary
Variable rewards
What are variable rewards?
Variable rewards differ in timing, occurrence, or amount. Those are different arrangements, so specify what varies. In reinforcement research, ratio schedules link reinforcement to responses, while interval schedules also involve elapsed time.
Fixed and variable schedules
These hypothetical arrangements illustrate the terminology used by Ferster and Skinner:
- Fixed ratio: a reward follows every fifth button press.
- Variable ratio: the required number of presses varies around an average. Three successive requirements might be 3, 7, and 5 presses, rather than five every time.
- Fixed interval: the first button press after 60 seconds have passed since the previous reward earns the next reward.
- Variable interval: that waiting period varies. Three successive intervals might be 30, 90, and 60 seconds; the first press after each interval earns the reward.
For interval schedules, elapsed time alone does not deliver the reward: a response is still required. These examples vary when a reward occurs while keeping the reward itself the same. Changing a prize's size or a feed's content is a different form of variability.
How reinforcement schedules are studied
Pérez and colleagues' 2019 experiment compared human responding under carefully matched ratio and interval contingencies. Response rates differed during a choice test even though they had not differed during training. Both schedule details and the point of measurement mattered.
Variable rewards and dopamine
Schultz, Dayan, and Montague described dopamine responses related to differences between expected and received rewards.
Example: a changing feed
The next item in a social-media feed might be interesting, funny, or dull. That illustrates variable content. It does not isolate why a person continues scrolling, or establish that the feed uses a particular reinforcement schedule.
Variable rewards in product design
Product frameworks such as the Hook Model discuss variable rewards as a design idea. Laboratory findings do not establish that random product rewards always produce the most persistent behavior, form habits, or cause addiction. Those claims need direct evidence about the product and behavior.
Our review of Hooked examines what the variable-reward studies establish and what they leave untested.
Why it matters
A reward's timing and conditions can be studied precisely. “Unpredictable” alone is too broad to explain an engaging product or determine whether it benefits its users.