I decided to test it.
I set my lens on a tripod and then attached both cams to it. Same aperture and ISO, focus on same spot. All shots are SOOC, auto WB and matrix metering. No post besides LR's lens correction and auto-sharpen/noise. What I noticed too is that I need twice the light with the D3300.
The D3300 shot suffered some very slight shake but nothing that hinders this comparison much. I can't put the mirror up and it likely picked up the trigger.
This is the D3300 in 24Mp DX mode.
View attachment 127737
Then the D750 In FX, 1.2x and 1.5x crop modes.
View attachment 127738
View attachment 127739
View attachment 127740
Now let's zoom in some:
View attachment 127741
View attachment 127742
View attachment 127743
View attachment 127744
My D3300 has more Mpix/area so logically it should translate into more detail/better shot.
I enlarged the FX mode until it had the same size as the D3300 shot. It's the worst pixel mode the D750 has compared to the D3300 sensor.
View attachment 127741
View attachment 127745
As you see it's not simply more Mpix/area = more detail or better quality. It depends entirely upon how good the lens/cam combination is. If the FX gets more out of the lens, there's no reason to shoot DX mode unless it's for practical reasons. There's also little reason to bird or macro with a DX if you have an FX. At least as long as that FX makes the lens better.
Btw, I always complained about the D3300 being too bright and here it shows clearly. In LR I need to lower the D3300's exposure by 1 full and 2/3ths of a stop to push the histogram until it equals the D750. The D750 shots also looks as I see them in front of me.