Keep in mind that there is no clear, rigid distinction between Pop. I,II and III stars... its a rough distinction that Pop. III stars formed from primordial element abundances, Pop. II stars formed from material mixed with ejecta of supernovae from Pop. III stars; and Pop. I stars from ejecta of Pop. II stars.
The additional important factor is that massive stars die younger.
In the galactic disk, the densities (and general conditions) are conducive to star formation, while the halo tends to have much less cold-gas which is necessary for star formation. Thus, in disks where there is star formation, there can still be massive stars (that live shorter); while the halo contains almost entirely older stars which must be less massive (because lower-mass stars live longer). Additionally, old stars in the disk have time to migrate (or be ejected) out of the disk --- while there aren't many methods of injecting halo stars into the disk.
From these effects, you end up with a natural, strong bias towards younger stars (Pop. I) in the disk, and older (Pop. II) stars in the Halo. This does not mean there are no stars from the opposite populations mixed in.
To my knowledge, the distribution of stars and their ages is completely consistent with population synthesis simulations, and there don't seem to be an excess of 'missing' Pop. II stars from the disk.