Concurrent performances: reinforcement by different doses of intravenous cocaine in rhesus monkeys.

Four rhesus monkeys lived in test chambers with two levers. Each lever delivered a different dose of cocaine through an IV line.

The schedule was concurrent VI-VI. Pressing on the left or right lever paid off on its own timer. Monkeys switched freely.

Sessions ran until the monkey had taken 40 total infusions. Scientists recorded every lever press and infusion.

Monkeys pressed the lever tied to the bigger dose about a large share of the time. Response ratios almost perfectly matched drug-intake ratios.

Even when the higher dose paid off less often, the monkeys still favored it. They worked harder to keep the richer option alive.

Macdonald et al. (1973) showed the same matching pattern in pigeons working for food. Catania et al. (1974) proves the rule holds when the reinforcer is an IV drug.

Aragona et al. (1975) later switched the schedule to concurrent VR-VR and still saw matching. The law survives when effort, not just time, drives payoff.

Davison et al. (1989) seems to disagree: they say overall reinforcer rate does not guide choice. The key difference is procedure. C et al. changed dose (magnitude), not the total payoff pool. Magnitude matters; global rate feedback does not.

The study tells us reinforcer magnitude, not just rate, steers choice. When you set up concurrent teaching tasks, weight the richer skill with stronger reinforcers—more tokens, longer iPad time, higher-quality praise. If two choices pay at the same rate, boost the quality on the target behavior and watch allocation shift without extra teaching time.