Which Statistical Test Should I Use? Interactive Test Chooser
Not sure whether you need a t-test, ANOVA, chi-square, Mann–Whitney or regression? Answer a few plain-language questions about your study and get the right test, the alternative if assumptions fail, R code, SPSS menu steps and a link to the matching free calculator.
- PriceFree, no login or sign-up needed
- Your dataProcessed in your browser, never uploaded
- Works onPhone, tablet or computer, any modern browser
- Made byDr. Muhammad Aammar Tufail, Codanics
- Published15 November 2026
- Last updated16 November 2026
📚 How to cite this tool
This tool is free. If it helped your thesis, paper, report or assignment, please cite it. It keeps the tools free and helps other students find them.
Tufail, M. A. (2026). Which statistical test should I use? Interactive test chooser [Online tool]. Codanics. Retrieved 16 November 2026, from https://codanics.com/tools/which-statistical-test/
In-text citation: (Tufail, 2026) · Narrative: Tufail (2026)
The access date is today’s date, filled in automatically. Author: Muhammad Aammar Tufail. Published 15 November 2026.
How to use the Which Statistical Test Should I Use?
- Choose what you want to find out: compare groups, look for a relationship, predict an outcome, measure agreement, compare counts or analyse survival.
- Answer the follow-up questions about your outcome variable, the number of groups and whether the groups are paired.
- If you are not sure whether your data are normal, choose Not sure and check a histogram with the descriptive statistics calculator.
- Read the recommended test, its non-parametric alternative and the assumptions, then copy the R code or follow the SPSS steps.
- Open the matching free Codanics calculator to run the test straight away.
Statistical test selection table
| Goal | Outcome | Groups / design | Parametric test | Non-parametric alternative |
|---|---|---|---|---|
| Compare | Continuous | 1 group vs value | One-sample t-test | Wilcoxon signed-rank |
| Compare | Continuous | 2 independent | Welch’s t-test | Mann–Whitney U |
| Compare | Continuous | 2 paired | Paired t-test | Wilcoxon signed-rank |
| Compare | Continuous | 3+ independent, 1 factor | One-way ANOVA + Tukey | Kruskal–Wallis + Dunn |
| Compare | Continuous | 3+ independent, 2 factors | Two-way ANOVA | Aligned rank transform ANOVA |
| Compare | Continuous | 3+ repeated | Repeated-measures ANOVA / mixed model | Friedman |
| Compare | Binary / categorical | 2+ independent | Chi-square test of independence (Fisher’s exact for small counts) | |
| Compare | Binary | 2 paired | McNemar’s test | |
| Relationship | Two continuous | – | Pearson r | Spearman ρ |
| Relationship | Two categorical | – | Chi-square + Cramér’s V | |
| Predict | Continuous / binary / count | Several predictors | Linear / logistic / Poisson regression | |
| Agreement | Categories / scores / methods | Raters or methods | Cohen’s or Fleiss’ kappa / ICC / Bland–Altman | |
| Survival | Time to event | Groups / covariates | Kaplan–Meier + log-rank / Cox regression | |
Parametric or non-parametric?
Parametric tests (t-test, ANOVA, Pearson) compare means and assume the data, or the residuals, are roughly normal. They are more powerful when that holds. Non-parametric tests (Mann–Whitney, Kruskal–Wallis, Spearman) use ranks, so they cope with skewed data, outliers and ordinal scales. With more than about 30 observations per group, t-tests and ANOVA are fairly robust to non-normality.
New to statistics? Start with the free ABC of statistics series and types of data analysis, then learn to run every test in R with the R with Aammar course.
Frequently asked questions
How do I know which statistical test to use?
It depends on your goal (comparing groups, relationships or prediction), the type of outcome variable (continuous, ordinal, binary, categorical or count), the number of groups and whether the groups are independent or paired. Answer those questions in the tool to get the right test.
When should I use a t-test instead of ANOVA?
Use a t-test for two groups and ANOVA for three or more. Running several t-tests instead of one ANOVA increases the chance of a false positive result.
What if my data are not normally distributed?
With small samples and clearly skewed data, use the non-parametric alternative: Mann–Whitney instead of an independent t-test, Wilcoxon instead of a paired t-test, Kruskal–Wallis instead of one-way ANOVA and Spearman instead of Pearson.
Should I use chi-square or Fisher's exact test?
Use chi-square when expected counts are 5 or more in at least 80% of cells. For small samples, especially 2×2 tables with low expected counts, use Fisher’s exact test.
Is correlation enough to compare two measurement methods?
No. Two methods can be highly correlated but disagree systematically. Use Bland–Altman analysis to measure the bias and limits of agreement.
Does this replace a statistician?
No. It covers the most common designs in theses and papers. For complex designs (nested, multilevel or longitudinal data), talk to a statistician or your supervisor before collecting data.
Do I need to log in or create an account to use the Which Statistical Test Should I Use??
No. There is no login, no sign-up and no email required. Open the page and start using it straight away, for free, on any device.
How do I cite this tool in my thesis or paper?
Use the citation box on this page. In APA 7th edition: Tufail, M. A. (2026). Which statistical test should I use? Interactive test chooser [Online tool]. Codanics. Retrieved [date you used it], from https://codanics.com/tools/which-statistical-test/. BibTeX, RIS (EndNote, Mendeley, Zotero), MLA, Harvard, Chicago and Vancouver formats are also available.
Free guides and courses from Codanics
More free tools
See all free Codanics tools → · Codanics on YouTube · WhatsApp channel