Title
fprank -- Two-Sample Fligner-Policello Robust Rank Order Test
Syntax
fprank varname [if] [in], by(groupvar)
options Description ------------------------------------------------------------------------- Main * by(groupvar) grouping variable ------------------------------------------------------------------------- * by(groupvar) is required. by may be used with the above commands; see by.
Description
fprank tests the hypothesis that two independent groups are sampled from the same population using the robust rank-order test which is a useful alternative of the Mann-Whitney-Wilcoxon test for non-normal populations with unequal variances. This test assumes neither normality, nor equal variances, nor equal shape.
fprank provides the asymptotic p-value which is based on the standard normal distribution. It should be noted, however, that the asymptotic p-value may not be adequate when both samples are <= 12. To obtain the significance of the statistic U when sample sizes are (<=12), refer to the table of critical Values of U for the robust rank order test found in the references below.
Examples
. fprank scores, by(group) References
Siegel, S. & Castellan, N.J.Jr.(1988). Nonparametric statistics for the behavioral sciences (2nd ed.). New York: McGraw-Hill.
Hollander, M & Wolfe, DA. (1999). Nonparametric statistical methods (2nd ed.). New York: Wiley.
Fligner, MA & Policello, GE. (1981). Robust rank procedures for the Behrens-Fisher problem. Journal of the American Statistical Association, 76(373), 162-168.
The two data files attached to this package are from Siegel & Castellan (1988, pp 138-143).
Author
Benmamoun, Mamoun, Saint Louis University , Saint Louis, MO, U.S.A. Email vmamoun@yahoo.com if you observe any problem or if you have any comment or question.
Also See
Online: ranksum, ttest