On February 1, the Iowa caucus will be the first vote determining the Republican and Democrat presidential candidates for 2016. For a brief history of the Iowa caucus, look at this interactive article from the Des Moines Register website. “The Iowa caucuses have a poor record of picking presidents, but they play the important role of winnowing the field. In only one instance in the history of the modern Iowa caucuses has someone finished lower than third and gone on to win a party’s nomination (John McCain finished fourth by a whisker in 2008).”
The polling numbers represent the percentage of contacted individuals who plan on voting for a specified candidate.
Though unrelated to the upcoming caucus, earlier this month the New York Times published an article regarding their experiment with using the internet in conjunction with telephone polling with successful results. “The Times has generally shied away from online polling […] But with large numbers of the public abandoning landline phones and response rates dropping, some form of online polling appears inevitable.”
Gallup published a comprehensive list chronicling history of election results contrasted with the polling predictions. Though the election results often mirror the polls, deviations do occur; notably with the 2012 election, where the polls slated Mitt Romney as one percent favored over President Barack Obama.
Charles F. Manski wrote an opinion piece for the New York Times regarding the fickle nature of polls back in October of 2000, a month before one of the closest elections in United States history. Manski describes a problem with the vague, conditional language that pollsters often use to pose questions; “Pollsters ask Americans to specify the candidate they ‘lean toward’ or are ‘more likely to vote for” or “would vote for if the election were held today.'”