Comparative Study On Multiple Comparisons Procedures In Proportions

Background and objectives: The comparison of proportions is a common problem in medical data processing. Comparing several proportions relies on multiple comparisons procedures. Although over 30 methods had been proposed, there has been no consensus on which one is worth advocating at present in biostatistics. Therefore, exploring the advantages and disadvantages of each method by simulations is the main task in this research. The author tries to recommend the best method and tries to conclude which method is better for a certain kind of data.Methods:Multiple comparisons procedures (MCP): This research deals with 31 methods altogether, including 22 pairwise comparisons procedures, 4 procedures of Multiple comparisons with the control, 3 all-contrast comparisons procedures (multiple contrasts among proportions) and 2 procedures based on repeating sampling. The classification and name of each MCP are as follows:Pairwise comparisons: they are divided into five groups. Based on test: Tukey-Levy(abbreviated LEVY, the same as follows), SNK-Zar(SNK), Scheffe-Marascuilo(MARA), Scheffe-Cohen(COHE), Seaman-Bonferroni(MABO),Ryan-tukey(RYAN); Adjusted critical value: Duyangzhi(DU), Luowenhai(LUO); Simple (One-step) adjust critical level: Simple Bonferroni(BON), Simple Sidak(SIDK);...