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.