Basic Design Concepts in Psychology

Introduction to the Study

• Speaker: Dr. Amber Ankowski, Psychology 100B

• Context: Review of research question, hypothesis, and experimentation basics.

Research Question and Hypothesis

• Research Question: Does our environment influence our well-being?

• Hypothesis: A cluttered environment causes stress.

Experiment Basics

• Hypothesis: A cluttered environment causes stress.

• Independent Variable (IV): Clutter

• Dependent Variable (DV): Stress

Defining the Clutter IV

• Initiative to operationally define what constitutes a "cluttered environment."

• Various definitions can include:
    - Visual clutter
    - Smelly clutter
    - Multiplicity of sensory stimuli (e.g. dirty kitchen)
    - Laboratory manipulations of clutter.

Defining the Chaos IV

• Selected Definition: Participants sitting in a room with dirty dishes scattered around.

• Contradiction: Instead of merely observing stress responses, multiple comparison groups are necessary to support findings.

Quantitative vs. Qualitative Measures

• Qualitative Measures: Assessment of types of clutter (e.g., dirty coffee cups vs. dirty plates).

• Quantitative Measures:
    - Manipulation of the number of dirty dishes (e.g., 0 vs. 7).
    - The necessity to specify operational definitions clearly.

Defining the Stress DV

• Types of Measures for Stress:
    - Self-report: Participants rate their stress level.
    - Behavioral Observations: Noting physical stress indicators (e.g., fidgeting, nail-biting).
    - Physiological Measures: Heart rate and cortisol levels.

• Measurement Scales:
    - Nominal: Categorizing participants as "Stressed" or "Not Stressed."
    - Ordinal: Ranking by stress levels (e.g., “not stressed” to “very stressed”).
    - Interval: Rate stress on a scale from 1 to 50.
    - Ratio: Count observed stress behaviors.

Considerations for Reliability

• Test-Retest Reliability: Measures should yield consistent results over time (e.g., stress surveys).

• Interrater Reliability: Consistency across different raters (e.g. multiple judges assessing stress).

Validity Considerations

• Construct Validity: Evaluating if the measure of clutter or stress adequately represents each concept.
    - Example: Are dirty dishes an accurate measure of clutter?

• External Validity: The generalizability of findings based on how variables and populations are defined.

• Internal Validity: Effects of the cluttered environment on stress must be conclusively established.

Extraneous Variables (EV)

• Definition: Any variable not classified as either IV or DV that could influence the outcome.

• Examples of EVs in Study:
    - Time of day
    - Color of dishes
    - Participant age

Controlling Extraneous Variables

• Methods to Control EVs:
    - Keep It Constant: Maintain the same conditions for all subjects (e.g., type of clutter).
    - Vary Randomly: Distributing levels randomly among participants.
    - Counterbalance: Balancing conditions across different trials to mitigate EV influences.

Example Study Context

• Topic: The influence of a drug on rat running speeds through a maze.

• Independent Variable (IV): Type of drug (Placebo vs. Drug).

• Dependent Variable (DV): Time (in seconds) for running the maze.

• Controlling EV Example: Monitor age of rats to ensure consistency among subjects.

Specific Strategies for Controlling EVs

• Keep Constant: Use rats of a consistent age to ensure uniformity but may limit external validity.

• Vary Randomly: Include a variety of ages, randomly assigning conditions but ensure spread across age groups is maintained.

• Counterbalance: Maintaining control across different trials ensures that age does not confound results.

Within-Subjects vs. Between-Subjects

• Within Subjects Design: Each participant experiences all levels of IV.
    - Example: All rats receive both placebo and active drug treatment.
    - Pros: Stronger test of hypothesis, less costly, fewer participants.
    - Cons: Possible order effects and complexities in managing extraneous variables.

• Between Subjects Design: Participants only experience one IV level.
    - Example: Different groups of rats receiving either the placebo or drug.
    - Pros: No order effects, easier design.
    - Cons: Less statistical power, higher participant costs, risk of unequal group characteristics due to sampling.

Addressing Participant and Experimenter Bias

• Experimenter Bias: Can be intentional or unintentional, influencing outcome.
    - Example: Experimenter’s knowledge of conditions could skew participant responses.
    - Mitigation Strategy: Researchers should remain blind to conditions to preserve objectivity.

• Subject Bias: Participants may alter behavior, consciously or unconsciously.
    - Double-Blind Design: Participants and experimenters are unaware of conditions.

Materials Bias

• Impacts of Bias in Assessing IV on DV:
    - Ceiling Effect: When a DV does not adequately resolve differences among IV levels due to high saturation.
    - Floor Effect: When a DV similarly lacks resolution at low performance levels.
    - Demand Characteristics: Participants respond based on perceived intent rather than true feeling/experience (e.g., social desirability).

Sampling Procedures

• Population Definition: Understanding which groups of people findings are intended to generalize to.

• Sampling Methods:
    - Simple Random Sampling: Equal selection opportunity from the entire population.
    - Convenience Sampling: Participants volunteer, influencing representativeness.
    - Cluster Sampling: Ensures representation from defined groups (clusters) within the population.
    - Stratified Proportional Random Sampling: Ensures subgroups are proportionate to their occurrence within the population.

Conclusion

• Random Sampling vs. Random Assignment: Crucial to differentiate study recruitment (random sampling) from participant allocation to various levels of IV (random assignment).

• Final Notes: Encourage questions and lab attendance to further explore concepts covered in the lecture.