Maximizing and matching on concurrent ratio schedules.

Researchers placed pigeons in front of two keys. Each key paid off on a variable-ratio schedule. The birds could hop between keys at any time.

The team tracked every peck. They asked: do birds match their responses to the payoff rates, or do they maximize total food?

The birds spread their pecks in a way that pulled in the most grain per minute. Their pattern fit a maximization rule, not just matching.

The data backed the matching-law framework, but only because maximizing and matching gave the same answer on these VR schedules.

Yuwiler et al. (1992) ran the same test with variable-interval schedules and saw the opposite: pigeons matched, they did not maximize. The clash looks real, yet it is schedule-specific. VR lets birds maximize; VI makes matching the better rule.

Macdonald et al. (1973) showed matching on VI-VI two years earlier. Davison et al. (1989) later showed that extra feedback does not push pigeons toward maximizing on VI. Together, these studies draw the boundary: maximization shows up on ratio, not interval, schedules.

Catania et al. (1974) stretched the idea to monkeys choosing cocaine doses. Their response split still tracked relative payoff, proving the matching law holds even when the reinforcer is a drug, not food.

When you set up concurrent reinforcement, check the schedule type. On VR-based token boards or response chains, clients may drift toward the richer side faster than the matching law alone predicts. Watch for maximization and be ready to thin the richer schedule or add response cost to keep both behaviors alive.