Keywords: Electoral polls in emerging/incomplete democracies

Polls has been questioned in the past years due to the unexpected elections results obtained in some processes such as the Brexit Referendum, Spanish General Elections, USA Presidential elections or the Colombian Plebiscite in 2016. However, polls have shown great accuracy in 2017 and have successfully anticipated results of the elections different countries such as: France, Germany or Austria.

If in 2016 research institutes surveyed dangerously in 2017, the research sector has proved that opinion polls work and are the appropiate tool to analyze public opinion, social trends and voting behavior. Nevertheless, methodology needs to be reviewed periodically.

In 2017, GAD3 has been able to anticipate the results of the regional elections in Catalonia, a process marked by the exceptionality of the circumstances: third elections in four years, president of the region fled in other country, political leaders in prison and elections held in a work day.

In this paper, we will explain how GAD3 has renewed its polling methodology, since the selection of the interviews to the application of the technology and big data. We will also identify the key factors in order to be the most accurate institute in anticipating the results of the Catalan elections the 21st of December and the principal problems found in the process.