I confirmed my statement by testing this activity on my wife's bike (a Suzuki LS650), basically because it's easier to measure, with it having the exposed rear suspension.
The bike is set at it's lowest/softest for her. I measured the distance between the top and bottom collars of the springs at 5.3 inches. I tightened them one notch - still 5.3 inches. Tightened them another notch - the springs very visibly were compressing this time, changing the setting was much harder, and the measurement was now 5.2 inches. I tightened them the remaining 2 notches (again, very visibly compressing) and the distance was just under 5 inches.
Of course these changes were after the point where the max suspension travel was reached. That's why it's called "PRE-load". And I stand by my statement that the spring rate is affected. You can watch the coils at the end compress against the "base" coil, shortening up the effective length. Of course, normal suspension action does this, as well, but tightening the preload changes the starting point.
None of this affects shock damping, of course. But changing the preload on my NT very definitely affects the degree of impact when hitting things like drain covers and such on the road.