An interresponse time analysis of variable-ratio punishment.

The team used a VR-10 punishment schedule. Every tenth lever press earned a brief electric shock.

They tracked the tiny gaps between responses. These gaps are called interresponse times, or IRTs.

Shocks did not slow the rats down. Instead, the animals produced quick bursts of presses right after each shock.

The burst pattern raised the overall rate. The next few responses came faster than before.

HOLZ et al. (1963) saw the opposite on a DRL schedule. Their shocks removed short IRTs and slowed the rats.

The difference is the schedule. DRL rewards long pauses, so removing short IRTs helps. VR rewards fast presses, so the burst helps even with shock.

Last et al. (1984) later showed timing matters. Shock right after a long IRT shortens that IRT. The 1981 paper shows shock after any VR response can create a burst chain.

If you use punishment on a ratio schedule, watch for brief speed-ups, not just slow-downs. A burst can look like the behavior is "working," but it is a side effect of the schedule. Check IRTs to see if the learner is responding faster right after the aversive event. If you see this pattern, rethink the plan or add reinforcement for slower, steady responding.