Polling of Likely Voters shows Romney leads Obama...

...according to Rasmussen.

Powerline explains why Rasmussens' polling may reflect reality better than the polling we keep seeing from the mainstream media:

In the past few days, some of these same news organizations have released polls purporting to show that Obama has pulled into a significant lead in several key states. But if you look deeper, you always find that the pollsters have over-sampled Democrats. It is no surprise that if a pollster asks 32% Democrats and 18% Republicans whom they favor for president, the Democrat will come out ahead. Strange as it seems, I think the pollsters and the news organizations they work for are doing this on purpose, in hopes of buoying Obama’s candidacy. Otherwise, wouldn’t they occasionally over-sample Republicans?

So one has to pierce through a lot of clutter to get any real sense of how the race is going. Given that actions speak louder than words, one thing you can do is look at where the candidates are spending their money: nearly all of Obama’s campaign spending (and Romney’s too) is directed at states that Obama carried in 2008. This means that notwithstanding the media cheerleading, Obama is playing defense, not offense.

You can also follow Rasmussen Reports. Rasmussen’s polls are more meaningful than most others because 1) he samples likely voters, and 2) he continuously tests voters’ party affiliation, and weights survey results to reflect the current composition of the electorate. (Republicans have a slight edge in party ID, but essentially the electorate is 1/3 Republican, 1/3 Democrat, and 1/3 independent.) Because of that methodology, Rasmussen’s polls do not jump around randomly depending on the sample composition, and they are better able to measure actual trends.