Dependent T-Test Cheat Sheet
The dependent t-test, also known as the paired-samples t-test or within-subjects t-test, is used to compare the means of two related groups or measurements from the same group to determine if there is a statistically significant difference.
Core Principles
- Dependence means measurements in one sample are directly related to, paired with, or matched to measurements in another sample (e.g., pre- and post-treatment, repeated measures on the same individual).
- The test focuses on the 'difference scores' (D) between paired observations to assess systematic change.
- It functions like a one-sample t-test performed on these difference scores, comparing the mean difference to a hypothesized population mean difference (usually zero).
Action Steps
- 1. State Hypotheses: Formulate null ($$H_0: \mu_D = 0$$) and alternative ($$H_A: \mu_D \neq 0$$ for two-tailed) hypotheses.
- 2. Set Significance Level and Degrees of Freedom: Choose alpha (e.g., 0.05) and calculate df = n - 1.
- 3. Calculate Difference Scores: For each pair, compute D = X1 - X2 (e.g., Treatment - Control).
- 4. Calculate Mean Difference ($$M_D$$): Sum all D scores and divide by n.
- 5. Calculate Variance/Standard Deviation of Differences ($$s_D^2$$ or $$s_D$$): Use the sum of squared differences (SS_D).
- 6. Calculate Standard Error of the Differences ($$s_{M_D}$$).
- 7. Calculate Observed t-value ($$t_{observed}$$).
- 8. Determine Critical t-value ($$t_{critical}$$) from a t-distribution table using alpha and df.
- 9. Make Decision: If |$$t_{observed}$$| > |$$t_{critical}$$|, reject $$H_0$$. Otherwise, fail to reject $$H_0$$.
- 10. Calculate Effect Size (Cohen's d) and Confidence Interval (CI) for interpretation.
Formulas
- Mean Difference: $$M_D = \frac{\sum D}{n}$$
- Sum of Squares for Differences: $$SS_D = \sum (D - M_D)^2 = \Sigma D^2 - \frac{(\Sigma D)^2}{n}$$
- Sample Variance of Differences: $$s_D^2 = \frac{SS_D}{n-1}$$
- Sample Standard Deviation of Differences: $$s_D = \sqrt{\frac{SS_D}{n-1}}$$
- Standard Error of the Differences: $$s_{M_D} = \frac{s_D}{\sqrt{n}} = \sqrt{\frac{s_D^2}{n}}$$
- Observed t-value: $$t_{observed} = \frac{M_D - \mu_D}{s_{M_D}}$$ (where $$\mu_D$$ is the hypothesized population mean difference, usually 0)
- Cohen's d (Effect Size): $$d = \left| \frac{M_D}{s_D} \right|$$ or $$d = \left| \frac{M_D}{s_D^2} \right|$$
- Variance Explained ($$r^2$$): $$r^2 = \frac{t^2}{t^2 + df}$$
- Confidence Interval: $$CI = M_D \pm (t_{critical})(s_{M_D})$$
Key Terms
- Dependent Samples: Observations where measurements in one sample are directly related to, paired with, or matched to measurements in another sample, such as repeated measures on the same individual.
- Difference Scores (D): The result of subtracting one paired score from another (e.g., pre-test score minus post-test score), used as the basis for analysis.
- Mean Difference ($$M_D$$): The average of all the calculated difference scores, representing the central tendency of the change.
- Standard Error of the Differences ($$s_{M_D}$$): A measure of the variability of the mean difference, indicating how much the sample mean difference is likely to vary from the true population mean difference.
- Cohen's d: An effect size measure that quantifies the magnitude of the difference between the two related means in terms of standard deviations of the difference scores, independent of sample size.
Real World Examples
- A clinical psychologist wants to know if a 10-week Parent Management Training (PMT) program changes oppositional behavior in adolescents diagnosed with ODD.: The psychologist measures oppositional behavior scores (0-40) for 12 adolescents before and after the PMT intervention. A dependent t-test is used on the difference scores to determine if the change in behavior is statistically significant.
- Researchers want to assess if a new medication reduces anxiety levels.: Anxiety scores are recorded for the same group of participants before and after they take the medication. A dependent t-test analyzes the difference scores to see if the medication caused a significant reduction in anxiety.
Timeline
- Week 12: Introduction to Dependent T-Test, assumptions, formulas, and example calculation.
- Week 12: Recitation Assignment 9 and Quiz 10 due.
- Friday 11/13 (Exam Window): Exam 3 covering One Sample T-Test, Independent T-Test, and Dependent T-Test.
- Week 13: Reading Chapter 11, lecture, and recitation on related topics.
- Friday 11/20: Recitation Assignment 10 and Quiz 11 due.
People
- Nate Caluda-Perdue: Instructor for Psychological Statistics 207C
Quiz
- Which of the following is NOT an assumption of the dependent t-test?: Homogeneity of variances between groups
- What does 'dependence' refer to in a dependent t-test?: The measurements in one sample are directly related to or paired with measurements in the other sample.
Related cheat sheets
More like this