Always consider the sample. The sample is everything.

And if the sample greatly differs from the electorate, the poll is worthless.

What does the electorate look like? Well, if Rasmussen is correct then there should be more Republicans than Democrats.

The number of Republicans in the country inched up half a percentage point in January, while the number of Democrats dipped to the lowest level ever recorded by Rasmussen Reports (32.5%).

During January, 35.9% of Americans considered themselves Republicans. That’s up from 35.4% in December and the highest number of Republicans measured since December 2010.

So a proper sample should be 36-33 Republican, with 31 percent independents – again, if Rasmussen is correct.

How do the latest presidential polls stack up to that model?

Well, a week ago there was an NBC/Wall Street Journal poll which had Barack Obama with a 48-42 lead over Mitt Romney. The sample on that poll? 44 percent Democrat, 37 percent Republican, 15 percent independent among the 1,000 respondents.

Just a 10-point swing to the Democrats on that sample.

And two weeks ago, Public Policy Polling put out a survey which gave Obama a 49-44 edge over Romney. That sample? 41 Democrat, 35 Republican, 24 independent. Just a nine-point swing from what the electorate actually is if Rasmussen is to be believed.

One has to be a premium subscriber to Rasmussen Reports in order to get the crosstabs on his latest tracking poll which has Obama ahead of Romney by a 47-43 count, and Gallup’s latest poll which has Obama and Romney tied at 48 doesn’t show its sample by party identification, but based on the other polls in the past couple of weeks it’s not unreasonable to question those samples as well.

But for future reference, any polls showing Romney – or Newt Gingrich or Rick Santorum, for that matter; we’re referencing the Romney results since at this point he seems to be performing better than the other two in a general election poll – trailing Obama needs to be scrutinized on the basis of its sample. Because if that poll shows more Democrats than Republicans when the actual electorate has 3.5 percent more Republicans than Democrats in it at present (again, if Rasmussen’s numbers are accurate), it’s not a reliable reflection of public opinion.