Abstract
Robust statistical analysis is essential for scientific validity and to ensure good scientific practice. Yet many researchers, especially in biomedical fields, struggle with checking assumptions, selecting the correct tests, and interpreting results. These obstacles can lead to misleading conclusions and undermine scientific progress. BioMedStatX explicitly addresses these issues by ensuring that the implemented workflows exclude the use of inadequate statistical tests. This Python-based desktop application features an intuitive graphical interface that automatically selects appropriate statistical tests based on the data and its characteristics, ensuring that users, even with minor statistical training, follow a statistically valid workflow. Users can import Excel or CSV files, select groups and let BioMedStatX manage the rest: from outlier detection, assumption checks and guided data transformations to test execution (parametric or non-parametric) and guided post-hoc analyses. Results are exported in a structured Excel workbook including a decision tree that visualizes each analytical step, and customizable plots are exported as SVG-/PNG-files. By embedding statistical expertise directly into the software, BioMedStatX prevents invalid analysis paths, increases transparency, and enables reproducibility.