The substrate could simply be flatter on one than the other. A distinct possibility not related to the coating material.
First: the process of making a flat can have the unintended effect of rolling off the edge of the mirror. They're held on the edge in a fixture that is not PERFECTLY rigid on the nanometer level. So the rotating table under the mirror can tilt the mirror ever so slightly.
Second: the process of coating the mirror with a large number of cavities of dielectric coatings also tends to roll off the edge of the mirror.
Roland Christen had a good article on this. It's because the coating thickness is not uniform over the mirror's surface and because of heat applied during coating.
Third: if you look at a lot of Danny's (PINBOUT) posts here on CN, he has run into a lot of secondary, flat, mirrors with slightly convex surfaces.
It is not uncommon. He shows the interference fringes on a lot of them, and it isn't pretty.
Fourth, one of the reasons a 2" diagonal is recommended on scopes small enough that the entire clear aperture of the diagonal is not illuminated is simply to avoid the edges of the mirror, where the majority of the optical problems (like TDE) lie. So, a good reason to use a slightly oversized secondary mirror in a reflector and an oversized mirror in a star diagonal.
If you hung out on the ATM Forum or REFLECTORS Forum a bit, you'd see all the reports on bad mirrors, mostly secondaries, but also star diagonals.
And, as I mentioned, a star test done with and without them usually reveals the problems caused by the star diagonal.
One of the reasons I recommend Antares Optics secondary mirrors is that they are tested AFTER coating, so WYSIWYG.
As for star diagonals, if you get a chance to test 4 of the same brand/model some time you'll see there is a variation from unit to unit just as there is with primary mirrors. All too often, people blame the primary mirror in inexpensive scopes for poor image quality when it is really the secondary at fault.