Study Notes on Experimental Design Concepts
Between-Subjects versus Within-Subjects Designs
Overview
Design Types: The main distinction is between between-subjects designs and within-subjects designs.
Between-Subjects Design
Definition: Different groups of participants are assigned to different levels of the independent variable (IV).
Experimental Group: Participants in Condition 1.
Control Group: Participants in Condition 2.
Participant Allocation Example: Participants labeled S1, S2, S3 in the experimental group and S4, S5, S6 in the control group.
Posttest-Only Design
**Characteristics: *Participants are randomly assigned to conditions after the independent variable is manipulated*.
**Procedure: **
Randomly assign participants to each level of the independent variable.
Measure the dependent variable for each group.
Example Illustrations:
Figure 10.9: Two groups with the independent variable manipulated at two levels, showing how each measurement on the dependent variable is conducted.
Figure 10.10: Depicts the effectiveness of different note-taking styles (laptop notes vs. longhand notes) on comprehension test scores.
Pretest/Posttest Design
Overview: Involves measuring the dependent variable before and after the introduction of an independent variable.
**Example: ** Participants randomly assigned to mindfulness or nutrition class, measuring Verbal GRE scores before and after.
Figure 10.11: Illustrates scoring methodology applied to both classes.
Within-Subjects Design
Definition: A single group of participants is tested across all treatment conditions, effectively acting as their own control group.
Example: Participants (S1, S2, S3, S4, S5, S6) experience both Condition 1 and Condition 2 of the independent variable.
Repeated-Measures Design
Definition: A specific form of a within-subjects design where the same participants experience all levels of the independent variable.
Example: A participant watches a TV show in two contexts (with friends and alone) and rates the show in both scenarios.
Advantages of Within-Subjects Designs
Participant Equivalence: Since the same individuals participate across conditions, there is less variability attributed to differences between participants.
Fewer Participants Required: More cost-effective in terms of participant recruitment and usage.
Disadvantages of Within-Subjects Designs
Order Effects: These can threaten internal validity and act as confounding variables. Participants may be influenced by the order in which they experience conditions, including:
Practice Effects: Improvement in performance due to repeated exposure.
Carry-Over Effects: The effects of one condition may influence performance in another condition.
Further Considerations in Within-Subjects Designs
Demand Characteristics: When participants know they are experiencing all conditions, it may alter their behavior, thus compromising the integrity of the findings.
Practicality Issues: In some cases, experiencing all levels of the IV might not be suitable or feasible.
Counterbalancing
Purpose: To reduce order effects by varying the order in which participants experience conditions.
Example: If participants watch a TV show both alone and with friends, order is randomized and balanced across test subjects:
One participant may watch alone first, whereas another may watch with friends first.
Factorial Designs
Definition: A factorial design includes two or more independent variables (IVs), each with two or more levels, allowing for complex interaction analyses.
Example Study - Kiesler & Baral (1970):
Research Question: What factors determine romantic behavior towards others?
Factor 1 (IV1): The attractiveness of the other person (good looking vs. moderately good looking).
Factor 2 (IV2): The self-esteem of the participant (high vs. low).
Manipulation of Self-Esteem: Participants were informed of their performance on an intelligence test in encouraging (high self-esteem) or discouraging (low self-esteem) terms.
Dependent Variable: The romantic behavior measured by whether participants asked the confederate out.
Kiesler & Baral Results
Graphical Representation:
A graph demonstrating romantic behavior as a function of the attractiveness of the confederate and the self-esteem of the participant, suggesting moderation of self-esteem on the attractiveness effect.
Results Interpretation: Indicate potential dependencies, categorized by varying self-esteem levels leading to differential romantic behavior responses to the attractiveness of others.