With less than a month to go before the first votes are cast, the 2016 cycle is heating up. It seems as if this year more than ever candidates are touting polls on the campaign trail. Coinciding with increased campaign rhetoric focused on polling, Pew Research recently published a fascinating study testing methods pollsters use to identify likely voters. Pew concluded that rather than relying on voters to self-identify, verifying whether or not someone actually voted by matching respondents back to the voter file is significantly more accurate.
Clarity’s work was featured in Section 2: Measuring the Likelihood to Vote. Pew Researchers used TargetSmart’s voting file, which aggregates official state-level data and voting records, to verify respondents voting history. Their data contains a voter likelihood score developed by us here at Clarity. The Washington Post praised Clarity for having a “strong likely voter model,” and Pew Research found that models such as ours provide more accurate predictions of the eventual electorate.
For more on likely voter models, read the full report by Pew Research Center here.
Be sure to check out the Washington Post's full article as well.