In-Depth Notes on the Repeated-Measures t-Test for Two Related Samples
Chapter 11 Learning Outcomes
Understand the structure of a research study for the repeated-measures t-test.
Test mean differences using the repeated-measures t statistic.
Evaluate effect size using Cohen’s d, r², and confidence intervals.
Explain the pros and cons of repeated-measures vs. independent-measures studies.
11-1 Introduction to Repeated-Measures Designs
Repeated-measures design (within-subjects design):
Each participant is tested in both conditions, reducing variability caused by individual differences.
Advantages:
Eliminates the risk of differences due to participant backgrounds.
11-2 The t Statistic for a Repeated-Measures Research Design
Structure:
Similar to single-sample t statistic.
Based on difference scores (D):
Formula: D = X2 - X1
Mean difference (MD) calculation is essential for hypothesis testing.
The Hypotheses for a Repeated-Measures t Test
Null Hypothesis (H0): μD = 0 (No mean difference in population)
Alternative Hypothesis (H1): μD ≠ 0 (There is a mean difference)
11-3 Hypothesis Tests for the Repeated-Measures Design
Steps of Hypothesis Testing:
State the hypotheses and select alpha level.
Locate critical region.
Calculate the t statistic.
Make a decision on H0.
Directional Hypotheses and One-Tailed Tests
Hypotheses can be directional:
H0: μD ≤ 0
H1: μD > 0
Critical region located in one tail of the distribution.
Assumptions of the Related-Samples t Test
Observations in each condition must be independent.
The distribution of difference scores must be normal (especially for small samples).
11-4 Effect Size, Confidence Intervals, Sample Size, and Variance
Effect Size:
Common measures: Cohen’s d and percentage of variance accounted for (r²).
Sample size and variance significantly impact the t statistic:
Larger mean differences increase the numerator (higher t).
Larger sample sizes reduce standard error (smaller standard error increases t).
Reporting the Results of a Repeated-Measures t Test
Include means and standard deviations.
Report t-values with degrees of freedom and significance level (p-value).
Example format: t(df) = value, p < .xx, r² = value.
Variability as a Measure of Consistency
Effect of variability:
Smaller variances (consistent treatment effect) increase statistical significance.
Larger variances may render significant effects non-significant.
11-5 Comparing Repeated- and Independent-Measures Designs
Advantages of repeated-measures:
Requires fewer subjects.
Can study changes over time.
Reduces individual difference variances, enhancing power.
Disadvantages:
Possible order effects (first treatment influencing the second).
Counterbalancing can help control these effects.
The Matched-Subjects Design
Combines benefits of both repeated and independent measures:
Two samples with matched subjects on specific variables (e.g., IQ).
Often requires twice as many participants as a repeated-measures design.
Learning Checks and True/False Statements
Evaluate statements concerning the efficacy and logistics of repeated-measures designs vs. independent measures.
For example:
True: Using the same subjects reduces individual differences across treatments.
SPSS Output for Repeated-Measures Hypothesis Test
Examples of SPSS Output:
Present paired sample statistics, means, standard deviations, and test results.
Showcase how to interpret confidence intervals and significance levels in the output.