a non-parametric rank test to compare two paired samples, whether values in one are bigger than in the other. This test is NOT the same as the sign test.

Differences between the two variables are ranked. The null hypothesis is that the mean of these ranks is zero, which implies that the distribution of differences in the population is symmetric about zero. The alternative hypothesis is that it is not so, and thus values of one of the variables tend to be higher. This tag is NOT to be confused with the [sign-test] tag.