### 61.5. Nonparametric rank statistic for detecting low and high numbers

Add an optional column that flags species for which the latest year's record as low (-) or high (+) in rank compared to the other years. This particular non-parametric statistic has been most useful to the author in his role as New Mexico regional editor for the CBCs.

To compute this statistic:

1. Divide the interval from 0.0 to 1.0 into `n` equal slices, where `n` is the number of columns.

For example, if there are five columns, the first slice would extend from 0.0 to 0.2, the second from 0.2 to 0.4, and so on.

2. Arrange the numbers of individuals in all the columns into ascending order, and assign each number to a slice.

3. Combine slices with the same count into groups.

4. The rank statistic for each group is the pair of numbers that brackets the group.

5. Next we test the rank statistic to see if it is high or low.

• If `n` is 10 or greater, the statistic is considered low if it is in the bottom decile (its upper limit is 0.1 or less). It is considered high if it is in the top decile (lower limit 0.9 or higher).

• If `n` is 5 or greater, we consider the statistic low if it is in the bottom quintile (its upper limit is 0.2 or less). It is considered high if it is in the top quintile (lower limit 0.8 or higher).

Here's an example. Suppose the counts of some species in all selected years are 0, 0, 5, 1, and 0. There are three groups: in order, [0,0,0], , and . The group for value 0 extends from 0.0 to 0.6; the group for value 1 extends from 0.6 to 0.8; and the group for value 5 extends from 0.8 to 1.0. For this example, if the latest year's count were zero, it would not be considered low, because three out of five years had zeroes. However, with the same distribution of values, if the latest year were the 5, that would be considered high because it is in the top quintile.

If this column is selected, it will be displayed just to the right of the last year's count column, and will contain “`+`” for high and “`-`” for low.