USE OF RANKING AND SCORING METHODS IN THE ANALYSIS OF ORDERED CATEGORICAL-DATA FROM FACTORIAL-EXPERIMENTS

The analysis of results from complete factorial experiments where responses are measured on an ordered categorical scale is considered. Ranking procedures to extract information about factorial effects such as main effects and interactions are discussed. In the proposed ranking method, main effects are investigated using a Friedman test statistic, and interaction is investigated using the idea of rank interaction due to de Kroon and van der Laan. Connections are made between this rank-based approach and methods based on scores using versions of Mantel-Haenszel statistics. The proposed method is applied to data from four experiments.