Are big data polls replacing traditional media polls?

The US presidential election in 2012 changed the way we think about media polls, says Jim Roberts, Executive Editor of Reuters Digital and former assistant managing editor of The New York Times.

He is referring to the forecasts of the American statistician Nate Silver, who predicted correctly the winner of all 50 states and the District of Columbia. In developing forecasts, Silver used a much bigger database than the traditional media polls. He used material from smaller and more un-familiar sources.

Roberts thinks that the essential question during the campaign was: Should the public believe in a poll made with a random selection of sources or should they believe the traditional polls?

"We need fundamental re-thinking about how traditional media polls work anymore if the industry gets better in predicting human behavior with these new kinds of polls that are based on big data."

Roberts wonders, whether traditional polls should merely be treated as a fuzzy snapshot of a moment in time, as David Brooks, a New York Times columnist, suggests.

Roberts attended the conference via Skype.