How do I know which statistical test to use?

Start with the research objective, outcome type, number and relationship of groups, study design, distribution, and assumptions. The interactive decision tree narrows these choices and presents suitable methods with requirements and limitations.

What is the difference between parametric and nonparametric tests?

Parametric tests make assumptions about distributions and parameters, while nonparametric tests generally rely on ranks, signs, or counts and can suit ordinal, skewed, or small-sample data. The correct choice depends on the design and assumptions, not normality alone.

Does the decision tree replace statistical advice?

No. It supports method selection but does not replace a protocol, subject-matter expertise, data-quality review, or specialist statistical advice for complex, regulated, or high-stakes research.

Statistical Test Selection Decision Tree

Choose your research objective to narrow the available tests, models, calculators, visualizations, and advanced research tools.

Summarize and Explore Data

Generate descriptive statistics, explore distributions, and understand your data characteristics.

Compare Group Differences

Choose tests for one, two, or three or more independent or related groups.

Predict Outcomes

Select models for continuous, binary, survival, count, nominal, ordinal, or time-series outcomes.

Examine Relationships

Explore correlations and associations among continuous, categorical, or mixed variables.

Visualize Data

Choose bar, pie, histogram, box, raincloud, error-bar, scatter, or Sankey visualizations.

Estimate Confidence Intervals

Calculate intervals for means, proportions, differences, correlations, regression, diagnostic accuracy, survival, and variances.

Plan Sample Size & Power

Choose one-sample, two-sample, paired, multi-group, or advanced study-design calculators.

Epidemiology Calculators

Navigate case-control, cohort, cross-sectional, diagnostic, survival, meta-analysis, outbreak, standardization, and screening tools.

Latent & Causal Modeling

Explore SEM, EFA, CFA, mediation, and moderation analysis.

Multivariate & Reliability Methods

Explore reliability, clustering, MANOVA/MANCOVA, multilevel models, multivariate linear models, and PLS.

Evidence, Causal & Quality Tools

Open difference-in-differences, effect conversion, risk-of-bias, variable-tree, and Six Sigma tools.