The multiple baseline design, a widely used single-subject research methodology, offers several variations to demonstrate experimental control by introducing an intervention sequentially across different baselines. These variations are primarily categorized by how the baselines are arranged and when the intervention is applied.
Core Categorizations: Concurrent and Nonconcurrent Designs
Multiple baseline studies are fundamentally categorized into two main types based on their temporal structure:
- Concurrent Designs: These are considered the traditional approach to multiple baseline studies. In a concurrent design, baseline measurements for all participants, behaviors, or settings begin at approximately the same moment in real time. The intervention is then introduced sequentially across these baselines, allowing for a clear demonstration of control as changes occur only when the intervention is applied to each respective baseline.
- Nonconcurrent Designs: In contrast, nonconcurrent designs do not require all baselines to start at the same time. Instead, the baseline data for each participant, behavior, or setting is collected and then the intervention is applied. Subsequent baselines are then started and intervened upon, often at different points in real time. This approach can be more flexible for practical application, though it may present challenges in ruling out certain historical confounds if significant time passes between the start of different baselines.
Principal Multiple Baseline Variations
Beyond the concurrent/nonconcurrent distinction, multiple baseline designs are typically differentiated by the dimension across which the intervention is staggered:
1. Multiple Baseline Across Individuals (Participants)
This is one of the most common variations, involving the application of an intervention to the same behavior exhibited by two or more different individuals.
- How it Works: Baseline data is collected on the target behavior for all participants. The intervention is then introduced to the first participant while others remain in baseline. Once a change is observed, the intervention is introduced to the second participant, and so on.
- Example: A study aiming to reduce disruptive classroom behavior might introduce a new positive reinforcement strategy to one student first, then to a second student, and finally to a third, all while tracking their disruptive behaviors.
- Practical Insight: This design is excellent for demonstrating the effectiveness of an intervention across different people, suggesting its generalizability.
2. Multiple Baseline Across Behaviors
This variation involves applying an intervention to two or more different behaviors exhibited by the same individual or group.
- How it Works: Baseline data is collected for several distinct behaviors of a single individual. The intervention is then applied to the first behavior, while other behaviors remain in baseline. After the first behavior changes, the intervention is applied to the second behavior, and so forth.
- Example: For an individual with autism, an intervention to improve social greetings might first be applied to "making eye contact," then to "saying hello," and finally to "initiating a brief conversation," all within the same social interaction context.
- Practical Insight: This variation helps determine if the intervention's effect is specific to the target behavior, rather than a general effect on the individual.
3. Multiple Baseline Across Settings
In this variation, the same behavior of the same individual (or group) is targeted across two or more different environmental settings.
- How it Works: Baseline data for a specific behavior is collected in various settings. The intervention is then introduced in the first setting, while the behavior in other settings remains in baseline. Once the behavior changes in the first setting, the intervention is introduced to the second setting, and so on.
- Example: A child's tantrum behavior might be targeted first in the home setting, then in the school setting, and finally in a community setting, with the intervention introduced sequentially across these environments.
- Practical Insight: This design demonstrates whether an intervention's effect generalizes across different environments or if environmental variables play a crucial role.
Summary Table of Multiple Baseline Variations
| Variation | Description | Primary Focus | Example Use Case |
|---|---|---|---|
| Concurrent Design | Baselines for all tiers start simultaneously in real-time. | Traditional and robust demonstration of experimental control. | Implementing a social skills program for multiple students starting on the same date. |
| Nonconcurrent Design | Baselines for different tiers start at different, non-overlapping times. | Flexibility in implementation when simultaneous starts are not feasible. | A therapist introduces a stress-reduction technique to clients as they join a program over several months. |
| Across Individuals | Intervention applied to the same behavior across different participants. | Effectiveness across people; generalizability. | Teaching three different students to follow instructions using the same technique. |
| Across Behaviors | Intervention applied to different behaviors of the same individual. | Specificity of intervention effect; avoiding co-variations. | Helping an individual improve eye contact, then verbal greetings, then conversational turn-taking. |
| Across Settings | Intervention applied to the same behavior of the same individual (or group) in different environments. | Generalization of effects across contexts; influence of environmental factors. | Reducing a child's screaming behavior first at home, then at school, and finally in the supermarket. |
These variations provide researchers and practitioners with flexible yet rigorous methods to evaluate interventions' effectiveness in real-world settings, offering strong internal validity without requiring a withdrawal or reversal of the intervention. For more detailed information on research designs, resources like Purdue OWL's guide to research methods can be helpful.
