ABA Fundamentals

Inappropriate speech in a severely retarded child: a case study in language conditioning and generalization.

Barton (1970) · Journal of applied behavior analysis 1970
★ The Verdict

Candy for good picture talk plus a quick magazine break for silly talk raised appropriate speech from 26% to 96% in one child.

✓ Read this if BCBAs teaching expressive language to children with severe intellectual disability in clinic or home rooms.
✗ Skip if Teams looking for classroom-wide or already-generalized language gains without extra planning.

01Research in Context

01

What this study did

One child with severe intellectual disability kept describing pictures with silly or off-topic words. The team set up a simple table-top game. When the child said something on-topic, he got a piece of candy. When he said something off-topic, the adult closed the picture magazine for ten seconds.

They ran an ABAB reversal. First they used the candy-plus-timeout rule, then they stopped it, then they brought it back, then they stopped it again. Each phase lasted about ten short sessions.

02

What they found

During the first candy-plus-timeout phase, appropriate picture descriptions jumped from 26% to 96%. When the rule was removed, the score dropped back. When the rule returned, the score shot up again. The reversal proved the candy and brief magazine break were driving the change.

The child kept the new skill only while the rule was in place. In new rooms or with new pictures, the gains faded. Generalization stayed limited.

03

How this fits with other research

Stocco et al. (2021) ran a similar test fifty years later. They also reinforced appropriate speech and cut odd speech. They dropped candy and used simple listener interest instead. Both tactics worked, showing the core idea holds across very different rewards.

Esposito et al. (2021) took the timeout idea further. They used red and green cards to signal when stereotypic speech would or would not earn attention. This added clear cues before the child spoke, building on the 1970 brief-removal idea but aiming at self-control rather than just suppression.

Drasgow et al. (2016) extended the work into real-world function. They taught children with autism to ask for items instead of making odd comments. Like the 1970 study, they used reversal designs, but they replaced the problem behavior with useful mands rather than simply cutting it.

04

Why it matters

You can flip off-topic talk to on-topic talk in under two weeks with candy and a ten-second magazine break. Keep the rule in place while you teach, then plan extra steps to help the skill travel to new places and people. Pair the lesson with mand training or stimulus-control cues if you want the change to last outside the table.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Run five picture cards, deliver one skittle for on-topic descriptions, and close the book for ten seconds after any off-topic line.

02At a glance

Intervention
discrete trial training
Design
reversal abab
Sample size
1
Population
intellectual disability
Finding
positive
Magnitude
large

03Original abstract

The modification of inappropriate speech, a class of behaviors rather than a limited number of specific examples, is little known in the severely retarded. In this study, operant techniques were used to modify the strikingly bizarre and inappropriate speech of a severely retarded boy. The boy's appropriate verbal responses to questions about magazine pictures were reinforced with candy. When he responded inappropriately, the magazine was withdrawn, and social interaction was discontinued for a 10-sec timeout period. Negative responses were ignored, the next picture displayed, and the next question asked immediately. In 10 sessions, appropriate responses increased from 26% to 86% of all responses. A reversal of reinforcement was then introduced, in which inappropriate responses were reinforced, appropriate responses resulted in timeout, and negative responses were treated as before. This reduced the percentage of appropriate responses to 24%. Subsequent sessions of reinforcement for appropriate responses increased appropriate responses to 96% of all responses. At significant stages in the experiment, a measure of possible generalization was attempted. Although some generalization was recorded, it was minimal: some explanations are discussed.

Journal of applied behavior analysis, 1970 · doi:10.1901/jaba.1970.3-299