Changeover behavior under pairs of fixed-ratio and variable-ratio schedules of reinforcement.

Burgess et al. (1971) tested pigeons on two side-by-side keys.

Each key gave grain on either a fixed-ratio (FR) or variable-ratio (VR) schedule.

The twist: on some trials the two schedules were independent; on others they were linked so a reinforcer waited on both keys after the VR requirement was met.

The authors counted how often birds hopped from one key to the other.

When the VR schedules were linked (non-independent), changeovers sky-rocketed.

Birds switched keys far more than under any other arrangement.

The result shows that the chance of payoff, not just the payoff itself, controls choice.

Van Houten et al. (1980) moved the same FR vs VR question into a classroom token system with deaf students.

They saw VR tokens raise attention and cut disruption, echoing the 1971 lab finding that VR schedules strengthen behavior.

Hawkins (1979) also ran concurrent FR VR with pigeons but left out the non-independent setup.

That paper saw less switching, so the two studies look opposite until you notice the missing link.

The 1971 paper therefore clarifies why later work found milder effects: schedule independence matters.

de Carvalho et al. (2018) later showed VR keeps two rats pressing a lever together more steadily than FR, again confirming VR’s special staying power.

If you run concurrent programs—like two work bins, two games, or two response options—remember that linked payoffs will make learners bounce between choices more often.

You can use this to keep engagement high during long tasks, or you can unlink payoffs when you want steadier, stay-put behavior.

Next time you design a token board or DRO system, decide first: do I want switching or staying? Then pick independent or linked schedules accordingly.