One-way ANOVA
One-way ANOVA compares the means of three or more independent groups (for example, stress in students from three majors) to see whether at least one is different. If the data aren’t normal, use Kruskal-Wallis. Paste each group’s scores in its own box; you can add up to 8 groups.
One per line, or separated by commas or spaces.
The calculation runs in your browser. We don’t store what you type.
How it’s calculated
Total variation is split into variation between groups and within each group. F = mean square between groups / mean square within groups, with k − 1 and N − k degrees of freedom.
Levene’s test (mean-centered, as in SPSS) checks whether variances are equal. If its p is below 0.05, Welch’s ANOVA is reported: it doesn’t assume equal variances.
Eta squared (η²) = sum of squares between groups / total sum of squares: the share of variation the grouping explains.
Pairwise comparisons use t with the ANOVA’s pooled variance and the Bonferroni correction (each p-value is multiplied by the number of pairs).
How to read the result
- p < 0.05: at least one group has a different mean. Check the pairwise comparisons to see which.
- Eta squared: 0.01 small, 0.06 medium and 0.14 large.
- If the ANOVA isn’t significant, don’t interpret the pairwise comparisons.
Frequently asked questions
- Does it match SPSS?
- Yes. F, degrees of freedom, p-value, Levene and the Welch statistic match Analyze > Compare Means > One-Way ANOVA. Pairwise comparisons match the Bonferroni Post Hoc option.
- What about Tukey?
- Tukey is another correction for pairwise comparisons. Bonferroni is slightly stricter, so a pair that is significant with Bonferroni will be with Tukey too.
- Can I use it with two groups?
- Yes, and it gives the same p-value as the t-test (F = t²). For two groups the t-test is more commonly reported.
Know what you want to ask?
Build the survey in SurvIQ, share it by link, WhatsApp or QR code, and watch results come in live. The Free plan is free forever.
Create a free survey