Wednesday, September 05, 2012

Are Polling Agencies Selectively Making The Race Look Tighter Than It Is? Why, Of course They Are

By Nathaniel Downes September 4, 2012 

It is traditional for a challenger to get a bump after their convention. This bump historically averages between 7 and 11 points, taking into account the margin of error. To get these estimates, polling companies use models and tools to sort through their data. Typically, the polling agencies are within the margin of error from each other as a general rule, due to their models and sampling data.

However, this year we are finding a wide variance between the polling from one agency to the next. Gallup finds Obama in the lead, with Public Policy Polling finding Romney ahead. The 9-point difference between the two is incredibly rare. This puts an alarm bell out, giving the possibility that some of the polling companies might be engaging in outright manipulation or misuse of the data in order to create an image of a tighter race than exists.

It gets even more unusual when you start to check on the individual states. Large population states, with heavy amounts of delegates, are typically polled at least once a month. You find California, New York, Florida, Ohio, Pennsylvania, Virginia, Georgia, Washington, New Jersey–in fact one can find polls for almost every large (over 10) electoral state having been conducted in the past 45 days. The lone exception for this is, fittingly, the lone star state, Texas. There have been a noticeable lack of polls conducted in Texas since May for the general presidential election, although polling for other races in the state have remained constant. With the weak showing of the Republican ticket among minority voters, and with Texas now solidly a quarter hispanic, the ignoring of it for polling comes as incredibly unusual. While it is not likely to be in play, the same can be said of other electoral delegate rich states, such as California, New York and Illinois, who all have been polled on a regular basis.

If one averages out the Romney post-convention bounce between all of the major polling agencies, we find an average of a 2.6 point bounce. If you drop the biggest changes in the polls for both the Romney and Obama directions, it reduces to only 2.3 points. Either of these puts the post-convention bounce for Romney as the second lowest for a challenger to an incumbent president since conventions were widely broadcast on national television starting in 1968, with only John Kerry’s 0.8% bounce coming in lower. This bodes very poorly for the Romney campaign as it is. But you would never know this if you turned on the news services.

We are looking at a race in which the challenger is doing very poorly, but in this fast paced media-driven world, that kind of race is not what the networks are looking for. They want a neck-and-neck race, and it seems that the polling agencies are setting out to deliver for them just that. That the various polling agencies are excluding states, and candidates, they and the media companies which own them, seem intent on setting the narrative. By not calling out the candidates on major issues, such as blatant lies in a convention speech, the mainstream media is clearly more interested in short-term ratings boosts than in doing its role for society.

Creative Commons License