Let's get this straight: SurveyUSA, an independent polling company, has Ken Buck leading by 16 points in Colorado.So Jane Norton hires a pollster for her campaign and finds that . . . well that Jane Norton is losing by 4 points.Now, of course Jane Norton is spinning this as a six point lead. Why? Well, her pollster asks people if they are likely to vote and says she leads with likely voters.However, objectively, likely voters tend to factor in voters who actually are likely to vote, i.e. voters who have voted in past Republican primaries. After all, if the best indicator of future performance is past action, you tend to want to know if someone has voted in a primary before.Among voters who have voted in a Republican primary before, Jane is losing by 4.I'm embarrassed for her.