The Unavoidable and Intentional Inaccuracies of Polling
The unavoidable and intentional inaccuracies of polling make results unreliable. Averaging many poll results leads to a more accurate estimate. Some of these inaccuracies are intentional, while others are the result of advancements in technology. The main stream media does a good job of researching the underlying data to rate the accuracy of polls. Using an average of only reliable polls produces the best results. We must always remember that polls rely on the honesty of the participants. Casting a vote is the only answer that cannot be retracted. Polls are predictive estimates of election results. Nothing, however, can placate those who only believe in polls showing them in the lead, and only believe election results are valid if they are the winner.
Technology is making polling increasingly difficult as they can only poll voters who still have landlines. Nowadays these people are becoming their own demographic, and less representative of the population in general. They are generally older, as the younger generation are allergic to wires of any kind. These voters are more conservative, the same people who own a VCR. While not everyone with a landline will fit this description, they are overly represented.
Manipulated results get the outcome a pollster wants. Reputable pollsters will publish the demographic information of those polled. If demographic data is not recorded and published; the pollster achieves a pro Democratic result through over representing a younger demographic; a pro Republican by over representing and older demographic. They get the same manipulation by over representing other demographics, such as gender or race.
In national polling, they get biased results by the states they over-poll. State polling it is even easier to manipulate. Results depend on whether they choose to poll in rural areas, or the population centers. In my home state of Pennsylvania, most countries are overwhelmingly Republican. They also have to a very small percentage of the state’s population. The two main population centers of Philadelphia and Pittsburg are enough to swing the Democratic in most presidential elections. This is why two polls can show Clinton with a 9 point lead, while the race is tied in another. One pollster made sure the demographics of those polls were representative of the state as a whole, the other did not.
How they ask a question will also change poll results. The options offered, wording of the question, or the setup questions that came before it which are rarely published. While they may not meet the level of bias to qualify as push polling, they still meet the result desired by the pollster.
Chief Political Editor