Generalization Strategies Flashcards
-Generalization:
the tendency for behavior to occur in situations different than the one in which the behavior was learned
-Discrimination:
refers to precise stimulus control
-Generalization:
refers to less precise stimulus control
Why is generalization so important?
- training usually occurs in on place and a narrow range of stimuli
- real world is not likely to give exact copies of the controlling stimuli that occurred during training
Strategies to Promote Generalization
-Reinforce Occurrences of Generalization
- when you see generalization occur, reinforce it
- initially, the training setting and the criterion setting should be quite similar and gradually become dissimilar
-Train skill that contact naturally occurring contingencies of reinforcement
- must analyze the natural contingencies at the outset of behaviour modification
- train skills that have their own naturally occurring reinforcers
- If the particular skills have no naturally occurring reinforcers then train the learner to solicit reinforcement (e.g., FCT)
-Modify contingencies of reinforcement and punishment in the natural environment
- requires control over the natural environment, if they do not have control, the trainer should teach others in the natural environment to reinforce the behavior
- choose, hospitals, correction facilities
-Incorporate a variety of relevant stimulus situations during training
- sample all relevant S^D and S^delta for the target behavior
- use a range of relevant and applicable stimuli in training, so the behavior is more likely to generalize across all situations
- i.e., train functionally equivalent stimuli
i. e., train sufficient “stimulus exemplars”
-Stimulus exemplars:
stimuli that represent the range of relevant stimulus situations in which the response should occur after training
-Incorporate common stimuli from form real world (criterion) setting into training setting
-i.e., make the two different contexts similar
bring in kids or something that will be a que in the natural environment
-Train a range of functionally equivalent responses
-Different S^D may require different responses to obtain the same reinforcer
- “general case programming”
- defined as using multiple training examples (stimulus exemplars) that sample the range of relevant stimulus situations and response variations
-Functionally equivalent
responses and stimuli need to be learned
- responses (e.g. lever presses vs nose-pokes) are “functionally equivalent”when they lead to the same outcome
- stimuli (levers vs nose-poke keys) are “functionally equivalent” when they lead to the same outcome “food”
-Incorporate stimuli from training into the criterion (i.e., Real World) setting
- may include people or aspects of the physical environment
- useful if irrelevant scenario (besides the desired SDs) are controlling the behaviour
-Provide prompts/cues in the criterion setting
- e.g., sings, reminders, pictures, etc
- Make use of “Self-generated mediators”
- self-instruction of general rules
- self-generated cues or prompts
- Final Thoughts
- “Teach Loosely”
- i.e., randomize non-important stimuli you don’t want controlling the behaviour
- e.g., rearrange furniture, move to different locations, have different people present
