# Analysis of categorical data from biological experiments with logistic regression and CMH tests

## TL;DR

This paper provides a practical guide for biologists on using logistic regression and CMH tests to analyze categorical data from experiments, using real examples and R code.

## Contribution

The paper introduces accessible tutorials and real-world examples to demonstrate the advantages of logistic regression over simpler tests for categorical data analysis.

## Key findings

- Logistic regression can handle complex multivariable datasets in biological experiments.
- Real biological examples show how logistic regression provides more comprehensive insights than simpler tests like CMH.
- Tutorials and R code are provided to improve the rigor and reproducibility of categorical data analysis.

## Abstract

The choice of appropriate statistical tests in experimental biology is critical for scientific rigor and can be challenging in the case of categorical data analysis. Using example datasets from Caenorhabditis elegans research, we conduct statistical analysis of (1) a rare cellular event involving the formation of a neuronal extrusion called an exopher and (2) a variable behavioral response across time. We employ the Cochran–Mantel–Haenszel (CMH) test and logistic regression for analysis. Recognizing there are potential accessibility issues using logistic regression, we provide step-by-step tutorials and example code. We emphasize that logistic regression can handle both simple and complex multivariable datasets; logistic regression can also provide more comprehensive insights into experimental outcomes when compared to simpler tests like CMH. By analyzing real biological examples and demonstrating their analysis with R code, we provide a practical guide for biologists to enhance the rigor and reproducibility of categorical data analysis in experimental studies.