Imagine you want to find out whether different learning methods affect students’ exam performance. You have three groups:
- Group A studies with flashcards,
- Group B studies with videos,
- Group C studies through group work.
The question is: Is there a difference in the average exam result between these three groups?
A conventional t-test only compares two groups at a time. But what do you do when there are three or more groups? This is where analysis of variance comes in.
What is Analysis of Variance (ANOVA)?
Analysis of Variance (ANOVA)
ANOVA is a statistical procedure used to test whether the means of several groups differ significantly from one another. It decomposes the total variance in the data into two components: between groups and within groups.
Core idea: If the group means differ, then part of the overall variation in the data should also be explainable by group membership.
Key Terms and Concepts
1. Group Means and Overall Mean
In ANOVA, we are interested in how much the means of the individual groups deviate from the overall mean. Large deviations suggest that group membership may have an effect.
2. Decomposition of variance
Between-group variance (SSB)
The variance of the group means relative to the overall mean. It shows how much of the total variance can be explained by group membership.
Within-group variance (SSW)
The variance within each group—that is, how much the individual data points vary within each group.
3. F-statistic
The test statistic in ANOVA is the so-called F-statistic:
A large F-statistic indicates that the group means differ more than would be expected by chance.
Assumptions of ANOVA
For ANOVA to be used meaningfully and correctly, certain assumptions must be met:
Assumptions of one-way ANOVA:
- Normal distribution of the dependent variable within each group.
- Homogeneous variances (equal variances) across the groups.
- Independence of observations (e.g., no repeated measurements of the same person).
- A categorical independent variable (group membership).
- A metric dependent variable (e.g., score, reaction time).
Example:
You want to find out whether people’s stress levels differ depending on their occupational field. You measure stress levels (on a scale from 0–100) in three occupational fields: nursing, IT, and education.
→ ANOVA is possible if the stress levels within each group are normally distributed, the variances are similar, and the groups consist of different people.
Hypotheses in ANOVA
Analysis of variance tests the following hypotheses:
Null hypothesis H0H_0H0: All group means are equal
Alternative hypothesis H1H_1H1: At least one group mean differs
ANOVA does not tell you which groups differ—it only shows whether a difference exists. Post hoc tests are necessary for this (e.g., Tukey’s test), which we will cover in a later section.
When is analysis of variance used?
ANOVA is useful whenever:
- You want to compare more than two groups.
- The dependent variable is metric.
- The independent variable is categorical (e.g., “Group A”, “Group B”).
- You want to know whether group means differ systematically.
Examples:
- Do reaction times differ depending on the time of day (morning, noon, evening)?
- Does income influence consumer behavior (low, medium, high)?
- Are there differences in test performance depending on the degree program?
Types of analysis of variance
One-Way ANOVA (one-way analysis of variance)
One independent variable with several levels (e.g., group membership) and one metric dependent variable.
Example: Comparing exam results across three study groups.
Two-Way ANOVA (two-factor analysis of variance)
Two independent variables. An interaction between the factors can also be tested.
Example: The effect of learning method (A/B/C) and gender (male/female) on exam performance.
Repeated-Measures ANOVA
Here, the same people are measured multiple times — e.g., before, during, and after an intervention.
Example: Measuring blood pressure in the same people at three points in time.
Mixed ANOVA
A combination of between-subjects factors (e.g., group membership) and within-subjects factors (e.g., points in time).
MANOVA (Multivariate ANOVA)
Several dependent variables simultaneously.
Example: Testing whether groups differ not only in stress levels, but also in sleep quality and ability to concentrate.
Limitations of ANOVA
- Mean differences only: ANOVA does not directly examine the distribution, only the means.
- Sensitive to violations of the assumptions (especially when variances are unequal).
- Does not indicate the direction of the difference — post-hoc tests are necessary.
Summary
ANOVA is a powerful tool for examining mean differences between several groups — as long as its assumptions are met. In practice, it is widely used in psychology, education research, medicine, and many other fields.
In the next article, we will take a practical look at how to conduct ANOVA in R, which models and formulas are used, and how to interpret the results.
Q&A – Test Your Knowledge
Question 1:
When should you use ANOVA instead of a t-test?
Answer:
When you want to compare more than two groups.
Question 2:
Which assumption must be met for ANOVA?
Answer:
The dependent variable should be normally distributed within each group.
Question 3:
What does a significant F-value mean?
Answer:
That at least one group mean differs significantly from the others.
Question 4:
Can ANOVA tell you which group differs?
Answer:
No, post hoc tests are required for that.
Question 5:
What is the difference between one-way and two-way ANOVA?
Answer:
In one-way ANOVA, there is only one explanatory factor, while in two-way ANOVA, there are two.