One of the most dangerous temptations of the political reporter is over-interpretation of polls, the need to explain every apparent movement in this week’s poll with reference to events that just happened. The result is a whole lot of utterly unsubstantiated claims explaining things lots of reporters don’t even understand or that may not actually have occurred at all. Only coverage of the stock market, where every news report confidently explains even the tiniest movement in share prices (“Apple shares fell one-tenth of a point today, with investors expressing concern after Billy Wilson of Saginaw, Michigan decided to buy a Droid to replace the iPhone he dropped in the toilet”), comes close. There are two reasons why: the first is that most reporters don’t understand, or willfully ignore, what a “margin of error” represents (meaning they talk about movement within the margin of error as though it represents something real, when it isn’t). The second is that when you have to write every day, you have to say something and explain to your readers/viewers/listeners what’s going on, so there’s an impulse to link effects (poll blips) with purported causes (events on the trail).