Nonparametric analysis of ordinal data in designed factorial experiments.

Area of Science:

• Plant pathology
• Statistical methodology
• Ordinal data analysis

Background:

• Plant disease severity is commonly measured using ordinal rating scales.
• Traditional parametric methods (e.g., ANOVA) are inappropriate for ordinal data.
• Limited nonparametric methods previously restricted experimental designs to simple one-way layouts.

Purpose of the Study:

• To illustrate the nonparametric analysis of ordinal data from complex experimental designs.
• To demonstrate the application of recent statistical advancements in plant pathology.
• To show how to quantify experimental factor effects on ordinal disease ratings.

Main Methods:

• Application of recent nonparametric statistical methods for hypothesis formulation and testing.
• Analysis of ordinal data from two-way factorial designs.
• Inclusion of repeated measures designs in the nonparametric framework.
• Quantification of factor effects using estimated relative marginal effects.

Main Results:

• Successful nonparametric analysis of ordinal data from complex designs is now feasible.
• Estimated relative marginal effects allow for quantification of experimental factor impacts.
• Advanced statistical techniques enable proper analysis of previously intractable data.

Conclusions:

• Recent advancements facilitate the appropriate nonparametric analysis of ordinal plant disease severity data.
• Complex experimental designs, including repeated measures, can now be effectively analyzed.
• This enables more accurate assessment and understanding of plant disease factors.