Chapter 11 In-Depth Notes on Research Designs

Other Types of Research Designs

• Types: Single-case, quasi-experimental, developmental research

• Purpose:

  • Single-case: For specific individual problems

  • Quasi-experimental: When random assignment is impractical or unethical

Single-case Experimental Designs

• Common Use: In behavioral research

• Focus: Influence of factors on target behaviors (e.g., child behavior, student learning, disorders)

• Process: Baseline → manipulation → reestablish baseline

• Note: Not to be confused with case studies (which are observational)

Example of Idealized Data

• Baseline Data Points:

  • Tantrums

  • Picking up after dog

  • Disruptions in class

  • Snacking behaviors

Study by Sil et al. (2012)

• Subject: Distraction for preschoolers during burn dressing changes

• Conditions: Watching vs. playing a video game (same child)

• Outcomes: OSBD Procedural Distress Scores across sessions

  • Passive vs. Interactive distraction

Single-case Experimental Designs (Additional Detail)

• Reversal Design: Addresses confounds with co-occurring events

  • A-B-A Design: Baseline (A) → Treatment (B) → Baseline (A)

  • Can extend to A-B-A-B or more

Examples of Potential Confounds

• Types of Confounds:

  • Tantrums linked to co-occurring parent actions

  • Phone use impacting interest in a movie scene

  • Classroom disruptions linked to shifts in teacher topics

Analyzing Trends

• Trend Lines can indicate changes independent of interventions

• Variability: Can reveal existing trends before interventions occur

ABA Design Applications

• Can test hypothetical scenarios such as the effects of removing distractions or changing environmental factors

Ethical Considerations for Reversal Designs

• Appropriate Uses: Non-dangerous situations like classroom behaviors

• Inappropriate Uses: Dangerous conditions like self-harm or addiction

Multiple Baseline Designs

• Concept: Introduce interventions at different times to observe effects

• Purpose: Supports causal inferences regarding behavior changes

• Example: Effects of smoking during pregnancy on fetuses noted through ultrasound changes

Study by Alvero & Austin (2004)

• Focus: Workplace safety behaviors across 8 students in a controlled environment

• Behaviors Emphasized: Proper lifting techniques, sitting posture, and phone use ergonomics

• Intervention Types:

  • Information sheet outlining safety procedures

  • Observation and checklist evaluations of behaviors

Results from Alvero & Austin Study

• Percentages of Safe Practices Monitored Over Time:

  • Data shown for interventions such as information sessions vs. observational learning

Behavior Change Strategies

• Communicative Approaches: Using positive reinforcement for desired behaviors, such as increased glove usage for safety

Study by Perrin et al. (2011)

• Investigated procrastination among graduate students over 10 weeks

• Access to Quizzes: Noncontingent vs. contingent access methods

• Findings: Significantly higher gains in scores were noted for students with contingent access to quizzes

Importance of Changing-Criterion Designs

• Purpose: Gradually shift criteria to improve behaviors over time, illustrated with token systems

Various Reinforcement Schedules

• Types:

  • Fixed Ratio: Predictable reward (e.g., jack-in-the-box)

  • Variable Ratio: Unpredictable reward (e.g., slot machines)

  • Fixed Interval: Time-based (e.g., toaster)

  • Variable Interval: Randomized timing (e.g., first behavior after a period)

Replications in Single-case Designs

• Important for establishing reliability and control in experimental procedures

Quasi-experimental Designs

• Lack of random assignment; valuable in applied settings (e.g., classrooms, workplaces)

One-group Designs

• Posttest-only design: Lacks comparison group; limited causal inferences

• Pretest-posttest design: Comparison before and after manipulation, though subject to various confounding problems

Challenges in One-group Designs

• Issue Types:

  • History effects: External events during the study

  • Maturation effects: Natural changes over time

  • Testing effects: Influence of pretests on participant behavior

  • Instrument decay: Changes in measurement effectiveness

  • Regression toward the mean: Trends towards average scores

Nonequivalent Control Group Designs

• Compares experimental group with a non-randomized control group; still faces confounding risks due to selection biases

Interrupted Time Series Design

• Monitors variable over time before and after manipulation; can suffer from misinterpretation biases

Developmental Research Designs

• Methods to study age-related changes:

  • Cross-sectional Method: Observing different age groups at one point in time

  • Longitudinal Method: Following the same individuals over a period

  • Sequential Method: Combination of both methods to enhance reliability

Longitudinal and Sequential Advantages

• Better control of confounding variables and provides insight into longitudinal changes of behavior across age groups.

• Attrition: Considerations and challenges of maintaining participant groups over time

Summary of Developmental Changes through Sequential Studies

• Notable trends in self-esteem and other behavioral attributes through observational designs.

Conclusion on Experimental Research Designs

• Each design has unique strengths and limitations depending on the context and ethical considerations of the research setting.