As I'm sometimes math-challenged, I went down to the shop and built a (full-scale) model.
I made measurements for three different turns: 7.5 degrees, 15 degrees. and 30 degrees.
As @
Trakhak mentions, the larger (degree) turns will only be made at very slow speeds.
The imagined hand position was at the hoods, and movement was measured at an imagined vertical line at the forward curve of the bars.
The model stem lengths were 25mm C-C and 100 mm C-C. 25mm is less than any real stem length could be, with normal steerer tubes and bar sizes.
It IS correct that with the longer stem, more movement at the ends of the bar is required for the same rotation of the steerer tube, compared to a short stem.
For a 7.5 degree turn, the longer stem setup will require my hands to move approximately) 40 mm (along an arc), the shorter stem 25 mm. For a 30 degree turn the longer stem will require a hand movement of 155 mm, the shorter stem 115 mm.
The longer stem will require more hand movement, and should seem 'less sensitive'.
So my initial thoughts and conceptual model were incorrect.
Either I adapt to different setups very quickly, or I'm not skilled enough to notice the difference with different stems.