The analysis of ranked data derived from completely randomized factorial designs

Area of Science:

• Statistics
• Experimental Design
• Nonparametric Methods

Background:

• Ranked data analysis is crucial in various scientific fields.
• Factorial designs are common in experiments but analyzing ranked data presents challenges.
• Existing methods may not adequately capture interaction effects in ranked factorial data.

Purpose of the Study:

• To present a new statistical method for analyzing ranked data from completely randomized factorial designs.
• To extend the capabilities of the Kruskal-Wallis ranks test for factorial experiments.
• To enable the calculation of interaction effects and linear contrasts for ranked data.

Main Methods:

• The proposed method is an extension of the Kruskal-Wallis ranks test.
• It is designed for completely randomized factorial designs.
• Statistical convergence was evaluated using a Monte Carlo simulation study.

Main Results:

• The extended Kruskal-Wallis test effectively analyzes ranked data in factorial designs.
• The method allows for the computation of interaction effects.
• The procedure also facilitates the calculation of linear contrasts.
• Monte Carlo simulations confirmed the test's convergence properties.

Conclusions:

• The presented method provides a robust approach for analyzing ranked data in factorial experiments.
• It enhances the Kruskal-Wallis test by incorporating interaction and contrast analysis.
• The method is suitable for situations where data are ranked and experimental designs are factorial.