Generalization of negatively reinforced mands in children with autism.
Teach "stop" with two yucky tasks and kids will use it everywhere, but start with goodies for compliance to get faster gains.
01Research in Context
What this study did
Two children with autism learned to sign "stop" when something felt bad.
The team trained the sign with only two unpleasant tasks.
They then checked if the kids used "stop" with new, untrained tasks.
What they found
Both kids quickly used "stop" in brand-new situations.
Problem behavior dropped at the same time.
Two teaching examples were enough for the skill to spread.
How this fits with other research
Slocum et al. (2025) now shows you can skip escape extinction at first.
Their big 2025 trial found positive reinforcement beats negative reinforcement for speed.
That update supersedes the 2014 focus on escape-based manding.
Au-Yeung et al. (2015) also saw positive reinforcement outperform breaks for compliance.
Yet Leung et al. (2014) still matters: it proves a tiny teaching set can generalize.
Use the newer speed tactics first, then borrow the tiny-set idea for generalization later.
Why it matters
You can teach one escape mand with just two examples and watch it pop up everywhere.
Pair this with faster positive-reinforcement starts from newer studies.
Your Monday move: run a quick two-task "stop" drill, then reinforce compliance with goodies, not breaks.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pick two short non-preferred tasks, teach the client to sign "stop", then reinforce compliance with a favorite snack instead of a break.
02At a glance
03Original abstract
Each day, people encounter stimuli they find unpleasant. Some children with autism may require systematic instruction to acquire the communication skills necessary to request the termination of such aversive stimuli. We taught 2 school-aged boys with autism a mand (e.g., signing "stop") that could be used to escape a variety of aversive stimuli. First, we employed a systematic assessment to identify aversive stimuli to use during training. We then conducted mand training sequentially across those stimuli until sufficient exemplars were trained for generalization to occur to untrained stimuli. For both participants, cross-stimulus generalization was observed after training with 2 stimuli. Participants manded for escape in the presence of aversive stimuli, but almost never manded in the presence of preferred stimuli or when the programmed stimuli were absent. In addition, we found an inverse relation between acquisition of the mand and engagement in problem behavior and evidence of generalization to nontraining contexts.
Journal of applied behavior analysis, 2014 · doi:10.1002/jaba.151