Assuming a 2020 deadline, I would say that Mars Direct style on-site generation of propellant (or anything else requiring a nuclear reactor being launched) is a nonstarter. We WILL need a heavy launch vehicle, however--whether it's not-Shuttle-C, Jupiter, Delta IV Superheavy (the version with 6+ Common Cores), Ares V, or whatever, we'll still need something that can launch at least eighty tonnes per flight. We'll also definitely need to do Earth Orbital Rendevous on it--the vessel is almost certain to use chemical rocket propulsion and thus will need a few hundred tonnes of fuel brought up to Low Earth Orbit for it.
Publically and politically, anybody who strongly opposes government-funded spaceflight or the Obama administration will probably not be swayed by ANY attempt to promote spaceflight--but then again these are the same people that already think that going to the moon is a bad idea, so the administration won't be making any NEW enemies from such a move.
On the plus side however, a strong and successful Mars program would allow the Obama administration to grab a piece of Kennedy-style glory, which would do wonders for its popularity among anybody who isn't already an opponent. It would also do a considerable amount to reverse the perception that America has spent the last thirty-odd years sliding into an anti-science attitude as exemplified by the amount of media and legislative attention we give to "faith before science" issues such as evolution vs. creation.
Politics aside, a Mars effort would create tens of thousands of engineering and IT jobs directly, as well as the tens of thousands of jobs necessary to support them (all those engineers need to be fed and cleaned up after, and somebody has to construct their factories, etc.). We would see a boost to the related industries comparable to that seen with the moon race. Furthermore, the increased demand for technically-educated people would in turn increase the number of people choosing to study such fields in college (and possibly an increase in overall college enrollment numbers). The magnitude of such increases will scale in proportion to the amount of money spent on non-administrative parts of the program.
For an actual mission architecture, I am in favor of one that uses a "core" vehicle consisting of the main habitat/service modules on one end connected by a boom/truss to a set of rocket engines on the other (say, two to four J-2X or other large restartable-in-orbit hydrogen engines). To this would be mated the Earth Reentry Capsule (something similar to the currently planned Orion Command Module) and the Mars Lander at the bow end, while a number of large fuel tanks would be clustered around the central boom/truss--these would be jetissoned in pairs after Trans-Mars insertion and Mars Orbit Insertion as they are used up. If sufficient delta-v budget can be spared for an Earth Orbit Insertion burn (or a heat shield provided for aerocapture) at the return to Earth, then the whole core unit can be reused.
.
