Statistics calculators – tests, confidence intervals, sample size
23 statistical tests and calculators in one tile, from the t-test to the randomisation list. Paste your data straight from Excel; everything is calculated in your browser – free, no account.
More about this tool
The tile covers the methods most often needed for theses, papers and grant applications:
- Comparing means: t-test for independent samples (Student, Welch), paired and one-sample t-test, one-way ANOVA.
- Nonparametric tests: Mann–Whitney U, Wilcoxon signed-rank, Kruskal–Wallis.
- Frequencies and tables: chi-square test (2×2 and r×c), Fisher’s exact test, McNemar test, difference of two proportions with Newcombe CI.
- Correlation and regression: Pearson and Spearman, linear regression with standard curve.
- Laboratory: Bland–Altman method comparison, Grubbs outlier test, coefficient of variation.
- Study design and effect measures: sample size for two groups, randomisation list (simple, block, stratified), 2×2 tables for diagnostic tests and for RR, OR and NNT, Cohen’s d and Hedges’ g, confidence interval for a proportion, descriptive statistics.
Data input and accuracy
Where a test needs raw data, you paste a data series – one column from Excel or values separated by spaces, with decimal comma or point. Invalid entries are ignored and counted. The p values come from the program’s own distribution functions, whose accuracy has been checked against reference values from SciPy; the tests themselves are checked against SciPy as well. Every calculator states its formula and source.
Not sure which test fits your data? The test selection assistant tile asks up to five short questions about your data and opens the matching calculator or tile.
The results are for orientation only and do not replace statistical advice; verify results for papers and grant applications with a second software.
Frequently asked questions
Are my data uploaded?
No. All calculations run in your browser and the values you enter are not transmitted. The last values used are stored only locally in your browser.
Which test should I use for two groups?
For roughly normally distributed measurements, the t-test (Welch if the variances may differ); for ordinal data or clearly skewed distributions, the Mann–Whitney U test; for paired measurements, the paired t-test or the Wilcoxon signed-rank test.
All statistics calculators by category
Study design
Diagnostic tests
Effect measures
Confidence intervals
Comparing means
- Two-sample t-test for independent samples (Student, Welch)
- Paired t-test (dependent samples)
- One-sample t-test
- One-way analysis of variance (ANOVA)
Frequencies and tables
- Chi-square test (2×2 and r×c table)
- Fisher’s exact test (2×2)
- McNemar test (paired proportions)
- Difference of two proportions (Newcombe CI)
Descriptive statistics
Nonparametric tests
Correlation and regression
Laboratory
pubmed.info sets no tracking cookies and builds no user profiles. Visits are counted anonymously (cookieless, shortened IP addresses). Ads from Google appear only according to your consent choice.