graph LR
pingouin_parametric["pingouin.parametric"]
pingouin_nonparametric["pingouin.nonparametric"]
pingouin_correlation["pingouin.correlation"]
pingouin_contingency["pingouin.contingency"]
pingouin_pairwise["pingouin.pairwise"]
pingouin_bayesian["pingouin.bayesian"]
pingouin_equivalence["pingouin.equivalence"]
pingouin_multivariate["pingouin.multivariate"]
pingouin_reliability["pingouin.reliability"]
pingouin_pairwise -- "is utilized by" --> pingouin_parametric
pingouin_pairwise -- "is utilized by" --> pingouin_nonparametric
pingouin_parametric -- "complements" --> pingouin_nonparametric
pingouin_nonparametric -- "complements" --> pingouin_parametric
pingouin_bayesian -- "complements" --> pingouin_parametric
pingouin_bayesian -- "complements" --> pingouin_nonparametric
The Core Statistical Tests subsystem is the heart of the pingouin library, providing a comprehensive suite of statistical analysis tools. It adheres to the "Statistical Library/Toolkit Pattern" by organizing functionalities into distinct, specialized modules.
Provides a comprehensive suite of parametric tests, including various ANOVA types (anova, mixed_anova, anova2, anovan, rm_anova) and t-tests (ttest). It serves as the foundation for statistical inference under parametric assumptions.
Related Classes/Methods:
Implements non-parametric statistical tests and related utility functions, such as Median Absolute Deviation (mad, madmedianrule), which are robust to distributional assumptions.
Related Classes/Methods:
Offers a versatile interface (corr) for various correlation analyses, dispatching to specialized methods like percentage bend correlation (percbend), biweight midcorrelation (bicor), Shepherd's pi correlation (shepherd), and partial correlation (partial_corr).
Related Classes/Methods:
Manages statistical analysis for categorical data organized in contingency tables. This includes foundational methods for creating and analyzing dichotomous tables (dichotomous_crosstab) and specific tests like the chi-squared McNemar test (chi2_mcnemar).
Related Classes/Methods:
Orchestrates and executes multiple pairwise statistical comparisons, a common requirement in post-hoc analysis. It provides high-level functions such as pairwise_tests and pairwise_ttests.
Related Classes/Methods:
Implements Bayesian statistical methods, providing tools for Bayesian inference and hypothesis testing, offering an alternative probabilistic framework for analysis.
Related Classes/Methods:
Focuses on statistical equivalence testing, allowing users to determine if two groups or conditions are practically equivalent within a defined margin, rather than just statistically different.
Related Classes/Methods:
Contains functions for multivariate statistical analysis, dealing with relationships among multiple variables simultaneously, such as MANOVA or canonical correlation.
Related Classes/Methods:
Provides tools for assessing the reliability and consistency of measurements, such as inter-rater reliability (e.g., Cohen's Kappa) or internal consistency (e.g., Cronbach's alpha).
Related Classes/Methods: