What Is External Validity?
External validity refers to the extent to which the findings of a study can be generalized beyond the specific conditions of that study. In ABA, this means asking whether an intervention that worked for one client in one setting will also work for other clients, in other settings, with other behaviors, and when implemented by other practitioners.
While internal validity asks "Did the intervention cause the change?", external validity asks "Will the same intervention produce similar results under different conditions?" Both are essential for building a robust evidence base in applied behavior analysis.
Internal Validity vs. External Validity
Internal and external validity are often described as existing in tension. Tightly controlled studies tend to have high internal validity but may sacrifice external validity because the controlled conditions do not represent real-world complexity. Conversely, studies conducted in natural settings may have greater external validity but face more threats to internal validity.
In ABA, the emphasis has traditionally been on internal validity (demonstrating functional relationships for individual clients). External validity is established gradually through replication across multiple studies, clients, settings, and behaviors.
Replication in ABA
Direct Replication
Direct replication involves repeating a study using the same procedures, setting, and similar participants. The goal is to determine whether the original findings are reliable. If the same intervention produces similar results when the study is repeated, confidence in the finding increases.
Within a single study, single-subject designs incorporate direct replication. For example, the second treatment phase (B2) of an ABAB design replicates the first treatment phase (B1). Each tier of a multiple baseline design also serves as a replication of the treatment effect.
Across studies, direct replication means that different researchers reproduce the original procedures as closely as possible and achieve similar results.
Systematic Replication
Systematic replication involves repeating a study while deliberately varying one or more aspects: the participants, the setting, the behavior, the implementer, or other conditions. The goal is to determine the generality of the findings. If an intervention produces positive results across different populations, settings, and variations, its external validity is strengthened.
Systematic replication is the primary mechanism through which external validity is established in ABA. No single study can demonstrate that an intervention works universally. Instead, a body of research that includes systematic replications builds the evidence base over time.
Generalization in ABA Practice
Generalization refers to the extent to which behavior change produced by an intervention extends beyond the specific training conditions. Stokes and Baer (1977) identified several types of generalization and strategies for promoting it. Understanding these is essential for both clinical practice and the BCBA exam.
Types of Generalization
Stimulus generalization: The behavior occurs in the presence of stimuli that are similar to, but different from, the training stimuli. For example, a child who learns to greet their therapist begins greeting other adults as well.
Response generalization: Changes in one behavior lead to changes in related but untrained behaviors. For example, teaching a child to request items using sign language might also increase their use of spoken words.
Temporal generalization (maintenance): The behavior change persists after the intervention has been removed. The behavior continues to occur over time without the specific intervention conditions in place.
Strategies for Programming Generalization
Behavior analysts should not assume that generalization will occur automatically. Instead, it should be actively programmed using evidence-based strategies:
- Train sufficient exemplars: Teach across multiple examples of the target skill until the learner generalizes to new, untrained examples. For instance, teach color identification with many different objects rather than just one set of colored blocks.
- Train loosely: Vary the training conditions (materials, instructions, setting, time of day) to prevent the behavior from becoming restricted to a narrow set of stimuli.
- Use common stimuli: Incorporate elements from the natural environment into the training setting so that the learner encounters familiar cues in both contexts.
- Program natural contingencies: Select target behaviors that will be reinforced by naturally occurring consequences in the learner's environment. If the behavior contacts reinforcement outside of training, it is more likely to generalize and maintain.
- Teach self-management: Help the learner monitor and manage their own behavior, reducing dependence on external contingencies provided by the therapist or trainer.
- Use indiscriminable contingencies: Thin the reinforcement schedule so the learner cannot easily distinguish between training and non-training conditions. Intermittent reinforcement during training promotes continued responding in other settings where reinforcement is also intermittent or unavailable.
- Introduce to natural maintaining contingencies: Ensure that the behavior will be supported by naturally occurring reinforcement in the generalization setting.
Generalization Across Settings
One of the most common goals in ABA is for behavior learned in the clinic or therapy room to generalize to home, school, and community settings. Strategies include conducting at least some training in the natural environment, involving caregivers and teachers in training, and ensuring that the skills taught are functional and relevant in those settings.
Generalization Across People
Behavior that occurs only in the presence of the therapist has limited functional value. Programming generalization across people involves training with multiple instructors, using naturally occurring social interactions as training opportunities, and ensuring that others in the learner's life are prepared to support the behavior.
BCBA Exam Considerations
The exam frequently tests your understanding of the difference between direct and systematic replication, the types of generalization, strategies for promoting generalization, and how external validity is established in ABA. Be prepared for scenario-based questions that ask you to recommend strategies for promoting generalization or to identify why an intervention did not generalize to a new setting.
Remember that external validity is not demonstrated by any single study. It is built over time through systematic replication. Every behavior analyst contributes to external validity by carefully documenting their procedures and outcomes, enabling others to evaluate and replicate their work.