Transfer tests of stimulus value in concurrent chains.
A stimulus keeps its reinforcing power no matter which schedule you drop it into.
01Research in Context
What this study did
Ladouceur et al. (1997) ran transfer tests with pigeons in a two-key chamber. Birds first learned two separate concurrent-chain schedules. Later, the team swapped the terminal-link stimuli to see which factor guided choice.
The question: does a stimulus keep its value from its own reinforcer, or from the schedule it now sits in?
What they found
Pigeons continued to prefer the stimulus that had delivered the better payoff, even after it moved to the worse schedule. Choice tracked the stimulus–reinforcer history, not the new alternative.
In plain words, the stimulus carried its own value tag.
How this fits with other research
Dougherty et al. (1996) showed that delay difference, not ratio size, controls choice. Ladouceur et al. (1997) add that the local stimulus–reinforcer link matters more than the molar schedule around it. Together, they tell the same story: pigeons look at the immediate relation, not the big picture.
Cowie et al. (2020) later extended the idea by signaling high- versus low-value grain. Preference followed the signaled value, backing up the 1997 claim that value sticks to the stimulus.
Iwata (1993) once found that choice can flip when alternatives are shown one at a time, seeming to credit molar rates. The 1997 transfer test clarifies the twist: single-alternative probes change the context, but the stimulus still keeps its own value.
Why it matters
When you program reinforcement, think stimulus first, schedule second. A highly preferred token, sound, or picture keeps its power even if you move it to a leaner patch. Use this to protect good stimuli during schedule thinning or when you shift a child to a new classroom. Keep the valued cue; the learner will follow it.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Keep the child’s favorite token board when you switch to a leaner DRL schedule.
02At a glance
03Original abstract
We report two experiments that use transfer tests to investigate whether in concurrent chains the value of a terminal‐link stimulus is affected by the alternate terminal link. In Experiment 1, two groups of pigeons were trained on multiple concurrent‐chains schedules in which switching between the schedules was via pecking a changeover key. For one group, the terminal links were fixed‐interval 8 s versus fixed‐interval 16 s in one component and fixed‐interval 16 s versus fixed‐interval 32 s in the other component. For a second group, the terminal links were variable‐interval 10 s versus variable‐interval 20 s in one component and variable‐interval 20 s versus variable‐interval 40 s in the other. After sufficient baseline training had been given so that performances had stabilized, transfer tests were conducted in which the two chains with equal terminal‐link schedules were presented together as a new concurrent pair. For 6 of the 7 subjects, initial‐link responding changed fairly rapidly during the test in the manner predicted if the values of the terminal links were equal. In Experiment 2, pigeons were trained on multiple concurrent chains using a two‐key procedure, and the terminal links were the same variable‐interval schedules as in Experiment 1. After baseline training, transfer tests were conducted that assessed (a) the relative reinforcing strength of the terminal‐link stimuli in a novel initial‐link situation and (b) the relative ability of those stimuli to evoke responding. The data from the reinforcing strength test were consistent with those from Experiment 1, but those from the evocation strength test were not. Although this discrepancy shows that responding in transfer tests is not solely a function of stimulus value, the results from both experiments suggest, overall, that value is determined by the stimulus—reinforcer relation independently of the alternative terminal link.
Journal of the experimental analysis of behavior, 1997 · doi:10.1901/jeab.1997.68-93