The Mann - Whitney test, compare the means of two independent samples. The null hypothesis implied that distribution frequencies are equals mean while the alternative give you the possibility of  set if the distribution are = or < or > (it is a one and two tails test). Then the positive or negatives sign of  " Z " it is in correspondence of what alternative hypothesis are you using and how you are making the comparison, if its sample A vs. B or B vs. A. 

Reicelis,

Thank you for your answer. However, i made a mistake in my question and i have correct that mistake. My question is about Wilcoxon Signed Ranks test and it's interpretation, but you answered me very nicely. 

Boban,

It is the same test, some author call the test: Mann - Whitney - Wilcoxon other ranks test or U test and even how you called in your response. It is a non - parametric test to compare means. I am giving you a link, that have an example and a little explanation of the test, even give you an example using MINITAB software for accomplish the statistical test

http://courses.wcupa.edu/rbove/Berenson/CD-ROM%20Topics/topice-10_5.pdf

I hope could be useful 

Wilcoxon gave their name to a number of different tests, so I can see where the confusion is. Plus the Mann-Whitney-Wilcoxon test was developed independently by these three folk, til one of them worked out they were actually the same thing! (so I'm told). So we often just call it the Mann-Whitney test to avoid confusion.

The Mann-Whitney test is the non-parametric equivalent of the two-sample t-test. Interpret it as a shift in the whole distribution, though present the estimate of the difference between the medians of each group, with confidence interval.

The Wilcoxon signed rank sum test is the non-parametric equivalent of the paired t-test. Again it's the whole distribution. Ignoring the signs of the numbers, rank the differences in order of magnitude, then sum the ranks separately for the positive differences and the negative differences. Under the null hypothesis you expect these sums of the ranks to be the same. How far apart they are leads to the p-value for the difference. But most useful to present the relevant estimate, the median difference between the groups, with confidence interval.

Is this what you wanted?

This is just what i wanted, both from Ms. Li and you. Thank you for your help.

will u please give me an example how to tabulate wilcoxon sign rank test and its interpretation?

Hi, just a quick note - Wilcoxon compares mean ranks rather than means, as it's a non-parametric test used where there isn't a normal distribution and/or where you are comparing ordinal data, e.g. numbers that stand for responses on a Likert scale, rather than interval data, e.g. temperatures.

Dear all,

I do hope you all can help me to confirm my interpretation whether it is correct one.

After I analyzed it showed that

Power - Interest; z= -5.679 ( b. Based on positive ranks); Asymp. Sig. (2-tailed)=.000

Based on z value (-5.679). Could I interpret that mean of Interest is significant greater than mean of Power?

Thank you

Mann-Whitney U test is different from Wilcoxon Signed Rank Test although both of them are among non-parametric statistics. Mann- Withney is applied for independent samples like male and female differences in sth but wilcoxon test is used for related and matched pairs which are compared with repeated measures or when subjects are measured on two occasions or conditions such as a group when answering two sets of questionnaires in Time 1 and time 2.

A Wilcoxon Signed Rank Test indicates that the Median post-test ranks or Time 2 scores are statistically significantly higher than the Median of pre-test or Time 1 scores and it has nothing to do with Mean!!!

@

This would help you I guess:

https://www.slideshare.net/plummer48/reporting-the-wilcoxon-signed-ranks-test