The difference between my two measurements at two different temperatures was 9mm AFTER the temperature compensation. Before the temperature compensation it was 47mm for 100m, or 141mm (5.5 inches) for a 300m cal course. So it is not insignificant.
But I would agree that with the current temperature compensation instructions it is very easy to make a mistake and end up subtracting distance from a course when you should be adding distance. So it might be better to eliminate the temperature compensation from the procedure.
A third option though, would be to just simplify the instructions:
If the temperature is above 68F, leave the course as measured.
If below 68F, add
(68-temp) x 0.00000645 x (course length)
to the course.