This paper is a contribution by the WIN Network (international association gathering leading independent market research and polling firms in their respective countries), to discussions on changes in the polling industry. Drawing on feedback from one of its members (bva-group.com), which tracked the 2017 French presidential election using various sources of data (web listening, an online community, and polls collected via POP2017 public opinion platform), it also presents the results of an experiment conducted with two behavioral scientists at the Santa Fe Institute (USA). The purpose of this joint research, which was conducted during the French presidential campaign, was to better understand the factors ? particularly social bubbles ? that influence voters as well as measure the impact of fake news on public opinion.
The accuracy of polling has been under the serious spotlight over the last couple of years. This paper is a summary of analysis conducted on an international database of 31310 polls from 473 elections and voting events across 40 countries around the world from 1936 to 2017 complied by Kantar.
Some would say that our industry started a century ago when in 1916 the Literary Digest ran the first national opinion poll among its readers, correctly predicting Woodrow Wilson as the next US President. The Digest correctly called the next four elections. How have we changed over the last hundred years and where will we be in the next hundred?
The eleven ESOMAR/WAPOR key requirements for opinion polls and published surveys.