“Choosing the right test begins with the right questions”: An interactive decision tree for statistical analysis

Selecting an appropriate statistical analysis is a central challenge in empirical research, requiring researchers to connect their question with the measurement level of variables, study design, dependence of observations, and statistical assumptions. Students and researchers may know how to run a procedure yet find it difficult to decide when that procedure is appropriate. We present the Statistical Decision Tree, a free, browser-based tool designed to make this reasoning process visible. The tool guides users through sequential questions about their analytical goal and data, retains their answers as an inspectable decision path, and recommends a compatible procedure. Users can also search directly for a procedure and view the associated path. Recommendations link to explanations of assumptions, examples from psychology and sport science, downloadable datasets, jamovi and R instructions, effect-size guidance, and reporting scaffolds. The tool covers a broad range of statistical methods, including classical hypothesis tests, regression and mixed models, factor analysis and other latent-variable methods, machine-learning approaches, and meta-analysis. The interface is available in English, German, French, Spanish, and Italian. This article describes the tool’s design rationale and development, then demonstrates its use through two worked examples: a paired-samples t-test and cluster analysis for exploratory discovery. Statistical method selection can often be organized into a constrained sequence of decisions, and the distinctive contribution of the Statistical Decision Tree is to connect those decisions with explanation, execution, and reporting.