Chapter 13: More About Multi-Factor Experiments

Chapter Overview

• Discusses multi-factor experiments in psychology.

Steps in the Research Process

• Step 1: Choose a research question.
• Step 2: Conduct a literature review.
• Step 3: Develop a hypothesis.
• Step 4: Design the study (experiments).
• Step 5: Conduct the study.
• Step 6: Analyze the data.
• Step 7: Report the results.

Factorial Designs

• Used by psychologists for experimenting with multiple independent variables.
• Advantages:
  • Efficiently tests effects of several independent variables.
  • Examines combined effects on the dependent variable.

Example: Design of Factorial Experiment with Daphne

• Independent Variables (IV):
  • Treatment Type: Worn shirt vs. Chew toy.
  • Length of Outing: Short vs. Long.
• Conducted 5 trials for each condition.

Interaction Effects

• Interaction effect is analyzed by comparing differences between rows or columns in data.
• Main Effects:
  • Test the independent effect of each independent variable on the dependent variable.
  • Conducted through ANOVA (Analysis of Variance).
  • Main Effects Examined By:
  • Comparing mean scores of each IV level.

ANOVA

• Tests interaction effects of one IV at each level of another IV to see how they affect the dependent variable.

Condition and Marginal Means Example

• Daphne's Experiment Mean Results:
  

• Each IV has levels that can be averaged to determine overall means and interaction effects.

Specific Factorial Design Example

• Types of therapy and their lengths can lead to different outcomes based on averages of means.

Examples of Interaction Effects

• Different types of interactions displayed through graphs comparing mean depression scores across various therapy types and durations.
• Graph Analysis: Key to identifying types of interactions.

Cognitive Example: Payne, 2006

• Participants judged whether an object was a tool or weapon based on the race of a face seen before the object.
• Findings showed biases based on race when participants were under pressure (snap judgments).

Example Multi-Factor Experiment: Darley and Batson (1973)

• Study focused on seminary students' behaviors after attending a lecture based on urgency and topic:
  • Independent Variables:
  • Topic of the lecture: "Good Samaritan" vs. "Getting a Job".
  • Urgency: Early vs. Running Late.
• Procedure:
  • Students encountered a confederate in need of help on their way to the next task.
  • Responses to help were coded on a scale from 0 (ignored) to 5 (full assistance).

Findings from Good Samaritan Study

• The topic of the lecture did not significantly affect helping behavior.
• Urgency was found to significantly impact whether students helped (more help from those in no hurry).

Common Pitfalls in Factorial Experiments

• Identifying Interactions:

  • Issue: Not recognizing interactions.
  • Solution: Graph condition means to detect interactions via non-parallel lines.

• Understanding Main Effects vs. Interactions:

  • Issue: Difficulty distinguishing between main effects and interactions.
  • Solution: Separate graphs can help clarify the impact of one independent variable on levels of others.

Summary and Conclusions

• Factorial designs allow for experiments with two or more independent variables.
• They are more efficient than separate studies for each effect.
• Significant interaction effects can be visualized through graphs to clarify interactions between variables.