Poll watching is a major past-time amongst the lamestream media. Up until a few weeks ago, most national pollsters (Rasmussen being the main exception, as I’ll explain in a bit) based their polling on “registered” voters, the largest potential population of voters. What is more telling, however, is the population of “likely” voters. These voters have shown a pattern of participating in the majority of past elections, including primaries. Rasmussen has consistently polled likely voters. Gallup, on the other hand, has been promoting polls of registered voters up until a few weeks ago. They’ve recently shifted to likely voters.
Here are Rasmussen’s polling results, which are based on a three-day, rolling average.
Now, take a look at Gallup’s poll of LIKELY voters. You’ll see a dramatically different result.
Likely voters are a much more accurate way to define voter sentiment. But I don’t think even it can predict what will happen. After all, it didn’t predict the dramatic results of 2010.