NEWS.md

In AssumpSure: 'shiny' Application for Statistical Test Assumption Checking and Guidance

AssumpSure 1.1.3

Bugs fixes

• Fixed detection of numeric-coded grouping variables for random effects (e.g., participant IDs coded as 1, 2, 3) in LMM input.
• Added a new helper function safe_run() to handle computation interruptions gracefully in Shiny, preventing errors when users stop a process manually.

AssumpSure 1.1.2

Bug fixes

• Corrected the name of the first category in Chi-square and Fisher tests.
• Corrected error messages for Mann–Whitney U test when fewer than 2 groups or more than 2 groups are detected.
• Corrected error message for one-way ANOVA when fewer than 3 groups are detected.
• Added error messages for Kruskal–Wallis test when only 1 or 2 groups are detected.

New features

• Added model metrics matrix for LM, LMM, GLM, and zero-inflated models showing key fit statistics (AIC, BIC, logLik, etc.).

AssumpSure 1.1.1

Bug fixes

• Fixed detection of numeric-coded categorical variables (e.g., groups coded as 1, 2, 3).
• Fixed incorrect detection of numeric-coded categorical variables in Poisson, negative binomial, and zero-inflated negative binomial models.
• Improved handling of categorical variables to accept up to 10 levels.
• Corrected error notifications to more clearly distinguish between numeric value columns and categorical group columns.
• Corrected the interpretation of the Shapiro–Wilk test to make it clearer.

New features and user guidance

• Added a clear message explaining that the "categorical" are limited to 10 levels to ensure ANOVA and Kruskal–Wallis results remain reliable and interpretable.
• Added a clear message explaining that the "categorical" are limited to 15 unique levels to improve validity and interpretability of Fisher’s exact and Chi-square tests.
• Added asterisks to the boxplots and a tooltip explaining their interpretation.

AssumpSure 1.1.0

Bug fixes

• Fixed incorrect statistical test recommendations in some cases.
• Logical columns are now detected correctly.
• Fixed a bug where continuous variables were sometimes not recognized, which blocked parametric and nonparametric tests.
• Prevented Chi-square tests when any categorical variable has only one level, with a clear user message.
• Correlation heatmap now accepts the prevalence filter input
• Corrected handling of covariates with perfect separation or near-zero variance, with informative messages and robust plotting.
• Fixed intermittent failures of the bicor correlation method.

New features and user guidance

• Added a clear message explaining that the "continuous" type does not accept count data. Only truly continuous variables are allowed.
• Added guidance for normality testing with small sample sizes.
• Reformatted test outputs to be easier to copy into spreadsheets.
• Added tooltips on how to interpret QQ plots and histograms.
• Improved test selection: if variances are unequal but normality holds, recommend Welch’s tests instead of Mann–Whitney or Kruskal–Wallis or One-Way ANOVA, depending on design.
• Added effect size estimates for t-tests, Mann-Whitney, Wilcoxon rank sum, one-way ANOVA, and Kruskal Wallis.
• Added guidance when bicor analysis fails.
• Added a tooltip explaining interpretation of correlation coefficients.
• Added a Shapiro–Wilk test below the histogram for checking the normality of the dependent variable, to help users with interpretation.
• Added tooltips for interpreting regression model assumptions.
• Added Poisson regression with a dedicated overdispersion test and guidance.
• Added a zero-inflation test for negative binomial regression and guidance.
• Added zero-inflated negative binomial regression.

AssumpSure documentation built on Nov. 12, 2025, 5:07 p.m.