If 2 tests are within 20% of each other, they are essentially the same number.
Ex
100 -> 80 - 120
90- > 78 - 108
The upper estimate of the lower number is greater than the lower estimate of the upper number, so these numbers may be considered the same.