Laboratory control of thumbsucking by withdrawal and re-presentation of reinforcement.

Three preschoolers watched cartoons while their thumbsucking was tracked. Each time a child sucked his thumb, the TV went dark for 10 seconds. The cartoon came back as soon as the thumb left the mouth.

Later the same cartoon shut off on a fixed time schedule, no matter what the child did. The lab compared how fast thumbsucking dropped in each condition.

When the cartoon disappeared only after thumbsucking, the behavior almost stopped within nine minutes. When the screen went dark on a timer, thumbsucking stayed high.

Only the response-contingent loss of cartoons worked. Non-contingent screen blackouts had no effect.

Saini et al. (2015) later showed the same rapid drop in vocal stereotypy using one-demand response interruption instead of cartoon loss. Both studies used an alternating-treatments design and hit the same quick suppression.

Conyers et al. (2004) moved the logic into a real preschool classroom. They compared response cost and DRO; response cost—where stickers were removed after disruptive acts—won over time, echoing the power of contingent removal seen in SIDMAN (1962).

Fritz et al. (2017) seems to disagree at first glance. They cut problem behavior without any extinction by giving non-contingent access to toys. The difference is the reinforcer source: M removed an outside reinforcer (cartoons), while Fritz flooded the session with free reinforcers, so extinction was unnecessary.

If you want to suppress a behavior, make the loss of the preferred item happen only when the behavior occurs. Randomly turning off music, videos, or iPads will not teach the child anything. Next session, try a brief 10-second blackout of the screen or music each time the target behavior appears; return access the instant the behavior stops.