JC <[EMAIL PROTECTED]> wrote: > Does anyone have a suggestion on what procedure to use in order to > report effect size from a Mann-Whitney?
There's a statistic called "Glass' rank biserial correlation" which works out to (U1 - U2) / (U1 + U2) and ranges from -1 to +1. See Glass, Gene V., "Note on Rank Biserial Correlation," _Educational and Psychological Measurement_, 1966, 26, 623-631. I've seen some textbooks recommend it as an effect size measure in this case. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
