The Z-Score report is like grading all your songs on a bell-curve. It really is the only way to properly compare songs that were tested in different cycles (by different groups of respondents) to each other.
The formula used to determine the Z-Score is:
(X - M) / SD
Where X = Positive Score, M = Mean and SD = Standard Deviation.
A Z-Score itself has a mean of zero. This essentially means that the average of all Z-scores for a given week or cycle would be zero. Some scores will be less than zero (below average); while others will be greater than zero (above average), and some will be right at (or very near) zero, indicating they are average.
Based on the Standard Deviation of Z-Scores, generally 1, we can then deduce that about 68% of your Z-Scores will fall between -1.00 and +1.00 (the remainder will fall outside this range). Once we calculate the Z-Score for a group of songs from a given test cycle (using the formula above), it's easy to do the same thing for those songs from previous test weeks (or cycles), and present the resultant Z-Scores side by side for comparison.
Z-Scores allow you to accurately compares "apples to oranges", whether it is merely the same songs cycle to cycle, or comparing the Z-Scores of Females vs. Males, or one demo group vs. another.