Stimulus Generalisation
When a behaviour learned in one context occurs in new settings, with new people, or in the presence of new stimuli.
Stimulus generalisation happens when the learner responds to new stimuli that are similar to the original training stimulus. Teaching a child to say 'dog' to your family's labrador and then having them say it to a poodle at the park - that's stimulus generalisation working correctly.
The problem is when it works too broadly or too narrowly. Too broad: a learner taught to ask for a break with one teacher asks for a break constantly with every adult because any adult has become a cue. Too narrow: a learner who can read sight words on one specific set of flashcards can't read them on a worksheet because the font changed.
Programming for stimulus generalisation means deliberately varying the training conditions - multiple teachers, multiple rooms, multiple materials - from the start rather than hoping it happens on its own. If you train to one exemplar and then test in a new context, you're not measuring generalisation, you're gambling on it.