Hi all, I'm stuck in a wilcoxon signed rank test z-score problem,can some one tell me how exactly command "signrank" calculate p.value and z , cause I found out the give different result compare to r when there are lot of zeros in the testing data,
here is the data , you can try on R (wilcox.test(Export,Export_U, alternative = "two.sided",paired = TRUE))and STATA (signrank Export=Export_U ), the result will be different , even the p.value
here is the data , you can try on R (wilcox.test(Export,Export_U, alternative = "two.sided",paired = TRUE))and STATA (signrank Export=Export_U ), the result will be different , even the p.value
Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input float(Export Export_U)
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
.02333072 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
.03802535 0
0 0
0 0
0 0
0 0
0 0
0 0
0 .012133468
0 0
0 0
.01906245 0
.2665769 0
0 0
0 0
0 0
.0013075314 0
0 0
0 0
0 0
0 0
.19016194 .026820315
0 0
0 0
0 0
0 0
0 0
0 .4625474
0 0
0 0
0 0
0 0
0 0
0 .5120914
0 0
.006225195 0
0 0
0 0
.24785425 .03528966
0 0
0 0
.03483107 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 .11168677
0 0
0 .05568282
0 .01983493
0 0
0 .3025498
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
1 0
0 .012808466
0 0
0 .406177
0 0
0 0
0 .0046266057
0 0
0 0
0 0
0 .0042638998
0 0
0 0
end