I must insist on this point. Not necessarily.
Imagine you're watching a video of someone jumping onto a crate. Now imagine that you see the same video, but with 2x zoom.
Now you can see this situation from two angles:
a) The crate in the original video is N pixels high, and in the zoomed video it is 2N pixels high. The number of pixels is doubled, therefore the jump in pixels/s must be twice the speed and the gravity must be adjusted to a new value in pixels/s².
Or:
b) The crate in the original video is N metres high, and same in the zoomed video. The speed of the jumping guy is in m/s and the gravity is in m/s², and they are the same in both videos. Each metre is PPU pixels, therefore the crate in the original video is N*PPU pixels high and in the zoomed video it is 2*N*PPU pixels high, but that doesn't change the speed or the gravity of the guy.
You seem to keep thinking in terms of (a); I'm suggesting you to look at it as in (b).