Parametric versus non-parametric statistics in the analysis of randomized trials with non-normally distributed data

Table 3 Relative efficiency of ANCOVA and Mann-Whitney for the moderate positive skew data. Values less than 1 indicate greater power of ANCOVA; greater than 1 indicates superiority of Mann-Whitney. In blank cells, the power of one or both tests was 100%.

Sample size	Correlation between baseline and post-treatment score
	0.1	0.2	0.3	0.4	0.5	0.6	0.7	0.8	0.9
10	0.8125	0.9773	0.6017	0.9403	0.9722	0.8834	1.2129	1.0851	1.1320
20	1.0160	1.1093	0.8172	0.9348	0.7742	0.9905	0.8124	1.0058	1.0231
30	1.0000	0.9167	0.7785	0.8617	0.8027	0.7756	0.8689	0.8760	1.0096
40	0.8866	0.8596	0.8120	0.7332	0.7365	0.7556	0.9172	0.9067	0.929
60	0.8925	0.8996	0.7752	0.7632	0.7418	0.8277	0.8728	0.8841	0.9892
100	0.8822	0.8594	0.7816	0.7259	0.7071	0.8277	0.8639	0.8702	0.9259
200	0.8484	0.8030	0.7611	0.6920	0.6979	0.7591	0.8793	0.8888	-
400	0.8512	0.8292	0.7392	0.7113	0.6707	0.8029	0.8336	-	-
800	0.8781	0.9087	-	-	-	-	-	-	-

ISSN: 1471-2288