The ARoHaN Lab Statistical Test Advisor is a specialized digital tool developed by Dr. Manik Ghosh to assist researchers in selecting and executing appropriate biostatistical analyses. Users can upload data in various formats to receive automated recommendations based on an internal audit of the dataset's characteristics, such as normality and variance. The platform supports 16 distinct types of evaluations, including parametric tests such as ANOVA and t-tests, as well as nonparametric alternatives such as the Kruskal-Wallis and Mann-Whitney U tests. Beyond mere selection, the software performs comprehensive computations, including post hoc testing, effect size determination, and the generation of APA-style summaries. It effectively bridges the gap between raw data collection and formal reporting by providing exact setup instructions for both wide and long data formats. This resource serves as a clinical and academic guide, ensuring statistical rigor in research fields ranging from pharmaceutical sciences to general medicine. 

The platform supports 16 distinct statistical evaluations, which are categorized by the shape of the data being analyzed (wide format versus long format).

Wide Format Tests (where each column equals one group)

1. One-way ANOVA: Used for multiple groups that have normal data and equal variances.

2. Welch's ANOVA: Applied when comparing groups that are normal but have significantly different (unequal) variances.

3. Kruskal-Wallis: A non-parametric test used when at least one group is heavily skewed and fails normality checks.

4. Independent Samples t-test: Used to compare two normally distributed groups with equal variances.

5. Welch's t-test: Used to compare two groups that have vastly different variances (e.g., one highly precise and one highly variable).

6. Mann-Whitney U: A non-parametric test for two groups that are skewed and fail normal distribution checks.

7. Paired t-test: Used when measuring the same subjects under two conditions (e.g., before and after) and the differences are normally distributed.

8. Wilcoxon Signed-Rank: Used for the same subjects before and after when the differences are skewed or contain outliers.

9. Repeated Measures ANOVA: Recommended for evaluating the same subjects across three or more time points, assuming the data is normally distributed.

10. Friedman Test: Used for measuring the same subjects under three or more conditions when the data is heavily skewed.

Long Format Tests (structured with an outcome column and a predictor column) 

11. Pearson Correlation: Measures the linear association between two approximately normal variables. 

12. Spearman Correlation: Measures the monotonic association between two variables when the data is skewed and fails normality. 

13. Multiple Linear Regression: Used to predict a continuous outcome based on multiple predictors (e.g., predicting blood pressure from age and BMI). 

14. Chi-Square: Evaluates categorical data (such as a 2×2 table) where all expected cell counts are 5 or greater. 

15. Fisher's Exact: Used for categorical data in small sample sizes where expected cell counts fall below 5. 

16. Logistic Regression: Recommended when predicting a binary disease outcome using a continuous predictor