Choice between constant and variable alternatives by rats: effects of different reinforcer amounts and energy budgets.
Hunger makes animals play it safe; fullness lets them gamble—so tailor reinforcer variability to the learner’s body state.
01Research in Context
What this study did
The team put eight rats in a chamber with two levers.
Pressing one lever always gave the same number of food pellets.
Pressing the other lever gave a random number of pellets that averaged the same.
They ran the test when the rats were lean (below 80 % of free-feed weight) and again when they were plump (above 90 %).
What they found
Hungry rats usually picked the fixed, sure-thing lever.
Well-fed rats flipped and chose the risky, variable lever.
The switch only happened when the two sides offered very different pellet amounts.
How this fits with other research
Madden et al. (2003) later showed the same energy-budget rule works with humans.
Their students picked fixed delays when time was plentiful and gambled on variable delays when time was tight.
Davison et al. (1995) looked like a contradiction: starlings shunned variable amounts.
The starlings were never made hungry or full; body state was not controlled.
Once hunger is added, the picture lines up—risky choice appears when the animal can afford to gamble.
Why it matters
Your client’s body state and reinforcer size can nudge preference as much as the schedule itself.
If a learner is satiated, sprinkling in occasional big rewards may keep them engaged.
If they are motivated and hungry, keep the payoff steady to protect responding.
Check motivation level before you decide to add variability to a program.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Before session, weigh or ask about recent food intake; if the learner just ate, slip a few larger prizes into the variable schedule—if not, stay fixed.
02At a glance
03Original abstract
Two experiments, using rats as subjects, investigated the effect of different reinforcer amounts and energy budgets on choice between constant and variable alternatives under a closed economy. Rats were housed in the chamber and were exposed to a modified concurrent-chains schedule in which the choice phase was separated from a rest phase during which the rats could engage in other activities. In the choice phase, a single variable-interval schedule arranged entry into one of two equal terminal links (fixed-interval schedules). The constant terminal link ended with the delivery of a fixed number of food pellets (two or three, depending on the condition), whereas the variable terminal link ended with a variable number of food pellets (means of two or three, depending on the condition). Energy budget was defined as positive when body weights were over 90% of free-feeding weights, and as negative when they were under 80% of free-feeding weights. The different body weights were produced by varying the duration of the equal terminal-link schedules within daily 3-hr sessions. In Experiment 1, rats chose between a constant and a variable three pellets under both energy budgets. Rats preferred the constant three pellets more under the positive energy budget, whereas they were indifferent under the negative energy budget. In Experiment 2, rats chose between a constant three pellets and a variable two pellets, and chose between a constant two pellets and a variable three pellets under both energy budgets. The rats strongly preferred the constant three pellets over the variable two pellets under both energy budgets. In contrast, rats preferred the variable three pellets over the constant two pellets only under the negative energy budget, whereas they were indifferent under the positive energy budget. These results indicate that rats choices are sensitive to the difference in reinforcer amounts and to the energy budgets defined by the level of body weight. The present results are consistent with those obtained with small granivorous birds as well as with the predictions of a recent risk-sensitive foraging theory.
Journal of the experimental analysis of behavior, 2000 · doi:10.1901/jeab.2000.73-79