A non-parametric test of the null hypothesis that the median is a specified value. It requires the assumption or knowledge that the distribution being sampled is symmetric. It is based on ranking the observations by their distance above and below the median and comparing the total of the rankings above and below. It is a more powerful test than the sign test which only counts the numbers of observations above and below the median and therefore discards quite a lot of information. The procedure is to rank the observations not equal to the hypothesized median by their distance from the median, and to sum the ranks above the median and below the median. There are tables of critical values for different values of n (the number of observations not equal to the hypothesized median) to conduct both one-tailed and two-tailed tests, but for even relatively small values of n, the distribution $N\left(\frac{1}{4}n\left(n+1\right),\frac{1}{24}n\left(n+1\right)\left(2n+1\right)\right)$ is a good approximation.