Related Experiment Videos
Bayesian analysis of square ordinal-ordinal tables.
1Department of Statistics, Chinese University of Hong Kong, Shatin, Hong Kong. wypoon@hp735.sta.cuhk.edu.hk
Summary
This study introduces a new Bayesian model for analyzing ordered categorical data in contingency tables. The model enables the estimation of location and dispersion parameters for underlying continuous variables, offering deeper insights into data relationships.
Area of Science:
- Statistics
- Data Analysis
- Ordinal Categorical Data
Background:
- Analyzing square contingency tables with ordered categories presents challenges in understanding underlying variable relationships.
- Traditional methods may not fully capture the nuances of ordinal data, limiting comparative analysis.
Purpose of the Study:
- To develop a novel statistical model for analyzing square contingency tables with ordered categories.
- To enable the comparison of location and dispersion parameters between underlying continuous variables.
Main Methods:
- Formulation of a model where ordinal categorical variables are seen as manifestations of underlying continuous variables.
- Imposition of stochastic constraints on thresholds to identify the model.
- Application of a Bayesian approach for parameter estimation.
Main Results:
- The proposed model allows for the estimation of previously non-estimable location and dispersion parameters of underlying continuous variables.
- Successful application demonstrated through illustrative examples on reported data sets.
Conclusions:
- The developed Bayesian model provides a robust framework for analyzing ordered categorical data.
- This approach enhances the understanding of variable locations and dispersions, offering valuable statistical insights.