Untitled
AN IMPLICIT TECHNOLOGY OF GENERALIZATION
Authors and Institutions
Trevor F. Stokes and Donald M. Baer
The University of Manitoba and The University of Kansas
Overview
Generalization has traditionally been viewed as a passive outcome, whereas discrimination has been actively developed as a technology in educational and psychological settings.
This review proposes that generalization should also be conceptualized actively and developed as a technology to improve behavior change across different settings, people, and behaviors.
Key Concepts
Generalization: Occurrence of relevant behavior under different, non-training conditions (e.g., across subjects, settings, time).
Generalization was traditionally seen as an accidental result of insufficient practice of discrimination technology.
A focus on the need for active programming of generalization techniques is emphasized.
Review Structure
The literature is categorized into nine general headings regarding techniques to assess or program generalization:
Train and Hope
Sequential Modification
Introduce to Natural Maintaining Contingencies
Train Sufficient Exemplars
Train Loosely
Use Indiscriminable Contingencies
Program Common Stimuli
Mediate Generalization
Train "To Generalize"
Summary of Implicit Technology
Generalization is positioned as equally important as discrimination in behavior analysis and educational practices.
The need for systematic approaches and programming to achieve generalization is highlighted.
Categories of Generalization Programming
1. Train and Hope
Description: Most common method in applied behavior analysis where generalization is noted but not actively pursued.
Example: Kifer et al. (1974) studied negotiation strategies for conflict resolution in a simulated classroom, finding that learned behaviors generalized to real-life situations.
Importance: Documents the extent and limits of generalization but often lacks depth in analysis.
2. Sequential Modification
Description: A systematic approach where after demonstrating a behavior change, further modifications are made in non-generalized conditions.
Example: Wahler (1969) showed behavior change in home settings was necessary for analogous changes in school settings.
This tactic formalizes the therapeutic process of programming generalization.
3. Introduce to Natural Maintaining Contingencies
Description: Transitioning behavioral control from educator to natural reinforcement in the environment.
Example: Ayllon and Azrin (1968) demonstrated that interventions could expose behaviors to natural contingencies for maintenance.
Research on dependent environments underscores the importance of understanding natural reinforcement settings.
4. Train Sufficient Exemplars
Description: Training multiple examples to facilitate generalization rather than focusing on a single instance.
Example: Stokes, Baer, and Jackson (1974) trained additional experimenters to enhance generalization beyond initial training conditions.
The diversity of examples serves as a crucial method for facilitating generalization across settings and subjects.
5. Train Loosely
Description: Encourages flexibility in examples and responses to optimize generalization potential.
Example: Schroeder and Baer (1972) found that looser training methods led to greater generalization of vocal imitations in children.
A more informal approach tends to yield broader applicability of learned skills.
6. Use Indiscriminable Contingencies
Description: Making reinforcement contingencies unpredictable so subjects cannot discriminate between reinforcement and non-reinforcement situations.
Example: Schwarz and Hawkins (1970) showed improved behavior in different classes under delayed reinforcement contingent on behaviors from a prior session.
This approach fosters generalization by minimizing perceived discrepancies in reinforcement conditions.
7. Program Common Stimuli
Description: Ensuring that salient stimuli common to both learning and generalization settings are present helps facilitate transfer.
Example: Stokes and Baer (1976) used peers as tutors to improve learning retention in both experimental and generalization environments.
Identifying and integrating common stimuli supports effective generalization programming.
8. Mediate Generalization
Description: Establishing a verbal or symbolic mediator (often language) to bridge learned material to novel contexts.
Example: Risley and Hart (1968) showed that verbal reports of play behavior increased actual engagement in that behavior.
Useful in creating connections between learned behavior and its application in new situations.
9. Train "To Generalize"
Description: Engaging in contingencies that reward the act of generalization itself, treating it as an operant.
Example: Goetz and Baer (1973) rewarded children for producing novel block forms to encourage a pattern of creating diverse forms through reinforcement.
This technique highlights generalization as an explicit target for teaching efforts.
Conclusion and Recommendations
The generalization literature, while indicating some instances of natural generalization, shows that programming techniques are often necessary for achieving desired behavioral changes longitudinally.
There is a call for further systematic research into each of these techniques to develop a coherent technology of generalization that can be applied in therapeutic and educational settings.
Recommended tactics include fostering natural community reinforcements, training diverse exemplars, loosening control during training sessions, obscuring contingency limits, and enhancing verbal capabilities to serve as mediators for generalization across scenarios.