Python Script for Simulating, Analyzing, and Evaluating Statistical Mirroring-Based Ordinalysis and Other Estimators

This presentation involves simulation and data generation processes, data analysis, and evaluation of classical and proposed methods of ordinal data analysis. All the parameters and metrics used are based on the method methodology presented in the article titled "Statistical Mirroring-Based Ordinalysis: A Sensitive, Robust, Efficient, and Ordinality-Preserving Method for Analyzing Ordinal Assessment Data" authored by Kabir Bindawa Abdullahi in 2024. For further details, you can follow the paper's publication submitted to MethodsX Elsevier Publishing.
The validation process of ordinal data analysis methods (estimators) has the following specifications:
• Simulation process: Monte Carlo simulation.
• Data generation distributions: categorical, normal and multivariate model distributions.
• Data analysis:
- Classical estimators: sum, average and median ordinal score.
- Proposed estimators: Kabirian coefficient of proximity, probability of proximity, probability of deviation.
• Evaluation metrics:
- Overall estimates average.
- Overall estimates median.
- Efficiency (by statistical absolute meanic deviation method).
- Sensitivity (by entropy method).
- Normality, Mann-Whitney-U test and others.