In our teaching and learning endeavors with children, we often are driven by the ultimate outcome and functionality of a skill without even realizing it. Behind this motivation for teaching is the value and importance of generalization, we want our students to be able to learn something in an instructional setting and apply it in a functional setting. Think back to the days when you learned the alphabet. Now think of how easily you were able to learn that A is A, no matter what color it is, how tall it is, what kind of paper it is on, if it was on the fridge or in a book, or who might be asking you about it. And notice how you did not forget that A is A once you mastered the skill. This is generalization.
Difficulties with generalization of skills are well-known in individuals with autism spectrum disorders (ASD) and to those who work with them. These difficulties often will mean that generalization will not just occur, but rather will need to be explicitly programmed and planned for in educational and therapeutic settings. Thus, it is important to think of generalization issues as being the responsibility of the teacher, rather than as a deficit in the child. Individuals with ASD frequently cannot functionally use what they have learned in a structured teaching situation and be able to apply it to other similar settings or with different materials and people. Often times children with ASD will need specific planning for maintenance of a skill and programming that can naturally embed learned skills into functional activities so that the skill is constantly and systematically reinforced over time. It is absolutely essential to program and plan for generalization, the "train and hope" approach (just teaching the skill and hoping it will generalize), is not sufficient.
If you are interested in more information on generalization, start with this article:
Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10, 349-367.
Here are some (not all of them though) strategies for programming for generalization from the Stokes & Baer article:
1. Use naturally reinforcing and occurring materials - Seek to change behaviors that receive reinforcement in the student’s natural environment. For example - learning colors because the child has a favorite color of Popsicle, M&Ms, and ice cream flavor is likely to be more maintained and generalized than learning colors by sorting colored blocks into color bowls.
2. Train Loosely - Adding variety to skills being taught. This will include using a variety of materials in a variety of ways and in a variety of situations. Ideas and approaches used in incidental teaching or naturalistic ABA tend to foster better generalization because these instructional environments more closely resemble the ultimate outcome. Studies have shown that the more naturalistic instructions and presentations of SDs tend to have better learning outcomes to intensive instruction.
3. Train Sufficient Exemplars - Providing many examples of the target response. An example of this is the computer-assisted instructional program, TeachTown: Basics , which has many examples incorporated into every lesson. You will notice many examples of one particular vocabulary word. You will also notice that pictures used in the pretest and posttest are different from the pictures used in the training lessons. Additionally, in the off computer activities there are many ideas that include the use of materials found around the house or classroom.
4. General Case Programming – Use many examples of stimuli, use many teachers, try different settings, and lots of materials.
a. Using a vending machine at local community center, using similar vending machine at school, using another similar one at the grocery store…
b. When teaching car, you would consider pictures of cars, different cars, toy cars, riding in family’s cars, labeling cars on the street, etc.
c. When teaching social skills like saying hi, saying hi to people where you know a name for them, saying hi to people when you don’t have a name for them, pretending to say hi to stuffed animals, pretending to say hi to pictures of friends, having dolls say hi to each other, etc.
Although the term "generalization" is often heard in the ABA (Applied Behavior Analysis) literature, there is no intervention in which generalization is not important, regardless of the philosophy. In seeking interventions for a particular child, it is essential to ask the treatment providers how they will program for and measure generalization, or real outcome. Regardless of the impressive statistics of a treatment program, if the children do not demonstrate generalization in the real world, the results of the treatment program may not be as impressive as they seem. A good resource for learning more about generalization, the research, and strategies for various interventions is Real Life, Real Progress for Children with Autism Spectrum Disorders: Strategies for Successful Generalization in Natural Environments (Whalen, 2009).