In-Depth Notes on One Way ANOVA Concepts

Chapter 1: Introduction

Types of One Way ANOVA

  • Between Subjects Design:
    • Involves different participants in each group.
    • Ex: Comparing Manila residents vs. Quezon City residents.
  • Within Subjects Design (Repeated Measures Design):
    • Involves the same participants tested under different conditions.
    • Ex: Comparing scores of the same group under different exam conditions.

Relation to T-tests

  • T-tests compare two means (average scores on a dependent variable or DV).
  • Independent t-tests for comparing different groups (two means).
  • Dependent t-tests (paired samples) where the same group is measured under different conditions (two means).

Analysis of Variance (ANOVA)

  • Used when comparing more than two groups or conditions.
  • One Way ANOVA can compare multiple groups, essential when t-tests are insufficient (only for two groups).

Requirements for One Way ANOVA

  1. One independent variable (IV) at a nominal level with two or more categories.
  2. One dependent variable (DV) at an interval or ratio level.
  3. Independent participants in different groups (between subjects design).

Key Concepts to Remember

  • F-Statistic: In One Way ANOVA, an F-statistic is calculated instead of a t-statistic.
  • The p-value from the F-statistic is compared against a significance level (usually alpha = 0.05) to determine if to reject the null hypothesis.
  • Null Hypothesis for ANOVA: All group means are equal (mu1 = mu2 = … = muk).

Chapter 2: T-test and ANOVA Procedure

T-test Summary

  • T-statistic computed from data derives the p-value for hypothesis testing.
  • Decision to reject or not reject the null is based on p-value > alpha.

One Way ANOVA Process

  • Calculates an F-statistic to compare variance among group means.
  • Procedure is analogous to the t-test, where F-statistic leads to the p-value to test the null hypothesis.

Null Hypothesis in ANOVA

  • H0: mu1 = mu2 = … = muk (all group means are equal).
  • Alternative Hypothesis (H1): At least two group means are different (more than just a pairwise comparison).
  • Correct phrasing for H1: "At least two population means are not equal."

Chapter 3: Post Hoc Tests

Post Hoc Analysis

  • Conducted after rejecting the overall F-test to determine which specific means differ.
  • Types of post hoc tests:
    • Tukey HSD Test
    • LSD (Least Significant Difference) Test

Importance of Post Hoc Tests

  • Required when H0 is rejected to explore specific differences among means.

Chapter 4: Example Data Analysis

Example: Analyzing Neighborliness Based on Social Class

  • Independent Variable (IV): Social class (lower, working, middle, upper).
  • Dependent Variable (DV): Neighborliness scores.

Steps in Reporting

  • Identify IV and DV, formulate hypotheses, compute F and p values.
  • Report findings in the format (F = x.xx, p = x.xx).
  • Conclusion based on comparison of p-value with alpha (0.05).

Chapter 5: Analysis of Variance

Variability Analysis

  • Between-Groups Variance: Differences due to the IV.
  • Within-Groups Variance (Error): Variability within each group that cannot be attributed to the IV.

Significance Testing

  • F-ratio compares the variance (BG variance / WG variance).
  • A higher F indicates a significant relationship, leading to the rejection of H0.

Chapter 6: Example of Rejected Null Hypothesis

Conditions of Analysis

  • Example with marital status as IV and life satisfaction as DV.
  • If p-value < alpha, reject H0, highlighting significant differences among groups.

Reporting Results

  • Emphasize significant differences found and recommend conducting post hoc tests.

Chapter 7: Conclusion

Review Key Takeaways

  • Know the differences between between subjects and within subjects ANOVA.
  • Be familiar with hypothesis formulation and testing, ANOVA table interpretation, and significance and post hoc testing procedures.

Importance of Understanding ANOVA

  • Essential for analyzing variance and making informed conclusions in psychological and social research.
  • Ability to distinguish among different experimental designs and properly analyze data.