Pigeons' choices between fixed-interval and random-interval schedules: utility of variability?
Pigeons sometimes prefer unpredictable reinforcement schedules even when they must wait longer on average.
01Research in Context
What this study did
Researchers let six pigeons choose between two keys. One key gave food every 30 seconds like clockwork. The other key gave food after random times that averaged 30 seconds but could be shorter or longer.
They made the random times more or less spread out across tests. Then they watched which key each bird picked most often.
What they found
The birds only slightly liked the steady 30-second key. When the random key became more unpredictable, they liked the steady key even less.
This means the birds did not hate waiting. They sometimes picked the messy schedule even when it meant longer average waits.
How this fits with other research
Soreth et al. (2009) later showed why this happens. They proved birds love the chance of a quick win. When random schedules have more short delays mixed in, birds pick them more often.
Davison et al. (1995) found the same pattern with starlings. Those birds liked variable wait times but hated variable food amounts. Together these studies show timing matters more than size.
Calamari et al. (1987) did the early groundwork. They first showed pigeons can rank schedules in a steady order. The 2005 study adds that the amount of messiness in a schedule can flip that order.
Why it matters
When you set up reinforcement schedules, remember that surprise itself can be a reward. A child might pick a variable token board over a fixed one if the board sometimes pays off early. Test both options and watch which one the learner chooses.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Put two identical tasks side by side: one with fixed 3-token payout, one with VR-3 that might pay after 1, 3, or 5 tokens. Let the client choose and track which they pick more often.
02At a glance
03Original abstract
Pigeons' choosing between fixed-interval and random-interval schedules of reinforcement was investigated in three experiments using a discrete-trial procedure. In all three experiments, the random-interval schedule was generated by sampling a probability distribution at an interval (and in multiples of the interval) equal to that of the fixed-interval schedule. Thus the programmed delays to reinforcement on the random alternative were never shorter and were often longer than the fixed interval. Despite this feature, the fixed schedule was not strongly preferred. Increases in the probability used to generate the random interval resulted in decreased preferences for the fixed schedule. In addition, the number of consecutive choices on the preferred alternative varied directly with preference, whereas the consecutive number of choices on the nonpreferred alternative was fairly constant. The probability of choosing the random alternative was unaffected by the immediately prior interval encountered on that schedule, even when it was very long relative to the average value. The results loosely support conceptions of a "preference for variability" from foraging theory and the "utility of behavioral variability" from human decision-making literatures.
Journal of the experimental analysis of behavior, 2005 · doi:10.1901/jeab.2005.30-04