π Statistical Test Selector
Answer 5 questions about your data and get a specific statistical test recommendation with the reasoning behind it, assumptions to check, and how to run it in common software. Runs entirely in your browser β no data uploaded.
Choosing the right statistical test: a guide for researchers
Selecting an inappropriate statistical test is one of the most common errors in published research β and one of the most frequently cited by peer reviewers. The choice depends on four key factors: the type of outcome variable (continuous, categorical, ordinal), the number of groups being compared, the relationship between the groups (independent or paired/matched), and whether the data meets normality assumptions. This tool walks through these factors step by step and recommends the specific test that matches your study design.
Parametric vs non-parametric tests
Parametric tests (t-test, ANOVA, Pearson correlation) assume that data are approximately normally distributed. Non-parametric tests (Mann-Whitney, Kruskal-Wallis, Spearman correlation) make no distributional assumptions and are appropriate when normality cannot be assumed β typically with small samples (n < 30), ordinal data, or data with outliers. With large samples, parametric tests are robust to non-normality due to the central limit theorem.
Multiple comparisons
When comparing more than two groups, always use a method that controls for multiple comparisons β ANOVA followed by a post-hoc test (Tukey, Bonferroni, Dunnett) rather than multiple t-tests. Running multiple t-tests at p < 0.05 inflates the type I error rate: with 5 comparisons, the probability of at least one false positive is 1 β 0.95β΅ = 23%.