Second, there has been pretty good evidence that people do not, on average, lie to pollsters.
Third, it greatly depends on the screening procedures for the pollster. For example, in the M-D poll I took, they asked the same question about four different ways, but placed throughout the interview so that at the end, they had a pretty good placement of my views. (One was on Iraq, and whether I "favored" US continued involvement or not. Well, I did IF we were in it to win. So I can't recall how I responded back then.
The trends are pretty sharp for Romney, if anyone cares to look: all the "Obama up" polls have massive oversamples of Dems---anywhere from 9-13 points---and often (not always) they are registered voters, not likely. All the recent polls of Romney being up are of likely voters with a small, if any, Dem oversample.
So when you combine the TRENDS: northern VA poll showed Zero up 4 when he won this region by 23; FL seniors AARP poll showing seniors favor Romney (oops. So much for Medicare cuts working); WI polling, twice in a row has Romney up by one; MI poll has him up five after being up one; two different "swing state" polls have him up everywhere but CO, where he is 1 down on RCP; poll out today of FL has him up huge (over 10), but with no internals; Cook Co. poll has Obama winning Cook Co., but by a much narrower margin than in 08 and not enough to carry even IL.
These TRENDS add up to a pretty big Romney win. Don't know if you'd call it a "landslide," but over 320 EVs.
The Northern Virginia trend is FLAT if you start in April, and maybe a little bit up for Romney since June.
Doesn't matter how you measure a trend though provided all your polling is done the same way ~ e.g. adults, registered voters, likely voters, etc.
All the polls ask people to self-define their political affiliation though ~ so unless they are the more expensive kind taken from a pre-determined base of semi-pro poll answering folks, there can be no "oversamples" by party.
Reporting is different though. If your question is "How are Democrats answering" you just strip out all the non-Democrat respondents.
Part of the clue to figuring out all of this is to note that few Republicans are attracted to any Democrat candidate, and vice versa, so if you want to win you must turn out your own base.
Romney's advisors abandoned that strategy long ago ~ guess they thought they could use Democrat voters or something.
Now all his cabinet appointees will have to be Conservatives, as will the judges. Anything less than that and it's gonna' be IMPEACHMENT TIME, and you know how that goes if the OTHER SIDE likes the idea too!