Learning 193THE PARTIAL REINFORCEMENT EFFECT Consider the following scenario: Alicia’s
mother agrees to give her a quarter every night she remembers to put her dirty clothes in
the clothes hamper. Bianca’s mother agrees to give her a dollar at the end of the week, but
only if she has put her clothes in the hamper every night. Alicia learns to put her clothes
in the hamper more quickly than does Bianca because responses that are reinforced each
time they occur are more easily and quickly learned. After a time, the mothers stop giving
the girls the money. Which child is more likely to stop putting her clothes in the hamper?
The answer might surprise you. It is more likely that Alicia, who has expected to
get a reinforcer (the quarter) after every single response, will stop putting her clothes in the
hamper. As soon as the reinforcers stop for her, the behavior is no longer reinforced and is
likely to extinguish. In contrast, Bianca has expected to get a reinforcer only after seven cor-
rect responses. When the reinforcers stop for her, she might continue to put the clothes in the
hamper for several more days or even another whole week, hoping that the reinforcer will
eventually come anyway. Bianca may have learned more slowly than Alicia, but once she
learned the connection between putting her clothes in the hamper and getting that dollar,
she is less likely to stop doing it—even when her mother fails to give the dollar as expected.
Bianca’s behavior illustrates the partial reinforcement effect (Skinner, 1956):
A response that is reinforced after some, but not all, correct responses will be more resistant to
extinction than a response that receives continuous reinforcement (a reinforcer for each and
every correct response). Although it may be easier to teach a new behavior using continuous
reinforcement, partially reinforced behavior is not only more difficult to suppress but also
more like real life. Imagine being paid for every hamburger you make or every report you
turn in. In the real world, people tend to receive partial reinforcement rather than continuous
reinforcement for their work.
Partial reinforcement can be accomplished according to different patterns or schedules.
For example, it might be a certain interval of time that’s important, such as an office safe that
can only be opened at a certain time of day. It wouldn’t matter how many times one tried
to open the safe if the effort didn’t come at the right time. On the other hand, it might be the
number of responses that is important, as it would be if one had to sell a certain number of
raffle tickets in order to get a prize. When the timing of the response is more important, it is
called an interval schedule. When it is the number of responses that is important, the schedule
is called a ratio schedule because a certain number of responses is required for each reinforcer
(e.g., 50 raffle tickets for each prize). The other way in which schedules of reinforcement can
CCWatch the Video Schedules of Reinforcement“Remember, every time he gives you
a pellet, reinforce that behavior by
pulling the lever.”
© The New Yorker Collection 2005 Joe
Dator from cartoonbank.com. All Rights
Reserved.partial reinforcement effect
the tendency for a response that is
reinforced after some, but not all, cor-
rect responses to be very resistant to
extinction.continuous reinforcement
the reinforcement of each and every
correct response.