Chapter 6: Withdrawal Designs

Introduction

• Withdrawal designs are basic experimental procedures for demonstrating treatment effects.
• Widely recognized across multiple disciplines, not just in ABA (Applied Behavior Analysis).

Learning Objectives

1. Identify mechanics of withdrawal designs.
2. Describe procedures and uses of ABA and ABAB designs.
3. Describe the principles of prediction, verification, and replication in ABAB designs.
4. Identify advantages and disadvantages of withdrawal designs.
5. Describe adaptations of typical withdrawal designs, including BAB and repeated withdrawal designs (ABABAB).

Withdrawal Designs Overview

• Also known as reversal designs.
• Involves the withdrawal of treatment during one or more phases to demonstrate effects on target behavior.
• Preferred term is "withdrawal" since it describes the mechanics rather than outcomes.
• Goal: Show functional relationship between target behavior and intervention.

ABAB Design

• A: Baseline - data collection with no intervention.
• B: Treatment/Intervention phase - implementing intervention and collecting data.
• A: Baseline - withdrawing intervention and collecting data.
• B: Treatment/Intervention phase - reintroducing intervention and collecting data.
• Helps rule out variables like history and maturation by showing behavior changes only with treatment introduction or withdrawal.

Steps in ABAB Design

1. Collect Baseline Data: Before intervention (labeled A or A1).
2. Introduce Intervention: For a specified time and collect data (labeled B or B1).
3. Withdraw Intervention: For a short period and collect data to see return to baseline levels.
4. Reintroduce Intervention: Collect data to see if intervention affects the target behavior.

Example Graphs

• Functional Relationship Demonstrated:
  • Behavior decreases during intervention and increases when withdrawn.
  • Clear cause & effect shown by data change in relation to intervention phases.
• No Functional Relationship Demonstrated:
  • No significant change in behavior level despite withdrawal of intervention.
  • Intervention may not be responsible for behavior change.

Considerations

• True reversal designs (with active reversal of intervention effects) are rare.
• Withdrawal designs are powerful for showing cause and effect in within-subject designs.
• ABAB is recommended over simple ABA designs in educational/clinical settings to avoid leaving subjects in non-treatment phases.
• Need to evaluate other potential factors affecting behavior change when no functional relationship is shown.