Assessing stimulus control in natural settings: an analysis of stimuli that acquire control during training.

Four young adults with intellectual disability were learning to say 'please' when they asked for items. The team first taught the word in a special-ed classroom using pictures of snacks and drinks.

After training, the researchers tested each picture alone. They wanted to see which pictures actually controlled the new 'please' response. They also tested pairs of pictures side-by-side.

One learner said 'please' only when items were shown one at a time. Two learners said 'please' only when items were shown in pairs. The fourth learner lost the response before testing even finished.

Mixed results like these show that the stimuli you train with do not automatically gain control over the new behavior.

Landry et al. (1989) saw the same problem from the other side. In their study, old stimuli that controlled swearing blocked new polite words from appearing. Together, the two papers warn that pre-existing stimulus control can either push out new skills or never let them in.

Wetherington (1979) and Palya (1985) showed that quick probe trials during teaching help transfer control faster. The 1991 paper keeps the probe idea but uses it after teaching to check which stimuli actually won control.

Borgen et al. (2017) later used a high-probability instruction sequence to build brand-new stimulus control with preschoolers. Their method gives one practical way to create strong control before you even need the post-training probe.

Before you plan generalization, run short single-stimulus and paired-stimulus probes with each item you trained. If the response shows up only in one format, teach with that format and then practice in new places. If no format works, a competing stimulus may be blocking you; run a quick competing-stimulus assessment like Landry et al. (1989) or switch to a high-p sequence like Borgen et al. (2017) to establish clean control first.