Non-Parametric Tests Calculators
This page presents tools that summarise ranked or non-normal data, providing numerical outputs for comparing groups without relying on distributional assumptions.
-
Friedman Test Calculator
Differences in ranked outcomes across several related conditions can be assessed by comparing how the ranks vary from one condition to another.
Example use: checking whether three different snack flavours receive noticeably different preference rankings from the same group of people.
Inputs: treatment 1 values, treatment 2 values, treatment 3 values, significance level
Outputs: result interpretation, friedman chi-square value, degrees of freedom, probability value, critical value, kendall's w, tie correction applied, comparison, difference in mean ranks, result
Visual: a display showing the chi-square curve with a marker for the observed value
-
Kruskal-Wallis Test Calculator
Ranked outcomes from several independent groups can be compared by examining how far their rank totals differ from what would be expected by chance.
Example use: comparing how long three unrelated groups take to finish a simple household task.
Inputs: group 1 values, group 2 values, group 3 values, significance level
Outputs: h statistic, degrees of freedom, probability value, total sample size, effect size, significant, step explanation, calculation process, ranked data, group rank totals, calculated h value, probability value
Visual: a display showing the chi-square curve with acceptance and rejection regions and a marker for the h statistic
-
Mann-Whitney U Test Calculator
Two independent groups can be compared by ranking all values together and assessing whether one group tends to have higher or lower ranks than the other.
Example use: comparing how many minutes two separate groups spend reading each day.
Inputs: group 1 data, group 2 data, significance level, alternative hypothesis
Outputs: test decision, minimum u value, maximum u value, z score, effect size, probability value, rank sum for group 1, rank sum for group 2, sample sizes
Visual: a display showing the normal curve with acceptance and rejection regions and a marker for the observed z value
-
Wilcoxon Signed-Rank Test Calculator
Paired measurements can be compared by ranking the size of their differences and examining whether positive and negative differences show a consistent direction.
Example use: comparing the time it takes a person to complete a short puzzle before and after a break.
Inputs: sample 1 data, sample 2 data, zero difference method, continuity correction, hypothesis, significance
Outputs: decision, test statistic, positive ranks, negative ranks, z score, probability value, effect size, valid pairs
Visual: a display showing the normal curve with a marker for the z score and the rejection region
Non-Parametric Tests FAQs
Non-parametric tests summarise ranked or ordinal data without assuming a specific distribution, offering alternatives to mean-based methods.
They are applied when data is skewed or sample sizes are small, providing results without requiring normality or equal variance.
Parametric tests rely on means and distributional assumptions, whereas non-parametric tests use ranks and median-based comparisons.
Common options include Mann-Whitney for two groups, Wilcoxon for paired data, and Kruskal-Wallis or Friedman for multiple groups.
Yes. Rank-based methods reduce the influence of extreme values, limiting their effect on the final test statistic.