I’m a big critic of the reliance on impressionistic data often peddled by the media. In classic high school essay form generally it’s rather clear that there’s a hypothesis, and that the journalist just goes looking for individuals who will validate that hypothesis. This is why survey data is so important, though it also makes a lot of political journalism worthless.<\/p>\n
But adding numbers into the mix doesn’t always improve the situation. Consider this piece in Poll: Romney leads 2012 field<\/a>. In reality, the pollster himself doesn’t put too much stock in the results:<\/p>\n \u201cIt\u2019s really impossible to say who the Republican nominee for President next time will be,\u201d said Dean Debnam, president of Public Policy Polling. \u201cThere are positive things each candidate can take from the current polling, but there are warning signs as well.\u201d<\/p><\/blockquote>\n A year closer to the 2008 election than we are to the 2012 election, we had headlines like this: Post-ABC Poll: Clinton, Giuliani Lead Primary Fields<\/a>. I think most people assumed that Giuliani’s lead was soft, and did not reflect the fact that he was far outside of the mainstream of the party on social issues. These sorts of polls lend the illusion of precision, of quantitative heft to what we already know qualitatively. But this far out quantitative assessments are worthless mostly because of the volatility of the trends. In fact, I think that these sorts of results tend to be cause problems because numbers often lend more of an air of authority to conventional wisdom which is dependent on current conditions (and so tends to underestimate volatility and changes in the future).<\/p>\n","protected":false},"excerpt":{"rendered":" I’m a big critic of the reliance on impressionistic data often peddled by the media. In classic high school essay form generally it’s rather clear that there’s a hypothesis, and that the journalist just goes looking for individuals who will … Continue reading