Here goes again , no need for rocket science here.
Mr A. uses a 100mm lens on his 1Ds MKII to take a picture of a model lying on the beach such that her head touches the left side of the "viewfinder" and her legs touches the right side of the "viewfinder". He selects 1/100 secs for the shutter speed. He managed to get a sharp picture of the model.
Mr B. seeing the outcome, wants to have the same result. But he is using a 350D and a 100mm lens. He dialed in 1/100 secs and stand at the same spot where Mr B. is. But hey, he can only see her shoulder to her thighs (which is just a crop). Therefore, he has to move back such that he gets the same composition as Mr A. He shoots using 1/100 secs and discovered that it's not as good. Is it because he stands further away? Well of course, the further away you are from the object, the shake is amplified and hence you would need to increase your shutter speed.
With the above example, it simply means that when using a crop-factored camera, we actually unknowingly shoot from an equivalent distance of (crop factor X [the distance when using a 35mm camera]). In other words, read my underlined reiterations
We take a picture wholly and think that it is a full-frame one even though we are using any crop factored camera. The crop factor is there and it's just that we do not think about it when we shoot.
Pixel density, imo, just simply means how much you can enlarge for the most satisfactory print. Handshake only affects the physical size of the sensor. Both the 1Ds and the 1Ds MK2 are full-frame sensors. The only difference is that one is about 11 mega pixels while the other is 16 mega pixels. One can print larger than the other.