Elimination of behavior of mental patients by response-produced extinction.

The team used timeout to stop problem behavior in mental-health patients.

They tested two set-ups: timeout plus an easy, reinforced task, or timeout with no other way to earn rewards.

Each person served as their own control, so the same person experienced both conditions.

Timeout wiped out the target response when clients could still get tokens by doing a simple alternate task.

When no alternate task was available, timeout had no effect; the problem behavior kept happening.

The study shows that timeout works only if reinforcement keeps flowing somewhere else.

Lejuez et al. (2001) later mapped response-class hierarchies and found that extinguishing mild forms also collapsed severe ones. Their work extends the 1963 idea by showing which behaviors to target first.

Eisenhower et al. (2006) used DRO plus extinction to stop preschool transition tantrums. They kept reinforcement active through DRO, echoing the 1963 requirement of an available alternative.

Kelley et al. (2017) showed dense noncontingent reinforcement can suppress desirable alternative responses. This warns us that if free reinforcement is too rich, clients may not engage in the alternate task and timeout could fail, aligning with the 1963 warning.

Powell et al. (1968) removed tantrums without any punishment by shifting task difficulty after each outburst. Their reinforcement-only approach conceptually replicates the 1963 outcome while avoiding timeout altogether.

Before you place a client in timeout, quickly check: can they still earn reinforcement in another way? If the answer is no, add an easy, reinforced task first or choose a different procedure. This simple step turns timeout from ineffective to fully eliminating the problem behavior.