Free Statistical Test Selector for Students
Not sure whether you need a t-test, ANOVA, chi-square or a correlation? Answer a few questions about your data and research question to find a suitable test, then check the assumptions before you run it.
- Free to use
- Runs in your browser
- Made for dissertation data
Answer a few questions about your data to shortlist a suitable statistical test.
How to use the statistical test selector
Choosing a test after data collection often causes problems. Ideally you decide during design, because your test affects your sample size and your questionnaire.
- Define your questionAre you comparing groups, testing a relationship or looking at association between categories?
- Identify your variablesNote whether each is categorical or continuous and how many groups you have.
- Check independenceDecide whether groups are separate people or the same people measured more than once.
- Confirm the assumptionsLook at normality, sample size and outliers before relying on the result.
Common tests at a glance
This table covers frequent dissertation scenarios. Your own data may need a different choice, so confirm with your supervisor.
| Question | Parametric test | Non-parametric alternative |
|---|---|---|
| Compare two independent groups | Independent samples t-test | Mann-Whitney U |
| Compare two related measurements | Paired samples t-test | Wilcoxon signed-rank |
| Compare three or more groups | One-way ANOVA | Kruskal-Wallis |
| Relationship between two continuous variables | Pearson correlation | Spearman rank correlation |
| Predict a continuous outcome | Multiple linear regression | Depends on which assumptions fail |
| Association between two categorical variables | Chi-square test of independence | Fisher’s exact test for small counts |
Non-parametric is not second best. It is the right choice when data are ordinal, heavily skewed or samples are very small. Justify the decision in your methodology.
Mistakes that cost marks
Forcing data to fit a test
Pick the test that suits your question and data, not the one you already know how to run.
Ignoring assumptions
Parametric tests assume conditions such as approximately normal residuals and similar variances. Report how you checked.
Running too many tests
Repeated testing raises the chance of false positives. Test what your hypotheses require.
Confusing significance with importance
A small p-value can accompany a tiny effect. Report effect sizes and confidence intervals where possible.
The tool suggests, you decide. A selector narrows your options. It cannot see your data, so check the assumptions yourself in SPSS, R, Jamovi or Excel.
Statistical test selector FAQs
Is the statistical test selector free?
Yes. The tool is free to use and you do not need to pay to find a suitable test.
Does the tool analyse my data?
No. It helps you choose a test. You then run the analysis in software such as SPSS, R, Jamovi or Excel.
How do I know whether my data are normally distributed?
Use histograms, Q-Q plots and a test such as Shapiro-Wilk together, and interpret them with your sample size in mind. Ask your supervisor how strictly your department applies this.
Should I treat Likert data as parametric?
It is debated. Single Likert items are ordinal, so non-parametric tests are often safer. Composite scales are frequently treated as continuous. Follow your department’s norms and justify your choice.
Do I need to justify my test in the methodology?
Yes. Explain why the test suits your question, variables and data, and how you checked the assumptions.
Building your analysis plan?
Pair your test with a clear hypothesis using the hypothesis builder, check your numbers with the sample size calculator, or read our guide to writing up dissertation results.
Get Dissertation Support