There are a couple of options.
1. use only unfaced batts (no kraft or foil backing), mastic-seal all of the seams in the subfloor, and either use a vinyl or some other highly vapor-retardent floor covering.
2. use only unfaced batts, put a sheet of 6-10 mil poly as an underlayment for the finish floor.
3. use only unfaced batts, and put a layer rigid foam insulation on the garage ceiling (taping or mastic-sealing the seams) before putting up the gypsum.
The highest performance would be #3, but it would also be the most expensive. The minimum R value you would need on the foam depends on the depth of the joists, but it doesn't have to be much to work. Installing exterior-side foam is protective of the wood from wintertime moisture accumulation by keeping the average temperature of the wood above the dew point of the interior space air. Since the only wood that is susceptible is the joists (no wood sheathing on the underside) and the average wintertime temp in the garage will be above that of the outdoor air on the stud edge is at risk. Without the foam the paper facer of the ceiling gypsum would be at risk, but with the foam it stays at the drier cooler environment of the garage.
Assuming the joists are 2x10s, the R value of the joist itself is around R9-R10. In Prattsburgh the mean outdoor temp in January is ~25F (according to weatherspark.com historical data, and it's safe to assume that the mean temp in the garage will be at least 5F warmer, call it 30F. The dew point of 35%RH 70F interior air is about 40F, and is the generally accepted dew point for residential air for design purposes, and as long as you aren't actively humidifying the air to a higher level (or you have a super-tight house with no active ventilation) it'll actually be drier than that. So if the R value of the foam is high enough to keep the mean temp at the joist edge at 40F or higher you're golden. The delta between 70F on the room side and 30F on the garage side is 40 degrees, and the difference between 40F (stud edge goal) and the 30F garage is 10 degrees, so to keep the joist edge at 40F the foam R has to be at least R5, which is with a single layer of 1" XPS sheathing. If the joists are 2x12s it's better to go to R6+ (1-1/2" XPS is R7.5, 2" EPS bead-board is ~R8, and often cheaper.)
You could also use foil-faced iso, but the foil facer is a double-edged sword- it keeps conditioned space moisture from reaching the paper facer on the ceiling gypsum, but it blocks the ability of the joists to dry toward the garage, whereas XPS (pink, blue, green , gray) is still semi-permeable and would allow seasonal drying into the garage of any moisture that got into the assembly.
Exterior foam is a standard method of controlling the temp of wood sheathed walls & roofs to be able to use only modestly vapor retardent materials on the interior, enshrined in building code ( IRC 2009 ). This is a less usual application, but the physics of water don't change just because it's a floor, and by being earth-coupled semi-conditioned space (garage) and has no wood sheathing on the cold side (ceiling) you don't need as much foam as you would need on a wall or roof. See:
http://www.buildingscience.com/documents/guides-and-manuals/irc-faqs/irc-faq-insulating-sheathing-vapor-retarder-requirements
and
http://www.buildingscience.com/documents/guides-and-manuals/irc-faqs/irc-faq-conditioned-attics
(Prattsburgh is either on the cold edge of climate zone 5, or the warm edge of zone 6, for reference on the minimum foam-Rs.)
In a garage you have the additional benefit of it being a vented space, and no rain/snow wetting of the cold side surface, making it far LESS susceptible than wall sheathing or roof decking that the IRC is primarily addressing. In the event that the garage door stayed open all winter it would be identical to the "vented cladding over gypsum" exception to vapor barrier requirements for zones 5 & 6, but with the door closed maybe not- safer to add the R5 of foam.
And adding the foam is FAR more reliable than an interior vapor barrier, since the odds of the floor staying truly air-tight forever are slim, and air-transported moisture through even a 1/4" hole in the floor exceeds the water volume that would get through 250-square feet of OSB subfloor via vapor diffusion alone. But with the foam it becomes a "who cares" situation- the air moisture will not be absorbed by the wood creating mold or rot conditions because the wood is too warm.