ANOVA Calculator | One-Way with p-value & Effect Size
Perform one-way ANOVA with up to 6 groups. Shows the full ANOVA table, exact p-value, effect size η², group statistics, and a significance verdict, no F-table look-up needed.
What Is the ANOVA Calculator | One-Way with p-value & Effect Size?
Analysis of Variance (ANOVA) tests whether the means of two or more independent groups are equal. It works by comparing the variability between groups to the variability within groups. If groups differ meaningfully, the between-group variance will be much larger than the within-group variance, producing a large F-statistic and a small p-value.
- ▸Up to 6 groups, add and remove groups dynamically; each group can have its own label.
- ▸p-value computed exactly, using the regularized incomplete beta function (not an approximation table).
- ▸Full ANOVA table, SS, df, MS, F, and p-value for both between-group and within-group sources.
- ▸Group statistics, n, mean, standard deviation, min, and max for each group.
- ▸Effect size η², tells you how much of total variance the factor explains (small / medium / large).
- ▸Significance level selector, test at α = 0.01, 0.05, or 0.10.
Formula
One-Way ANOVA decomposes total variance into two components, between-group variation (explained by the factor) and within-group variation (random error). Their ratio forms the F-statistic.
SSB = Σᵢ nᵢ(x̄ᵢ − x̄)²
How far each group mean is from the grand mean, weighted by group size.
SSW = ΣᵢΣⱼ (xᵢⱼ − x̄ᵢ)²
How spread out the values within each group are around their group mean.
MSB = SSB/(k−1) MSW = SSW/(N−k) F = MSB / MSW
k = number of groups, N = total observations. Large F → groups differ.
η² = SSB / SST
η² < 0.06 = small, 0.06–0.14 = medium, > 0.14 = large effect.
How to Use
- 1
Enter your group data
Type the values for Group 1 and Group 2 as comma-separated numbers. Click "+ Add Group" to add up to 6 groups. Each group needs at least 2 values.
- 2
Label your groups (optional)
Click the group name (e.g. "Group 1") to rename it, useful for documenting what each group represents (Diet A, Treatment B, etc.).
- 3
Choose a significance level
Select α = 0.01, 0.05 (default), or 0.10. This is the threshold below which you reject the null hypothesis that all group means are equal.
- 4
Click Run ANOVA
The calculator computes SSB, SSW, MSB, MSW, the F-statistic, and the exact p-value. It also shows descriptive statistics for each group.
- 5
Interpret the results
If p < α, the difference is statistically significant. Check η² for effect size. If significant, use post-hoc tests (Tukey HSD or Bonferroni) to find which groups differ.
Example Calculation
Example, Comparing three diet plans
Weight loss (kg) after 8 weeks for three groups on different diets.
Understanding ANOVA | One-Way with p-value & Effect Size
What Is ANOVA?
ANOVA (Analysis of Variance) is one of the most widely used statistical tests in research. It determines whether the means of three or more independent groups differ more than would be expected by random chance. Despite its name containing "variance," ANOVA is ultimately a test about means, it uses variance as a tool to assess whether observed differences in group averages are real or just noise.
How ANOVA Decomposes Variance
The core insight of ANOVA is that total variance in a dataset can be split into two meaningful parts:
- ▸Between-group variance (SSB): Reflects how much group means differ from the overall mean. If all groups have the same mean, SSB ≈ 0.
- ▸Within-group variance (SSW): Reflects natural variability within each group, random measurement error and individual differences. This is the noise floor.
The F-statistic is the ratio MSB/MSW. If between-group variation is large relative to within-group variation, F will be large and the p-value will be small, indicating that the group differences are real.
When to Use ANOVA
| Scenario | Appropriate test |
|---|---|
| Comparing 2 independent groups | Independent samples t-test |
| Comparing 3+ independent groups | One-Way ANOVA ← this calculator |
| Comparing 2 factors simultaneously | Two-Way ANOVA |
| Repeated measures (same subjects) | Repeated Measures ANOVA |
| Non-normal data, 3+ groups | Kruskal-Wallis test (non-parametric) |
| Unequal variances, 3+ groups | Welch's ANOVA |
ANOVA Assumptions and How to Check Them
- ▸Independence: Each observation must come from a separate subject. Check your study design, repeated measurements on the same person violate this.
- ▸Normality: Each group should follow an approximately normal distribution. Test with Shapiro-Wilk (small samples) or check Q-Q plots. With n > 30 per group, the Central Limit Theorem means mild non-normality is acceptable.
- ▸Homoscedasticity (equal variances): All groups should have similar variance. Test with Levene's test. If violated, use Welch's ANOVA instead.
Effect Size, What η² Tells You
A statistically significant ANOVA result does not tell you how large the effect is. With large enough samples, even trivially small differences become significant. Effect size measures how meaningful the difference is in practice.
| η² range | Interpretation | Example context |
|---|---|---|
| 0.01 – 0.06 | Small effect | Subtle differences, may need large N to detect |
| 0.06 – 0.14 | Medium effect | Practically noticeable in real-world settings |
| > 0.14 | Large effect | Strong, obvious group differences |
Post-Hoc Tests After a Significant ANOVA
When ANOVA is significant, run a post-hoc test to identify which specific pairs of groups differ:
- ▸Tukey HSD: Compares all pairs; controls family-wise error rate. Best for equal group sizes and balanced designs.
- ▸Bonferroni correction: Divides α by the number of comparisons. Conservative but widely applicable.
- ▸Scheffé's test: Most conservative; suitable for complex contrasts beyond simple pairwise comparisons.
- ▸Games-Howell: Like Tukey HSD but works when variances are unequal across groups.
Real-World Applications of ANOVA
- ▸Medical research: comparing the effectiveness of three or more drug treatments or dosage levels
- ▸Agricultural science: testing crop yields under different fertilizer conditions (the original use case by Ronald Fisher in the 1920s)
- ▸Education: comparing student performance across multiple teaching methods or curricula
- ▸Manufacturing quality control: checking whether product measurements differ across production lines or shifts
- ▸Marketing: testing consumer response to multiple ad variants or pricing strategies (A/B/C testing)
- ▸Psychology: comparing behavioral outcomes across experimental conditions
Frequently Asked Questions
Why use ANOVA instead of multiple t-tests?
Running k(k-1)/2 individual t-tests inflates the family-wise error rate. With 3 groups and α = 0.05, three t-tests give a ~14% chance of at least one false positive. ANOVA controls the error rate at exactly α regardless of the number of groups by testing all means simultaneously.
What does the F-statistic actually measure?
F = MSB / MSW is the ratio of between-group variance to within-group variance. When groups are truly equal (H₀ true), both numerator and denominator estimate the same population variance and F ≈ 1. When group means differ, MSB becomes larger while MSW stays the same, pushing F above 1. The larger F, the stronger the evidence against H₀.
What are the assumptions of one-way ANOVA?
Three key assumptions: (1) Independence, observations within and across groups must be independent. (2) Normality, the data in each group should be approximately normally distributed (by the Central Limit Theorem, this matters less with large samples). (3) Homoscedasticity, the variances of all groups should be roughly equal. Violations can be addressed with Welch's ANOVA (for unequal variances) or Kruskal-Wallis test (non-parametric alternative).
What is eta-squared (η²) and how do I interpret it?
η² = SSB / SST measures the proportion of total variance explained by the grouping factor. η² = 0.10 means the factor accounts for 10% of all variation in the outcome. Conventional benchmarks: < 0.06 = small effect, 0.06–0.14 = medium, > 0.14 = large. Unlike p-value, η² does not depend on sample size.
My ANOVA is significant, which groups are different?
ANOVA only tells you that at least one pair of group means differs, not which pair. You need a post-hoc multiple comparison test. Tukey HSD (Honest Significant Difference) is the most common, it compares all pairs while controlling the family-wise error rate. Bonferroni correction divides α by the number of comparisons and is more conservative. Scheffé's method is the most conservative but also the most flexible.
What is the difference between one-way and two-way ANOVA?
One-way ANOVA tests one independent variable (factor) with multiple levels. Two-way ANOVA tests two factors simultaneously and can detect interaction effects, whether the effect of Factor A depends on the level of Factor B. For example, does a drug's effect differ by gender?
Can ANOVA handle unequal group sizes?
Yes. The standard one-way ANOVA formula naturally handles unequal group sizes (unbalanced design) because SSB = Σ nᵢ(x̄ᵢ − x̄)² weights each group mean by its sample size. However, unequal sizes make ANOVA more sensitive to violations of the homoscedasticity assumption.
What is Welch's ANOVA?
Welch's ANOVA is a modification that does not assume equal variances across groups. It uses a corrected F-statistic and adjusted degrees of freedom. It is recommended when Levene's or Bartlett's test suggests the variances are significantly different across groups.